The option can be used to suppress output if we only want to test the
return value of the command.
Also, mention this option in the documentation.
Signed-off-by: Michael Chang <mchang@suse.com>
Signed-off-by: Maxim Suhanov <dfirblog@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Switching to another EFI boot application while there are secrets in
RAM is dangerous, because not all firmware is wiping memory on free.
To reduce the attack surface, wipe the passphrase acquired when
unlocking an encrypted volume.
Signed-off-by: Maxim Suhanov <dfirblog@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
This commit adds the grub_cryptodisk_erasesecrets() function to wipe
master keys from all cryptodisks. This function is EFI-only.
Since there is no easy way to "force unmount" a given encrypted disk,
this function renders all mounted cryptodisks unusable. An attempt to
read them will return garbage.
This is why this function must be used in "no way back" conditions.
Currently, it is used when unloading the cryptodisk module and when
performing the "exit" command (it is often used to switch to the next
EFI application). This function is not called when performing the
"chainloader" command, because the callee may return to GRUB. For this
reason, users are encouraged to use "exit" instead of "chainloader" to
execute third-party boot applications.
This function does not guarantee that all secrets are wiped from RAM.
Console output, chunks from disk read requests and other may remain.
This function does not clear the IV prefix and rekey key for geli disks.
Also, this commit adds the relevant documentation improvements.
Signed-off-by: Maxim Suhanov <dfirblog@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Document the --cryptodisk-only argument. Also, document the
"cryptocheck" command invoked when that argument is processed.
Signed-off-by: Maxim Suhanov <dfirblog@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
When the --cryptodisk-only argument is given, also check the target
device using the "cryptocheck" command, if available.
This extends the checks to common layouts like LVM-on-LUKS, so the
--cryptodisk-only argument transparently handles such setups.
Signed-off-by: Maxim Suhanov <dfirblog@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
This command examines a given diskfilter device, e.g., an LVM disk,
and checks if underlying disks, physical volumes, are cryptodisks,
e.g., LUKS disks, this layout is called "LVM-on-LUKS".
The return value is 0 when all underlying disks (of a given device)
are cryptodisks (1 if at least one disk is unencrypted or in an
unknown state).
Users are encouraged to include the relevant check before loading
anything from an LVM disk that is supposed to be encrypted.
This further supports the CLI authentication, blocking bypass
attempts when booting from an encrypted LVM disk.
Signed-off-by: Maxim Suhanov <dfirblog@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
This allows users to restrict the "search" command's scope to
encrypted disks only.
Typically, this command is used to "rebase" $root and $prefix
before loading additional configuration files via "source" or
"configfile". Unfortunately, this leads to security problems,
like CVE-2023-4001, when an unexpected, attacker-controlled
device is chosen by the "search" command.
The --cryptodisk-only argument allows users to ensure that the
file system picked is encrypted.
This feature supports the CLI authentication, blocking bypass
attempts.
Signed-off-by: Maxim Suhanov <dfirblog@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
This further mitigates potential misuse of the CLI after the
root device has been successfully unlocked via TPM.
Fixes: CVE-2025-4382
Signed-off-by: Maxim Suhanov <dfirblog@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
When large extent counter / NREXT64 support was added to GRUB, it missed
a couple of direct reads of nextents which need to be changed to the new
NREXT64-aware helper as well. Without this, we'll have mis-reads of some
directories with this feature enabled.
The large extent counter fix likely raced on merge with commit 07318ee7e
(fs/xfs: Fix XFS directory extent parsing) which added the new direct
nextents reads just prior, causing this issue.
Fixes: aa7c132267 (fs/xfs: Add large extent counters incompat feature support)
Signed-off-by: Eric Sandeen <sandeen@redhat.com>
Reviewed-by: Anthony Iliopoulos <ailiop@suse.com>
Reviewed-by: Jon DeVree <nuxi@vault24.org>
Reviewed-by: Vladimir Serbinenko <phcoder@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
RAID 4 requires the same recovery module as RAID 5. Extend the condition to
cover both RAID levels.
Signed-off-by: Egor Ignatov <egori@altlinux.org>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The code goes on to allocate memory in another region on failure, hence
it should discard the error.
Signed-off-by: Vladimir Serbinenko <phcoder@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Other platforms specify license in platform-specific files but corresponding
code for emu is in kernel, so datetime ends up without license section.
Signed-off-by: Vladimir Serbinenko <phcoder@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Without this option compiler sometimes emits R_RISCV_ALIGN relocs.
Unlike other relocs this one requires the linker to do NOP deletions
and we can't ignore them. Just instruct compiler not to emit them.
Signed-off-by: Vladimir Serbinenko <phcoder@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Add the long options of tpm2_key_protect_init along with the short options.
Signed-off-by: Gary Lin <glin@suse.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The TPM2 key protector tests require two external packages: swtpm-tools
and tpm2-tools. Add those two packages to the INSTALL file to inform
the user to install those packages before starting the TPM2 key protector
tests.
Signed-off-by: Gary Lin <glin@suse.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Reviewed-by: Stefan Berger <stefanb@linux.ibm.com>
This commit updates the NV index mode section and the grub-protect
section to reflect the recent changes in TPM2 key protector and
grub-protect.
Signed-off-by: Gary Lin <glin@suse.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Two more NV index test cases are added to test key sealing and
unsealing with the NV index handle 0x1000000.
Signed-off-by: Gary Lin <glin@suse.com>
Reviewed-by: Stefan Berger <stefanb@linux.ibm.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Reset "ret" to 0 when a test case fails so that the other test cases
could continue.
Also set the exit status to 1 when encountering a failure to reflect the
test result.
Signed-off-by: Gary Lin <glin@suse.com>
Reviewed-by: Stefan Berger <stefanb@linux.ibm.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Since grub-protect already supports NV index mode, tpm2_seal_nv() is
replaced with one grub-protect command to simplify the test script.
"tpm2_evictcontrol" is also replaced with "grub-protect --tpm2-evict".
Signed-off-by: Gary Lin <glin@suse.com>
Reviewed-by: Stefan Berger <stefanb@linux.ibm.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
This commit implements the missing NV index mode support in grub-protect.
NV index mode stores the sealed key in the TPM non-volatile memory (NVRAM)
instead of a file. There are two supported types of TPM handles.
1. Persistent handle (0x81000000~0x81FFFFFF)
Only the raw format is supported due to the limitation of persistent
handles. This grub-protect command seals the key into the
persistent handle 0x81000000.
# grub-protect \
--protector=tpm2 \
--action=add \
--tpm2-bank=sha256 \
--tpm2-pcrs=7,11 \
--tpm2-keyfile=luks-key \
--tpm2-nvindex=0x81000000
2. NV index handle (0x1000000~0x1FFFFFF)
Both TPM 2.0 Key File format and the raw format are supported by NV
index handles. Here is the grub-protect command to seal the key in
TPM 2.0 Key File format into the NV index handle 0x1000000.
# grub-protect \
--protector=tpm2 \
--action=add \
--tpm2key \
--tpm2-bank=sha256 \
--tpm2-pcrs=7,11 \
--tpm2-keyfile=luks-key \
--tpm2-nvindex=0x1000000
Besides the "add" action, the corresponding "remove" action is also
introduced. To remove the data from a persistent or NV index handle,
just use "--tpm2-nvindex=HANDLE" combining with "--tpm2-evict". This
sample command removes the data from the NV index handle 0x1000000.
# grub-protect \
--protector=tpm2 \
--action=remove \
--tpm2-evict \
--tpm2-nvindex=0x1000000
Also set and check the boolean variables with true/false instead of 1/0.
Signed-off-by: Gary Lin <glin@suse.com>
Reviewed-by: Stefan Berger <stefanb@linux.ibm.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Previously, NV index mode only supported persistent handles which are
only for TPM objects.
On the other hand, the "NV index" handle allows the user-defined data,
so it can be an alternative to the key file and support TPM 2.0 Key
File format immediately.
The following tpm2-tools commands store the given key file, sealed.tpm,
in either TPM 2.0 Key File format or the raw format into the NV index
handle 0x1000000.
# tpm2_nvdefine -C o \
-a "ownerread|ownerwrite" \
-s $(stat -c %s sealed.tpm) \
0x1000000
# tpm2_nvwrite -C o -i sealed.tpm 0x1000000
To unseal the key in GRUB, add the "tpm2_key_protector_init" command to
grub.cfg:
tpm2_key_protector_init --mode=nv --nvindex=0x1000000
cryptomount -u <UUID> --protector tpm2
To remove the NV index handle:
# tpm2_nvundefine -C o 0x1000000
Signed-off-by: Gary Lin <glin@suse.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Extract the logic to handle the file buffer from the SRK recover
function to prepare to load the sealed key from the NV index handle,
so the NV index mode can share the same code path in the later patch.
The SRK recover function now only reads the file and sends the file
buffer to the new function.
Besides this, to avoid introducing more options for the NV index mode,
the file format is detected automatically before unmarshaling the data,
so there is no need to use the command option to specify the file format
anymore. In other words, "-T" and "-k" are the same now.
Also update grub.text to address the change.
Signed-off-by: Gary Lin <glin@suse.com>
Reviewed-by: Stefan Berger <stefanb@linux.ibm.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The following TPM 2.0 commands are introduced to tss2 to access the
TPM non-volatile memory associated with the NV index handles:
- TPM2_NV_DefineSpace,
- TPM2_NV_UndefineSpace,
- TPM2_NV_ReadPublic,
- TPM2_NV_Read,
- TPM2_NV_Write.
The related marshal/unmarshal functions are also introduced.
Signed-off-by: Gary Lin <glin@suse.com>
Reviewed-by: Stefan Berger <stefanb@linux.ibm.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
grub_tpm2_readpublic() and grub_tpm2_testparms() didn't check
authCommand when marshaling the input data buffer. Currently, there is
no caller using non-NULL authCommand. However, to avoid the potential
issue, the conditional check is added to insert authCommand into the
input buffer if necessary.
Also fix a few pointer checks.
Signed-off-by: Gary Lin <glin@suse.com>
Reviewed-by: Stefan Berger <stefanb@linux.ibm.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The user may need to inspect the TPM 2.0 PCR values with the GRUB shell,
so the new tpm2_dump_pcr command is added to print all PCRs of the
specified bank.
Also update the document for the new command.
Signed-off-by: Gary Lin <glin@suse.com>
Tested-by: Stefan Berger <stefanb@linux.ibm.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
PCR mismatch is one common cause of TPM key unsealing fail. Since the
system may be compromised, it is not safe to boot into OS to get the PCR
values and TPM eventlog for the further investigation.
To provide some hints, GRUB now dumps PCRs on policy fail, so the user
can check the current PCR values. PCR 0~15 are chosen to cover the
firmware, bootloader, and OS.
The sample output:
PCR Mismatch! Check firmware and bootloader before typing passphrase!
TPM PCR [sha256]:
00: 17401f37710984c1d8a03a81fff3ab567ae9291bac61e21715b890ee28879738
01: 7a114329ba388445a96e8db2a072785937c1b7a8803ed7cc682b87f3ff3dd7a8
02: 11c2776849e8e24b7d80c926cbc4257871bffa744dadfefd3ed049ce25143e05
03: 6c33b362073e28e30b47302bbdd3e6f9cee4debca3a304e646f8c68245724350
04: 62d38838483ecfd2484ee3a2e5450d8ca3b35fc72cda6a8c620f9f43521c37d1
05: d8a85cb37221ab7d1f2cc5f554dbe0463acb6784b5b8dc3164ccaa66d8fff0e1
06: 9262e37cbe71ed4daf815b4a4881fb7251c9d371092dde827557d5368121e10e
07: 219d542233be492d62b079ffe46cf13396a8c27e520e88b08eaf2e6d3b7e70f5
08: de1f61c973b673e505adebe0d7e8fb65fde6c24dd4ab4fbaff9e28b18df6ecd3
09: c1de7274fa3e879a16d7e6e7629e3463d95f68adcfd17c477183846dccc41c89
10: 0000000000000000000000000000000000000000000000000000000000000000
11: 0000000000000000000000000000000000000000000000000000000000000000
12: 0000000000000000000000000000000000000000000000000000000000000000
13: 0000000000000000000000000000000000000000000000000000000000000000
14: 9ab9ebe4879a7f4dd00c04f37e79cfd69d0dd7a8bcc6b01135525b67676a3e40
15: 0000000000000000000000000000000000000000000000000000000000000000
16: 0000000000000000000000000000000000000000000000000000000000000000
17: ffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffff
18: ffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffff
19: ffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffff
20: ffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffff
21: ffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffff
22: ffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffffff
23: 0000000000000000000000000000000000000000000000000000000000000000
error: failed to unseal sealed key (TPM2_Unseal: 0x99d).
error: no key protector provided a usable key for luks (af16e48f-746b-4a12-aae1-c14dcee429e0).
If the user happens to have the PCR values for key sealing, the PCR dump
can be used to identify the changed PCRs and narrow down the scope for
closer inspection.
Please note that the PCR dump is trustworthy only if the GRUB binary is
authentic, so the user has to check the GRUB binary thoroughly before
using the PCR dump.
Signed-off-by: Gary Lin <glin@suse.com>
Reviewed-by: Stefan Berger <stefanb@linux.ibm.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Update linux_kernel_params to match the v6.13.7 upstream version of boot_params.
Refactor most things out into structs, as the Linux kernel does.
edid_info should be a struct with "unsigned char dummy[128]" and efi_info should
be a struct as well, starting at 0x1c0. However, for backwards compatibility,
GRUB can have efi_systab at 0x1b8 and padding at 0x1bc (or padding at both spots).
This cuts into the end of edid_info. Make edid_info inline and only make it go
up to 0x1b8.
Signed-off-by: Patrick Colp <patrick.colp@oracle.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
In grub_xnu_load_kext_from_dir(), when the call to grub_device_open()
failed, it simply cleaned up previously allocated memory and returned
GRUB_ERR_NONE. However, it neglected to free ctx->newdirname which is
allocated before the call to grub_device_open().
Fixes: CID 473859
Signed-off-by: Lidong Chen <lidong.chen@oracle.com>
Reviewed-by: Vladimir Serbinenko <phcoder@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
In grub_cmd_initrd(), initrd_ctx is allocated before calling
grub_relocator_alloc_chunk_align(). When that function fails,
initrd_ctx should be freed before exiting grub_cmd_initrd().
Fixes: CID 473852
Signed-off-by: Lidong Chen <lidong.chen@oracle.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Fix memory leaks in make_vg() with new helper functions, free_pv()
and free_lv(). Additionally, correct a check after allocating
comp->segments->nodes that mistakenly checked lv->segments->nodes
instead, likely due to a copy-paste error.
Fixes: CID 473878
Fixes: CID 473884
Fixes: CID 473889
Fixes: CID 473890
Signed-off-by: Lidong Chen <lidong.chen@oracle.com>
Reviewed-by: Vladimir Serbinenko <phcoder@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
A regression was introduced recently as a part of the series of
filesystem related patches to address some CVEs found in GRUB.
This issue may cause either an infinite loop at startup when
accessing certain valid NTFS filesystems, or may cause a crash
due to a NULL pointer dereference on systems where NULL address
is invalid (such as may happen when calling grub-mount from
the operating system level).
Correct this issue by checking that at->attr_cur is within bounds
inside find_attr().
Fixes: https://savannah.gnu.org/bugs/?66855
Fixes: aff263187 (fs/ntfs: Fix out-of-bounds read)
Signed-off-by: B Horn <b@horn.uk>
Signed-off-by: Andrew Hamilton <adhamilt@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com
Reviewed-by: Vladimir Serbinenko <phcoder@gmail.com>
The grub_malloc() has been inadvertently removed from the code after it
has been modified to use safe math functions.
Fixes: 4beeff8a (net: Use safe math macros to prevent overflows)
Signed-off-by: Nicolas Frayer <nfrayer@redhat.com>
Tested-by: Marta Lewandowska <mlewando@redhat.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Change RMA size from 512 MB to 768 MB which will result in more memory
at boot time for PowerPC. When vTPM, Secure Boot or FADump are enabled
on PowerPC the 512 MB RMA memory is not sufficient for boot. With this
512 MB RMA, GRUB runs out of memory and fails to boot the machine.
Sometimes even usage of CDROM requires more memory for installation and
along with the options mentioned above exhausts the boot memory which
results in boot failures. Increasing the RMA size will resolves multiple
out of memory issues observed on PowerPC machines.
Failure details (GRUB debug console dump):
kern/ieee1275/init.c:550: mm requested region of size 8513000, flags 1
kern/ieee1275/init.c:563: Cannot satisfy allocation and retain minimum runtime space
kern/ieee1275/init.c:550: mm requested region of size 8513000, flags 0
kern/ieee1275/init.c:563: Cannot satisfy allocation and retain minimum runtime space
kern/file.c:215: Closing `/ppc/ppc64/initrd.img' ...
kern/disk.c:297: Closing `ieee1275//vdevice/v-scsi@30000067/disk@8300000000000000'...
kern/disk.c:311: Closing `ieee1275//vdevice/v-scsi@30000067/disk@8300000000000000' succeeded.
kern/file.c:225: Closing `/ppc/ppc64/initrd.img' failed with 3.
kern/file.c:148: Opening `/ppc/ppc64/initrd.img' succeeded.
error: ../../grub-core/kern/mm.c:552:out of memory.
Signed-off-by: Avnish Chouhan <avnish@linux.ibm.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Without this fix the GRUB failed to boot linux with "out of memory" after
trying to run a "search --fs-uuid..." on a system that has 7 ZFS pools
across about 80 drives.
Signed-off-by: Stuart Hayes <stuart.w.hayes@gmail.com>
Reviewed-by: Vladimir Serbinenko <phcoder@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
This removes a lot of empty grub-shell working directories in the TMPDIR directory.
Signed-off-by: Thomas Schmitt <scdbackup@gmx.net>
Signed-off-by: Glenn Washburn <development@efficientek.com>
Tested-by: Thomas Schmitt <scdbackup@gmx.net>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
grub_cmd_cryptomount creates a directory per subtest. If a subtest is
successful and debugging is not on, the directory should be empty.
So, it can be deleted.
Signed-off-by: Thomas Schmitt <scdbackup@gmx.net>
Signed-off-by: Glenn Washburn <development@efficientek.com>
Tested-by: Thomas Schmitt <scdbackup@gmx.net>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
This fixes behavior where grub_cmd_cryptomount temporary files, which are
some times not cleaned up, are left in the / directory. Set TMPDIR if your
system does not have /tmp or it can not be used for some reason.
Reported-by: Thomas Schmitt <scdbackup@gmx.net>
Signed-off-by: Glenn Washburn <development@efficientek.com>
Tested-by: Thomas Schmitt <scdbackup@gmx.net>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
This fixes an issue where the grub_cmd_cryptomount test leaves a file
with an ambiguous name in the / directory when TMPDIR is not set.
Reported-by: Thomas Schmitt <scdbackup@gmx.net>
Signed-off-by: Glenn Washburn <development@efficientek.com>
Tested-by: Thomas Schmitt <scdbackup@gmx.net>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
grub-shell-luks-tester only cleans up generated files when the test it
runs returns success. Sometimes tests are run that should fail. Add
a --xfail argument to grub-shell-luks-tester and pass it from
grub_cmd_cryptomount when invoking a test that is expected to fail.
Reported-by: Thomas Schmitt <scdbackup@gmx.net>
Signed-off-by: Glenn Washburn <development@efficientek.com>
Tested-by: Thomas Schmitt <scdbackup@gmx.net>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Set the RET variable to the exit status of the script, as was assumed in
the cleanup() function.
Reported-by: Thomas Schmitt <scdbackup@gmx.net>
Signed-off-by: Glenn Washburn <development@efficientek.com>
Tested-by: Thomas Schmitt <scdbackup@gmx.net>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
GRUB has the capability to search all the disks for a cryptodisk of a
given UUID. Use this instead of hardcoding which disk is the cryptodisk,
which can change when devices are added or removed, or potentially when
QEMU is upgraded. This can not be done for the detached header tests
because the header contains the UUID.
Also, capitalize comment lines for consistency.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Fix a regression where qemuopts was mistakenly defaulted to the empty
string. This prevents the sending of arbitrary QEMU options to tests,
which may be desirable for overriding the machine type. There was a
concern that allowing the tester to accept arbitrary options would add
headaches for another developer trying to diagnose why a test failed on
the testers machine because he could not be sure if any additional
options were passed to make the test fail. However, all the options are
recorded in the run.sh generated script, so this concern is unwarranted.
Fixes: 6d729ced70 (tests/util/grub-shell: Add $GRUB_QEMU_OPTS to run.sh to easily see unofficial QEMU arguments)
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Mark cachevol LV's as ignored features, which is true only if they are
configured as "writethrough". This patch does not let GRUB boot from
"writeback" cache-enabled LV's.
Signed-off-by: Patrick Plenefisch <simonpatp@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The LV matching must be done after processing the ignored feature
indirections, as integrity volumes & caches may have several levels
of indirection that the segments must be shifted through.
Signed-off-by: Patrick Plenefisch <simonpatp@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The PV matching must be completely finished before validating a volume,
otherwise referenced RAID stripes may not have PV data applied yet.
This change is required for integrity & cachevol support.
Signed-off-by: Patrick Plenefisch <simonpatp@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The cache_pool is never read or used, remove it.
Signed-off-by: Patrick Plenefisch <simonpatp@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
This patch isn't necessary by itself, but when combined with subsequent
patches it enhances readability as ignored_features_lv is then used for
multiple types of extra LV's, not just cache LV's.
Signed-off-by: Patrick Plenefisch <simonpatp@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Like the GNU ls, first print a line with the directory path before printing
files in the directory, which will not have a directory component, but only
if there is more than one argument.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
For arguments that are paths to files, print the full path of the file.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The modification time for paths to files was not being printed because
the grub_dirhook_info, which contains the mtime, was initialized to NULL.
Instead of calling print_file() directly, use fs->fs_dir() to call
print_file() with a properly filled in grub_dirhook_info. This has the
added benefit of reducing code complexity.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Simplify the code by removing logic around which file printer to call.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Update documentation to capture that all memrw commands, the minicmd
dump command, and raw memory dumping via hexdump are restricted when
lockdown is enabled. This aligns to recent GRUB code updates.
Signed-off-by: Andrew Hamilton <adhamilt@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Document which filesystems are not allowed when lockdown
is enabled to align to recent GRUB changes.
Signed-off-by: Andrew Hamilton <adhamilt@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
FreeBSD loader always passes "elf kernel". We currently pass "elf64 kernel"
when loading 64-bit kernel. The -CURRENT, HEAD, kernel accepts only
"elf kernel". Older kernel accepts either.
Tested with FreeBSD and DragonFlyBSD.
Reference: https://cgit.freebsd.org/src/commit/?id=b72ae900d4348118829fe04abdc11b620930c30f
Signed-off-by: Vladimir Serbinenko <phcoder@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
When using syntax "hd0,gtp3,dfly1" then ptr points to trailing part, ",dfly1".
So, it's improper to consider it as an invalid partition.
Signed-off-by: Vladimir Serbinenko <phcoder@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
grub-core/lib/tss2/tss2_structs.h contains a duplicate typedef as follows:
typedef TPMS_SCHEME_HASH_t TPMS_SCHEME_KDF2_t;
This causes a build failure when compiling with clang. Remove the
duplicate typedef which allows successfully building GRUB with clang.
Signed-off-by: Andrew Hamilton <adhamilt@gmail.com>
Reviewed-by: Ross Philipson <ross.philipson@oracle.com>
Reviewed-by: Gary Lin <glin@suse.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The grub_script_execute_sourcecode() parses and executes code one line
at a time, updating the return code each time because only the last line
determines the final status. However, trailing new lines were also
executed, masking any failure on the previous line. Fix this by only
trying to execute the command when there is actually one present.
This has presumably never been noticed because this code is not used by
regular functions, only in special cases like eval and menu entries. The
latter generally don't return at all, having booted an OS. When failing
to boot, upstream GRUB triggers the fallback mechanism regardless of the
return code.
We noticed the problem while using Red Hat's patches, which change this
behaviour to take account of the return code. In that case, a failure
takes you back to the menu rather than triggering a fallback.
Signed-off-by: James Le Cuirot <jlecuirot@microsoft.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The commit 504058e8 (libtasn1: Compile into asn1 module) generates files
into the grub-core/lib/libtasn1-grub directory and commit 99cda678
(asn1_test: Test module for libtasn1) generates files into the
grub-core/tests/asn1/tests directory. Ignore these directories as they
are not under revision control.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
On systems which support multiple boot platforms such as BIOS and
EFI, it makes no sense to show menu entries which are not supported
by the current boot platform. Menu entries generated from os-prober
"chain" boot type use boot sector chainloading which is supported
on PC BIOS platform only.
Show "chain" menu entries only if boot platform is PC BIOS.
Show "efi" menu entries only if boot platform is EFI.
This is aimed to allow os-prober to report both EFI and PC BIOS
boot loaders regardless of the current boot mode on x86 systems
which support both EFI and legacy BIOS boot, in order to generate
a config file which can be used with either BIOS or EFI boot.
Signed-off-by: Pascal Hambourg <pascal@plouf.fr.eu.org>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
GRUB documentation states:
GRUB_OS_PROBER_SKIP_LIST
List of space-separated FS UUIDs of filesystems to be ignored from
os-prober output. For efi chainloaders it’s <UUID>@<EFI FILE>
But the actual behaviour does not match this description.
GRUB_OS_PROBER_SKIP_LIST="<UUID>"
does nothing. In order to skip non-EFI bootloaders, you must set
GRUB_OS_PROBER_SKIP_LIST="<UUID>@<DEVICE>"
which is both absurd, <UUID> and <DEVICE> are redundant, and wrong,
<DEVICE> such as /dev/sd* may not be persistent across boots.
Also, any non-word character is accepted as a separator, including "-"
and "@" which may be present in UUIDs. This can cause false positives
because of partial UUID match.
This patch fixes these flaws while retaining some backward compatibility
with previous behaviour which may be expected by existing setups:
- also accept <UUID>@/dev/* (with warning) for non-EFI bootloaders,
- also accept comma and semicolon as separator.
Fixes: 55e706c9 (Add GRUB_OS_PROBER_SKIP_LIST to selectively skipping systems)
Signed-off-by: Pascal Hambourg <pascal@plouf.fr.eu.org>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
This appears to be a relic from GRUB legacy that used a --dumb option for
its terminal command. The proper way to do this in GRUB2 is to set the
terminal to "dumb" via the terminfo command.
Fixes: https://savannah.gnu.org/bugs/?66302
Reported-by: Jernej Jakob <jernej.jakob+savgnu@gmail.com>
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Support for @lbracechar{} and @rbracechar{} was added in GNU Texinfo 5.0
but many older systems may have versions lower than this. Use @{ and @}
to support a wider range of GNU Texinfo versions.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Commit ef7850c757 (fs/xfs: Fix issues found while fuzzing the XFS
filesystem) introduced multiple boundary checks in grub_xfs_iterate_dir()
but handled the error incorrectly returning error code instead of 0.
Fix it. Also change the error message so that it doesn't match the
message in grub_xfs_read_inode().
Fixes: ef7850c757 (fs/xfs: Fix issues found while fuzzing the XFS filesystem)
Signed-off-by: Egor Ignatov <egori@altlinux.org>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The Linux port of XFS added a few new features in 2024. The existing
GRUB driver doesn't attempt to read or write any of the new metadata,
so, all three can be added to the incompat allowlist.
On the occasion align XFS_SB_FEAT_INCOMPAT_NREXT64 value.
Signed-off-by: Darrick J. Wong <djwong@kernel.org>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Previously, the number of extent entries was not properly capped based
on the actual available space. This could lead to insufficient reads for
external extents since the computation was based solely on the inline
extent layout.
In this patch, when processing the extent header we determine whether
the header is stored inline, i.e. at inode->blocks.dir_blocks, or in an
external extent block. We then clamp the number of entries accordingly
(using max_inline_ext for inline extents and max_external_ext for
external extent blocks).
This change ensures that only the valid number of extent entries is
processed preventing out-of-bound reads and potential filesystem
corruption.
Fixes: 7e2f750f0a (fs/ext2: Fix out-of-bounds read for inline extents)
Signed-off-by: Michael Chang <mchang@suse.com>
Tested-by: Christian Hesse <mail@eworm.de>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The conditional makes no sense when the two possible expressions have
the same value, so, remove it (perhaps the compiler does it for us but
better to remove it). This change makes spinup argument unused. So, drop
it as well.
Signed-off-by: Leo Sandoval <lsandova@redhat.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The Microsoft spec for SPCR says "The base address of the Serial Port
register set described using the ACPI Generic Address Structure, or
0 if console redirection is disabled". So, return early if redirection
is disabled (base address = 0). If this check is not done we may get
invalid ports on machines with redirection disabled and boot may hang
when reading the grub.cfg file.
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Reviewed-by: Leo Sandoval <lsandova@redhat.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The pointer returned by grub_elf_file() is not checked to verify it is
not NULL before use. A NULL pointer may be returned when the given file
does not have a valid ELF header.
Fixes: https://savannah.gnu.org/bugs/?61960
Signed-off-by: Glenn Washburn <development@efficientek.com>
Signed-off-by: Lukas Fink <lukas.fink1@gmail.com>
Reviewed-by: Ross Philipson <ross.philipson@oracle.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
%s/hueristic/heuristic/
Signed-off-by: Heinrich Schuchardt <heinrich.schuchardt@canonical.com>
Reviewed-by: Ross Philipson <ross.philipson@oracle.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The EFI Boot Services can be used after ExitBootServices() call because
the GRUB code still may allocate memory.
An example call stack is:
grub_multiboot_boot
grub_multiboot2_make_mbi
grub_efi_finish_boot_services
b->exit_boot_services
normal_boot
grub_relocator32_boot
grub_relocator_alloc_chunk_align_safe
grub_relocator_alloc_chunk_align
grub_malloc
grub_memalign
grub_mm_add_region_fn
[= grub_efi_mm_add_regions]
grub_efi_allocate_any_pages
grub_efi_allocate_pages_real
b->allocate_pages
This can lead to confusing errors. After ExitBootServices() call
b->allocate_pages may point to the NULL address resulting in something like:
!!!! X64 Exception Type - 01(#DB - Debug) CPU Apic ID - 00000000 !!!!
RIP - 000000000000201F, CS - 0000000000000038, RFLAGS - 0000000000200002
RAX - 000000007F9EE010, RCX - 0000000000000001, RDX - 0000000000000002
RBX - 0000000000000006, RSP - 00000000001CFBEC, RBP - 0000000000000000
RSI - 0000000000000000, RDI - 00000000FFFFFFFF
R8 - 0000000000000006, R9 - 000000007FEDFFB8, R10 - 0000000000000000
R11 - 0000000000000475, R12 - 0000000000000001, R13 - 0000000000000002
R14 - 00000000FFFFFFFF, R15 - 000000007E432C08
DS - 0000000000000030, ES - 0000000000000030, FS - 0000000000000030
GS - 0000000000000030, SS - 0000000000000030
CR0 - 0000000080010033, CR2 - 0000000000000000, CR3 - 000000007FC01000
CR4 - 0000000000000668, CR8 - 0000000000000000
DR0 - 0000000000000000, DR1 - 0000000000000000, DR2 - 0000000000000000
DR3 - 0000000000000000, DR6 - 00000000FFFF0FF0, DR7 - 0000000000000400
GDTR - 000000007F9DE000 0000000000000047, LDTR - 0000000000000000
IDTR - 000000007F470018 0000000000000FFF, TR - 0000000000000000
FXSAVE_STATE - 00000000001CF840
Ideally we would like to avoid all memory allocations after exiting EFI
Boot Services altogether but that requires significant code changes. This
patch adds a simple workaround that resets grub_mm_add_region_fn to NULL
after ExitBootServices() call, so:
- Memory allocations have a better chance of succeeding because grub_memalign()
will try to reclaim the disk cache if it sees a NULL in grub_mm_add_region_fn.
- At worst it will fail to allocate memory but it will explicitly tell users
that it's out of memory, which is still much better than the current
situation where it fails in a fairly random way and triggers a CPU fault.
Signed-off-by: Ruihan Li <lrh2000@pku.edu.cn>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
This patch is used to fix GRUB menu gets stuck in server AC
poweron/poweroff stress test of x86_64, which is reproduced with
1/200 ratio. The root cause analysis as below:
Q: What's the code logic?
A: The grub_tsc_init() function will init tsc by setting grub_tsc_rate,
which call stack is:
grub_tsc_init() -> grub_tsc_calibrate_from_pmtimer() -> grub_divmod64()
Among, grub_divmod64() function needs tsc_diff as the second parameter.
In grub_pmtimer_wait_count_tsc(), we will call grub_get_tsc() function
to get time stamp counter value to assign to start_tsc variable, and
get into while (1) loop space to get end_tsc variable value with same
function, after 3580 ticks, return "end_tsc - start_tsc". Actually,
rdtsc instruction will be called in grub_get_tsc, but rdtsc instruction
is not reliable (for the reason see the next question), which will cause
tsc_diff to be a very big number larger than (1UL << 32) or a negative
number, so that grub_tsc_rate will be zero. When run_menu() function is
startup, and calls grub_tsc_get_time_ms() function to get current time
to check if timeout time reach, at this time, grub_tsc_get_time_ms()
function will return zero due to zero grub_tsc_rate variable, then GRUB
menu gets stuck...
Q: What's the difference between rdtsc and rdtscp instructions in x86_64
architecture? Here is more explanations from Intel® 64 and IA-32
Architectures Software Developer’s Manual Volume 2B (December 2024):
https://cdrdv2.intel.com/v1/dl/getContent/671241
A: In page 4-558 -> RDTSC—Read Time-Stamp Counter:
The RDTSC instruction is not a serializing instruction. It does not
necessarily wait until all previous instructions have been executed
before reading the counter. Similarly, subsequent instructions may
begin execution before the read operation is performed. The following
items may guide software seeking to order executions of RDTSC:
- If software requires RDTSC to be executed only after all previous
instructions have executed and all previous loads are globally
visible, it can execute LFENCE immediately before RDTSC.
- If software requires RDTSC to be executed only after all previous
instructions have executed and all previous loads and stores are
globally visible, it can execute the sequence MFENCE;LFENCE
immediately before RDTSC.
- If software requires RDTSC to be executed prior to execution of any
subsequent instruction (including any memory accesses), it can execute
the sequence LFENCE immediately after RDTSC.
A: In page 4-560 -> RDTSCP—Read Time-Stamp Counter and Processor ID:
The RDTSCP instruction is not a serializing instruction, but it does wait
until all previous instructions have executed and all previous loads are
globally visible. But it does not wait for previous stores to be globally
visible, and subsequent instructions may begin execution before the read
operation is performed. The following items may guide software seeking to
order executions of RDTSCP:
- If software requires RDTSCP to be executed only after all previous
stores are globally visible, it can execute MFENCE immediately before
RDTSCP.
- If software requires RDTSCP to be executed prior to execution of any
subsequent instruction (including any memory accesses), it can execute
LFENCE immediately after RDTSCP.
Q: Why there is a cpuid serializing instruction before rdtsc instruction,
but "grub_get_tsc" still cannot work as expect?
A: From Intel® 64 and IA-32 Architectures Software Developer's Manual
Volume 2A: Instruction Set Reference, A-L (December 2024):
https://cdrdv2.intel.com/v1/dl/getContent/671199
In page 3-222 -> CPUID—CPU Identification:
CPUID can be executed at any privilege level to serialize instruction execution.
Serializing instruction execution guarantees that any modifications to flags,
registers, and memory for previous instructions are completed before
the next instruction is fetched and executed.
So we only kept the instruction rdtsc and its previous instruction in order
currently. But it is still out-of-order possibility between rdtsc instruction
and its subsequent instruction.
Q: Why do we do this fix?
A: In the one hand, add cpuid instruction after rdtsc instruction to make sure
rdtsc instruction to be executed prior to execution of any subsequent instruction,
about serializing execution that all previous instructions have been executed
before rdtsc, there is a cpuid usage in original code. In the other hand, using
cpuid instruction rather than lfence can make sure a forward compatibility for
previous HW.
Base this fix, we did 1500 cycles power on/off stress test, and did not reproduce
this issue again.
Fixes: https://savannah.gnu.org/bugs/?66257
Signed-off-by: Duan Yayong <duanyayong@bytedance.com>
Signed-off-by: Li Yongqiang <liyongqiang@huaqin.com>
Signed-off-by: Sun Ming <simon.sun@huaqin.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The grub_divmod64() may return 0 but grub_tsc_calibrate_from_pmtimer()
still returns 1 saying calibration succeeded. Of course it is not true.
So, return 0 when grub_divmod64() returns 0. This way other calibration
functions can be called subsequently.
Signed-off-by: Duan Yayong <duanyayong@bytedance.com>
Signed-off-by: Li Yongqiang <liyongqiang@huaqin.com>
Signed-off-by: Sun Ming <simon.sun@huaqin.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Simply returning from grub_cmd_linux() doesn't free "file" resource nor
calls grub_dl_ref(my_mod). Jump to "fail" label for proper cleanup like
other error checks do.
Signed-off-by: Sergii Dmytruk <sergii.dmytruk@3mdeb.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The operation kern_end - kern_start may underflow when we input it into
grub_relocator_alloc_chunk_addr() call. To avoid this we can use safe
math for this subtraction.
Fixes: CID 73845
Signed-off-by: Alec Brown <alec.r.brown@oracle.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The Coverity complains that we might overflow into a negative value when
setting linux_params.kernel_alignment to (1 << align). We can remedy
this by casting it to grub_uint32_t.
Fixes: CID 473876
Signed-off-by: Alec Brown <alec.r.brown@oracle.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
When the format string, fmt0, includes a positional argument
grub_strtoul() or grub_strtoull() is called to extract the argument
position. However, the returned argument position isn't fully validated.
If the format is something like "%0$x" then these functions return
0 which leads to an underflow in the calculation of the args index, curn.
The fix is to add a check to ensure the extracted argument position is
greater than 0 before computing curn. Additionally, replace one
grub_strtoull() with grub_strtoul() and change curn type to make code
more correct.
Fixes: CID 473841
Signed-off-by: Lidong Chen <lidong.chen@oracle.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The current code incorrectly assumes that both the input and the values
returned by grub_strtoul() are always valid which can lead to potential
errors. This fix ensures proper validation to prevent any unintended issues.
Fixes: CID 473843
Signed-off-by: Lidong Chen <lidong.chen@oracle.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The Coverity indicates that the variable current_entry might overflow.
To prevent this use safe math when adding GRUB_MENU_PAGE_SIZE to current_entry.
On the occasion fix limiting condition which was broken.
Fixes: CID 473853
Signed-off-by: Alec Brown <alec.r.brown@oracle.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The Coverity indicates that GRUB_EHCI_TOGGLE is an int that contains
a negative value and we are using it for the variable token which is
grub_uint32_t. To remedy this we can cast the definition to grub_uint32_t.
Fixes: CID 473851
Signed-off-by: Alec Brown <alec.r.brown@oracle.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Update the overflow error messages to make them consistent
across the GRUB code.
Signed-off-by: Lidong Chen <lidong.chen@oracle.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The entry_len is initialized in grub_find_root_devices_from_mountinfo()
to 0 before the while loop iterates through /proc/self/mountinfo. If the
file is empty or contains only invalid entries entry_len remains
0 causing entry_len - 1 in the subsequent for loop initialization
to underflow. To prevent this add a check to ensure entry_len > 0 before
entering the for loop.
Fixes: CID 473877
Signed-off-by: Lidong Chen <lidong.chen@oracle.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Reviewed-by: Ross Philipson <ross.philipson@oracle.com>
The result is initialized to 0 in grub_script_arglist_to_argv().
If the for loop condition is not met both result.args and result.argc
remain 0 causing result.argc - 1 to underflow and/or result.args NULL
dereference. Fix the issues by adding relevant checks.
Fixes: CID 473880
Signed-off-by: Lidong Chen <lidong.chen@oracle.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
When using grub_zalloc(), if we are out of memory, this function can fail.
After allocating memory, we should check if grub_zalloc() returns NULL.
If so, we should handle this error.
Fixes: CID 473856
Signed-off-by: Alec Brown <alec.r.brown@oracle.com>
Reviewed-by: Ross Philipson <ross.philipson@oracle.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
When using grub_malloc(), the function can fail if we are out of memory.
After allocating memory we should check if this function returned NULL
and handle this error if it did.
Signed-off-by: Alec Brown <alec.r.brown@oracle.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Use grub_calloc() when allocating memory for arrays to ensure proper
overflow checks are in place.
Signed-off-by: Lidong Chen <lidong.chen@oracle.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Replace direct arithmetic operations with macros from include/grub/safemath.h
to prevent potential overflow issues when calculating the memory sizes.
Signed-off-by: Lidong Chen <lidong.chen@oracle.com>
Signed-off-by: Alec Brown <alec.r.brown@oracle.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
When using grub_malloc() or grub_zalloc(), these functions can fail if
we are out of memory. After allocating memory we should check if these
functions returned NULL and handle this error if they did.
Signed-off-by: Lidong Chen <lidong.chen@oracle.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Use grub_calloc() when allocating memory for arrays to ensure proper
overflow checks are in place.
Signed-off-by: Lidong Chen <lidong.chen@oracle.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Replace direct arithmetic operations with macros from include/grub/safemath.h
to prevent potential overflow issues when calculating the memory sizes.
Signed-off-by: Lidong Chen <lidong.chen@oracle.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The direct assignment of the unsigned long long value returned by
read_number() can potentially lead to an overflow on a 32-bit systems.
The fix replaces the direct assignments with calls to grub_cast()
which detects the overflows and safely assigns the values if no
overflow is detected.
Signed-off-by: Lidong Chen <lidong.chen@oracle.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Use grub_calloc() when allocating memory for arrays to ensure proper
overflow checks are in place.
The HFS+ and squash4 security vulnerabilities were reported by
Jonathan Bar Or <jonathanbaror@gmail.com>.
Fixes: CVE-2025-0678
Fixes: CVE-2025-1125
Signed-off-by: Lidong Chen <lidong.chen@oracle.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Replace direct arithmetic operations with macros from include/grub/safemath.h
to prevent potential overflow issues when calculating the memory sizes.
Signed-off-by: Lidong Chen <lidong.chen@oracle.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
In the dev_iterate() function a handle is opened but isn't closed when
grub_malloc() returns NULL. We should fix this by closing it on error.
Signed-off-by: Alec Brown <alec.r.brown@oracle.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
When using grub_malloc(), grub_zalloc() or grub_calloc(), these functions can
fail if we are out of memory. After allocating memory we should check if these
functions returned NULL and handle this error if they did.
On the occasion make a NULL check in ATA code more obvious.
Signed-off-by: Alec Brown <alec.r.brown@oracle.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Use grub_calloc() when allocating memory for arrays to ensure proper
overflow checks are in place.
Signed-off-by: Alec Brown <alec.r.brown@oracle.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Replace direct arithmetic operations with macros from include/grub/safemath.h
to prevent potential overflow issues when calculating the memory sizes.
Signed-off-by: Alec Brown <alec.r.brown@oracle.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The idea is to permit the following: btrfs, cpio, exfat, ext, f2fs, fat,
hfsplus, iso9660, squash4, tar, xfs and zfs.
The JFS, ReiserFS, romfs, UDF and UFS security vulnerabilities were
reported by Jonathan Bar Or <jonathanbaror@gmail.com>.
Fixes: CVE-2025-0677
Fixes: CVE-2025-0684
Fixes: CVE-2025-0685
Fixes: CVE-2025-0686
Fixes: CVE-2025-0689
Suggested-by: Daniel Axtens <dja@axtens.net>
Signed-off-by: Daniel Axtens <dja@axtens.net>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The BFS is not fuzz-clean. Don't allow it to be loaded under lockdown.
This will also disable the AFS.
Fixes: CVE-2024-45778
Fixes: CVE-2024-45779
Reported-by: Nils Langius <nils@langius.de>
Signed-off-by: Daniel Axtens <dja@axtens.net>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
With the rest of module being blocked in lockdown mode it does not make
a lot of sense to leave memory reading enabled. This also goes in par
with disabling the dump command.
Reported-by: B Horn <b@horn.uk>
Signed-off-by: B Horn <b@horn.uk>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The dump enables a user to read memory which should not be possible
in lockdown mode.
Fixes: CVE-2025-1118
Reported-by: B Horn <b@horn.uk>
Reported-by: Jonathan Bar Or <jonathanbaror@gmail.com>
Signed-off-by: B Horn <b@horn.uk>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The test_parse() evaluates test expression recursively. Due to lack of
recursion depth check a specially crafted expression may cause a stack
overflow. The recursion is only triggered by the parentheses usage and
it can be unlimited. However, sensible expressions are unlikely to
contain more than a few parentheses. So, this patch limits the recursion
depth to 100, which should be sufficient.
Reported-by: Nils Langius <nils@langius.de>
Signed-off-by: Lidong Chen <lidong.chen@oracle.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The grub_getline() function currently has a signed integer variable "i"
that can be overflown when user supplies more than 2^31 characters.
It results in a memory corruption of the allocated line buffer as well
as supplying large negative values to grub_realloc().
Fixes: CVE-2025-0690
Reported-by: Jonathan Bar Or <jonathanbaror@gmail.com>
Signed-off-by: Jonathan Bar Or <jonathanbaror@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The size calculation of the translation buffer in
grub_gettext_getstr_from_position() may overflow
to 0 leading to heap OOB write. This patch fixes
the issue by using grub_add() and checking for
an overflow.
Fixes: CVE-2024-45777
Reported-by: Nils Langius <nils@langius.de>
Signed-off-by: Lidong Chen <lidong.chen@oracle.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Reviewed-by: Alec Brown <alec.r.brown@oracle.com>
Calculation of ctx->grub_gettext_msg_list size in grub_mofile_open() may
overflow leading to subsequent OOB write or read. This patch fixes the
issue by replacing grub_zalloc() and explicit multiplication with
grub_calloc() which does the same thing in safe manner.
Fixes: CVE-2024-45776
Reported-by: Nils Langius <nils@langius.de>
Signed-off-by: Lidong Chen <lidong.chen@oracle.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Reviewed-by: Alec Brown <alec.r.brown@oracle.com>
The gettext module does not entirely cleanup after itself in
its GRUB_MOD_FINI() leaving a few variables hooks in place.
It is not possible to unload gettext module because normal
module depends on it. Though fix the issues for completeness.
Fixes: CVE-2025-0622
Reported-by: B Horn <b@horn.uk>
Signed-off-by: B Horn <b@horn.uk>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The normal module does not entirely cleanup after itself in
its GRUB_MOD_FINI() leaving a few variables hooks in place.
It is not possible to unload normal module now but fix the
issues for completeness.
On the occasion replace 0s with NULLs for "pager" variable
hooks unregister.
Fixes: CVE-2025-0622
Reported-by: B Horn <b@horn.uk>
Signed-off-by: B Horn <b@horn.uk>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
If the hooks are not removed they can be called after the module has
been unloaded leading to an use-after-free.
Fixes: CVE-2025-0622
Reported-by: B Horn <b@horn.uk>
Signed-off-by: B Horn <b@horn.uk>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The grub_strrchr() may return NULL when the dirname do not contain "/".
This can happen on broken filesystems.
Reported-by: B Horn <b@horn.uk>
Signed-off-by: B Horn <b@horn.uk>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The grub_extcmd_dispatcher() calls grub_arg_list_alloc() to allocate
a grub_arg_list struct but it does not verify the allocation was successful.
In case of failed allocation the NULL state pointer can be accessed in
parse_option() through grub_arg_parse() which may lead to a security issue.
Fixes: CVE-2024-45775
Reported-by: Nils Langius <nils@langius.de>
Signed-off-by: Lidong Chen <lidong.chen@oracle.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Reviewed-by: Alec Brown <alec.r.brown@oracle.com>
The grub_dl_relocate_symbols() iterates through the sections in
an ELF looking for relocation sections. According to the spec [1]
the SHF_INFO_LINK flag should be set if the sh_info field is meant
to be a section index.
[1] https://refspecs.linuxbase.org/elf/gabi4+/ch4.sheader.html
Reported-by: B Horn <b@horn.uk>
Signed-off-by: B Horn <b@horn.uk>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The previous code would never actually call grub_update_mem_attrs()
as sh_info will always be zero for the sections that exist in memory.
Reported-by: B Horn <b@horn.uk>
Signed-off-by: B Horn <b@horn.uk>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
It was possible to overflow the value of mod->ref_count, a signed
integer, by repeatedly invoking insmod on an already loaded module.
This led to a use-after-free. As once ref_count was overflowed it became
possible to unload the module while there was still references to it.
This resolves the issue by using grub_add() to check if the ref_count
will overflow and then stops further increments. Further changes were
also made to grub_dl_unref() to check for the underflow condition and
the reference count was changed to an unsigned 64-bit integer.
Reported-by: B Horn <b@horn.uk>
Signed-off-by: B Horn <b@horn.uk>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Otherwise a subsequent header could change the height and width
allowing future OOB writes.
Fixes: CVE-2024-45774
Reported-by: Nils Langius <nils@langius.de>
Signed-off-by: Daniel Axtens <dja@axtens.net>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
An overly long filename can be passed to tftp_open() which would cause
grub_normalize_filename() to write out of bounds.
Fixed by adding an extra argument to grub_normalize_filename() for the
space available, making it act closer to a strlcpy(). As several fixed
strings are strcpy()'d after into the same buffer, their total length is
checked to see if they exceed the remaining space in the buffer. If so,
return an error.
On the occasion simplify code a bit by removing unneeded rrqlen zeroing.
Reported-by: B Horn <b@horn.uk>
Signed-off-by: B Horn <b@horn.uk>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The function included a call to grub_strcpy() which copied data from an
environment variable to a buffer allocated in grub_cmd_normal(). The
grub_cmd_normal() didn't consider the length of the environment variable.
So, the copy operation could exceed the allocation and lead to an OOB
write. Fix the issue by replacing grub_strcpy() with grub_strlcpy() and
pass the underlying buffers size to the grub_net_search_config_file().
Fixes: CVE-2025-0624
Reported-by: B Horn <b@horn.uk>
Signed-off-by: B Horn <b@horn.uk>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The grub_net_network_level_interface_unregister(), previously
implemented in a header, did not remove the variables hooks that
were registered in grub_net_network_level_interface_register().
Fix this by implementing the same logic used to register the
variables and move the function into the grub-core/net/net.c.
Signed-off-by: B Horn <b@horn.uk>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The net module is a dependency of normal. So, it shouldn't be possible
to unload the net. Though unregister variables hooks as a precaution.
It also gets in line with unregistering the other net module hooks.
Signed-off-by: B Horn <b@horn.uk>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
If unbounded recursion is allowed it becomes possible to collide the
stack with the heap. As UEFI firmware often lacks guard pages this
becomes an exploitable issue as it is possible in some cases to do
a controlled overwrite of a section of this heap region with
arbitrary data.
Reported-by: B Horn <b@horn.uk>
Signed-off-by: B Horn <b@horn.uk>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The part_iterate() is used by grub_partition_iterate() as a callback in
the partition iterate functions. However, part_iterate() may also call
the partition iterate functions which may lead to recursion. Fix potential
issue by limiting the recursion depth.
Signed-off-by: B Horn <b@horn.uk>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The grub_disk_read() may trigger other disk reads, e.g. via loopbacks.
This may lead to very deep recursion which can corrupt the heap. So, fix
the issue by limiting reads depth.
Reported-by: B Horn <b@horn.uk>
Signed-off-by: B Horn <b@horn.uk>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
It was possible to delete a loopback while there were still references
to it. This led to an exploitable use-after-free.
Fixed by implementing a reference counting in the grub_loopback struct.
Reported-by: B Horn <b@horn.uk>
Signed-off-by: B Horn <b@horn.uk>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The GRUB may use TPM to verify the integrity of boot components and the
result can determine whether a previously sealed key can be released. If
everything checks out, showing nothing has been tampered with, the key
is released and GRUB unlocks the encrypted root partition for the next
stage of booting.
However, the liberal Command Line Interface (CLI) can be misused by
anyone in this case to access files in the encrypted partition one way
or another. Despite efforts to keep the CLI secure by preventing utility
command output from leaking file content, many techniques in the wild
could still be used to exploit the CLI, enabling attacks or learning
methods to attack. It's nearly impossible to account for all scenarios
where a hack could be applied.
Therefore, to mitigate potential misuse of the CLI after the root device
has been successfully unlocked via TPM, the user should be required to
authenticate using the LUKS password. This added layer of security
ensures that only authorized users can access the CLI reducing the risk
of exploitation or unauthorized access to the encrypted partition.
Fixes: CVE-2024-49504
Signed-off-by: Michael Chang <mchang@suse.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The grub_file_open() and grub_file_close() should be the only places
that allow a reference to a filesystem to stay open. So, add grub_dl_t
to grub_fs_t and set this in the GRUB_MOD_INIT() for each filesystem to
avoid issues when filesystems forget to do it themselves or do not track
their own references, e.g. squash4.
The fs_label(), fs_uuid(), fs_mtime() and fs_read() should all ref and
unref in the same function but it is essentially redundant in GRUB
single threaded model.
Signed-off-by: B Horn <b@horn.uk>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
This is to avoid a generic issue were some filesystems would not set
data and also not set a grub_errno. This meant it was possible for many
filesystems to grub_dl_unref() themselves multiple times resulting in
it being possible to unload the filesystems while there were still
references to them, e.g., via a loopback.
Reported-by: B Horn <b@horn.uk>
Signed-off-by: B Horn <b@horn.uk>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
It was previously possible for grub_xfs_mount() to return NULL without
setting grub_errno if the XFS version was invalid. This resulted in it
being possible for grub_dl_unref() to be called twice allowing the XFS
module to be unloaded while there were still references to it.
Fixing this problem in general by ensuring a grub_errno is set if the
fail label is reached.
Reported-by: B Horn <b@horn.uk>
Signed-off-by: B Horn <b@horn.uk>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The number of records in the root key array read from disk was not being
validated against the size of the root node. This could lead to an
out-of-bounds read.
This patch adds a check to ensure that the number of records in the root
key array does not exceed the expected size of a root node read from
disk. If this check detects an out-of-bounds condition the operation is
aborted to prevent random errors due to metadata corruption.
Reported-by: Daniel Axtens <dja@axtens.net>
Signed-off-by: Michael Chang <mchang@suse.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
It was possible to read OOB when an attribute had a size that exceeded
the allocated buffer. This resolves that by making sure all attributes
that get read are fully in the allocated space by implementing
a function to validate them.
Defining the offsets in include/grub/ntfs.h but they are only used in
the validation function and not across the rest of the NTFS code.
Signed-off-by: B Horn <b@horn.uk>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Right now to access the next attribute the code reads the length of the
current attribute and adds that to the current pointer. This is error
prone as bounds checking needs to be performed all over the place. So,
implement a helper and ensure its used across find_attr() and read_attr().
This commit does *not* implement full bounds checking. It is just the
preparation work for this to be added into the helper.
Signed-off-by: B Horn <b@horn.uk>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The end of the attribute buffer should be stored alongside the rest of
the attribute struct as right now it is not possible to implement bounds
checking when accessing attributes sequentially.
This is done via:
- updating init_attr() to set at->end and check is is not initially out of bounds,
- implementing checks as init_attr() had its type change in its callers,
- updating the value of at->end when needed.
Signed-off-by: B Horn <b@horn.uk>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
When parsing NTFS file records the presence of the 0xFF marker indicates
the end of the attribute list. This value signifies that there are no
more attributes to process.
However, when the end marker is missing due to corrupted metadata the
loop continues to read beyond the attribute list resulting in out-of-bounds
reads and potentially entering an infinite loop.
This patch adds a check to provide a stop condition for the loop ensuring
it stops at the end of the attribute list or at the end of the Master File
Table. This guards against out-of-bounds reads and prevents infinite loops.
Reported-by: Daniel Axtens <dja@axtens.net>
Signed-off-by: Michael Chang <mchang@suse.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
When inline extents are used, i.e. the extent tree depth equals zero,
a maximum of four entries can fit into the inode's data block. If the
extent header states a number of entries greater than four the current
ext2 implementation causes an out-of-bounds read. Fix this issue by
capping the number of extents to four when reading inline extents.
Reported-by: Daniel Axtens <dja@axtens.net>
Signed-off-by: Michael Chang <mchang@suse.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The getblk() returns a value of type grub_int64_t which is assigned to
iagblk and inoblk, both of type grub_uint64_t, in grub_jfs_read_inode()
via grub_jfs_blkno(). This patch fixes the type mismatch in the
functions. Additionally, the getblk() will return 0 instead of -1 on
failure cases. This change is safe because grub_errno is always set in
getblk() to indicate errors and it is later checked in the callers.
Signed-off-by: Lidong Chen <lidong.chen@oracle.com>
Reviewed-by: Alec Brown <alec.r.brown@oracle.com>
Reviewed-by: Ross Philipson <ross.philipson@oracle.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
An extent's logical offset and address are represented as a 40-bit value
split into two parts: the most significant 8 bits and the least
significant 32 bits. Currently the JFS code uses only the least
significant 32 bits value for offsets and addresses assuming the data
size will never exceed the 32-bit range. This approach ignores the most
significant 8 bits potentially leading to incorrect offsets and
addresses for larger values. The patch fixes it by incorporating the
most significant 8 bits into the calculation to get the full 40-bits
value for offsets and addresses.
https://jfs.sourceforge.net/project/pub/jfslayout.pdf
"off1,off2 is a 40-bit field, containing the logical offset of the first
block in the extent.
...
addr1,addr2 is a 40-bit field, containing the address of the extent."
Signed-off-by: Lidong Chen <lidong.chen@oracle.com>
Reviewed-by: Alec Brown <alec.r.brown@oracle.com>
Reviewed-by: Ross Philipson <ross.philipson@oracle.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
While fuzz testing JFS with ASAN enabled an OOB read was detected in
grub_jfs_opendir(). The issue occurred due to an invalid directory slot
index in the first entry of the sorted directory slot array in the inode
directory header. The fix ensures the slot index is validated before
accessing it. Given that an internal or a leaf node in a directory B+
tree is a 4 KiB in size and each directory slot is always 32 bytes, the
max number of slots in a node is 128. The validation ensures that the
slot index doesn't exceed this limit.
[1] https://jfs.sourceforge.net/project/pub/jfslayout.pdf
JFS will allocate 4K of disk space for an internal node of the B+ tree.
An internal node looks the same as a leaf node.
- page 10
Fixed number of Directory Slots depending on the size of the node. These are
the slots to be used for storing the directory slot array and the directory
entries or router entries. A directory slot is always 32 bytes.
...
A Directory Slot Array which is a sorted array of indices to the directory
slots that are currently in use.
...
An internal or a leaf node in the directory B+ tree is a 4K page.
- page 25
Signed-off-by: Lidong Chen <lidong.chen@oracle.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Reviewed-by: Alec Brown <alec.r.brown@oracle.com>
The JFS fuzzing revealed an OOB read in grub_jfs_getent(). The crash
was caused by an invalid leaf nodes count, diro->dirpage->header.count,
which was larger than the maximum number of leaf nodes allowed in an
inode. This fix is to ensure that the leaf nodes count is validated in
grub_jfs_opendir() before calling grub_jfs_getent().
On the occasion replace existing raw numbers with newly defined constant.
Signed-off-by: Lidong Chen <lidong.chen@oracle.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Reviewed-by: Alec Brown <alec.r.brown@oracle.com>
The ctx->filename can point to either a string literal or a dynamically
allocated string. The ctx->filename_alloc field is used to indicate the
type of allocation.
An issue has been identified where ctx->filename is reassigned to
a string literal in susp_iterate_dir() but ctx->filename_alloc is not
correctly handled. This oversight causes a memory leak and an invalid
free operation later.
The fix involves checking ctx->filename_alloc, freeing the allocated
string if necessary and clearing ctx->filename_alloc for string literals.
Reported-by: Daniel Axtens <dja@axtens.net>
Signed-off-by: Michael Chang <mchang@suse.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
It was possible for a grub_errno to not be set if mount of an ISO 9660
filesystem failed when set_rockridge() returned 0.
This isn't known to be exploitable as the other filesystems due to
filesystem helper checking the requested file type. Though fixing
as a precaution.
Reported-by: B Horn <b@horn.uk>
Signed-off-by: B Horn <b@horn.uk>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
It was possible for mount to fail but not set grub_errno. This led to
a possible double decrement of the module reference count if the NULL
page was mapped.
Fixing in general as a similar bug was fixed in commit 61b13c187
(fs/hfsplus: Set grub_errno to prevent NULL pointer access) and there
are likely more variants around.
Fixes: CVE-2024-45783
Reported-by: B Horn <b@horn.uk>
Signed-off-by: B Horn <b@horn.uk>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
It was previously possible for grub_errno to not be set when
grub_f2fs_mount() failed if nat_bitmap_ptr() returned NULL.
This issue is solved by ensuring a grub_errno is set in the fail case.
Reported-by: B Horn <b@horn.uk>
Signed-off-by: B Horn <b@horn.uk>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Both namesize and linksize are derived from hd.size, a 12-digit octal
number parsed by read_number(). Later direct arithmetic calculation like
"namesize + 1" and "linksize + 1" may exceed the maximum value of
grub_size_t leading to heap OOB write. This patch fixes the issue by
using grub_add() and checking for an overflow.
Fixes: CVE-2024-45780
Reported-by: Nils Langius <nils@langius.de>
Signed-off-by: Lidong Chen <lidong.chen@oracle.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Reviewed-by: Alec Brown <alec.r.brown@oracle.com>
It was possible to iterate through grub_cpio_find_file() without
allocating name and not setting mode to GRUB_ARCHELP_ATTR_END, which
would cause the uninitialized value for name to be used as an argument
for canonicalize() in grub_archelp_dir().
Reported-by: B Horn <b@horn.uk>
Signed-off-by: B Horn <b@horn.uk>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
grub_strcpy() was used to copy a symlink name from the filesystem
image to a heap allocated buffer. This led to a OOB write to adjacent
heap allocations. Fix by using grub_strlcpy().
Fixes: CVE-2024-45781
Reported-by: B Horn <b@horn.uk>
Signed-off-by: B Horn <b@horn.uk>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
grub_strlcpy() acts the same way as strlcpy() does on most *NIX,
returning the length of src and ensuring dest is always NUL
terminated except when size is 0.
Signed-off-by: B Horn <b@horn.uk>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Follow recent extensions of EFI support providing a TCG2 driver with
a public API for getting the maximum TPM command size and passing a TPM
command through to the TPM 2. Implement this functionality using ieee1275
PowerPC firmware API calls. Move tcg2.c into the TCG2 driver.
Signed-off-by: Stefan Berger <stefanb@linux.ibm.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Move tpm_get_tpm_version() into grub_ieee1275_tpm_init() and invalidate
grub_ieee1275_tpm_ihandle in case no TPM 2 could be detected. Try the
initialization only once so that grub_tpm_present() will always return
the same result. Use the grub_ieee1275_tpm_ihandle as indicator for an
available TPM instead of grub_ieee1275_tpm_version, which can now be
removed.
Signed-off-by: Stefan Berger <stefanb@linux.ibm.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Move common initialization functions from the ibmvtpm driver module into
tcg2.c that will be moved into the new TCG2 driver in a subsequent patch.
Make the functions available to the ibmvtpm driver as public functions
and variables.
Signed-off-by: Stefan Berger <stefanb@linux.ibm.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Consolidate repeated definitions of IEEE1275_IHANDLE_INVALID that are cast
to the type grub_ieee1275_ihandle_t. On the occasion add "GRUB_" prefix to
the constant name.
Signed-off-by: Stefan Berger <stefanb@linux.ibm.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Cast 0 to proper type grub_ieee1275_ihandle_t. This type is
used for struct grub_serial_port's handle that assigns or
compares with IEEE1275_IHANDLE_INVALID.
Signed-off-by: Stefan Berger <stefanb@linux.ibm.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The TPM bit fields need to be in reverse order for big endian targets,
such as ieee1275 PowerPC platforms that run GRUB in big endian mode.
Signed-off-by: Stefan Berger <stefanb@linux.ibm.com>
Reviewed-by: Gary Lin <glin@suse.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Update the user manual to address TPM2 key protector including the two
related commands, tpm2_key_protector_init and tpm2_key_protector_clear,
and the user-space utility: grub-protect.
Signed-off-by: Gary Lin <glin@suse.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Reviewed-by: Stefan Berger <stefanb@linux.ibm.com>
Tested-by: Stefan Berger <stefanb@linux.ibm.com>
For the tpm2_key_protector module, the TCG2 command submission function
is the only difference between a QEMU instance and grub-emu. To test
TPM2 key unsealing with a QEMU instance, it requires an extra OS image
to invoke grub-protect to seal the LUKS key, rather than a simple
grub-shell rescue CD image. On the other hand, grub-emu can share the
emulated TPM2 device with the host, so that we can seal the LUKS key on
host and test key unsealing with grub-emu.
This test script firstly creates a simple LUKS image to be loaded as a
loopback device in grub-emu. Then an emulated TPM2 device is created by
"swtpm chardev" and PCR 0 and 1 are extended.
There are several test cases in the script to test various settings. Each
test case uses grub-protect or tpm2-tools to seal the LUKS password
with PCR 0 and PCR 1. Then grub-emu is launched to load the LUKS image,
try to mount the image with tpm2_key_protector_init and cryptomount, and
verify the result.
Based on the idea from Michael Chang.
Cc: Michael Chang <mchang@suse.com>
Cc: Stefan Berger <stefanb@linux.ibm.com>
Cc: Glenn Washburn <development@efficientek.com>
Signed-off-by: Gary Lin <glin@suse.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Reviewed-by: Stefan Berger <stefanb@linux.ibm.com>
Tested-by: Stefan Berger <stefanb@linux.ibm.com>
As a preparation to test tpm2_key_protector with grub-emu, the new
option, --tpm-device, is introduced to specify the TPM device for
grub-emu so that grub-emu can access an emulated TPM device from
the host.
Since grub-emu can directly access the device on host, it's easy to
implement the essential TCG2 command submission function with the
read/write functions and enable tpm2_key_protector module for grub-emu,
so that we can further test TPM2 key unsealing with grub-emu.
Signed-off-by: Gary Lin <glin@suse.com>
Reviewed-by: Stefan Berger <stefanb@linux.ibm.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Tested-by: Stefan Berger <stefanb@linux.ibm.com>
When using disk auto-unlocking with TPM 2.0, the typical grub.cfg may
look like this:
tpm2_key_protector_init --tpm2key=(hd0,gpt1)/boot/grub/sealed.tpm
cryptomount -u <PART-UUID> -P tpm2
search --fs-uuid --set=root <FS-UUID>
Since the disk search order is based on the order of module loading, the
attacker could insert a malicious disk with the same FS-UUID root to
trick GRUB to boot into the malicious root and further dump memory to
steal the unsealed key.
Do defend against such an attack, we can specify the hint provided by
"grub-probe" to search the encrypted partition first:
search --fs-uuid --set=root --hint='cryptouuid/<PART-UUID>' <FS-UUID>
However, for LVM on an encrypted partition, the search hint provided by
"grub-probe" is:
--hint='lvmid/<VG-UUID>/<LV-UUID>'
It doesn't guarantee to look up the logical volume from the encrypted
partition, so the attacker may have the chance to fool GRUB to boot
into the malicious disk.
To minimize the attack surface, this commit tweaks the disk device search
in diskfilter to look up cryptodisk devices first and then others, so
that the auto-unlocked disk will be found first, not the attacker's disk.
Cc: Fabian Vogt <fvogt@suse.com>
Signed-off-by: Gary Lin <glin@suse.com>
Reviewed-by: Stefan Berger <stefanb@linux.ibm.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Tested-by: Stefan Berger <stefanb@linux.ibm.com>
An attacker may insert a malicious disk with the same crypto UUID and
trick GRUB to mount the fake root. Even though the key from the key
protector fails to unlock the fake root, it's not wiped out cleanly so
the attacker could dump the memory to retrieve the secret key. To defend
such attack, wipe out the cached key when we don't need it.
Cc: Fabian Vogt <fvogt@suse.com>
Signed-off-by: Gary Lin <glin@suse.com>
Reviewed-by: Stefan Berger <stefanb@linux.ibm.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Tested-by: Stefan Berger <stefanb@linux.ibm.com>
If a protector is specified, but it fails to unlock the disk, fall back
to asking for the passphrase.
Before requesting the passphrase, the error from the key protector(s)
has to be cleared, or the later code, e.g., LUKS code, may stop as
grub_errno is set. This commit prints error from the key protector(s)
and sets grub_errno to GRUB_ERR_NONE to have a fresh start.
Signed-off-by: Patrick Colp <patrick.colp@oracle.com>
Signed-off-by: Gary Lin <glin@suse.com>
Reviewed-by: Stefan Berger <stefanb@linux.ibm.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Tested-by: Stefan Berger <stefanb@linux.ibm.com>
Currently with the TPM2 protector, only SRK mode is supported and
NV index support is just a stub. Implement the NV index option.
Note: This only extends support on the unseal path. grub-protect
has not been updated. tpm2-tools can be used to insert a key into
the NV index.
An example of inserting a key using tpm2-tools:
# Get random key.
tpm2_getrandom 32 > key.dat
# Create primary object.
tpm2_createprimary -C o -g sha256 -G ecc -c primary.ctx
# Create policy object. `pcrs.dat` contains the PCR values to seal against.
tpm2_startauthsession -S session.dat
tpm2_policypcr -S session.dat -l sha256:7,11 -f pcrs.dat -L policy.dat
tpm2_flushcontext session.dat
# Seal key into TPM.
cat key.dat | tpm2_create -C primary.ctx -u key.pub -r key.priv -L policy.dat -i-
tpm2_load -C primary.ctx -u key.pub -r key.priv -n sealing.name -c sealing.ctx
tpm2_evictcontrol -C o -c sealing.ctx 0x81000000
Then to unseal the key in GRUB, add this to grub.cfg:
tpm2_key_protector_init --mode=nv --nvindex=0x81000000 --pcrs=7,11
cryptomount -u <UUID> --protector tpm2
Signed-off-by: Patrick Colp <patrick.colp@oracle.com>
Signed-off-by: Gary Lin <glin@suse.com>
Reviewed-by: Stefan Berger <stefanb@linux.ibm.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Tested-by: Stefan Berger <stefanb@linux.ibm.com>
This commit handles the TPM2_PolicyAuthorize command from the key file
in TPM 2.0 Key File format.
TPM2_PolicyAuthorize is the essential command to support authorized
policy which allows the users to sign TPM policies with their own keys.
Per TPM 2.0 Key File [1], CommandPolicy for TPM2_PolicyAuthorize
comprises "TPM2B_PUBLIC pubkey", "TPM2B_DIGEST policy_ref", and
"TPMT_SIGNATURE signature". To verify the signature, the current policy
digest is hashed with the hash algorithm written in "signature", and then
"signature" is verified with the hashed policy digest and "pubkey". Once
TPM accepts "signature", TPM2_PolicyAuthorize is invoked to authorize the
signed policy.
To create the key file with authorized policy, here are the pcr-oracle [2]
commands:
# Generate the RSA key and create the authorized policy file
$ pcr-oracle \
--rsa-generate-key \
--private-key policy-key.pem \
--auth authorized.policy \
create-authorized-policy 0,2,4,7,9
# Seal the secret with the authorized policy
$ pcr-oracle \
--key-format tpm2.0 \
--auth authorized.policy \
--input disk-secret.txt \
--output sealed.key \
seal-secret
# Sign the predicted PCR policy
$ pcr-oracle \
--key-format tpm2.0 \
--private-key policy-key.pem \
--from eventlog \
--stop-event "grub-file=grub.cfg" \
--after \
--input sealed.key \
--output /boot/efi/efi/grub/sealed.tpm \
sign 0,2,4,7,9
Then specify the key file and the key protector to grub.cfg in the EFI
system partition:
tpm2_key_protector_init -a RSA --tpm2key=(hd0,gpt1)/efi/grub/sealed.tpm
cryptomount -u <PART_UUID> -P tpm2
For any change in the boot components, just run the "sign" command again
to update the signature in sealed.tpm, and TPM can unseal the key file
with the updated PCR policy.
[1] https://www.hansenpartnership.com/draft-bottomley-tpm2-keys.html
[2] https://github.com/okirch/pcr-oracle
Signed-off-by: Gary Lin <glin@suse.com>
Reviewed-by: Stefan Berger <stefanb@linux.ibm.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Tested-by: Stefan Berger <stefanb@linux.ibm.com>
To utilize the key protectors framework, there must be a way to protect
full-disk encryption keys in the first place. The grub-protect tool
includes support for the TPM2 key protector but other protectors that
require setup ahead of time can be supported in the future.
For the TPM2 key protector, the intended flow is for a user to have
a LUKS 1 or LUKS 2-protected fully-encrypted disk. The user then creates
a new LUKS key file, say by reading /dev/urandom into a file, and creates
a new LUKS key slot for this key. Then, the user invokes the grub-protect
tool to seal this key file to a set of PCRs using the system's TPM 2.0.
The resulting sealed key file is stored in an unencrypted partition such
as the EFI System Partition (ESP) so that GRUB may read it. The user also
has to ensure the cryptomount command is included in GRUB's boot script
and that it carries the requisite key protector (-P) parameter.
Sample usage:
$ dd if=/dev/urandom of=luks-key bs=1 count=32
$ sudo cryptsetup luksAddKey /dev/sdb1 luks-key --pbkdf=pbkdf2 --hash=sha512
To seal the key with TPM 2.0 Key File (recommended):
$ sudo grub-protect --action=add \
--protector=tpm2 \
--tpm2-pcrs=0,2,4,7,9 \
--tpm2key \
--tpm2-keyfile=luks-key \
--tpm2-outfile=/boot/efi/efi/grub/sealed.tpm
Or, to seal the key with the raw sealed key:
$ sudo grub-protect --action=add \
--protector=tpm2 \
--tpm2-pcrs=0,2,4,7,9 \
--tpm2-keyfile=luks-key \
--tpm2-outfile=/boot/efi/efi/grub/sealed.key
Then, in the boot script, for TPM 2.0 Key File:
tpm2_key_protector_init --tpm2key=(hd0,gpt1)/efi/grub/sealed.tpm
cryptomount -u <SDB1_UUID> -P tpm2
Or, for the raw sealed key:
tpm2_key_protector_init --keyfile=(hd0,gpt1)/efi/grub/sealed.key --pcrs=0,2,4,7,9
cryptomount -u <SDB1_UUID> -P tpm2
The benefit of using TPM 2.0 Key File is that the PCR set is already
written in the key file, so there is no need to specify PCRs when
invoking tpm2_key_protector_init.
Signed-off-by: Hernan Gatta <hegatta@linux.microsoft.com>
Signed-off-by: Gary Lin <glin@suse.com>
Reviewed-by: Stefan Berger <stefanb@linux.ibm.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Tested-by: Stefan Berger <stefanb@linux.ibm.com>
Add a new parameter to cryptomount to support the key protectors framework: -P.
The parameter is used to automatically retrieve a key from specified key
protectors. The parameter may be repeated to specify any number of key
protectors. These are tried in order until one provides a usable key for any
given disk.
Signed-off-by: Hernan Gatta <hegatta@linux.microsoft.com>
Signed-off-by: Michael Chang <mchang@suse.com>
Signed-off-by: Gary Lin <glin@suse.com>
Reviewed-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Stefan Berger <stefanb@linux.ibm.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Tested-by: Stefan Berger <stefanb@linux.ibm.com>
The TPM2 key protector is a module that enables the automatic retrieval
of a fully-encrypted disk's unlocking key from a TPM 2.0.
The theory of operation is such that the module accepts various
arguments, most of which are optional and therefore possess reasonable
defaults. One of these arguments is the keyfile/tpm2key parameter, which
is mandatory. There are two supported key formats:
1. Raw Sealed Key (--keyfile)
When sealing a key with TPM2_Create, the public portion of the sealed
key is stored in TPM2B_PUBLIC, and the private portion is in
TPM2B_PRIVATE. The raw sealed key glues the fully marshalled
TPM2B_PUBLIC and TPM2B_PRIVATE into one file.
2. TPM 2.0 Key (--tpm2key)
The following is the ASN.1 definition of TPM 2.0 Key File:
TPMPolicy ::= SEQUENCE {
CommandCode [0] EXPLICIT INTEGER
CommandPolicy [1] EXPLICIT OCTET STRING
}
TPMAuthPolicy ::= SEQUENCE {
Name [0] EXPLICIT UTF8STRING OPTIONAL
Policy [1] EXPLICIT SEQUENCE OF TPMPolicy
}
TPMKey ::= SEQUENCE {
type OBJECT IDENTIFIER
emptyAuth [0] EXPLICIT BOOLEAN OPTIONAL
policy [1] EXPLICIT SEQUENCE OF TPMPolicy OPTIONAL
secret [2] EXPLICIT OCTET STRING OPTIONAL
authPolicy [3] EXPLICIT SEQUENCE OF TPMAuthPolicy OPTIONAL
description [4] EXPLICIT UTF8String OPTIONAL,
rsaParent [5] EXPLICIT BOOLEAN OPTIONAL,
parent INTEGER
pubkey OCTET STRING
privkey OCTET STRING
}
The TPM2 key protector only expects a "sealed" key in DER encoding,
so "type" is always 2.23.133.10.1.5, "emptyAuth" is "TRUE", and
"secret" is empty. "policy" and "authPolicy" are the possible policy
command sequences to construct the policy digest to unseal the key.
Similar to the raw sealed key, the public portion (TPM2B_PUBLIC) of
the sealed key is stored in "pubkey", and the private portion
(TPM2B_PRIVATE) is in "privkey".
For more details: https://www.hansenpartnership.com/draft-bottomley-tpm2-keys.html
This sealed key file is created via the grub-protect tool. The tool
utilizes the TPM's sealing functionality to seal (i.e., encrypt) an
unlocking key using a Storage Root Key (SRK) to the values of various
Platform Configuration Registers (PCRs). These PCRs reflect the state
of the system as it boots. If the values are as expected, the system
may be considered trustworthy, at which point the TPM allows for a
caller to utilize the private component of the SRK to unseal (i.e.,
decrypt) the sealed key file. The caller, in this case, is this key
protector.
The TPM2 key protector registers two commands:
- tpm2_key_protector_init: Initializes the state of the TPM2 key
protector for later usage, clearing any
previous state, too, if any.
- tpm2_key_protector_clear: Clears any state set by tpm2_key_protector_init.
The way this is expected to be used requires the user to, either
interactively or, normally, via a boot script, initialize/configure
the key protector and then specify that it be used by the "cryptomount"
command (modifications to this command are in a different patch).
For instance, to unseal the raw sealed key file:
tpm2_key_protector_init --keyfile=(hd0,gpt1)/efi/grub/sealed-1.key
cryptomount -u <PART1_UUID> -P tpm2
tpm2_key_protector_init --keyfile=(hd0,gpt1)/efi/grub/sealed-2.key --pcrs=7,11
cryptomount -u <PART2_UUID> -P tpm2
Or, to unseal the TPM 2.0 Key file:
tpm2_key_protector_init --tpm2key=(hd0,gpt1)/efi/grub/sealed-1.tpm
cryptomount -u <PART1_UUID> -P tpm2
tpm2_key_protector_init --tpm2key=(hd0,gpt1)/efi/grub/sealed-2.tpm --pcrs=7,11
cryptomount -u <PART2_UUID> -P tpm2
If a user does not initialize the key protector and attempts to use it
anyway, the protector returns an error.
Before unsealing the key, the TPM2 key protector follows the "TPMPolicy"
sequences to enforce the TPM policy commands to construct a valid policy
digest to unseal the key.
For the TPM 2.0 Key files, "authPolicy" may contain multiple "TPMPolicy"
sequences, the TPM2 key protector iterates "authPolicy" to find a valid
sequence to unseal key. If "authPolicy" is empty or all sequences in
"authPolicy" fail, the protector tries the one from "policy". In case
"policy" is also empty, the protector creates a "TPMPolicy" sequence
based on the given PCR selection.
For the raw sealed key, the TPM2 key protector treats the key file as a
TPM 2.0 Key file without "authPolicy" and "policy", so the "TPMPolicy"
sequence is always based on the PCR selection from the command
parameters.
This commit only supports one policy command: TPM2_PolicyPCR. The
command set will be extended to support advanced features, such as
authorized policy, in the later commits.
Cc: James Bottomley <jejb@linux.ibm.com>
Signed-off-by: Hernan Gatta <hegatta@linux.microsoft.com>
Signed-off-by: Gary Lin <glin@suse.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Reviewed-by: Stefan Berger <stefanb@linux.ibm.com>
Tested-by: Stefan Berger <stefanb@linux.ibm.com>
A Trusted Platform Module (TPM) Software Stack (TSS) provides logic to
compose and submit TPM commands and parse responses.
A limited number of TPM commands may be accessed via the EFI TCG2
protocol. This protocol exposes functionality that is primarily geared
toward TPM usage within the context of Secure Boot. For all other TPM
commands, however, such as sealing and unsealing, this protocol does not
provide any help, with the exception of passthrough command submission.
The SubmitCommand method allows a caller to send raw commands to the
system's TPM and to receive the corresponding response. These
command/response pairs are formatted using the TPM wire protocol. To
construct commands in this way, and to parse the TPM's response, it is
necessary to, first, possess knowledge of the various TPM structures, and,
second, of the TPM wire protocol itself.
As such, this patch includes implementations of various grub_tpm2_* functions
(inventoried below), and logic to write and read command and response
buffers, respectively, using the TPM wire protocol.
Functions:
- grub_tpm2_create(),
- grub_tpm2_createprimary(),
- grub_tpm2_evictcontrol(),
- grub_tpm2_flushcontext(),
- grub_tpm2_load(),
- grub_tpm2_pcr_read(),
- grub_tpm2_policygetdigest(),
- grub_tpm2_policypcr(),
- grub_tpm2_readpublic(),
- grub_tpm2_startauthsession(),
- grub_tpm2_unseal(),
- grub_tpm2_loadexternal(),
- grub_tpm2_hash(),
- grub_tpm2_verifysignature(),
- grub_tpm2_policyauthorize(),
- grub_tpm2_testparms().
Signed-off-by: Hernan Gatta <hegatta@linux.microsoft.com>
Signed-off-by: Gary Lin <glin@suse.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Reviewed-by: Stefan Berger <stefanb@linux.ibm.com>
Tested-by: Stefan Berger <stefanb@linux.ibm.com>
This commit adds the necessary TPM2 types and structs as the preparation
for the TPM2 Software Stack (TSS2) support. The Marshal/Unmarshal
functions are also added to handle the data structure to be submitted to
TPM2 commands and to be received from the response.
Signed-off-by: Hernan Gatta <hegatta@linux.microsoft.com>
Signed-off-by: Gary Lin <glin@suse.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Reviewed-by: Stefan Berger <stefanb@linux.ibm.com>
Tested-by: Stefan Berger <stefanb@linux.ibm.com>
As the preparation to support TPM2 Software Stack (TSS2), this commit
implements the TPM2 buffer handling functions to pack data for the TPM2
commands and unpack the data from the response.
Cc: Stefan Berger <stefanb@linux.ibm.com>
Signed-off-by: Hernan Gatta <hegatta@linux.microsoft.com>
Signed-off-by: Gary Lin <glin@suse.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Tested-by: Stefan Berger <stefanb@linux.ibm.com>
A key protector encapsulates functionality to retrieve an unlocking key
for a fully-encrypted disk from a specific source. A key protector
module registers itself with the key protectors framework when it is
loaded and unregisters when unloaded. Additionally, a key protector may
accept parameters that describe how it should operate.
The key protectors framework, besides offering registration and
unregistration functions, also offers a one-stop routine for finding and
invoking a key protector by name. If a key protector with the specified
name exists and if an unlocking key is successfully retrieved by it, the
function returns to the caller the retrieved key and its length.
Cc: Vladimir Serbinenko <phcoder@gmail.com>
Signed-off-by: Hernan Gatta <hegatta@linux.microsoft.com>
Signed-off-by: Gary Lin <glin@suse.com>
Reviewed-by: Stefan Berger <stefanb@linux.ibm.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Tested-by: Stefan Berger <stefanb@linux.ibm.com>
Document libtasn1 in docs/grub-dev.texi and add the upgrade steps.
Also add the patches to make libtasn1 compatible with GRUB code.
Signed-off-by: Gary Lin <glin@suse.com>
Reviewed-by: Vladimir Serbinenko <phcoder@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Tested-by: Stefan Berger <stefanb@linux.ibm.com>
Import tests from libtasn1 that use functionality we import.
This test module is integrated into functional_test so that the
user can run the test in GRUB shell.
This doesn't test the full decoder but that will be exercised in
test suites for coming patch sets.
Add testcase target in accordance with commit 5e10be48e5 (tests: Add
check-native and check-nonnative make targets).
Cc: Vladimir Serbinenko <phcoder@gmail.com>
Signed-off-by: Daniel Axtens <dja@axtens.net>
Signed-off-by: Gary Lin <glin@suse.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Tested-by: Stefan Berger <stefanb@linux.ibm.com>
Create a wrapper file that specifies the module license.
Set up the makefile so it is built.
Signed-off-by: Daniel Axtens <dja@axtens.net>
Signed-off-by: Gary Lin <glin@suse.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Tested-by: Stefan Berger <stefanb@linux.ibm.com>
There is a testcase to test the values larger than "int" but smaller
than "long". However, for some architectures, "long" and "int" are the
same and the compiler may issue a warning like this:
grub-core/tests/asn1/tests/Test_overflow.c:48:50: error: left shift of negative value [-Werror=shift-negative-value]
unsigned long num = ((long) GRUB_UINT_MAX) << 2;
^~
To avoid unnecessary error the testcase is enabled only when
GRUB_LONG_MAX is larger than GRUB_INT_MAX.
Signed-off-by: Daniel Axtens <dja@axtens.net>
Signed-off-by: Gary Lin <glin@suse.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Tested-by: Stefan Berger <stefanb@linux.ibm.com>
This commit replaces printf() and fprintf() with grub_printf() to print
the error messages for the testcases. Besides, asn1_strerror() is used
to convert the result code to strings instead of asn1_perror().
Signed-off-by: Daniel Axtens <dja@axtens.net>
Signed-off-by: Gary Lin <glin@suse.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Tested-by: Stefan Berger <stefanb@linux.ibm.com>
This commit removes the "verbose" variables and the unnecessary printf()
to simplify the output.
Signed-off-by: Daniel Axtens <dja@axtens.net>
Signed-off-by: Gary Lin <glin@suse.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Tested-by: Stefan Berger <stefanb@linux.ibm.com>
Some testcases use exit() to end the test. Since all the asn1 testcases
are invoked as functions, this commit replaces exit() with return to
reflect the test results, so that the main test function can check the
results.
Signed-off-by: Daniel Axtens <dja@axtens.net>
Signed-off-by: Gary Lin <glin@suse.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Tested-by: Stefan Berger <stefanb@linux.ibm.com>
This commit changes the main functions in the testcases to the test
names so that the real "main" test function can invokes them.
Signed-off-by: Daniel Axtens <dja@axtens.net>
Signed-off-by: Gary Lin <glin@suse.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Tested-by: Stefan Berger <stefanb@linux.ibm.com>
This commit removes all the headers and only uses asn1_test.h.
To avoid including int.h from grub-core/lib/libtasn1-grub/lib,
CONST_DOWN is defined in reproducers.c.
Signed-off-by: Daniel Axtens <dja@axtens.net>
Signed-off-by: Gary Lin <glin@suse.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Tested-by: Stefan Berger <stefanb@linux.ibm.com>
In _asn1_tag_der(), the first while loop for the long form may end up
with a "k" value with "ASN1_MAX_TAG_SIZE" and cause the buffer overrun
in the second while loop. This commit tweaks the conditional check to
avoid producing a too large "k".
This is a quick fix and may differ from the official upstream fix.
libtasn1 issue: https://gitlab.com/gnutls/libtasn1/-/issues/49
Signed-off-by: Gary Lin <glin@suse.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Tested-by: Stefan Berger <stefanb@linux.ibm.com>
Replace a 64-bit division with a call to grub_divmod64(), preventing
creation of __udivdi3() calls on 32-bit platforms.
Signed-off-by: Daniel Axtens <dja@axtens.net>
Signed-off-by: Gary Lin <glin@suse.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Tested-by: Stefan Berger <stefanb@linux.ibm.com>
Since libtasn1.h is the header to be included by users, including the
standard POSIX headers in libtasn1.h would force the user to add the
CFLAGS/CPPFLAGS for the POSIX headers.
This commit adjusts the header paths to use the grub headers instead of
the standard POSIX headers, so that users only need to include
libtasn1.h to use libtasn1 functions.
Signed-off-by: Daniel Axtens <dja@axtens.net>
Signed-off-by: Gary Lin <glin@suse.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Tested-by: Stefan Berger <stefanb@linux.ibm.com>
strcat() is not available in GRUB. This commit replaces strcat() and
_asn1_strcat() with the bounds-checking _asn1_str_cat().
Signed-off-by: Daniel Axtens <dja@axtens.net>
Signed-off-by: Gary Lin <glin@suse.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Tested-by: Stefan Berger <stefanb@linux.ibm.com>
strcat() is not available in GRUB. This commit replaces strcat() with
strcpy() in _asn1_str_cat() as the preparation to replace other strcat()
with the bounds-checking _asn1_str_cat().
Signed-off-by: Daniel Axtens <dja@axtens.net>
Signed-off-by: Gary Lin <glin@suse.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Tested-by: Stefan Berger <stefanb@linux.ibm.com>
We don't expect to be able to write ASN.1, only read it,
so we can disable some code.
Do that with #if 0/#endif, rather than deletion. This means
that the difference between upstream and GRUB is smaller,
which should make updating libtasn1 easier in the future.
With these exclusions we also avoid the need for minmax.h,
which is convenient because it means we don't have to
import it from gnulib.
Signed-off-by: Daniel Axtens <dja@axtens.net>
Signed-off-by: Gary Lin <glin@suse.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Tested-by: Stefan Berger <stefanb@linux.ibm.com>
Cc: Vladimir Serbinenko <phcoder@gmail.com>
Signed-off-by: Daniel Axtens <dja@axtens.net>
Signed-off-by: Gary Lin <glin@suse.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Tested-by: Stefan Berger <stefanb@linux.ibm.com>
Unlike files accessed via a normal file system, the file->read_hook() is
not honoured when using blocklist notation.
This means that when trying to use a dedicated, 1 KiB, raw partition
for the environment block and hence does something like
save_env --file=(hd0,gpt9)0+2 X Y Z
this fails with "sparse file not allowed", which is rather unexpected,
as I've explicitly said exactly which blocks should be used. Adding
a little debugging reveals that grub_file_size(file) is 1024 as expected,
but total_length is 0, simply because the callback was never invoked, so
blocklists is an empty list.
Fix that by honouring the ->read_hook() set by the caller, also when
a "file" is specified with blocklist notation.
Signed-off-by: Rasmus Villemoes <rasmus.villemoes@prevas.dk>
Reviewed-by: Vladimir Serbinenko <phcoder@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Correct the documentation for the grub.cfg searching via network that
will be done based on ethernet type, -01, which was missing, and a given
MAC address.
Fixes: https://savannah.gnu.org/bugs/?65152
Signed-off-by: Andrew Hamilton <adhamilt@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Correct documentation for UEFI secure boot to remove statement that
chainloader does not work with secure boot. This was fixed by the commit
6d05264 (kern/efi/sb: Add chainloaded image as shim's verifiable object).
Fixes: https://savannah.gnu.org/bugs/?62004
Signed-off-by: Andrew Hamilton <adhamilt@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Correct documentation for pxe_default_server, pxe_default_gatway and
pxe_blksize. Only pxe_default_server is actually used (alias for
net_default_server). So, capture this and remove the other two.
Fixes: https://savannah.gnu.org/bugs/?54480
Signed-off-by: Andrew Hamilton <adhamilt@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Multiboot modules that could not be read successfully, e.g. via network,
should not be added to the list of modules to forward to the operating
system that is to be booted subsequently.
This patch is necessary because even if a grub.cfg checks whether or not
a module was successfully downloaded, it is futile to retry a failed
download as the corrupted module will be forwarded either way.
Signed-off-by: Valentin Gehrke <valentin.gehrke@kernkonzept.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The SBAT metadata is read from CSV file and transformed into an ELF note
with the -s option.
Signed-off-by: Daniel Axtens <dja@axtens.net>
Signed-off-by: Sudhakar Kuppusamy <sudhakar@linux.ibm.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
In order to store the SBAT data we create a new ELF note. The string
".sbat", zero-padded to 4 byte alignment, shall be entered in the name
field. The string "SBAT"'s ASCII values, 0x53424154, should be entered
in the type field.
Signed-off-by: Daniel Axtens <dja@axtens.net>
Signed-off-by: Sudhakar Kuppusamy <sudhakar@linux.ibm.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Rename has been skipped by mistake in the original commit.
Fixes: 94649c026 (nx: Set page permissions for loaded modules)
Signed-off-by: Daniel Kiper <daniel.kiper@oracle.com>
Tested-by: Sudeep Holla <sudeep.holla@arm.com>
Reviewed-by: Ross Philipson <ross.philipson@oracle.com>
The calculation of the size of the table was incorrect (copy/pasta from
grub_acpi_rsdt_find_table() I assume...). The entries are 64-bit long.
This causes us to access beyond the end of the table which is causing
crashes during boot on some systems. Typically this is causing a crash
on VMWare when using UEFI and enabling serial autodetection, as
grub_acpi_find_table (GRUB_ACPI_SPCR_SIGNATURE);
will goes past the end of the table (the SPCR table doesn't exits).
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Reviewed-by: Vladimir Serbinenko <phcoder@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Reviewed-by: Ross Philipson <ross.philipson@oracle.com>
Tested-by: Renata Ravanelli <rravanel@redhat.com>
For NX the GRUB binary has to announce that it is compatible with the
NX feature. This implies that when loading the executable GRUB image
several attributes are true:
- the binary doesn't need an executable stack,
- the binary doesn't need sections to be both executable and writable,
- the binary knows how to use the EFI Memory Attributes Protocol on code
it is loading.
This patch:
- adds a definition for the PE DLL Characteristics flag GRUB_PE32_NX_COMPAT,
- changes grub-mkimage to set that flag.
Signed-off-by: Peter Jones <pjones@redhat.com>
Signed-off-by: Jan Setje-Eilers <jan.setjeeilers@oracle.com>
Signed-off-by: Mate Kukri <mate.kukri@canonical.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
For NX we need to set write and executable permissions on the sections
of GRUB modules when we load them. All allocatable sections are marked
readable. In addition:
- SHF_WRITE sections are marked as writable,
- and SHF_EXECINSTR sections are marked as executable.
Where relevant for the platform the tramp and GOT areas are marked non-writable.
Signed-off-by: Peter Jones <pjones@redhat.com>
Signed-off-by: Robbie Harwood <rharwood@redhat.com>
Signed-off-by: Laszlo Ersek <lersek@redhat.com>
Signed-off-by: Jan Setje-Eilers <jan.setjeeilers@oracle.com>
Signed-off-by: Mate Kukri <mate.kukri@canonical.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
For NX we need to set the page access permission attributes for write
and execute permissions. This patch adds two new primitives, grub_set_mem_attrs()
and grub_clear_mem_attrs(), and associated constants definitions used
for that purpose. For most platforms it adds a dummy implementation.
On EFI platforms it implements the primitives using the EFI Memory
Attribute Protocol, defined in UEFI 2.10 specification.
Signed-off-by: Peter Jones <pjones@redhat.com>
Signed-off-by: Jan Setje-Eilers <jan.setjeeilers@oracle.com>
Signed-off-by: Mate Kukri <mate.kukri@canonical.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Currently we load module sections at whatever alignment gcc+ld happened
to dump into the ELF section header which is often less then the page
size. Since NX protections are page based this alignment must be rounded
up to page size on platforms supporting NX protections. This patch
switches EFI platforms to load module sections at 4 KiB page-aligned
addresses. It then changes the allocation size computation and the
loader code in grub_dl_load_segments() to align the locations and sizes
up to these boundaries and fills any added padding with zeros. All of
this happens before relocations are applied, so the relocations factor
that in with no change.
Signed-off-by: Peter Jones <pjones@redhat.com>
Signed-off-by: Laszlo Ersek <lersek@redhat.com>
Signed-off-by: Jan Setje-Eilers <jan.setjeeilers@oracle.com>
Signed-off-by: Mate Kukri <mate.kukri@canonical.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Currently when loading GRUB modules we allocate space for all sections
including those without SHF_ALLOC set. We then copy the sections that
/do/ have SHF_ALLOC set into the allocated memory leaving some of our
allocation untouched forever. Additionally, on platforms with GOT fixups
and trampolines we currently compute alignment round-ups for the
sections and sections with sh_size = 0. This patch removes the extra
space from the allocation computation and makes the allocation
computation loop skip empty sections as the loading loop does.
Signed-off-by: Peter Jones <pjones@redhat.com>
Signed-off-by: Jan Setje-Eilers <jan.setjeeilers@oracle.com>
Signed-off-by: Mate Kukri <mate.kukri@canonical.com>
Reviewed-By: Vladimir Serbinenko <phcoder@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Currently GRUB modules built with Clang or GCC have several sections
which we don't actually need or support. We already have a list of
sections to skip in genmod.sh and this patch adds the following
sections to that list (as well as a few newlines):
- .note.gnu.property
- .llvm*
Note that the glob there won't work without a new enough linker but the
failure is just reversion to the status quo. So, that's not a big problem.
Signed-off-by: Peter Jones <pjones@redhat.com>
Signed-off-by: Jan Setje-Eilers <jan.setjeeilers@oracle.com>
Signed-off-by: Mate Kukri <mate.kukri@canonical.com>
Reviewed-By: Vladimir Serbinenko <phcoder@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Currently .module_license is set writable, that is, the section has the
SHF_WRITE flag set, in the module's ELF headers. This probably never
actually matters but it can't possibly be correct. The patch sets that
data as "const" which causes that flag not to be set.
Signed-off-by: Peter Jones <pjones@redhat.com>
Signed-off-by: Jan Setje-Eilers <jan.setjeeilers@oracle.com>
Signed-off-by: Mate Kukri <mate.kukri@canonical.com>
Reviewed-By: Vladimir Serbinenko <phcoder@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
This is an x86-specific thing and should be available globally.
Signed-off-by: Daniel Kiper <daniel.kiper@oracle.com>
Signed-off-by: Sergii Dmytruk <sergii.dmytruk@3mdeb.com>
Reviewed-by: Ross Philipson <ross.philipson@oracle.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
This fixes naming inconsistency that goes against coding style as well
as helps to avoid potential conflicts and confusion as this constant is
used in multiple places.
Signed-off-by: Daniel Kiper <daniel.kiper@oracle.com>
Signed-off-by: Sergii Dmytruk <sergii.dmytruk@3mdeb.com>
Reviewed-by: Ross Philipson <ross.philipson@oracle.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Currently rdmsr and wrmsr commands have own MSR support detection code.
This code is the same. So, it is duplicated. Additionally, this code
cannot be reused by others. Hence, extract this code to a function and
make it public. By the way, improve a code a bit.
Additionally, use GRUB_ERR_BAD_DEVICE instead of GRUB_ERR_BUG to signal
an error because errors encountered by this new routine are not bugs.
Signed-off-by: Daniel Kiper <daniel.kiper@oracle.com>
Signed-off-by: Sergii Dmytruk <sergii.dmytruk@3mdeb.com>
Reviewed-by: Ross Philipson <ross.philipson@oracle.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Use more obvious names which match corresponding instructions:
* grub_msr_read() => grub_rdmsr(),
* grub_msr_write() => grub_wrmsr().
Signed-off-by: Daniel Kiper <daniel.kiper@oracle.com>
Signed-off-by: Sergii Dmytruk <sergii.dmytruk@3mdeb.com>
Reviewed-by: Ross Philipson <ross.philipson@oracle.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
It does not make sense to have separate headers for individual static
functions. So, make one common place to store them.
Signed-off-by: Daniel Kiper <daniel.kiper@oracle.com>
Signed-off-by: Sergii Dmytruk <sergii.dmytruk@3mdeb.com>
Reviewed-by: Ross Philipson <ross.philipson@oracle.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The loopback image is configured to function as a disk by being mapped
as a block device. Instead of measuring the entire block device we
should focus on tracking the individual files accessed from it. For
example, we do not directly measure block devices like hd0 disk but the
files opened from it.
This method is important to avoid running out of memory since loopback
images can be very large. Trying to read and measure the whole image at
once could cause out of memory errors and disrupt the boot process.
Signed-off-by: Michael Chang <mchang@suse.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Similarly to the issue described in commit c52ae4057 (efinet: skip
virtual IPv4 and IPv6 devices during card enumeration) the UEFI PXE
driver creates additional VLAN child devices when a VLAN ID is
configured on a network interface associated with a physical NIC. These
virtual VLAN devices must be skipped during card enumeration to ensure
that the subsequent SNP exclusive open operation targets the correct
physical card instances. Otherwise packet transfer would fail.
A device path example with VLAN nodes:
/MAC(123456789ABC,0x1)/Vlan(20)/IPv4(0.0.0.0,0x0,DHCP,0.0.0.0,0.0.0.0,0.0.0.0)
Signed-off-by: Michael Chang <mchang@suse.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
A regression in GRUB 2.12 causes the GRUB screen to become cluttered
with artifacts from the previous screen whether it's the UEFI post UI,
UEFI shell or any graphical UI running before GRUB. This issue occurs
in situations like booting GRUB from the UEFI shell and going straight
to the rescue or command shell causing visual discomfort.
The regression was introduced by commit 2d7c3abd8 (efi/console: Do not
set text-mode until it is actually needed). To address the screen
flickering issue this commit suppresses the text-mode setting until the
first output is requested. Before text-mode is set any attempt to clear
the screen has no effect. This inactive period renders the clear screen
ineffective in early boot stages, potentially leaving leftover artifacts
that will clutter the GRUB console display, as there is no guarantee
there will always be a clear screen after the first output.
The issue is fixed by ensuring grub_console_cls() to work through lazy
mode-setting, while also avoiding screen clearing for the hidden menu
which the flicker-free patch aims to improve.
Fixes: 2d7c3abd8 (efi/console: Do not set text-mode until we actually need it)
Signed-off-by: Michael Chang <mchang@suse.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The cycle register is not guaranteed to count at constant frequency.
If it is counting at all depends on the state the performance monitoring
unit. Use the time register to measure time.
Signed-off-by: Heinrich Schuchardt <heinrich.schuchardt@canonical.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Avoid dangling pointer. Code should not be reached but better safe than sorry.
Signed-off-by: Frediano Ziglio <frediano.ziglio@cloud.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
They are single 64-bit values. Used in other assembly files too.
Signed-off-by: Frediano Ziglio <frediano.ziglio@cloud.com>
Reviewed-by: Vladimir Serbinenko <phcoder@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The instruction uses a 64-bit immediate.
Signed-off-by: Frediano Ziglio <frediano.ziglio@cloud.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The function called is grub_utf8_to_utf16().
Signed-off-by: Frediano Ziglio <frediano.ziglio@cloud.com>
Reviewed-by: Vladimir Serbinenko <phcoder@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Assembly code looks for modules at __bss_start. Make this position explicit
rather than matching BSS alignment and module alignment.
Signed-off-by: Vladimir Serbinenko <phcoder@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Module structure has natural alignment of 4. Respect it explicitly
rather than relying on the fact that _end is usually aligned.
Signed-off-by: Vladimir Serbinenko <phcoder@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Otherwise it breaks the decompressors for MIPS targets.
Signed-off-by: Vladimir Serbinenko <phcoder@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Without it compiler generates GPREL16 references which do not work
with our memory layout.
Signed-off-by: Vladimir Serbinenko <phcoder@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
BCJ is not available for all platforms hence arguments may end up unused.
Signed-off-by: Vladimir Serbinenko <phcoder@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Otherwise depending on compiler we end up with umoddi3 reference and
failed module dependency resolution.
Signed-off-by: Vladimir Serbinenko <phcoder@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Users have no reason to see this and it can break graphical boot.
Signed-off-by: Mate Kukri <mate.kukri@canonical.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Add documentation for all GRUB modules contained in the source code tree.
When possible, cross-references to additional detail on commands was added
from their corresponding module documentation. In addition, documentation
for the file command was added.
Signed-off-by: Andrew Hamilton <adhamilt@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The get_part_uuid() function made an assumption that the target GRUB
device is a partition device and accessed device->disk->partition
without checking for NULL. There are four situations where this
assumption is problematic:
1. The device is a net device instead of a disk.
2. The device is an abstraction device, like LVM, RAID, or CRYPTO, which
is mostly logical "disk" ((lvmid/<UUID>) and so on).
3. Firmware RAID may present the ESP to GRUB as an EFI disk (hd0) device
if it is contained within a Linux software RAID.
4. When booting from a CD-ROM, the ESP is a VFAT image indexed by the El
Torito boot catalog. The boot device is set to (cd0), corresponding
to the CD-ROM image mounted as an ISO 9660 filesystem.
As a result, get_part_uuid() could lead to a NULL pointer dereference
and trigger a synchronous exception during boot if the ESP falls into
one of these categories. This patch fixes the problem by adding the
necessary checks to handle cases where the ESP is not a partition device.
Additionally, to avoid disrupting the boot process, this patch relaxes
the severity of the errors in this context to non-critical. Errors will
be logged, but they will not prevent the boot process from continuing.
Fixes: e0fa7dc84 (bli: Add a module for the Boot Loader Interface)
Signed-off-by: Michael Chang <mchang@suse.com>
Reviewed-By: Oliver Steffen <osteffen@redhat.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
As reported by Victoriia Egorova in bug 65880, grub-mkrescue does not
verify that the expected argument of an option like -d or -k does really
exist in argv. So, check the loop counter before incrementing it inside
the loop which copies argv to argp_argv. Issue an error message similar
to what older versions of grub-mkrescue did with a missing argument,
e.g. 2.02.
Fixes: https://savannah.gnu.org/bugs/index.php?65880
Signed-off-by: Thomas Schmitt <scdbackup@gmx.net>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The fdtdump command allows dumping arbitrary device tree properties
and saving them to a variable similar to the smbios command.
This is useful in scripts where further actions such as selecting
a kernel or loading another device tree depend on the compatible
or model values of the device tree provided by the firmware.
For now only the root level properties of the dtb are exposed.
Signed-off-by: Tobias Heider <tobias.heider@canonical.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
First they're use macros so they can't be translated as-is.
Second there is no point in translating them as they're too technical.
Signed-off-by: Vladimir Serbinenko <phcoder@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Current code works only if package matches binary name transformation rules.
It's often true but is not guaranteed.
Fixes: https://savannah.gnu.org/bugs/?64410
Signed-off-by: Vladimir Serbinenko <phcoder@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Silently keeping entries in the list if the address matches, but the
page count doesn't is a bad idea, and can lead to double frees.
grub_efi_free_pages() have already freed parts of this block by this
point, and thus keeping the whole block in the list and freeing it again
at exit can lead to double frees.
Signed-off-by: Mate Kukri <mate.kukri@canonical.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
If the firmware happens to return 0 as an address of allocated pages,
grub_efi_allocate_pages_real() tries to allocate a new set of pages,
and then free the ones at address 0.
However at that point grub_efi_store_alloc() wasn't yet called, so
freeing the pages at 0 using grub_efi_free_pages() which calls
grub_efi_drop_alloc() isn't necessary, so let's call b->free_pages()
instead.
The call to grub_efi_drop_alloc() doesn't seem particularly harmful,
because it seems to do nothing if the allocation it is asked to drop
isn't on the list, but the call to it is obviously unnecessary here.
Signed-off-by: Mate Kukri <mate.kukri@canonical.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
If the map was too big for the initial allocation, it was freed and replaced
with a bigger one, but the free call still used the hard-coded size.
Seems like this wasn't hit for a long time, because most firmware maps
fit into 12K.
This bug was triggered on Project Mu firmware with a big memory map, and
results in the heap getting trashed and the firmware ASSERTING on
corrupted heap guard values when GRUB exits.
Signed-off-by: Mate Kukri <mate.kukri@canonical.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
mkfs.erofs with version < 1.6 does not support the -L option.
Let's detect the version of mkfs.erofs and skip the label tests
if it is not supported.
Suggested-by: Glenn Washburn <development@efficientek.com>
Signed-off-by: Yifan Zhao <zhaoyifan@sjtu.edu.cn>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The current Debian stable, now 12, has dropped the exfat-utils package
that the exfat filesystem test requires to run. There is an exfatprogs
package that replaces exfat-utils, though it is not a drop-in replacement
because mkfs.exfat has differing command line option names. Note, that
we're not yet switching to using the exfat kernel module because this
allows the testings on kernels that do not have the module.
Update mkfs.exfat usage to adhere to the different exfatprogs usage. Also,
the exfatprogs mkfs.exfat, following the exfat specification more closely,
only allows a maximum of 22 bytes of UTF-16 characters in the volume label
compared to 30 bytes from exfat-utils. So the exfat label test is updated
accordingly.
Update documentation to note that exfatprogs is now needed and also
exfat-fuse, which is needed do the fuse mount.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
When $detached_header was set 1, $luksdiskfile was set to the LUKS header
file path with "${detached_header:-$luksfile}" appended, which evaluates
to "1". Fix this by using two statements to set $luksdiskfile. The first
sets it to the header file if $detached_header is set, otherwise leave it
unset. The second statement sets it to itself if it is already set,
otherwise it is set to $luksfile.
Fixes: a7b540e6e (tests: Add cryptomount functional test)
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
First look for firmware files in the source directory and then, if not
found, look for them in locations where Debian installs them. Prefer to
use the unified firmware file and, if not found, use the pflash firmware
files split in to code and variables. By looking for files in the source
directory first, system firmware files can be overridden and it can be
ensured that the tests can be run regardless of the distro or where the
system firmware files are stored. If no firmware files are found, print
an error message and exit with error.
If a firmware VARS file is found, use it with snapshot mode enabled, which
makes the VARS writable to the virtual machine, but does not write back
the changes to the file. This allows using the readonly system VARS file
without copying it or using it in readonly mode, which causes the ARM
machine to fail. This also gives tests effectively their own ephemeral VARS
file that can be written to without causing side-effects for other tests.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
According to the OVMF whitepaper [1]:
IMPORTANT: Never pass OVMF.fd to qemu with the -bios option. That option
maps the firmware image as ROM into the guest's address space, and forces
OVMF to emulate non-volatile variables with a fallback driver that is
bound to have insufficient and confusing semantics.
Use the pflash interface instead. Currently the unified firmware file is
used, which contains both firmware code and variable sections. By enabling
snapshot on the pflash device, the firmware can be loaded in such a way
that variables can be written to without writing to the backing file.
Since pflash does no searching for firmware paths that are not absolute,
unlike the -bios option, also make firmware paths absolute. Additionally,
update the previous firmware paths or file names that did not correspond to
ones installed by Debian.
Use the q35 machine, instead of the default i440fx, for i386-efi because
the default machine type does not emulate a flash device, which is now
needed to load the firmware.
[1] http://www.linux-kvm.org/downloads/lersek/ovmf-whitepaper-c770f8c.txt
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Allow using GDB to debug a failing QEMU test. This output does not cause
issues for tests because it happens before the trim line, and so will be
ignored.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Type 0x01 was introduced with the ACPI DBGP table and type 0x12 was introduced
with the ACPI DBG2 table. Type 0x12 is used by the ACPI SPCR table on recent
AWS bare-metal instances (c6i/c7i). Also give each debug type a proper name.
Signed-off-by: Udo Steinberg <udo@hypervisor.org>
Reviewed-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Current code in some codepaths neither discards nor reports errors.
Properly surface the error.
While on it split 2 cases of unrelated variables both named err.
Signed-off-by: Vladimir Serbinenko <phcoder@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
When GRUB image is netbooted on ppc64le, the keyboard input exhibits
significant latency, reports even say that characters are processed
about once per second. This issue makes interactively trying to debug
a ppc64le config very difficult.
It seems that the latency is largely caused by a 200 ms timeout in the
idle event loop, during which the network card interface is consistently
polled for incoming packets. Often, no packets arrive during this
period, so the timeout nearly always expires, which blocks the response
to key inputs.
Furthermore, this 200 ms timeout might not need to be enforced at this
basic layer, considering that GRUB performs synchronous reads and its
timeout management is actually handled by higher layers, not directly in
the card instance. Additionally, the idle polling, which reacts to
unsolicited packets like ICMP and SLAAC, would be fine at a less frequent
polling interval, rather than needing a timeout for receiving a response.
For these reasons, we believe the timeout in get_card_packet() should be
effectively removed. According to test results, the delay has disappeared,
and it is now much easier to use interactively.
Signed-Off-by: Michael Chang <mchang@suse.com>
Tested-by: Tony Jones <tonyj@suse.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The measurements for confidential computing has been introduced in the
commit 4c76565b6 (efi/tpm: Add EFI_CC_MEASUREMENT_PROTOCOL support).
Recently the patch 30708dfe3 (tpm: Disable the tpm verifier if the TPM
device is not present) has been introduced to optimize the memory usage
when a TPM device is not available on platforms. This fix prevents the
tpm module to be loaded on confidential computing platforms, e.g. Intel
machines with TDX enabled, where the TPM device is not available.
In this patch, we propose to load the tpm module for this use case by
generalizing the tpm feature detection in order to cover CC platforms.
Basically, we do it by detecting the availability of the
EFI_CC_MEASUREMENT_PROTOCOL EFI protocol.
Fixes: https://savannah.gnu.org/bugs/?65821
Fixes: 30708dfe3 (tpm: Disable the tpm verifier if the TPM device is not present)
Signed-off-by: Hector Cao <hector.cao@canonical.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Reviewed-by: Kuppuswamy Sathyanarayanan <sathyanarayanan.kuppuswamy@linux.intel.com>
Allocate memory if needed, while saving the corresponding release
operation, reducing the amount of code and code complexity.
Signed-off-by: Tianjia Zhang <tianjia.zhang@linux.alibaba.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
This patch adds support for Radix, Xive and Radix_gtse in Options
vector5 which is required for KVM LPARs. KVM LPARs ONLY support
Radix and not the Hash. Not enabling Radix on any PowerVM KVM LPARs
will result in boot failure.
Signed-off-by: Avnish Chouhan <avnish@linux.ibm.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
We don't need any actual adjustments as we don't use the affected structures.
Signed-off-by: Vladimir Serbinenko <phcoder@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Without them, e.g., 0x80LL on 64-bit target is 32-bit byte-swapped to
0xffffffff80000000 instead of correct 0x80000000.
Signed-off-by: Vladimir Serbinenko <phcoder@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
ARRAY_SIZE() is the count of elements, but the element size is 4 bytes, so
this was only initing the first 1/4th of the table. Detected with valgrind.
This should only matter in error paths, and I've not been able to identify
any actual misbehaviour that results from reading in-bounds but uninited data.
Signed-off-by: Daniel Axtens <dja@axtens.net>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
This isn't intended to be a functional change, but it makes a lot of failures a lot
faster, which is extremely helpful for fuzzing.
Without this change, we keep trying and trying to read more bytes into our buffer,
never being able to (read always returns 0) and so we just return old buffer contents
over and over until the decompression process fails some other way.
Signed-off-by: Daniel Axtens <dja@axtens.net>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Add functionality to disable command line interface access and editing of GRUB
menu entries if GRUB image is built with --disable-cli.
Signed-off-by: Alec Brown <alec.r.brown@oracle.com>
Reviewed-by: Vladimir Serbinenko <phcoder@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The EROFS [1] is a lightweight read-only filesystem designed for performance
which has already been shipped in most Linux distributions as well as widely
used in several scenarios, such as Android system partitions, container
images and rootfs for embedded devices.
This patch brings in the EROFS uncompressed support. Now, it's possible to
boot directly through GRUB with an EROFS rootfs.
Support for the EROFS compressed files will be added later.
[1] https://erofs.docs.kernel.org
Signed-off-by: Yifan Zhao <zhaoyifan@sjtu.edu.cn>
Tested-by: Daniel Axtens <dja@axtens.net>
Signed-off-by: Gao Xiang <hsiangkao@linux.alibaba.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The following EROFS patch will use this helper to handle
ALIGN_UP() overflow.
Signed-off-by: Gao Xiang <hsiangkao@linux.alibaba.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
This is required if the pathname contains spaces or GRUB shell
metacharacters else the generated config file check will fail.
Signed-off-by: Pascal Hambourg <pascal@plouf.fr.eu.org>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
When handling a regular LVM volume, GRUB can fail with the message:
error: disk `lvmid/******-****-****-****-****-****-****/******-****-****-****-****-****-******' not found.
If the condition which triggers this exists, grub-probe will report the
error mentioned above. Similarly, the GRUB boot code will fail to detect
LVM volumes, resulting in a failure to boot off of LVM disks/partitions.
The condition can be created on any LVM VG by an LVM configuration change,
so any system with /boot on LVM can become unbootable at "any" time (after
any LVM configuration change).
The problem is caused by an incorrect computation of mda_end in disk/lvm.c,
when the metadata area wraps around. Apparently, this can start happening at
around 220 metadata changes to the VG.
Fixes: 879c4a834 (lvm: Fix two more potential data-dependent alloc overflows)
Fixes: https://savannah.gnu.org/bugs/?61620
Signed-off-by: Rogier <rogier777@gmail.com>
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Tested-By: Michael Chang <mchang@suse.com>
Give the user a chance to re-enter their cryptodisk passphrase after a typo,
rather than immediately failing (and likely dumping them into a GRUB shell).
By default, we allow 3 tries before giving up. A value in the
cryptodisk_passphrase_tries environment variable will override this default.
The user can give up early by entering an empty passphrase, just as they
could before this patch.
Signed-off-by: Forest <forestix@nom.one>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The test corpus for version-1 RAID generated an infinite recursion
in grub_partition_iterate() while attempting to read the superblock.
The reason for the issue was that the data region overlapped with
the superblock.
The infinite call loop looks like this:
grub_partition_iterate() -> partmap->iterate() ->
-> grub_disk_read() -> grub_disk_read_small() ->
-> grub_disk_read_small_real() -> grub_diskfilter_read() ->
-> read_lv() -> read_segment() -> grub_diskfilter_read_node() ->
-> grub_disk_read() -> grub_disk_read_small() -> ...
The fix adds checks for both the superblock region and the data
region when parsing the superblock metadata in grub_mdraid_detect().
Signed-off-by: Lidong Chen <lidong.chen@oracle.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The "ground truth" stack protector cookie value is kept in a global
variable, and loaded in every function prologue and epilogue to store
it into resp. compare it with the stack slot holding the cookie.
If the comparison fails, the program aborts, and this might occur
spuriously when the global variable changes values between the entry and
exit of a function. This implies that assigning the global variable at
boot should not involve any instrumented function calls, unless special
care is taken to ensure that the live call stack is synchronized, which
is non-trivial.
So avoid any function calls, including grub_memcpy(), which is
unnecessary given that the stack cookie is always a suitably aligned
variable of the native word size.
While at it, leave the last byte 0x0 to avoid inadvertent unbounded
strings on the stack.
Note that the use of __attribute__((optimize)) is described as
unsuitable for production use in the GCC documentation, so let's drop
this as well now that it is no longer needed.
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Add a new keyword, "depends", to the module definition syntax
used in Makefile.core.def. This allows specifying explicit module
dependencies together with the module definition.
Do not track the "extra_deps.lst" file in the repository anymore,
it is now auto-generated.
Make use of this new keyword in the bli module definition.
Signed-off-by: Oliver Steffen <osteffen@redhat.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Otherwise the GRUB cannot start due to missing symbols when stack
protector is enabled on EFI platforms.
Signed-off-by: Daniel Kiper <daniel.kiper@oracle.com>
Reviewed-by: Vladimir Serbinenko <phcoder@gmail.com>
_split_longopt() was the bash-completion private API and removed since
bash-completion 2.12. This commit initializes the bash-completion
general variables with _init_completion() to avoid the potential
"command not found" error.
Although bash-completion 2.12 introduces _comp_initialize() to deprecate
_init_completion(), _init_completion() is still chosen for the better
backward compatibility.
Signed-off-by: Gary Lin <glin@suse.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The CMOS actually exists on most EFI platforms and in some cases is used to
store useful data that makes it justifiable for GRUB to read/write it.
As for date and time keep using EFI API and not CMOS one.
Signed-off-by: Vladimir Serbinenko <phcoder@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
No alignment is guaranteed and in fact on my IA-64 SAPIC is aligned
to 4 bytes instead of 8 and causes a trap. It affects only rarely used
lsacpi command and so went unnoticed.
Signed-off-by: Vladimir Serbinenko <phcoder@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
When enabling gfxmenu and choosing to boot the Xen hypervisor from its
menu, an error occurred:
error: ../../grub-core/video/bitmap_scale.c:42:null src bitmap in grub_video_create_scaled.
The error is returned by grub_video_bitmap_create_scaled() when the
source pixmap is not there. The init_background() uses it to scale up
the background image so it can fully fit into the screen resolution.
However not all backgrounds are set by a image, i.e. the "desktop-image"
property of the theme file. Instead a color code may be used, for
example OpenSUSE's green background uses "desktop-color" property:
desktop-color: "#0D202F"
So it is absolutely fine to call init_background() without a raw pixmap
if color code is used. A missing check has to be added to ensure the
grub_errno will not be erroneously set and gets in the way of ensuing
boot process.
The reason it happens sporadically is due to grub_errno is reset to
GRUB_ERR_NONE in other places if a function's error return can be
ignored. In particular this hunk in grub_gfxmenu_create_box() does the
majority of the reset of grub_errno returned by init_background(), but
the path may not be always chosen.
grub_video_bitmap_load (&box->raw_pixmaps[i], path);
grub_free (path);
/* Ignore missing pixmaps. */
grub_errno = GRUB_ERR_NONE;
In any case, we cannot account on such random behavior and should only
return grub_errno if it is justified.
On the occasion move the grub_video_bitmap struct definition to the
beginning of the function.
Signed-off-by: Michael Chang <mchang@suse.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The directory extent list does not have to be a continuous list of data
blocks. When GRUB tries to read a non-existant member of the list,
grub_xfs_read_file() will return a block of zero'ed memory. Checking for
a zero'ed magic number is sufficient to skip this non-existant data block.
Prior to commit 07318ee7e (fs/xfs: Fix XFS directory extent parsing)
this was handled as a subtle side effect of reading the (non-existant)
tail data structure. Since the block was zero'ed the computation of the
number of directory entries in the block would return 0 as well.
Fixes: 07318ee7e (fs/xfs: Fix XFS directory extent parsing)
Fixes: https://bugzilla.redhat.com/show_bug.cgi?id=2254370
Signed-off-by: Jon DeVree <nuxi@vault24.org>
Reviewed-By: Vladimir Serbinenko <phcoder@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
We reinstated these functions before the 2.12 release with a warning
such that users upgrading to 2.12 who had custom scripts using them
would not get broken in the upgrade and agreed to remove them after
the 2.12 release. This removes them accordingly.
This reverts commit e7a831963 (templates: Reinstate unused version
comparison functions with warning).
Cc: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Cc: Daniel Kiper <daniel.kiper@oracle.com>
Signed-off-by: Julian Andres Klode <julian.klode@canonical.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
There are two system directories for bash-completion scripts. One is
/usr/share/bash-completion/completions/ and the other is
/etc/bash_completion.d/. The "etc" scripts are loaded in advance and
for backward compatibility while the "usr" scripts are loaded on demand.
To load scripts on demand it requires a corresponding script for every
command. So, the main bash-completion script is split into several
subscripts for different "grub-*" commands. To share the code the real
completion functions are still implemented in "grub" and each
subscript sources "grub" and invokes the corresponding function.
Signed-off-by: Gary Lin <glin@suse.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The 64-bit ABI defines ld.so to be /lib/ld-x86-64.so.1.
Signed-off-by: Samuel Thibault <samuel.thibault@ens-lyon.org>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The file_get_fs_options() takes a mach_msg_type_number_t, 32-bit,
not a size_t, 64-bit on 64-bit platforms.
Signed-off-by: Samuel Thibault <samuel.thibault@ens-lyon.org>
Reviewed-by: Vladimir Serbinenko <phcoder@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
In grub-core/loader/i386/multiboot_mbi.c, Coverity spotted redundant code where
the variable err was being set to GRUB_ERR_NONE and then being overwritten
later without being used. Since this is unnecessary, we can remove the code
that sets err to GRUB_ERR_NONE.
Fixes: CID 428877
Signed-off-by: Alec Brown <alec.r.brown@oracle.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
In grub-core/osdep/unix/getroot.c, Coverity spotted redundant code where the
double pointer os_dev was being set to 0 and then being overwritten later
without being used. Since this is unnecessary, we can remove the code that
sets os_dev to 0.
Fixes: CID 428875
Signed-off-by: Alec Brown <alec.r.brown@oracle.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
In grub-core/fs/jfs.c, Coverity spotted redundant code where the pointer diro
was being set to 0 and then being overwritten later without being used. Since
this is unnecessary, we can remove the code that sets diro to 0.
Fixes: CID 428876
Signed-off-by: Alec Brown <alec.r.brown@oracle.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
When adding/changing the password for the luks2 partition, cryptsetup
may reject the command due to the weak password. Since this is only for
testing, add "--force-password" to switch password quality check off to
avoid the unexpected failure.
Signed-off-by: Gary Lin <glin@suse.com>
Reviewed-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Fixes build failure due to the extra_deps.lst file not existing in the
tarball. Found while trying to package GRUB 2.12 for Gentoo.
make[3]: *** No rule to make target '/var/tmp/portage/sys-boot/grub-2.12/work/grub-2.12/grub-core/extra_deps.lst', needed by 'syminfo.lst'. Stop.
Fixes: 89fbe0cac (grub-core/Makefile.am: Make path to extra_deps.lst relative to $(top_srcdir)/grub-core)
Fixes: 154dcb1ae (build: Allow explicit module dependencies)
Signed-off-by: Oskari Pirhonen <xxc3ncoredxx@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Having randomly generated bytes in the binary output breaks reproducible
builds. Since build timestamps are usually the source of irreproducibility
there is a standard which defines an environment variable SOURCE_DATE_EPOCH
to be used when set for build timestamps. According to the standard [1], the
value of SOURCE_DATE_EPOCH is a base-10 integer of the number of seconds
since the UNIX epoch. Currently, this is a 10 digit number that fits into
32-bits, but will not shortly after the year 2100. So to be future-proof
only use the least significant 32-bits. On 64-bit architectures, where the
canary is also 64-bits, there is an extra 32-bits that can be filled to
provide more entropy. The first byte is NUL to filter out string buffer
overflow attacks and the remaining 24-bits are set to static random bytes.
[1] https://reproducible-builds.org/specs/source-date-epoch
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Generating the canary at build time allows the canary to be different for
every build which could limit the effectiveness of certain exploits.
Fallback to the statically generated random bytes if /dev/urandom is not
readable, e.g. Windows.
On 32-bit architectures, which use a 32-bit canary, reduce the canary to
4 bytes with one byte being NUL to filter out string buffer overflow attacks.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The canary, __stack_chk_guard, is in the BSS and so will get initialized to
zero if it is not explicitly initialized. If the UEFI firmware does not
support the RNG protocol, then the canary will not be randomized and will
be zero. This seems like a possibly easier value to write by an attacker.
Initialize canary to static random bytes, so that it is still random when
there is no RNG protocol. Set at least one byte to NUL to protect against
string buffer overflow attacks [1]. Code that writes NUL terminated strings
will terminate when a NUL is encountered in the input byte stream. So the
attacker will not be able to forge the canary by including it in the input
stream without terminating the string operation and thus limiting the
stack corruption.
[1] https://www.sans.org/blog/stack-canaries-gingerly-sidestepping-the-cage/
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
In grub-core/gfxmenu/gui_image.c, Coverity detected a double free in the
function load_image(). The function checks if self->bitmap and self->raw_bitmap
aren't NULL and then frees them. In the case self->bitmap and self->raw_bitmap
are the same, only self->raw_bitmap is freed which would also free the memory
used by self->bitmap. However, in this case self->bitmap isn't being set to NULL
which could lead to a double free later in the code. After self->raw_bitmap is
freed, it gets set to the variable bitmap. If this variable is NULL, the code
could have a path that would free self->bitmap a second time in the function
rescale_image().
Fixes: CID 292472
Signed-off-by: Alec Brown <alec.r.brown@oracle.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
According to the ACPI specification the XSDT Entry field contains an array
of 64-bit physical addresses which points to other DESCRIPTION_HEADERs. However,
the entry_ptr iterator is defined as a 32-bit pointer. It means each 64-bit
entry in the XSDT table is treated as two separate 32-bit entries then. Fix the
issue by using correct addresses sizes when processing RSDT and XSDT tables.
Signed-off-by: Qiumiao Zhang <zhangqiumiao1@huawei.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
NetBSD uses slightly different function names for the same functions.
Signed-off-by: Vladimir Serbinenko <phcoder@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
NetBSD gettext is older than the check but we don't actually need 0.18.3,
older one works fine. This is needed to make bootstrap work on NetBSD.
Signed-off-by: Vladimir Serbinenko <phcoder@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
There is some variance in how compiler treats sizeof() especially
on 32-bit platforms where it can be naturally either int or long.
Explicit cast solves the issue.
Signed-off-by: Vladimir Serbinenko <phcoder@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Without this build-time mkfont fails dynamic linking. This is not ideal
but improves the situation until a better solution is available.
Signed-off-by: Vladimir Serbinenko <phcoder@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
*BSD puts fonts in other places. Add them to the list.
Signed-off-by: Vladimir Serbinenko <phcoder@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The HAVE_LIBZFS is defined by libzfs test and hence conflicts with
manual definition. On NetBSD it ends up detecting zfs but not detecting
nvpair and creates confusion. Split them.
Signed-off-by: Vladimir Serbinenko <phcoder@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
It's not critical, -Werror on it is inappropriate. We don't want to
modify gnulib too much. This warning is pretty much irrelevant.
Signed-off-by: Vladimir Serbinenko <phcoder@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
It's not available on NetBSD outside of syslog. Using strerror() is more
reliable as we retrieve errno immediately rather than down the stack.
Signed-off-by: Vladimir Serbinenko <phcoder@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Wrong function and variable name cause a stupid compilation error on
NetBSD and OpenBSD. Only NetBSD and OpenBSD use this file. No other
platform is affected.
Additionally, define RAW_FLOPPY_MAJOR constant if it is missing.
Signed-off-by: Vladimir Serbinenko <phcoder@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
After recent change in blocklist types we have a type mismatch. Fixing it
requires a wrapper or large changes. I feel like wrapper makes more sense.
Without this patch we end up with a compilation problem and without wrapping
callback data is not passed properly anymore.
Signed-off-by: Vladimir Serbinenko <phcoder@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Variable e is set but never used. We can just remove it now.
Signed-off-by: Vladimir Serbinenko <phcoder@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
We don't really control the small aspects of generated files and NetBSD
version has an unused variable that is then detected by gcc as warning
that is then promoted to error.
Signed-off-by: Vladimir Serbinenko <phcoder@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Current code implicitly assumes that aligning chunk_size + *kern_end is
the same as aligning on curload which is not the case because
chunk_size starts at zero even if *kern_end is unaligned and ALIGN_PAGE
moved curload to an aligned position but not *kern_end + chunk_size.
This fixes booting of FreeBSD with zfs module.
Signed-off-by: Vladimir Serbinenko <phcoder@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The commit 154dcb1ae (build: Allow explicit module dependencies) broke
out of tree builds by introducing the extra_deps.lst file into the
source tree but referencing it just by name in grub-core/Makefile.am.
Fix it by adding $(top_srcdir)/grub-core to the path.
Fixes: 154dcb1ae (build: Allow explicit module dependencies)
Signed-off-by: Mate Kukri <mate.kukri@canonical.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The commit 3f9eace2d (util/grub-install: Delay copying files to
{grubdir,platdir} after install_device was validated) delaying
copying of files caused a regression when installing without an
existing directory structure.
This patch ensures that the platform directory actually exists by the
time the code tries to canonicalize its filename.
Fixes: 3f9eace2d (util/grub-install: Delay copying files to {grubdir,platdir} after install_device was validated)
Signed-off-by: Mate Kukri <mate.kukri@canonical.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The add_tar_files() function currently iterates through a directory's
content using readdir(), which doesn't guarantee a specific order. This
lack of deterministic behavior impacts reproducibility in the build process.
This commit resolves the issue by introducing sorting functionality.
The list retrieved by readdir() is now sorted alphabetically before
incorporation into the tar archive, ensuring consistent and predictable
file ordering within the archive.
On the occasion fix tfp memory leak.
Signed-off-by: Michael Chang <mchang@suse.com>
Signed-off-by: Bernhard Wiedemann <bwiedemann@suse.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
This change mirrors a previous fix [1] but is specific to images
generated by grub-mkstandalone.
The former fix, commit 85a7be241 (util/mkimage: Use stable timestamp
when generating binaries.), focused on utilizing a stable timestamp
during binary generation in the util/mkimage context. This commit
extends that approach to the images produced by grub-mkstandalone,
ensuring consistency and stability in timestamps across all generated
binaries.
[1] 85a7be241 util/mkimage: Use stable timestamp when generating binaries.
Signed-off-by: Michael Chang <mchang@suse.com>
Signed-off-by: Bernhard Wiedemann <bwiedemann@suse.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Replace definition of HTTP_PORT with a pre-processor macro that converts
the constant to the correct grub_uint16_t type.
Change "port" local variable definition in http_establish() to have the
same type.
Signed-off-by: Mate Kukri <mate.kukri@canonical.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com
Revert the commit a79c567f6 (templates: Remove unused version comparison
functions) and add a warning to the functions that they are deprecated.
Removing the functions directly caused a lot of upgrade issues
with custom user scripts that called the functions. In Debian and
Ubuntu, grub-mkconfig is invoked as a post-installation script
and would fail, causing upgrades to fail halfway through and
putting the package manager into an inconsistent state.
FWIW, we get one bug per 2 weeks basically, for an interim Ubuntu
release which generally does not receive much usage, that is a high
number.
The proposal is to pick this for 2.12 and directly after the release
remove it again. Then users will have time to fix their scripts without
systems breaking immediately.
This reverts commit a79c567f6 (templates: Remove unused version
comparison functions).
Cc: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Cc: Daniel Kiper <daniel.kiper@oracle.com>
Signed-off-by: Julian Andres Klode <julian.klode@canonical.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Previously grub-install copied modules to grubdir before doing any
validation on the install_device.
When grub-install was called with an invalid install_device, modules
were already copied to /boot before it found out and was forced to rely
on atexit() rollback.
This patch delays copying the modules after at least some install_device
validation was done, and thus reduces reliance on successful rollback.
Signed-off-by: Mate Kukri <mate.kukri@canonical.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
If validation has been disabled via MokSbState, secure boot on the
firmware is still enabled, and the kernel fails to boot.
This is a bit hacky, because shim_lock is not *fully* enabled, but
it triggers the right code paths.
Ultimately, all this will be resolved by shim gaining it's own image
loading and starting protocol, so this is more a temporary workaround.
Fixes: 6425c12cd (efi: Fallback to legacy mode if shim is loaded on x86 archs)
Cc: Peter Jones <pjones@redhat.com>
Cc: Michael Chang <mchang@suse.com>
Signed-off-by: Julian Andres Klode <julian.klode@canonical.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Improve the documentation of the bli module and explain in more detail what
it does. Make clear that GPT formatted drives are expected and other
partition formats are ignored. Also reorder and reword this section a bit.
Signed-off-by: Oliver Steffen <osteffen@redhat.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The bli module has a "hidden" dependency on the part_gpt module, which
is not picked up automatically by the build system. One purpose of the
bli module is to communicate the GPT UUID of the partition GRUB was
launched from to Linux user-space (systemd-gpt-auto-generator).
Without the part_gpt module, bli is not able to obtain the UUID. Since
bli does its work in the module initialization function, the order in
which the modules are loaded is also important: part_gpt needs to be
loaded before the bli module.
To solve this, track this dependency explicitly.
Note that the Boot Loader Interface specification, which bli aims to
implement, requires GPT formatted drives. The bli module ignores all
other partition formats.
Signed-off-by: Oliver Steffen <osteffen@redhat.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The build system deduces inter-module dependencies from the symbols
required and exported by the modules. This works well, except for some
rare cases where the dependency is indirect or hidden. A module might
not make use of any function of some other module, but still expect its
functionality to be available to GRUB.
To solve this, introduce a new file, currently empty, called extra_deps.lst
to track these cases manually. This file gets processed in the same way
as the automatically generated syminfo.lst, making it possible to inject
data into the dependency resolver.
Since *.lst files are set to be ignored by git, add an exception for
extra_deps.lst.
Additionally, introduce a new keyword for the syminfo.lst syntax:
"depends" allows specifying a module dependency directly:
depends <module> <depdendency>...
Signed-off-by: Oliver Steffen <osteffen@redhat.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Display upper_mem_limit and its rounded-down value in MiB.
Signed-off-by: Stefan Berger <stefanb@linux.ibm.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
On PowerVM and KVM on Power use the new memory allocation function that
honors restrictions on which memory GRUB can actually use. In the request
structure indicate the request for a single memory block along with
address alignment restrictions. Request direct usage of the memory block
by setting init_region to false (prevent it from being added to GRUB's
heap). Initialize the found addr to -1, so that -1 will be returned
to the loader in case no memory could be allocated.
Report an out-of-memory error in case the initrd could not be loaded.
Signed-off-by: Stefan Berger <stefanb@linux.ibm.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Cc: Hari Bathini <hbathini@linux.ibm.com>
Cc: Pavithra Prakash <pavrampu@in.ibm.com>
Cc: Michael Ellerman <mpe@ellerman.id.au>
Cc: Carolyn Scherrer <cpscherr@us.ibm.com>
Cc: Mahesh Salgaonkar <mahesh@linux.ibm.com>
Cc: Sourabh Jain <sourabhjain@linux.ibm.com>
Introduce flags to identify PowerVM and KVM on Power and set them where
each type of host has been detected.
Signed-off-by: Stefan Berger <stefanb@linux.ibm.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Cc: Hari Bathini <hbathini@linux.ibm.com>
Cc: Pavithra Prakash <pavrampu@in.ibm.com>
Cc: Michael Ellerman <mpe@ellerman.id.au>
Cc: Carolyn Scherrer <cpscherr@us.ibm.com>
Cc: Mahesh Salgaonkar <mahesh@linux.ibm.com>
Cc: Sourabh Jain <sourabhjain@linux.ibm.com>
Rename regions_claim() to grub_regions_claim() to make it available for
memory allocation. The ieee1275 loader will use this function on PowerVM
and KVM on Power and thus avoid usage of memory that it is not allowed
to use.
Signed-off-by: Stefan Berger <stefanb@linux.ibm.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Cc: Hari Bathini <hbathini@linux.ibm.com>
Cc: Pavithra Prakash <pavrampu@in.ibm.com>
Cc: Michael Ellerman <mpe@ellerman.id.au>
Cc: Carolyn Scherrer <cpscherr@us.ibm.com>
Cc: Mahesh Salgaonkar <mahesh@linux.ibm.com>
Cc: Sourabh Jain <sourabhjain@linux.ibm.com>
Add support for memory alignment requirements and adjust a candidate
address to it before checking whether the block is large enough. This
must be done in this order since the alignment adjustment can make
a block smaller than what was requested.
None of the current callers has memory alignment requirements but the
ieee1275 loader for kernel and initrd will use it to convey them.
Signed-off-by: Stefan Berger <stefanb@linux.ibm.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Cc: Hari Bathini <hbathini@linux.ibm.com>
Cc: Pavithra Prakash <pavrampu@in.ibm.com>
Cc: Michael Ellerman <mpe@ellerman.id.au>
Cc: Carolyn Scherrer <cpscherr@us.ibm.com>
Cc: Mahesh Salgaonkar <mahesh@linux.ibm.com>
Cc: Sourabh Jain <sourabhjain@linux.ibm.com>
Return the allocated address of the memory block in the request structure
if a memory allocation was actually done. Leave the address untouched
otherwise. This enables a caller who wants to use the allocated memory
directly, rather than adding the memory to the heap, to see where memory
was allocated. None of the current callers need this but the converted
ieee1275 loader will make use of it.
Signed-off-by: Stefan Berger <stefanb@linux.ibm.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Cc: Hari Bathini <hbathini@linux.ibm.com>
Cc: Pavithra Prakash <pavrampu@in.ibm.com>
Cc: Michael Ellerman <mpe@ellerman.id.au>
Cc: Carolyn Scherrer <cpscherr@us.ibm.com>
Cc: Mahesh Salgaonkar <mahesh@linux.ibm.com>
Cc: Sourabh Jain <sourabhjain@linux.ibm.com>
Let the regions_claim() request structure's init_region determine whether
to call grub_mm_init_region() on it. This allows for adding memory to
GRUB's memory heap if init_region is set to true, or direct usage of the
memory otherwise. Set all current callers' init_region to true since they
want to add memory regions to GRUB's heap.
Signed-off-by: Stefan Berger <stefanb@linux.ibm.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Cc: Hari Bathini <hbathini@linux.ibm.com>
Cc: Pavithra Prakash <pavrampu@in.ibm.com>
Cc: Michael Ellerman <mpe@ellerman.id.au>
Cc: Carolyn Scherrer <cpscherr@us.ibm.com>
Cc: Mahesh Salgaonkar <mahesh@linux.ibm.com>
Cc: Sourabh Jain <sourabhjain@linux.ibm.com>
The regions_claim() function limits the allocation of memory regions
by excluding certain memory areas from being used by GRUB. This for
example includes a gap between 640MB and 768MB as well as an upper
limit beyond which no memory may be used when an fadump is present.
However, the ieee1275 loader for kernel and initrd currently does not
use regions_claim() for memory allocation on PowerVM and KVM on Power
and therefore may allocate memory in those areas that it should not use.
To make the regions_claim() function more flexible and ultimately usable
for the ieee1275 loader, introduce a request structure to pass various
parameters to the regions_claim() function that describe the properties
of requested memory chunks. In a first step, move the total and flags
variables into this structure.
Signed-off-by: Stefan Berger <stefanb@linux.ibm.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Cc: Hari Bathini <hbathini@linux.ibm.com>
Cc: Pavithra Prakash <pavrampu@in.ibm.com>
Cc: Michael Ellerman <mpe@ellerman.id.au>
Cc: Carolyn Scherrer <cpscherr@us.ibm.com>
Cc: Mahesh Salgaonkar <mahesh@linux.ibm.com>
Cc: Sourabh Jain <sourabhjain@linux.ibm.com>
XFS introduced 64-bit extent counters for inodes via a series of
upstream commits and the feature was marked as stable in v6.5 via
commit 61d7e8274cd8 (xfs: drop EXPERIMENTAL tag for large extent
counts).
Further, xfsprogs release v6.5.0 switched this feature on by default
in mkfs.xfs via commit e5b18d7d1d96 (mkfs: enable large extent counts
by default).
Filesystems formatted with large extent count support, nrext64=1, are
thus currently not recognizable by GRUB, since this is an incompat
feature. Add the required support so that those filesystems and inodes
with large extent counters can be read by GRUB.
Signed-off-by: Anthony Iliopoulos <ailiop@suse.com>
Reviewed-by: Andrey Albershteyn <aalbersh@redhat.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Tested-by: Marta Lewandowska <mlewando@redhat.com>
Tested-by: Sebastian Andrzej Siewior <sebastian@breakpoint.cc>
With new alignment specification it's easy to screw up. Fortunately if it
happens the size will be bigger than intended. Compile time assert will catch
this.
Signed-off-by: Vladimir Serbinenko <phcoder@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
On ia64 alignment requirements are strict. When we pass a pointer to
UUID it needs to be at least 4-byte aligned or EFI will crash.
On the other hand in device path there is no padding for UUID, so we
need 2 types in one formor another. Make 4-byte aligned and unaligned types
The code is structured in a way to accept unaligned inputs
in most cases and supply 4-byte aligned outputs.
Efiemu case is a bit ugly because there inputs and outputs are
reversed and so we need careful casts to account for this
inversion.
Signed-off-by: Vladimir Serbinenko <phcoder@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
gpt_partition contains grub_guid. We need to decide whether the whole
structure is unaligned and then we need to use packed_guid. But we never
have unaligned part entries as we read them in an aligned buffer from disk.
Hence just make it all aligned.
Signed-off-by: Vladimir Serbinenko <phcoder@gmail.com>
We do table search in many places doing exactly the same algorithm.
The only minor variance in users is which table is used if several entries
are present. As specification mandates uniqueness and even if it ever isn't,
first entry is good enough, unify this code and always use the first entry.
Signed-off-by: Vladimir Serbinenko <phcoder@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
known_protocols isn't used anywhere else and even misses grub_ prefix, so
let's make it local (static).
Signed-off-by: Vladimir Serbinenko <phcoder@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The function argp_parser() in util/grub-mount.c lacks a check on the
sanity of the file path when parsing parameters. This results in
a segmentation fault if a partition is mounted to a non-existent path.
Signed-off-by: Qiumiao Zhang <zhangqiumiao1@huawei.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Font might be located in different location, the default font might
not be available on all systems or other font might be preferred.
Signed-off-by: Richard Marko <srk@48.io>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Font might be located in different location, the default font might
not be available on all systems or other font might be preferred.
Signed-off-by: Mads Kiilerich <mads@kiilerich.com>
Signed-off-by: Richard Marko <srk@48.io>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The XFS directory entry parsing code has never been completely correct
for extent based directories. The parser correctly handles the case
where the directory is contained in a single extent, but then mistakenly
assumes the data blocks for the multiple extent case are each identical
to the single extent case. The difference in the format of the data
blocks between the two cases is tiny enough that its gone unnoticed for
a very long time.
A recent change introduced some additional bounds checking into the XFS
parser. Like GRUB's existing parser, it is correct for the single extent
case but incorrect for the multiple extent case. When parsing a directory
with multiple extents, this new bounds checking is sometimes (but not
always) tripped and triggers an "invalid XFS directory entry" error. This
probably would have continued to go unnoticed but the /boot/grub/<arch>
directory is large enough that it often has multiple extents.
The difference between the two cases is that when there are multiple
extents, the data blocks do not contain a trailer nor do they contain
any leaf information. That information is stored in a separate set of
extents dedicated to just the leaf information. These extents come after
the directory entry extents and are not included in the inode size. So
the existing parser already ignores the leaf extents.
The only reason to read the trailer/leaf information at all is so that
the parser can avoid misinterpreting that data as directory entries. So
this updates the parser as follows:
For the single extent case the parser doesn't change much:
1. Read the size of the leaf information from the trailer
2. Set the end pointer for the parser to the start of the leaf
information. (The previous bounds checking set the end pointer to the
start of the trailer, so this is actually a small improvement.)
3. Set the entries variable to the expected number of directory entries.
For the multiple extent case:
1. Set the end pointer to the end of the block.
2. Do not set up the entries variable. Figuring out how many entries are
in each individual block is complex and does not seem worth it when
it appears to be safe to just iterate over the entire block.
The bounds check itself was also dependent upon the faulty XFS parser
because it accidentally used "filename + length - 1". Presumably this
was able to pass the fuzzer because in the old parser there was always
8 bytes of slack space between the tail pointer and the actual end of
the block. Since this is no longer the case the bounds check needs to be
updated to "filename + length + 1" in order to prevent a regression in
the handling of corrupt fliesystems.
Notes:
* When there is only one extent there will only ever be one block. If
more than one block is required then XFS will always switch to holding
leaf information in a separate extent.
* B-tree based directories seems to be parsed properly by the same code
that handles multiple extents. This is unlikely to ever occur within
/boot though because its only used when there are an extremely large
number of directory entries.
Fixes: ef7850c75 (fs/xfs: Fix issues found while fuzzing the XFS filesystem)
Fixes: b2499b29c (Adds support for the XFS filesystem.)
Fixes: https://savannah.gnu.org/bugs/?64376
Signed-off-by: Jon DeVree <nuxi@vault24.org>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Tested-by: Sebastian Andrzej Siewior <sebastian@breakpoint.cc>
Tested-by: Marta Lewandowska <mlewando@redhat.com>
After parsing of the current entry, the entry pointer is advanced
to the next entry at the end of the "for" loop. In case where the
last entry is at the end of the data boundary, the advanced entry
pointer can point off the data boundary. The subsequent boundary
check for the advanced entry pointer can cause a failure.
The fix is to include the boundary check into the "for" loop
condition.
Signed-off-by: Lidong Chen <lidong.chen@oracle.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Tested-by: Sebastian Andrzej Siewior <sebastian@breakpoint.cc>
Tested-by: Marta Lewandowska <mlewando@redhat.com>
Original commit is wrong because grub_file_get_device_name() may return NULL
if we use implicit $root. Additionally, the grub_errno is guaranteed to be
GRUB_ERR_NONE at the beginning of a command. So, everything should work as
expected and Coverity report, CID 73668, WRT to this code should be treated
as false positive.
This reverts commit 7aab03418 (zfsinfo: Correct a check for error allocating memory).
Fixes: 7aab03418 (zfsinfo: Correct a check for error allocating memory)
Signed-off-by: Vladimir Serbinenko <phcoder@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Current code imposes limitations on the amount of sectors read in
a single call according to CHS layout of the disk even in LBA
read mode. There's no need to obey CHS layout restrictions for
LBA reads on LBA disks. It only slows down booting process.
See: https://lore.kernel.org/grub-devel/d42a11fa-2a59-b5e7-08b1-d2c60444bb99@valdikss.org.ru/
Signed-off-by: ValdikSS <iam@valdikss.org.ru>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The code flushes the cache on VIA processors unconditionally which
is excessive. Check for cpuid family and execute wbinvd only on C3
and earlier.
Fixes: https://savannah.gnu.org/bugs/?45149
Fixes: 25492a0f0 (Add wbinvd around bios call.)
Signed-off-by: ValdikSS <iam@valdikss.org.ru>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Implicit holes in file data need to be zeroed explicitly, instead of
just leaving the data in the buffer uninitialized.
This led to kernels randomly failing to boot in "fun" ways when loaded
from btrfs with the no_holes feature enabled, because large blocks of
zeros in the kernel file contained random data instead.
Signed-off-by: Fabian Vogt <fvogt@suse.de>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Reviewed-by: Qu Wenruo <wqu@suse.com>
When a kernel dump is present then restrict the high memory regions to
avoid allocating memory where the kernel dump resides. Use the
ibm,kernel-dump node under /rtas to determine whether a kernel dump
exists and up to which limit GRUB can use available memory. Set the
upper_mem_limit to the size of the kernel dump section of type
REAL_MODE_REGION and therefore only allow GRUB's memory usage for high
addresses from RMO_ADDR_MAX to upper_mem_limit. This means that GRUB can
use high memory in the range of RMO_ADDR_MAX (768MB) to upper_mem_limit
and the kernel-dump memory regions above upper_mem_limit remain
untouched. This change has no effect on memory allocations below
linux_rmo_save (typically at 640MB).
Also, fall back to allocating below rmo_linux_save in case the chunk of
memory there would be larger than the chunk of memory above RMO_ADDR_MAX.
This can for example occur if a free memory area is found starting at 300MB
extending up to 1GB but a kernel dump is located at 768MB and therefore
does not allow the allocation of the high memory area but requiring to use
the chunk starting at 300MB to avoid an unnecessary out-of-memory condition.
Signed-off-by: Stefan Berger <stefanb@linux.ibm.com>
Reviewed-by: Hari Bathini <hbathini@linux.ibm.com>
Cc: Pavithra Prakash <pavrampu@in.ibm.com>
Cc: Michael Ellerman <mpe@ellerman.id.au>
Cc: Carolyn Scherrer <cpscherr@us.ibm.com>
Cc: Mahesh Salgaonkar <mahesh@linux.ibm.com>
Cc: Sourabh Jain <sourabhjain@linux.ibm.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
In certain firmwares, e.g. OVMF, the RNG protocol is not enabled unless
there is an RNG device. When not enabled, GRUB fails to initialize the
stack guard with random bytes. For testing, this is not a big issue, but
there have been bugs found in the initialization. So turn this on for EFI
platforms to catch any regressions.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
GCC is electing to instrument grub_efi_init() to give it stack smashing
protection when configuring with --enable-stack-protector on the x86_64-efi
target. In the function prologue, the canary at the top of the stack frame
is set to the value of the stack guard. And in the epilogue, the canary is
checked to verify if it is equal to the guard and if not to call the stack
check fail function. The issue is that grub_efi_init() sets up the guard
by initializing it with random bytes, if the firmware supports the RNG
protocol. So in its prologue the canary will be set with the value of the
uninitialized guard, likely NUL bytes. Then the guard is initialized, and
finally the epilogue checks the canary against the guard, which will almost
certainly be different. This causes the code path for a smashed stack to be
taken, causing the machine to print out a message that stack smashing was
detected, wait 5 seconds, and then reboot. Disable grub_efi_init()
instrumentation so there is no stack smashing false positive generated.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The sector size in bytes is added to each line and it is allowed to be
6 decimal digits long, which covers the most common cases of 512 and 4096
byte sectors with space for two additional digits as future-proofing. The
size allocation is updated to reflect this additional field. Also make
clearer the size allocation calculation.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Use the return value of grub_snprintf() to move the string pointer forward,
instead of incrementing the string pointer iteratively until a NULL byte is
reached. Move the space out of the format string argument, a small
optimization, but also makes the spacing clearer. Also, use the new
PRIxGRUB_OFFSET instead of PRIuGRUB_UINT64_T to accurately reflect the
format string for this type.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
A large enough argument to the --port option could cause a string buffer
to be not NULL terminated because grub_strncpy() does not guarantee NULL
termination if copied string is longer than max characters to copy.
Fixes: 712309eaae (term/serial: Use grub_strncpy() instead of grub_snprintf() when only copying string)
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
E.g. 2.10 instead of 00020064 and 2.3.1 instead of 0002001f.
See UEFI 2.10 specification, chapter 4.2.1 EFI_TABLE_HEADER.
Signed-off-by: Heinrich Schuchardt <heinrich.schuchardt@canonical.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Move some calls used to access NTFS attribute header fields into
functions with human-readable names.
Suggested-by: Daniel Kiper <daniel.kiper@oracle.com>
Signed-off-by: Maxim Suhanov <dfirblog@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
This fix introduces checks to ensure that an NTFS volume label is always
read from the corresponding file record segment.
The current NTFS code allows the volume label string to be read from an
arbitrary, attacker-chosen memory location. However, the bytes read are
always treated as UTF-16LE. So, the final string displayed is mostly
unreadable and it can't be easily converted back to raw bytes.
The lack of this check is a minor issue, likely not causing a significant
data leak.
Reported-by: Maxim Suhanov <dfirblog@gmail.com>
Signed-off-by: Maxim Suhanov <dfirblog@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
This fix introduces checks to ensure that bitmaps for directory indices
are never read beyond their actual sizes.
The lack of this check is a minor issue, likely not exploitable in any way.
Reported-by: Maxim Suhanov <dfirblog@gmail.com>
Signed-off-by: Maxim Suhanov <dfirblog@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
This fix introduces checks to ensure that index entries are never read
beyond the corresponding directory index.
The lack of this check is a minor issue, likely not exploitable in any way.
Reported-by: Maxim Suhanov <dfirblog@gmail.com>
Signed-off-by: Maxim Suhanov <dfirblog@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
When reading a file containing resident data, i.e., the file data is stored in
the $DATA attribute within the NTFS file record, not in external clusters,
there are no checks that this resident data actually fits the corresponding
file record segment.
When parsing a specially-crafted file system image, the current NTFS code will
read the file data from an arbitrary, attacker-chosen memory offset and of
arbitrary, attacker-chosen length.
This allows an attacker to display arbitrary chunks of memory, which could
contain sensitive information like password hashes or even plain-text,
obfuscated passwords from BS EFI variables.
This fix implements a check to ensure that resident data is read from the
corresponding file record segment only.
Fixes: CVE-2023-4693
Reported-by: Maxim Suhanov <dfirblog@gmail.com>
Signed-off-by: Maxim Suhanov <dfirblog@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
When parsing an extremely fragmented $MFT file, i.e., the file described
using the $ATTRIBUTE_LIST attribute, current NTFS code will reuse a buffer
containing bytes read from the underlying drive to store sector numbers,
which are consumed later to read data from these sectors into another buffer.
These sectors numbers, two 32-bit integers, are always stored at predefined
offsets, 0x10 and 0x14, relative to first byte of the selected entry within
the $ATTRIBUTE_LIST attribute. Usually, this won't cause any problem.
However, when parsing a specially-crafted file system image, this may cause
the NTFS code to write these integers beyond the buffer boundary, likely
causing the GRUB memory allocator to misbehave or fail. These integers contain
values which are controlled by on-disk structures of the NTFS file system.
Such modification and resulting misbehavior may touch a memory range not
assigned to the GRUB and owned by firmware or another EFI application/driver.
This fix introduces checks to ensure that these sector numbers are never
written beyond the boundary.
Fixes: CVE-2023-4692
Reported-by: Maxim Suhanov <dfirblog@gmail.com>
Signed-off-by: Maxim Suhanov <dfirblog@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
During attempts to configure a serial console, a Page Fault Exception
and system reset were encountered, specifically on release 2.12~rc1.
This issue was not present in prior versions and seemed to affect only
a specific machine, potentially pointing to hardware or firmware flaw.
After investigation, it was discovered that the invalid page access
occurred during the discovery of serial MMIO ports as specified by
ACPI's SPCR table [1]. The recent change uncovered an issue in GRUB's
ACPI driver.
In certain cases, the XSDT/RSDT root table might contain a NULL entry as
a terminator, depending on how the tables are assembled. GRUB cannot
blindly trust the address in the root table to be valid and should
perform a sanity check for NULL entries. This patch introduces this
simple check.
This fix is also inspired by a related Linux kernel fix [2].
[1] 7b192ec4c term/ns8250: Use ACPI SPCR table when available to configure serial
[2] 0f929fbf0 ACPICA: Tables: Add new mechanism to skip NULL entries in RSDT and XSDT.
Signed-off-by: Michael Chang <mchang@suse.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
When grub-install is run with the verbose option, it will print a log
message indicating the grub-mkimage command and arguments used.
GRUB no longer calls the grub-mkimage binary internally, however the
command logged is a command that if run should effectively be what
grub-install used. However, as this has changed some of the newer
options have been incorrectly added so that the printed command fails
when run separately. This change makes the displayed command run as
intended.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
This is a preparatory patch to make the following patch less cluttered. The
only visible change made here is to not print extra spaces when either or
both --note or --disable-shim-lock are not given and to not print an extra
space at the end of the command. The latter is done by constructing the
trailing argument string with spaces in front of each argument rather than
trailing. The allocation of the argument string is made precise, which has
the benefit of saving a few bytes, but more importantly self-documenting
what the needed allocated bytes are. Also, unneeded braces are removed from
an if block.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The commit 80948f532d (lib/i386/relocator64: Build fixes for i386) has
broken 64-bit FreeBSD boot on BIOS. This patch fixes the issue.
Fixes: 80948f532d (lib/i386/relocator64: Build fixes for i386)
Signed-off-by: Vladimir Serbinenko <phcoder@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
It turns out that setting $xen_version in linux_entry_xsm() override
$xen_version in the loop over $reverse_sorted_xen_list. This means
that only one entry per Xen version is going to enable XSM, but all
further entries are going to have "(XSM enabled)" in their titles
without enabling XSM.
When a "xenpolicy-$xen_version" file was found for the current
$xen_version, it would overwrite $xen_version to add "(XSM enabled)" to
the menu entry title. Once updated, the next call to linux_entry_xsm()
would also have this modified $xen_version and would look for the file
"xenpolicy-*(XSM enabled)" and fail.
Signed-off-by: Anthony PERARD <anthony.perard@citrix.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
In the configure phase, the "-mcmodel=large" CFLAGS passed the test, but
because it has not been implemented in gcc, the following warning will
appear when compiling:
gcc: warning: 'large' is not supported, now cmodel is set to 'normal'
Signed-off-by: Xiaotian Wu <wuxiaotian@loongson.cn>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The backtrace module is written assuming that the frame pointer is in %ebp.
By default, -Os optimization level is used, which enables the gcc option
-fomit-frame-pointer. This breaks the backtrace functionality. Enabling
this may cause an unnoticeable performance cost and virtually no size increase.
The backtrace command on x86_64 and probably i386 is broken due to the
above rationale. I've not verified, but presumably the backtrace that used
to be printed for an unhandled CPU exception is also broken. Do any distros
handle this?
Considering that, to my knowledge, no one has complained about this in the
over 13 years that -Os has been used, has this code actually been useful?
Is it worth disabling -fomit-frame-pointer? Though, I don't see much downside
right now in disabling it. Alternatively, we could disable/remove the
backtrace code. I think it would be nice to keep it and have it working.
Nowadays, presumably QEMU makes the GDB stub rarely used as I imagine most
are developing in a virtual machines. Also, the GDB stub does not work in UEFI.
So, if anyone is using it on real hardware, they are doing so on pretty old
machines. The lack of a GDB stub does not seem to be a pain point because
no one has got it working on UEFI.
This patch gets the backtrace command working on x86_64-efi in QEMU for me.
However, it hangs when run on my laptop. Not sure what's going on there.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Recent mixed-mode Linux kernels, i.e., v4.0 or newer, can access EFI
runtime services at OS runtime even when the OS was not entered via the
EFI stub. This is because, instead of reverting back to the firmware's
segment selectors, GDTs and IDTs, the 64-bit kernel simply calls 32-bit
runtime services using compatibility mode, i.e., the same mode used for
32-bit user space, without taking down all interrupt handling, exception
handling, etc.
This means that GRUB's legacy x86 boot mode is sufficient to make use of
this: 32-bit i686 builds of GRUB can already boot 64-bit kernels in EFI
enlightened mode, but without going via the EFI stub, and provide all
the metadata that the OS needs to map the EFI runtime regions and call
EFI runtime services successfully.
It does mean that GRUB should not attempt to invoke the firmware's
LoadImage()/StartImage() methods on kernel builds that it knows cannot
be started natively. So, add a check for this in the native EFI boot
path and fall back to legacy x86 mode in such cases.
Note that in the general case, booting non-native images of the same
native word size, e.g., x64 EFI apps on arm64 firmware, might be
supported by means of emulation. So, let's only disallow images that use
a non-native word size. This will also permit booting i686 kernels on
x86_64 builds, although without access to runtime services, as this is
not supported by Linux.
This change on top of 2.12-rc1 is sufficient to boot ordinary Linux
mixed mode builds and get full access to the EFI runtime services.
Cc: Daniel Kiper <daniel.kiper@oracle.com>
Cc: Steve McIntyre <steve@einval.com>
Cc: Julian Andres Klode <julian.klode@canonical.com>
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
Acked-by: Dimitri John Ledkov <dimitri.ledkov@canonical.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The x86_64 Linux kernel can be booted in 32-bit mode, in which case the
startup code creates a set of preliminary page tables that map the first
4 GiB of physical memory 1:1 and enables paging. This is a prerequisite
for 64-bit execution and can therefore only be implemented in 32-bit code.
The x86_64 Linux kernel can also be booted in 64-bit mode directly: this
implies that paging is already enabled and it is the responsibility of
the bootloader to ensure that the active page tables cover the entire
loaded image, including its BSS space, the size of which is described in
the image's setup header.
Given that the EFI spec mandates execution in long mode for x86_64 and
stipulates that all system memory is mapped 1:1, the Linux/x86
requirements for 64-bit entry can be met trivially when booting on
x86_64 via EFI. So, enter via the 64-bit entry point in this case.
This involves inspecting the xloadflags field in the setup header to
check whether the 64-bit entry point is supported. This field was
introduced in Linux version v3.8 (early 2013).
This change ensures that all EFI firmware tables and other assets passed
by the firmware or bootloader in memory remain mapped and accessible
throughout the early startup code.
Avoiding the drop out of long mode will also be needed to support
upcoming CPU designs that no longer implement 32-bit mode at all
(as recently announced by Intel [0]).
[0] https://www.intel.com/content/www/us/en/developer/articles/technical/envisioning-future-simplified-architecture.html
Cc: Daniel Kiper <daniel.kiper@oracle.com>
Cc: Julian Andres Klode <julian.klode@canonical.com>
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Some dnodes are shared with properties zap. This is used
e.g. for quotas. Then dnode type is 0xc4 and GRUB stumbles on
this. Check bonus type and if it's ok then ignore dnode type mismatch
Signed-off-by: Vladimir Serbinenko <phcoder@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Reading them is harmless but useless as they are empty by definition
Signed-off-by: Vladimir Serbinenko <phcoder@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
We ended up comparing over unset values as we had dnode_phys on one side
and dnode on another
Signed-off-by: Vladimir Serbinenko <phcoder@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
This is a speedup used in some ZFS version. This trips GRUB and makes it
unable to access directories. Just skip it for now and revisit
if we ever need this speedup.
Signed-off-by: Vladimir Serbinenko <phcoder@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
If the EFI graphics pixel format is PixelBltOnly, we cannot write directly
to the frame buffer. We need the shadow frame buffer which we copy via
the BitBlt operation to the hardware.
If the pixel format is PixelBltOnly and allocation of the shadow frame
buffer fails, we must raise an error to signal that the EFI GOP protocol
is not usable.
Signed-off-by: Heinrich Schuchardt <heinrich.schuchardt@canonical.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
It has been reported that makeinfo version 4.13a complains and returns
error when menus for chapter structuring commands are not present. It
is also known that newer makeinfos, such as version 6.7, will create
default menus when needed. Since the menu will be created regardless,
explicitly create it to support older makeinfo versions. This also
enables building to be successful when an older makeinfo is installed
because in that case info files are attempted to be generated with the
"all" target.
Reported-by: Olaf Hering <olaf@aepfle.de>
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Tested-by: Olaf Hering <olaf@aepfle.de>
The @xref command is meant to be used at the beginning of a sentence
because its expansion creates a "See " prefix on all output formats, and
on older makeinfo versions is strict about enforcing a "." or "," after
the command. The @ref command has no such restriction and is just the
link, which allows more control over output. This also fixes an issue
where there was a repeated "see" in the output.
Reported-by: Olaf Hering <olaf@aepfle.de>
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Tested-by: Olaf Hering <olaf@aepfle.de>
Allow using the envvar GRUB_SHELL_LUKS_TIMEOUT to change the default
timeout. If not specified, use value of GRUB_SHELL_DEFAULT_TIMEOUT. And
if that is not specified, fallback to original 600s timeout.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
This was causing the cryptomount command to return failure even though
the crypto device was successfully added. Of course, this meant that any
script using the return code would behave unexpectedly.
Fixes: 3cf2e848bc (disk/cryptodisk: Allows UUIDs to be compared in a dash-insensitive manner)
Suggested-by: Olaf Hering <olaf@aepfle.de>
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Patrich Steinhardt <ps@pks.im>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
To comply with C99 and POSIX standards, snprintf() should return the
number of bytes that would be written to the string (excluding the
terminating NUL byte) if the buffer size was big enough. Before this
change, the return value was the minimum of the standard return and the
length of the buffer. Rarely is the return value of grub_snprintf() or
grub_vsnprintf() used with current code, and the few places where it is
used do not need to be changed.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
This test is meant to test output via various serial devices. Currently,
only the PCI serial device is tested.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
While here, move "-qemu=*" case to be next to the "--qemu-opts=*" case.
This causes no change in logic, but is more logically located.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
In long list mode, if the file can not be opened, the file is not printed.
Instead, print the file but print the size as "????????????".
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
For each non-directory path argument to the ls command, the full path was
being sent to the print functions, instead of the dirname. The long output
print function expected dirname to be the directory containing the file
and so could not open the file to get the file size because the generated
path was incorrect. This caused the output to be a blank line.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Specifically, return GRUB_ERR_BAD_FILE_TYPE because this is what is
expected by the ls command when it is given a path to a non-directory.
This fixes a bug where calling ls with a list of non-directory paths
outputs a blank line for each such argument.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The videoinfo command will initialize all non-active video adapters. Video
drivers tend to zero out the global framebuffer object on initialization.
This is not a problem when there is no active video adapter. However, when
there is, then outputting to the video adapter will cause a crash because
methods in the framebuffer object are reinitialized. For example, this
command sequence will cause a crash.
terminal_output --append gfxterm; videoinfo
When running in a QEMU headless with GRUB built for the x86_64-efi target,
the first command initializes the Bochs video adapter, which, among
other things, sets the set_page() member function. Then when videoinfo is
run, all non-Bochs video adapters will be initialized, each one wiping
the framebuffer and thus setting set_page to NULL. Soon after the videoinfo
command finishes there will be a call to grub_refresh(), which will
ultimately call the framebuffer's set_page which will be NULL and cause
a crash when called.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
A list of improvements:
* Remove reference to "initial ramdisk" and replace with "initrd". This
then covers the case of ramdisk and ramfs, which is the usual method
with kernels 2.6 and newer.
* Add sentence with URL to initrd documentation Linux kernel.
* Add a section documenting how to have the initrd command generate
a new-style initrd via a specially crafted argument and include an example.
* Update initrd16 to refer to the initrd section and make note that
initrd16 is only on the pc platform.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Oskari Pirhonen <xxc3ncoredxx@gmail.com>
Reviewed-by: Paul Menzel <pmenzel@molgen.mpg.de>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
According to commit 0231d00082 (ACPI: SPCR: Make SPCR available to x86)
to the Linux kernel, "On x86, many systems have a valid SPCR table but the
table version is not 2 so the table version check must be a warning."
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The cat command should not be used to print binary data because it can
show bytes not in the binary data and not show bytes that are in the data,
which can lead to confusion. This happens because cat does some processing
of the data stream, namely trying to decode substrings as UTF-8.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Oskari Pirhonen <xxc3ncoredxx@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
During configuration of SDL2, variable enable_grub_emu_sdl2 is checked
whether to throw an error message. However, error could not happen
because two unequal strings were compared. Fix this by referencing
value of enable_grub_emu_sdl2, not name.
Fixes: 17d6ac1a7 (emu: Add SDL2 support)
Signed-off-by: Michał Grzelak <mchl.grzlk@gmail.com>
Reviewed-by: Julian Andres Klode <julian.klode@canonical.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Reviewed-by: Paul Menzel <pmenzel@molgen.mpg.de>
Also reword a prior sentence to be more clear.
Fixes: 5a3d2b4742 (docs: Add debugging chapter to development documentation)
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Oskari Pirhonen <xxc3ncoredxx@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
On an unmerged-usr system, grub-mkconfig errors out with the following
error due to /usr/bin/sh not existing:
/usr/sbin/grub-mkconfig: /etc/grub.d/25_bli: /usr/bin/sh: bad interpreter: No such file or directory
Use a /bin/sh shebang to fix the error as well as match the other
existing files.
Fixes: 158a6583e (util/grub.d/25_bli.in: Activate bli module on EFI)
Signed-off-by: Oskari Pirhonen <xxc3ncoredxx@gmail.com>
Reviewed-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Reviewed-by: Oliver Steffen <osteffen@redhat.com>
This allows an invocation of grub-shell to set the value of debug regardless
of the global default environment variable GRUB_SHELL_DEFAULT_DEBUG.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Since this is fairly verbose output, do not enable first level of debug
is turned on.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The LoadImage() provided by the shim does not consult MOK when loading
an image. So, simply signature verification fails when it should not.
This means we cannot use Linux EFI stub to start the kernel when the
shim is loaded. We have to fallback to legacy mode on x86 architectures.
This is not possible on other architectures due to lack of legacy mode.
This is workaround which should disappear when the shim provides
LoadImage() which looks up MOK during signature verification.
On the occasion align constants in include/grub/efi/sb.h.
Signed-off-by: Daniel Kiper <daniel.kiper@oracle.com>
Reviewed-by: Ard Biesheuvel <ardb@kernel.org>
... because (surprisingly) it does not use specific EFI calling convention...
Fixes: 6a080b9cd (efi: Add calling convention annotation to all prototypes)
Signed-off-by: Daniel Kiper <daniel.kiper@oracle.com>
Reviewed-by: Ard Biesheuvel <ardb@kernel.org>
acpi actually needs to access PCI, while pci-arbiter will not be making
use of ACPI, so we need to start acpi first.
Signed-off-by: Samuel Thibault <samuel.thibault@ens-lyon.org>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
During configuration of SDL, variable enable_grub_emu_sdl is checked
whether to throw an error message. However, error could not happen
because two unequal strings were compared. Fix this by referencing
value of enable_grub_emu_sdl, not name.
Fixes: 17d6ac1a7 (emu: Add SDL2 support)
Signed-off-by: Michał Grzelak <mchl.grzlk@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
When referring to initrd16 the link for initrd16 should be used, not a link
for initrd. Also, correct the spelling of additionally and add a comma after
it to correct its grammatical usage.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Using grub_strncpy() instead of grub_snprintf() is less overhead and
indicates clearly that the dest should be the same string as the source.
Also fix indentation.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
These are currently always the same as PRI*GRUB_UINT64_T, but they may
not be in the future.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Also add parenthesis to nested ternary operator to improve clarity.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Python bytecode files, which end in .pyc, may be generated by the build
system as needed and should not go into the git repository.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
When creating at runtime a newc initrd via arguments to initrd with "newc:"
prefixes, only emit a directory path record once. The original code
intended to do that by bailing out of emitting the record when the record
to be created matches an existing record. However, this does not happen
because grub_memcmp() is improperly checked.
Generating duplicate newc directory records does not cause any problems
because the Linux unpacker will skip it once it sees the directory already
exists. This fix saves a little processing and makes the generated newc
cpio archive a little smaller.
Fixes: 92750e4c60 (Add ability to generate newc additions on runtime.)
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
EFI status codes are of different classes depending on the first byte and
all error status codes defined in appendix D of the main spec start from
1 and have the high bit set. When printing as a uint, the decimal is a very
large number that needs have the high bit cleared get the spec error code.
This can be easily visually done by a human if the number is printed as hex.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Small set of wording and grammatical edits which did not make it in time
for the original review of the chapter.
Signed-off-by: Oskari Pirhonen <xxc3ncoredxx@gmail.com>
Reviewed-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Most of leftover code blindly assumes GRUB_RELOCATOR_FIRMWARE_REQUESTS_QUANT
divisibility by 8. So, enforce this at compile time.
Signed-off-by: Daniel Kiper <daniel.kiper@oracle.com>
Reviewed-by: Vladimir Serbinenko <phcoder@gmail.com>
So all we did with the surface in SDL1 was split into window,
surface, renderer and texture. Instead of drawing into the
surface and then flipping, you build your pixels, then update
a texture and then copy the texture to the renderer.
Here we use an empty RGB surface to hold our pixels, which enables
us to keep most of the code the same. The SDL1 code has been adjusted
to refer to "surface" instead of "window" when trying to access the
properties of the surface.
This approaches the configuration by adding a new --enable-grub-emu-sdl2
argument. If set to yes, or auto detected, it disables SDL1 support
automatically.
This duplicates the sdl module block in Makefile.core.def which may
be something to be aware of, but we also don't want to build separate
module.
Fixes: https://bugs.debian.org/1038035
Signed-off-by: Julian Andres Klode <julian.klode@canonical.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
These should be quite obvious and will make the SDL2 patch easier
to read then doing it inline there.
Signed-off-by: Julian Andres Klode <julian.klode@canonical.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Currently booting the system is prevented when call to EFI firmware
hash_log_extend_event() returns unknown error. Solve this by following
convention used in commit a4356538d (commands/tpm: Don't propagate
measurement failures to the verifiers layer).
Let the system to be bootable by default when unknown TPM error is
encountered. Check environment variable tpm_fail_fatal to fallback to
previous behaviour.
Signed-off-by: Michał Grzelak <mchl.grzlk@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Currently bootstrap complains in the following way when
patching gnulib files:
patching file argp-help.c
Hunk #1 succeeded at 52 (offset 1 line).
Hunk #2 succeeded at 1548 (offset 115 lines).
patching file mbswidth.c
patching file mbswidth.h
Hunk #1 succeeded at 40 (offset -5 lines).
Let's fix it by amending line numbers in the patch.
Signed-off-by: Daniel Kiper <daniel.kiper@oracle.com>
Reviewed-by: Alec Brown <alec.r.brown@oracle.com>
The commit bb4aa6e06 (efi: Drop all uses of efi_call_XX() wrappers) did
not add some __grub_efi_api attributes to the EFI calls. Lack of them
led to hangs on x86_64-efi target. So, let's add missing __grub_efi_api
attributes.
Fixes: bb4aa6e06 (efi: Drop all uses of efi_call_XX() wrappers)
Reported-by: Christian Hesse <list@eworm.de>
Reported-by: Robin Candau <antiz@archlinux.org>
Signed-off-by: Daniel Kiper <daniel.kiper@oracle.com>
Tested-by: Robin Candau <antiz@archlinux.org>
Tested-by: Christian Hesse <list@eworm.de>
Reviewed-by: Peter Jones <pjones@redhat.com>
Move the constant from grub-core/osdep/linux/getroot.c to
include/grub/disk.h and then reuse it in place of the
hardcoded 1024 limit in diskfilter.
Fixes: 2a5e3c1f2 (disk/diskfilter: Don't make a RAID array with more than 1024 disks)
Cc: Daniel Axtens <dja@axtens.net>
Cc: Kees Cook <keescook@chromium.org>
Signed-off-by: Julian Andres Klode <julian.klode@canonical.com>
Reviewed-by: Kees Cook <keescook@chromium.org>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
A working GRUB cannot be built with upcoming binutils and GCC, because linker
relaxation was added [1] causing new unsupported relocations to appear in modules.
So we pass -mno-relax to GCC if it is supported, to disable relaxation and make
GRUB forward-compatible with new toolchains.
While similar code already exists for sparc64 in configure.ac, sparc64 sets
LDFLAGS while LoongArch requires CFLAGS to be set. If we only set LDFLAGS on
LoongArch, GCC will still generate relaxation relocations in the .o files, so
the sparc64 code cannot be reused.
[1] https://sourceware.org/git/?p=binutils-gdb.git;a=commit;h=56576f4a722b7398d35802ecf7d4185c27d6d69b
Signed-off-by: Xiaotian Wu <wuxiaotian@loongson.cn>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
According to the relocation documentation, the following function names are
renamed to show their exact meaning:
- from grub_loongarch64_xxx64_hi12() to grub_loongarch64_abs64_hi12(),
- from grub_loongarch64_xxx64_hi12() to grub_loongarch64_abs64_lo20().
Signed-off-by: Xiaotian Wu <wuxiaotian@loongson.cn>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
We already have the pc variable, no need to calculate it again.
Signed-off-by: Xiaotian Wu <wuxiaotian@loongson.cn>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Use PRIxGRUB_INT64_T format specifier for grub_int64_t type
and drop redundant casts.
Signed-off-by: Xiaotian Wu <wuxiaotian@loongson.cn>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
According to the ACPI specification, in ACPI 2.0 or later, an
ACPI-compatible OS must use the XSDT if present. So, we should
use xsdt_addr instead of rsdt_addr if xsdt_addr is valid.
Signed-off-by: Qiumiao Zhang <zhangqiumiao1@huawei.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
According to the ACPI specification, in ACPI 2.0 or later, an
ACPI-compatible OS must use the XSDT if present. So, we should
use xsdt_addr instead of rsdt_addr if xsdt_addr is valid.
Signed-off-by: Qiumiao Zhang <zhangqiumiao1@huawei.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Implemented a boundary check before advancing the allocation
descriptors pointer.
Signed-off-by: Lidong Chen <lidong.chen@oracle.com>
Reviewed-by: Darren Kenny <darren.kenny@oracle.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Debugging GRUB can be tricky and require arcane knowledge. This will
help those unfamiliar with the process to get started debugging GRUB
with less effort.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
While performing fuzz testing with XFS filesystem images with ASAN
enabled, several issues were found where the memory accesses are made
beyond the data that is allocated into the struct grub_xfs_data
structure's data field.
The existing structure didn't store the size of the memory allocated into
the buffer in the data field and had no way to check it. To resolve these
issues, the data size is stored to enable checks into the data buffer.
With these checks in place, the fuzzing corpus no longer cause any crashes.
Signed-off-by: Darren Kenny <darren.kenny@oracle.com>
Signed-off-by: Robbie Harwood <rharwood@redhat.com>
Signed-off-by: Marta Lewandowska <mlewando@redhat.com>
Signed-off-by: Lidong Chen <lidong.chen@oracle.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Ensure the generated unidata.c file is deterministic by sorting the
keys of the dict.
Signed-off-by: Alexander Kanavin <alex@linutronix.de>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The output in moddep.lst generated from syminfo.lst using genmoddep.awk
is not deterministic since the order of the dependencies on each line
can vary depending on how awk sorts the values in the array.
Be deterministic in the output by sorting the dependencies on each line.
Signed-off-by: Alexander Kanavin <alex@linutronix.de>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The output of the SOURCES lines in grub-core/Makefile.core.am, generated
from grub-core/Makefile.core.def with gentpl.py is not deterministic due to
missing sorting of the list used to generate it. Add such a sort.
Signed-off-by: Alexander Kanavin <alex@linutronix.de>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
EFI firmware determines where to load the GRUB EFI at runtime, and so the
addresses of debug symbols are not known ahead of time. There is a command
defined in the gdb_grub script which will load the debug symbols at the
appropriate addresses, if given the application load address for GRUB.
So add a command named "gdbinfo" to allow the user to print this GDB command
string with the application load address on-demand. For the outputted GDB
command to have any effect when entered into a GDB session, GDB should have
been started with the script as an argument to the -x option or sourced into
an active GDB session before running the outputted command.
Documentation for the gdbinfo command is also added.
Co-developed-by: Peter Jones <pjones@redhat.com>
Signed-off-by: Peter Jones <pjones@redhat.com>
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The EFI chainloader checks that a device path can be created for the $root
device before allowing chainloading to a given file. This is probably to
ensure that the given file can be accessed and loaded by the firmware.
However, since GRUB is loading the image itself, the firmware need not
be able to access the file location of the image. So remove this check.
Also, this fixes an issue where chainloading an image file on a location
that is accessible by the firmware, e.g. (hd0,1)/efi/boot.efi, would
fail when root is a location inaccessible by the firmware, e.g. memdisk.
Use GRUB_EFI_BYTES_TO_PAGES() instead of doing the calculation explicitly.
Add comment noting the section where the load options for the chainloaded
EFI application is constructed.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Ard Biesheuvel <ardb@kernel.org>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Extra arguments given to chainloader on EFI platforms will be sent to
the chainloaded application. Also, minor edit in the chainloading section
to note that chainloading can be a jump via the firmware and not
necessarily in real mode (which does not exist on some architectures).
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Ard Biesheuvel <ardb@kernel.org>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Add a new configuration drop-in file that loads the bli module and runs
the command if booting on the EFI platform.
Signed-off-by: Oliver Steffen <osteffen@redhat.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Add a new module named bli. It implements a small but quite useful part
of the Boot Loader Interface [0]. This interface uses EFI variables for
communication between the boot loader and the operating system.
When loaded, this module sets two EFI variables under the vendor GUID
4a67b082-0a4c-41cf-b6c7-440b29bb8c4f:
- LoaderInfo: contains GRUB + <version number>.
This allows the running operating system to identify the boot loader
used during boot.
- LoaderDevicePartUUID: contains the partition UUID of the EFI System
Partition (ESP). This is used by systemd-gpt-auto-generator [1] to
find the root partitions (and others too), via partition type IDs [2].
This module is available on EFI platforms only. The bli module relies on
the part_gpt module which has to be loaded beforehand to make the GPT
partitions discoverable.
Update the documentation, add a new chapter "Modules" and describe the
bli module there.
[0] https://systemd.io/BOOT_LOADER_INTERFACE/
[1] https://www.freedesktop.org/software/systemd/man/systemd-gpt-auto-generator.html
[2] https://uapi-group.org/specifications/specs/discoverable_partitions_specification/
Signed-off-by: Oliver Steffen <osteffen@redhat.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Add checks for NULL pointers to grub_device_close() and
grub_disk_close() to make these functions more robust.
Signed-off-by: Oliver Steffen <osteffen@redhat.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Reword some section headings, remove "The List of" from titles. While
grammatically correct, this phrase can be omitted to increase
readability, especially in the table of contents.
Signed-off-by: Oliver Steffen <osteffen@redhat.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Add a function that sets an EFI variable to a string value.
The string is converted from UTF-8 to UTF-16.
Signed-off-by: Oliver Steffen <osteffen@redhat.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Create a new function for UTF-8 to UTF-16 conversion called
grub_utf8_to_utf16_alloc() in the grub-code/kern/misc.c and replace
charset conversion code used in some places in the EFI code. It is
modeled after the grub_utf8_to_ucs4_alloc() like functions in
include/grub/charset.h. It can't live in include/grub/charset.h,
because it needs to be reachable from the kern/efi code.
Add a check for integer overflow and remove redundant NUL-termination.
Signed-off-by: Oliver Steffen <osteffen@redhat.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
In the same way as GRUB_SIZE_MAX, add GRUB_SSIZE_MAX.
Signed-off-by: Oliver Steffen <osteffen@redhat.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Use the new printf format specifier %pG.
Fixes the text representation of GUIDs in the output of the lsefisystab
command (missing 4th dash).
Signed-off-by: Oliver Steffen <osteffen@redhat.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Extend the printf format specifier for pointers (%p) to accept a suffix
specifier G to print GUIDs: %pG can be used to print grub_guid structs.
This does not interfere with the -Wformat checking of gcc. Note that
the data type is not checked though (%p accepts void *).
Signed-off-by: Oliver Steffen <osteffen@redhat.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
There are 3 implementations of a GUID in GRUB. Replace them with
a common one, placed in types.h.
It uses the "packed" flavor of the GUID structs, the alignment attribute
is dropped, since it is not required.
Signed-off-by: Oliver Steffen <osteffen@redhat.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Add a function to the EFI module that allows setting EFI variables
with specific attributes.
This is useful for marking variables as volatile, for example.
Signed-off-by: Oliver Steffen <osteffen@redhat.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
In grub-core/kern/efi/mm.c, grub_efi_finish_boot_services() has an instance
where the memory for the variable finish_mmap_buf is freed, but on the next
iteration of a while loop, grub_efi_get_memory_map() uses finish_mmap_buf. To
prevent this, we can set finish_mmap_buf to NULL after the free.
Signed-off-by: Alec Brown <alec.r.brown@oracle.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The EFI spec mandates that the handle produced by the LoadImage boot
service has a LoadedImage protocol instance installed on it, but for
robustness, we should still deal with a NULL return value from the
helper routine that obtains this protocol pointer.
If this happens, don't try to start the image but unload it and return
an error.
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Switch the x86 based EFI platform builds to the generic EFI loader,
which exposes the initrd via the LoadFile2 protocol instead of the
x86-specific setup header. This will launch the Linux kernel via its EFI
stub, which performs its own initialization in the EFI boot services
context before calling ExitBootServices() and performing the bare metal
Linux boot.
Given that only Linux kernel versions v5.8 and later support this initrd
loading method, the existing x86 loader is retained as a fallback, which
will also be used for Linux kernels built without the EFI stub. In this
case, GRUB calls ExitBootServices() before entering the Linux kernel,
and all EFI related information is provided to the kernel via struct
boot_params in the setup header, as before.
Note that this means that booting EFI stub kernels older than v5.8 is
not supported even when not using an initrd at all. Also, the EFI
handover protocol, which has no basis in the UEFI specification, is not
implemented.
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The call wrappers are no longer needed now that GCC can generate
function calls using MS calling convention, so let's get rid of them.
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Now that GCC can generate function calls using the correct calling
convention for us, we can stop using the efi_call_XX() wrappers, and
just dereference the function pointers directly.
This avoids the untyped variadic wrapper routines, which means better
type checking for the method calls.
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
UEFI mandates MS calling convention on x86_64, which was not supported
on GCC when UEFI support was first introduced into GRUB. However, now we
can use the ms_abi function type attribute to annotate functions and
function pointers as adhering to the MS calling convention, and the
compiler will generate the correct instruction sequence for us.
So let's add the appropriate annotation to all the function prototypes.
This will allow us to drop the special call wrappers in a subsequent patch.
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The grub_efi_pxe_t struct definition has placeholders for the various
protocol method pointers, given that they are never called in the code,
and the prototypes have been omitted, and therefore do not comply with
the UEFI spec.
So let's convert them into void* pointers, so they cannot be called
inadvertently.
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
In grub-core/loader/multiboot_elfxx.c, we need to make sure that the program
header offset is less than the file size along with the MULTIBOOT_SEARCH
constant. We can do so by setting the variable phlimit to the minimum value of
the two limits and check it each time we change program header index to insure
that the program header offset isn't outside of the limits.
Fixes: CID 314029
Fixes: CID 314038
Signed-off-by: Alec Brown <alec.r.brown@oracle.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
In grub-core/loader/multiboot_elfxx.c, space is being allocated for the section
header region, but isn't verifying if the region is within the file's size.
Before calling grub_calloc(), we can add a conditional to check if the section
header region is smaller than the file size.
Fixes: CID 314029
Fixes: CID 314038
Signed-off-by: Alec Brown <alec.r.brown@oracle.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
In grub-core/loader/multiboot_elfxx.c, the code is filling an area of memory
with grub_memset() but doesn't check if there is space in the allocated memory
before doing so. To make sure we aren't zeroing memory past the allocated memory
region, we need to check that the offset into the allocated memory region plus
the memory size of the program is smaller than the allocated memory size.
Fixes: CID 314029
Fixes: CID 314038
Signed-off-by: Alec Brown <alec.r.brown@oracle.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The return value of grub_loongarch64_stack_pop() is unsigned, so -1 should
not be used in the first place. Replacing with 0 is enough to avoid the
UB in this edge case.
Technically though, proper error handling is needed throughout the
management of the reloc stack, so no unexpected behavior will happen
even in case of malformed object code input (right now, pushes become
no-ops when the stack is full, and garbage results if the stack does not
contain enough operands for an op). The refactor would touch some more
places so would be best done in a separate series.
Fixes: CID 407777
Fixes: CID 407778
Signed-off-by: WANG Xuerui <git@xen0n.name>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Glenn suggested to rename the existing PCI_CLASS defines to have
explicit class and subclass names.
Suggested-by: Glenn Washburn <development@efficientek.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Loosely based on early_pci_serial_init() from Linux, allow GRUB to make
use of PCI serial devices.
Specifically, my Alderlake NUC exposes the Intel AMT SoL UART as a PCI
enumerated device but doesn't include it in the EFI tables.
Tested and confirmed working on a "Lenovo P360 Tiny" with Intel AMT
enabled. This specific machine has (from lspci -vv):
00:16.3 Serial controller: Intel Corporation Device 7aeb (rev 11) (prog-if 02 [16550])
DeviceName: Onboard - Other
Subsystem: Lenovo Device 330e
Control: I/O+ Mem+ BusMaster- SpecCycle- MemWINV- VGASnoop- ParErr- Stepping- SERR- FastB2B- DisINTx-
Status: Cap+ 66MHz+ UDF- FastB2B+ ParErr- DEVSEL=fast >TAbort- <TAbort- <MAbort- >SERR- <PERR- INTx-
Interrupt: pin D routed to IRQ 19
Region 0: I/O ports at 40a0 [size=8]
Region 1: Memory at b4224000 (32-bit, non-prefetchable) [size=4K]
Capabilities: [40] MSI: Enable- Count=1/1 Maskable- 64bit+
Address: 0000000000000000 Data: 0000
Capabilities: [50] Power Management version 3
Flags: PMEClk- DSI+ D1- D2- AuxCurrent=0mA PME(D0-,D1-,D2-,D3hot-,D3cold-)
Status: D0 NoSoftRst+ PME-Enable- DSel=0 DScale=0 PME-
Kernel driver in use: serial
From which the following config (/etc/default/grub) gets a working
serial setup:
GRUB_CMDLINE_LINUX="console=tty0 earlyprintk=pciserial,00:16.3,115200 console=ttyS0,115200"
GRUB_SERIAL_COMMAND="serial --port=0x40a0 --speed=115200"
GRUB_TERMINAL="serial console"
Documentation is added to note that serial devices found on the PCI bus will
be exposed as "pci,XX:XX.X" and how to find serial terminal logical names.
Also, some minor documentation improvements were added.
This can be tested in QEMU by adding a pci-serial device, e.g. using the option
"-device pci-serial".
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Tested-by: Glenn Washburn <development@efficientek.com>
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
On some systems /usr/share/dict/american-english can be larger than the
available space on the filesystem being tested (e.g. vfat12a). This
causes a failure of the filesystem test and is not a real test failure.
Instead, use dd to copy at most 1 MiB of data to the filesystem, which is
enough for our purposes and will not fill any of the tested filesystems.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The menu entry commands now have their own section. Change the wording in
the section that they were in to reflect this.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
On failure to allocate from grub_relocator_firmware_alloc_region() in
malloc_in_range() the function would stop enforcing the alignment, and
the following was returned:
lib/relocator.c:431: trying to allocate in 0x200000-0xffbf9fff aligned 0x200000 size 0x406000
lib/relocator.c:1197: allocated: 0x74de2000+0x406000
lib/relocator.c:1407: allocated 0x74de2000/0x74de2000
Fix this by making sure that target always contains a suitably aligned
address. After the change the return from the function is:
lib/relocator.c:431: trying to allocate in 0x200000-0xffb87fff aligned 0x200000 size 0x478000
lib/relocator.c:1204: allocated: 0x74c00000+0x478000
lib/relocator.c:1414: allocated 0x74c00000/0x74c00000
Fixes: 3a5768645c (First version of allocation from firmware)
Signed-off-by: Roger Pau Monné <roger.pau@citrix.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The access_size is part of a union, so doesn't technically exist for
a PIO port (i.e., not MMIO), but we set it anyways.
This doesn't cause a bug today because the other leg of the union
doesn't have anything overlapping with it now, but it's bad, I will
punish myself for writing it that way :-) In the meantime, fix this
and actually name the struct inside the union for clarity of intent
and to avoid such issue in the future.
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Commit f23bc6510 (Transform -C option to grub-mkstandalone to
--core-compress available in all grub-install flavours.) declared
a new long option for specifying the compression method to use for
the core image.
However, the option key has not been replaced in the parser function,
it still expects the old one formerly used by grub-mkstandalone.
Because of this the option is not recognized by any of the utils for
which it is listed as supported.
Signed-off-by: Ákos Nagy <nagyakos@outlook.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
When an invalid node size is detected in grub_hfsplus_mount(), data
pointer is freed. Thus, file->data is not set. The code should also
set the grub_errno when that happens to indicate an error and to avoid
accessing the uninitialized file->data in grub_file_close().
Signed-off-by: Lidong Chen <lidong.chen@oracle.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
A corrupted hfsplus can have a catalog key that is out of range. This
can lead to out of bound access when advancing the pointer to access
catalog file info. The valid range of a catalog key is specified in
HFS Plus Technical Note TN1150 [1].
[1] https://developer.apple.com/library/archive/technotes/tn/tn1150.html
Signed-off-by: Lidong Chen <lidong.chen@oracle.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The invalid btree node size can cause crashes when parsing the btree.
The fix is to ensure the btree node size is within the valid range
defined in the HFS Plus technical note, TN1150 [1].
[1] https://developer.apple.com/library/archive/technotes/tn/tn1150.html
Signed-off-by: Lidong Chen <lidong.chen@oracle.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The exfat-utils package is an older package complementing exfat-fuse, and
was the only exfat tools for a long time. The exfat filesystem testing code
was written with these tools in mind. A newer project exfatprogs appears to
be of better quality and functionality and was written to complement the
somewhat new exfat kernel module. Ideally we should be using the newer
exfatprogs. However, the command line interface for mkfs.exfat is different
between the two. So we can't use the exfatprogs tools until the test scripts
have been updated to account for this. Recommend installing exfat-utils
instead of exfatprogs for now.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
When trying to resolve DNS names into IP addresses, the DNS code fails
from time to time with the following error:
-------- 8< ---------------- 8< ---------------- 8< ---------------- 8< --------
error: ../../grub-core/net/dns.c:688:no DNS record found.
-------- 8< ---------------- 8< ---------------- 8< ---------------- 8< --------
This happens when both IPv4 and IPv6 queries are performed against the
DNS server (e.g. 8.8.8.8) but there is no IP returned for IPv6 query, as
shown below:
-------- 8< ---------------- 8< ---------------- 8< ---------------- 8< --------
grub> net_del_dns 192.168.122.1
grub> net_add_dns 8.8.8.8
grub> net_nslookup ipv4.test-ipv6.com
error: ../../grub-core/net/dns.c:688:no DNS record found.
grub> net_nslookup ipv4.test-ipv6.com
216.218.228.115
-------- 8< ---------------- 8< ---------------- 8< ---------------- 8< --------
The root cause is the code exiting prematurely when the data->addresses
buffer has been allocated in recv_hook(), even if there was no address
returned last time recv_hook() executed.
Signed-off-by: Renaud Métrich <rmetrich@redhat.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
When deleting the DNS server, we get the following error message:
-------- 8< ---------------- 8< ---------------- 8< ---------------- 8< --------
grub> net_del_dns 192.168.122.1
error: ../../grub-core/net/dns.c:646:no DNS reply received.
-------- 8< ---------------- 8< ---------------- 8< ---------------- 8< --------
This happens because the implementation is broken, it does a "add"
internally instead of a "delete".
Signed-off-by: Renaud Métrich <rmetrich@redhat.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
I ran the test suite on a 3A5000 desktop, a LoongArch architecture machine,
using Archlinux for LoongArch distro, see https://github.com/loongarchlinux.
Some software versions are:
* linux 6.3.0-rc4
* gcc 13.0.1 20230312
* binutils 2.40
* qemu 7.2.0
The test results of running "make check" with qemu 7.2 are as follows:
=================================
GRUB 2.11: ./test-suite.log
=================================
# TOTAL: 85
# PASS: 73
# SKIP: 8
# XFAIL: 0
# FAIL: 2
# XPASS: 0
# ERROR: 2
.. contents:: :depth: 2
ERROR: f2fs_test
================
mount: /tmp/grub-fs-tester.20230418175640563815408.f2fs.UDs/f2fs_rw: unknown filesystem type 'f2fs'.
dmesg(1) may have more information after failed mount system call.
MOUNT FAILED.
ERROR f2fs_test (exit status: 99)
FAIL: hfs_test
==============
recode: Request `utf8..macroman' is erroneous
mkfs.hfs: name required with -v option
FAIL hfs_test (exit status: 1)
ERROR: zfs_test
===============
zpool not installed; cannot test zfs.
ERROR zfs_test (exit status: 99)
SKIP: pata_test
===============
SKIP pata_test (exit status: 77)
SKIP: ahci_test
===============
SKIP ahci_test (exit status: 77)
SKIP: uhci_test
===============
SKIP uhci_test (exit status: 77)
SKIP: ohci_test
===============
SKIP ohci_test (exit status: 77)
SKIP: ehci_test
===============
SKIP ehci_test (exit status: 77)
SKIP: fddboot_test
==================
SKIP fddboot_test (exit status: 77)
SKIP: netboot_test
==================
SKIP netboot_test (exit status: 77)
SKIP: pseries_test
==================
SKIP pseries_test (exit status: 77)
FAIL: grub_func_test
====================
WARNING: Image format was not specified for '/tmp/grub-shell.HeTAD8Ty3U/grub.iso' and probing guessed raw.
Automatically detecting the format is dangerous for raw images, write operations on block 0 will be restricted.
Specify the 'raw' format explicitly to remove the restrictions.
Functional test failure: shift_test:
...
gfxterm_menu_640x480xi16:3 failed: 0xce34981e vs 0xd9f04953
tests/video_checksum.c:checksum:615: assert failed: 0 Checksum
gfxterm_menu_640x480xi16:2 failed: 0xa8fb749d vs 0xbf3fa5d0
tests/video_checksum.c:checksum:615: assert failed: 0 Checksum
gfxterm_menu_640x480xi16:1 failed: 0xce34981e vs 0xd9f04953
gfxterm_menu: FAIL
...
videotest_checksum:
videotest_checksum: PASS
exfctest:
exfctest: PASS
TEST FAILURE
FAIL grub_func_test (exit status: 1)
We got 2 errors:
* f2fs_test
The kernel uses 16k pages, causing failures when loading the f2fs kernel module,
see https://github.com/torvalds/linux/blob/master/fs/f2fs/super.c#L4670
This error can be ignored.
* zfs_test
zfs does not support the LoongArch architecture and is not compatible with the
6.3 kernel.
This error can be ignored.
We got 2 failures:
* hfs_test
I use recode 3.7.14-1 on Archlinux, running `recode -l` gives no output `MacRoman`,
so we get this error.
On Linux systems that support LoongArch, there is currently no need to use HFS,
so this failure can be ignored.
* grub_func_test
I don't know the reason for this failure. I guess it may be related to qemu's edk2.
In the previous review, I was told that the failure here is the expected behavior.
So, we can ignore this failure.
Signed-off-by: Xiaotian Wu <wuxiaotian@loongson.cn>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
This patch adds LoongArch to the GRUB build system and various tools,
so GRUB can be built on LoongArch as a UEFI application.
Signed-off-by: Zhou Yang <zhouyang@loongson.cn>
Signed-off-by: Xiaotian Wu <wuxiaotian@loongson.cn>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Add support for manipulating architectural cache and timers, and EFI
memory maps.
Signed-off-by: Zhou Yang <zhouyang@loongson.cn>
Signed-off-by: Xiaotian Wu <wuxiaotian@loongson.cn>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
A new set of relocation types was added in the LoongArch ELF psABI v2.00
spec [1], [2] to replace the stack-based scheme in v1.00. Toolchain
support is available from binutils 2.40 and gcc 13 onwards.
This patch adds support for the new relocation types, that are simpler
to handle (in particular, stack operations are gone). Support for the
v1.00 relocs are kept for now, for compatibility with older toolchains.
[1] https://github.com/loongson/LoongArch-Documentation/pull/57
[2] https://loongson.github.io/LoongArch-Documentation/LoongArch-ELF-ABI-EN.html#_appendix_revision_history
Signed-off-by: Xiaotian Wu <wuxiaotian@loongson.cn>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
This patch adds support of the stack-based LoongArch relocations
throughout GRUB, including tools, dynamic linkage, and support for
conversion of ELF relocations into PE ones. A stack machine is required
to handle these per the spec [1] (see the R_LARCH_SOP types), of which
a simple implementation is included.
These relocations are produced by binutils 2.38 and 2.39, while the newer
v2.00 relocs require more recent toolchain (binutils 2.40+ & gcc 13+, or
LLVM 16+). GCC 13 has not been officially released as of early 2023, so
support for v1.00 relocs are expected to stay relevant for a while.
[1] https://loongson.github.io/LoongArch-Documentation/LoongArch-ELF-ABI-EN.html#_relocations
Signed-off-by: Zhou Yang <zhouyang@loongson.cn>
Signed-off-by: Xiaotian Wu <wuxiaotian@loongson.cn>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
On entry, we need to save the system table pointer as well as our image
handle. Add an early startup file that saves them and then brings us
into our main function.
Signed-off-by: Zhou Yang <zhouyang@loongson.cn>
Signed-off-by: Xiaotian Wu <wuxiaotian@loongson.cn>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
This patch adds a setjmp implementation for LoongArch.
Signed-off-by: Zhou Yang <zhouyang@loongson.cn>
Signed-off-by: Sun Haiyong <sunhaiyong@loongson.cn>
Signed-off-by: Xiaotian Wu <wuxiaotian@loongson.cn>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
GRUB since 93a786a00 (kern/efi/sb: Enforce verification of font files)
has enforced verification of font files in secure boot mode. In order to
continue to be able to load some default fonts, vendors may bundle them
with their signed EFI image by adding them to the built-in memdisk.
This change makes the font loader try loading fonts from the memdisk
before the prefix path when attempting to load a font file by specifying
its filename, which avoids having to make changes to GRUB configurations
in order to accommodate memdisk bundled fonts. It expects the directory
structure to be the same as fonts stored in the prefix path,
i.e. /fonts/<name>.pf2.
Signed-off-by: Chris Coulson <chris.coulson@canonical.com>
Reviewed-by: Steve McIntyre <93sam@debian.org>
Tested-by: Steve McIntyre <93sam@debian.org>
Reviewed-by: Robbie Harwood <rharwood@redhat.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Allow specifying port numbers for http and tftp paths and allow IPv6
addresses to be recognized with brackets around them, which is required
to specify a port number.
Co-authored-by: Aaron Miller <aaronmiller@fb.com>
Signed-off-by: Aaron Miller <aaronmiller@fb.com>
Co-authored-by: Peter Jones <pjones@redhat.com>
Signed-off-by: Peter Jones <pjones@redhat.com>
Signed-off-by: Robbie Harwood <rharwood@redhat.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The notation introduced in ac8a37dda (net/http: Allow use of non-standard
TCP/IP ports) contradicts that used in downstream distributions including
Fedora, RHEL, Debian, Ubuntu, and others. Revert it and apply the downstream
notation which was originally proposed to the GRUB in 2016.
This reverts commit ac8a37dda (net/http: Allow use of non-standard TCP/IP ports).
Signed-off-by: Robbie Harwood <rharwood@redhat.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
This fixes the GRUB on Coreboot on HP EliteBooks by implementing
a 200 ms timeout. The GRUB used to hang.
Fixes: https://ticket.coreboot.org/issues/141
Signed-off-by: Riku Viitanen <riku.viitanen@protonmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
In filesystem timestamp test, a check is done to verify that the timestamp
for a file as reported in Linux by the filesystem is within a few seconds
of the timestamp as reported by GRUB. This is done by grepping the output
of GRUB's ls command for the timestamp as reported by the filesystem in
Linux and for each of 3 seconds past that timestamp. All of these checks
except one redirect the output of grep to /dev/null. Fix this exception
to behave as the other checks.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The transform_sector() function is not very clear in what it's doing
and confusing. The GRUB already has a function which is doing the same
thing in a very self explanatory way, i.e., grub_disk_to_native_sector().
So, it's much better to use self explanatory one than transform_sector().
Signed-off-by: Mukesh Kumar Chaurasiya <mchauras@linux.vnet.ibm.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The ISO filesystem image iso9660_early_ce.iso exposes the unusual
situation that the Rock Ridge name entry of its only file is located
after a CE entry which points to the next continuation area.
The correct behavior is to read the Rock Ridge name and to only then
load the next continuation area. If GRUB performs this correctly, then
the name "RockRidgeName:x" will be read and reported by grub-fstest.
If GRUB wrongly performs the CE hop immediately when encountering the CE
entry, then the dull ISO 9660 name "rockridg" will not be overridden and
be put out by grub-fstest.
Signed-off-by: Thomas Schmitt <scdbackup@gmx.net>
Tested-by: Lidong Chen <lidong.chen@oracle.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The SUSP specs demand that the reading of the next SUSP area which is
depicted by a CE entry shall be delayed until reading of the current
SUSP area is completed. Up to now GRUB immediately ends reading of the
current area and loads the new one. So, buffer the parameters of a found
CE entry and perform checks and reading of new data only after the
reader loop has ended.
Signed-off-by: Thomas Schmitt <scdbackup@gmx.net>
Tested-by: Lidong Chen <lidong.chen@oracle.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
This patch enables multiple options in Vec5 which are required and
solves the boot issues seen on some machines which are looking for
these specific options.
1. LPAR: Client program supports logical partitioning and
associated hcall()s.
2. SPLPAR: Client program supports the Shared
Processor LPAR Option.
3. DYN_RCON_MEM: Client program supports the
“ibm,dynamic-reconfiguration-memory” property and it may be
presented in the device tree.
4. LARGE_PAGES: Client supports pages larger than 4 KB.
5. DONATE_DCPU_CLS: Client supports donating dedicated processor cycles.
6. PCI_EXP: Client supports PCI Express implementations
utilizing Message Signaled Interrupts (MSIs).
7. CMOC: Enables the Cooperative Memory Over-commitment Option.
8. EXT_CMO: Enables the Extended Cooperative Memory Over-commit Option.
9. ASSOC_REF: Enables “ibm,associativity” and
“ibm,associativity-reference-points” properties.
10. AFFINITY: Enables Platform Resource Reassignment Notification.
11. NUMA: Supports NUMA Distance Lookup Table Option.
12. HOTPLUG_INTRPT: Supports Hotplug Interrupts.
13. HPT_RESIZE: Enable Hash Page Table Resize Option.
14. MAX_CPU: Defines maximum number of CPUs supported.
15. PFO_HWRNG: Supports Random Number Generator.
16. PFO_HW_COMP: Supports Compression Engine.
17. PFO_ENCRYPT: Supports Encryption Engine.
18. SUB_PROCESSORS: Supports Sub-Processors.
19. DY_MEM_V2: Client program supports the “ibm,dynamic-memory-v2” property in the
“ibm,dynamic-reconfiguration-memory” node and it may be presented in the device tree.
20. DRC_INFO: Client program supports the “ibm,drc-info” property definition and it may be
presented in the device tree.
Signed-off-by: Avnish Chouhan <avnish@linux.vnet.ibm.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
This patch converts the plain numbers used in Vec5 properties to constants.
1. LPAR: Client program supports logical partitioning and
associated hcall()s.
2. SPLPAR: Client program supports the Shared
Processor LPAR Option.
3. CMO: Enables the Cooperative Memory Over-commitment Option.
4. MAX_CPU: Defines maximum number of CPUs supported.
Signed-off-by: Avnish Chouhan <avnish@linux.vnet.ibm.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Per systemctl(1), it "is asynchronous; it will return after the reboot
operation is enqueued, without waiting for it to complete". This differs
from kexec(8), which calls reboot(2) and therefore does not return.
When not using fallback, this confusingly results in:
error trying to perform 'systemctl kexec': 0
Aborted. Press any key to exit.
on screen for a bit, followed by successful kexec.
To reduce the likelihood of hitting this case, add a delay on successful
return. Ultimately, the systemd interface is racy: we can't avoid it
entirely unless we never fallback on success.
Signed-off-by: Robbie Harwood <rharwood@redhat.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
When the tpm module is loaded, the verifier reads entire file into
memory, measures it and uses verified content as a backing buffer for
file accesses. However, this process may result in high memory
utilization for file operations, sometimes causing a system to run out
of memory which may finally lead to boot failure. To address this issue,
among others, the commit 887f98f0d (mm: Allow dynamically requesting
additional memory regions) have optimized memory management by
dynamically allocating heap space to maximize memory usage and reduce
threat of memory exhaustion. But in some cases problems may still arise,
e.g., when large ISO images are mounted using loopback or when dealing
with embedded systems with limited memory resources.
Unfortunately current implementation of the tpm module doesn't allow
elimination of the back buffer once it is loaded. Even if the TPM device
is not present or it has been explicitly disabled. This may unnecessary
allocate a lot memory. To solve this issue, a patch has been developed
to detect the TPM status at module load and skip verifier registration
if the device is missing or deactivated. This prevents allocation of
memory for the back buffer, avoiding wasting memory when no real measure
boot functionality is performed. Disabling the TPM device in the system
can reduce memory usage in the GRUB. It is useful in scenarios where
high memory utilization is a concern and measurements of loaded
artifacts are not necessary.
Signed-off-by: Michael Chang <mchang@suse.com>
Signed-off-by: Stefan Berger <stefanb@linux.ibm.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Now that the gdb_grub script uses the Python API in GDB, a GDB with Python
support must be used. Note that this means a GDB with version greater than
7.0 must be used. This should not be an issue since that was released over
a decade ago. Also, the minimum version of Python must be 3.5, which was
released around 8 years ago.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
RISC-V doesn't have to do anything very different from other architectures
to loader EFI stub linux kernel. As a result, just use the common linux
loader instead of defining a RISC-V specific linux loader.
Signed-off-by: Atish Patra <atishp@rivosinc.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The arch specific image header details are not very useful as most of
the GRUB just looks at the PE/COFF spec parameters (PE32 magic and
header offset).
Remove the arch specific images headers and define a generic arch
headers that provide enough PE/COFF fields for the GRUB to parse
kernel images correctly.
Signed-off-by: Atish Patra <atishp@rivosinc.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
ARM64 linux loader code is written in such a way that it can be reused
across different architectures without much change. Move it to common
code so that RISC-V doesn't have to define a separate loader.
Signed-off-by: Atish Patra <atishp@rivosinc.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
In grub-module-verifierXX.c, the function grub_module_verifyXX() performs an
initial check that the ELF section headers are within the module's size, but
doesn't check if the sections being accessed have contents that are within the
module's size. In particular, we need to check that sh_offset and sh_size are
less than the module's size. However, for some section header types we don't
need to make these checks. For the type SHT_NULL, the section header is marked
as inactive and the rest of the members within the section header have undefined
values, so we don't need to check for sh_offset or sh_size. In the case of the
type SHT_NOBITS, sh_offset has a conceptual offset which may be beyond the
module size. Also, this type's sh_size may have a non-zero size, but a section
of this type will take up no space in the module. This can all be checked in the
function get_shdr(), but in order to do so, the parameter module_size must be
added to functions so that the value of the module size can be used in
get_shdr() from grub_module_verifyXX().
Also, had to rework some for loops to ensure the index passed to get_shdr() is
within bounds.
Signed-off-by: Alec Brown <alec.r.brown@oracle.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Add symbols for boot.image, disk.image, and lzma_decompress.image if the
target is i386-pc. This is only done for i386-pc because that is the only
target that uses the images. By loading the symbols for these images,
these images can be more easily debugged by allowing the setting of break-
points in that code and to see easily get the value of data symbols.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
This will let users know that the GDB session is using the GRUB gdb scripts.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
A new command, run_on_start, for things to do before GRUB starts executing.
Currently, this is setting up the loading of module symbols as they are
loaded and allowing user-defined script to be run if a command named
"onstart" exists.
On some platforms, notably x86, software breakpoints set in GDB before
the GRUB image is loaded will be cleared when the image is loaded. This
is because the breakpoints work by overwriting the memory of the break-
point location with a special instruction which when hit will cause the
debugger to stop execution. Just before execution is resumed by the
debugger, the original instruction bytes are put back. When a breakpoint
is set before the GRUB image is loaded, the special debugger instruction
will be written to memory and when the GRUB image is loaded by the
firmware, which has no knowledge of the debugger, the debugger instruction
is overwritten. To the GDB user, GDB will show the breakpoint as set, but
it will never be hit. Furthermore, GDB now becomes confused, such that
even deleting and re-setting the breakpoint after the GRUB image is loaded
will not allow for a working breakpoint.
To work around this, in run_on_start, first a watchpoint is set on _start,
which will be triggered when the firmware starts loading the GRUB image.
When the _start watchpoint is hit, the current breakpoints are saved to a
file and then deleted by GDB before they can be overwritten by the firmware
and confuse GDB. Then a temporary software breakpoint is set on _start,
which will get triggered when the firmware hands off to GRUB to execute. In
that breakpoint load the previously saved and deleted breakpoints now that
there is no worry of them getting overwritten by the firmware. This is
needed for runtime_load_module to work when it is run before the GRUB image
is loaded.
Note that watchpoints are generally types of hardware breakpoints on x86, so
its deleted as soon as it gets triggered so that a minimal set of hardware
breakpoints are used, allowing more for the user.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Many targets, such as EFI, load GRUB at addresses that are determined at
runtime. So the load addresses in kernel.exec will almost certainly be
wrong. Given the address of the start of the text segment, these
functions will tell GDB to load the symbols at the proper locations. It
is left up to the user to determine how to get the text address of the
loaded GRUB image.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Remove gmodule.pl and rewrite as a python in gdb_helper.py. This removes
Perl dependency for the GRUB GDB script, but adds Python as a dependency.
This is more desirable because Python is tightly integrated with GDB and
can do things not even available to GDB native scripting language. GDB must
be built with Python, however this is not a major limitation because every
major distro non-end-of-life versions build GDB with Python support. And GDB
has had support for Python since around 7.1-ish, which is about a decade.
This re-implementation has an added feature. If there is a user defined
command named "onload_<module name>", then that command will be executed
after the symbols for the specified module are loaded. When debugging a
module it can be desirable to set break points on code in the module.
This is difficult in GRUB because, at GDB start, the module is not loaded
and on EFI platforms its not known ahead of time where the module will
be loaded. So allow users to create an "onload_<modname>" command which
will be run when the module with name "modname" is loaded.
Another addition is a new convenience function is defined
$is_user_command(), which returns true if its string argument is
the name of a user-defined command.
A secondary benefit of these changes is that the script does not write
temporary files and has better error handling capabilities.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The gdb_grub script was originally meant to be run once when GDB first
starts up via the -x argument. So it runs commands unconditionally
assuming that the script has not been run before. Its nice to be able
to source the script again when developing the script to modify/add
commands. So only run the commands not defined in user-defined commands,
if a variable $runonce has already been set and when those commands have
been run to set $runonce.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
There are broadly two classes of targets to consider when loading symbols
for GRUB, targets that determine where to load GRUB at runtime
(dynamically positioned) and those that do not (statically positioned).
For statically positioned targets, symbol loading is determined at link
time, so nothing more needs to be known to load the symbols. For
dynamically positioned targets, such as EFI targets, at runtime symbols
should be offset by an amount that depends on where the runtime chose to
load GRUB.
It is important to not load symbols statically for dynamic targets
because then when subsequently loading the symbols correctly one must
take care to remove the existing static symbols, otherwise there will be
two sets of symbols and GDB seems to prefer the ones loaded first (i.e.
the static ones).
Use autoconf variables to generate a gdb_grub for a particular target,
which conditionally run startup code depending on if the target uses
static or dynamic loading.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
By moving this code into a function, it can be run re-utilized while gdb is
running, not just when loading the script. This will also be useful in
some following changes which will make a separate script path for targets
which statically vs dynamically position GRUB code.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The gcc build has failed for 32-bit host (e.g. i386-emu and arm-emu)
due to mismatch between format specifier and data type.
../grub-core/osdep/devmapper/getroot.c: In function
'grub_util_pull_devmapper':
../grub-core/osdep/devmapper/getroot.c:265:75: error: format '%lu'
expects argument of type 'long unsigned int', but argument 2 has type
'int' [-Werror=format=]
../grub-core/osdep/devmapper/getroot.c:276:80: error: format '%lu'
expects argument of type 'long unsigned int', but argument 2 has type
'int' [-Werror=format=]
This patch fixes the problem by casting the type of calculated offset to
grub_size_t and use platform PRIuGRUB_SIZE as format specifier.
Signed-off-by: Michael Chang <mchang@suse.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Add support for trusted boot using a vTPM 2.0 on the IBM IEEE1275
PowerPC platform. With this patch grub now measures text and binary data
into the TPM's PCRs 8 and 9 in the same way as the x86_64 platform
does.
This patch requires Daniel Axtens's patches for claiming more memory.
Note: The tpm_init() function cannot be called from GRUB_MOD_INIT() since
it does not find the device nodes upon module initialization and
therefore the call to tpm_init() must be deferred to grub_tpm_measure().
For vTPM support to work on PowerVM, system driver levels 1010.30
or 1020.00 are required.
Note: Previous versions of firmware levels with the 2hash-ext-log
API call have a bug that, once this API call is invoked, has the
effect of disabling the vTPM driver under Linux causing an error
message to be displayed in the Linux kernel log. Those users will
have to update their machines to the firmware levels mentioned
above.
Cc: Eric Snowberg <eric.snowberg@oracle.com>
Signed-off-by: Stefan Berger <stefanb@linux.ibm.com>
Signed-off-by: Daniel Axtens <dja@axtens.net>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Tested-by: Nageswara R Sastry <rnsastry@linux.ibm.com>
Reviewed-by: Robbie Harwood <rharwood@redhat.com>
When working on memory, it's nice to be able to test your work.
Add a memtest module. When compiled with --enable-mm-debug, it exposes
3 commands:
* lsmem - print all allocations and free space in all regions
* lsfreemem - print free space in all regions
* stress_big_allocs - stress test large allocations:
- how much memory can we allocate in one chunk?
- how many 1MB chunks can we allocate?
- check that gap-filling works with a 1MB aligned 900kB alloc + a
100kB alloc.
Signed-off-by: Daniel Axtens <dja@axtens.net>
Signed-off-by: Stefan Berger <stefanb@linux.ibm.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Tested-by: Nageswara R Sastry <rnsastry@linux.ibm.com>
Reviewed-by: Robbie Harwood <rharwood@redhat.com>
As a legacy support, if the vector 5 is not implemented, Power Hypervisor will
consider the max CPUs as 64 instead 256 currently supported during
client-architecture-support negotiation.
This patch implements the vector 5 and set the MAX CPUs to 256 while setting the
others values to 0 (default).
Signed-off-by: Diego Domingos <diegodo@linux.vnet.ibm.com>
Acked-by: Daniel Axtens <dja@axtens.net>
Signed-off-by: Stefan Berger <stefanb@linux.ibm.com>
Signed-off-by: Avnish Chouhan <avnish@linux.vnet.ibm.com>
Tested-by: Nageswara R Sastry <rnsastry@linux.ibm.com>
Reviewed-by: Robbie Harwood <rharwood@redhat.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
On powerpc-ieee1275, we are running out of memory trying to verify
anything. This is because:
- we have to load an entire file into memory to verify it. This is
difficult to change with appended signatures.
- We only have 32MB of heap.
- Distro kernels are now often around 30MB.
So we want to be able to claim more memory from OpenFirmware for our heap
at runtime.
There are some complications:
- The grub mm code isn't the only thing that will make claims on
memory from OpenFirmware:
* PFW/SLOF will have claimed some for their own use.
* The ieee1275 loader will try to find other bits of memory that we
haven't claimed to place the kernel and initrd when we go to boot.
* Once we load Linux, it will also try to claim memory. It claims
memory without any reference to /memory/available, it just starts
at min(top of RMO, 768MB) and works down. So we need to avoid this
area. See arch/powerpc/kernel/prom_init.c as of v5.11.
- The smallest amount of memory a ppc64 KVM guest can have is 256MB.
It doesn't work with distro kernels but can work with custom kernels.
We should maintain support for that. (ppc32 can boot with even less,
and we shouldn't break that either.)
- Even if a VM has more memory, the memory OpenFirmware makes available
as Real Memory Area can be restricted. Even with our CAS work, an LPAR
on a PowerVM box is likely to have only 512MB available to OpenFirmware
even if it has many gigabytes of memory allocated.
What should we do?
We don't know in advance how big the kernel and initrd are going to be,
which makes figuring out how much memory we can take a bit tricky.
To figure out how much memory we should leave unused, I looked at:
- an Ubuntu 20.04.1 ppc64le pseries KVM guest:
vmlinux: ~30MB
initrd: ~50MB
- a RHEL8.2 ppc64le pseries KVM guest:
vmlinux: ~30MB
initrd: ~30MB
So to give us a little wriggle room, I think we want to leave at least
128MB for the loader to put vmlinux and initrd in memory and leave Linux
with space to satisfy its early allocations.
Allow other space to be allocated at runtime.
Tested-by: Stefan Berger <stefanb@linux.ibm.com>
Signed-off-by: Daniel Axtens <dja@axtens.net>
Tested-by: Nageswara R Sastry <rnsastry@linux.ibm.com>
Reviewed-by: Robbie Harwood <rharwood@redhat.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
This was apparently "required by some firmware": commit dc94685009
(2007-02-12 Hollis Blanchard <hollis@penguinppc.org>).
It's not clear what firmware that was, and what platform from 14 years ago
which exhibited the bug then is still both in use and buggy now.
It doesn't cause issues on qemu (mac99 or pseries) or under PFW for Power8.
I don't have access to old Mac hardware, but if anyone feels especially
strongly we can put it under some feature flag. I really want to disable
it under pseries because it will mess with region merging.
Signed-off-by: Daniel Axtens <dja@axtens.net>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Tested-by: Nageswara R Sastry <rnsastry@linux.ibm.com>
Reviewed-by: Robbie Harwood <rharwood@redhat.com>
On PowerVM, the first time we boot a Linux partition, we may only get
256MB of real memory area, even if the partition has more memory.
This isn't enough to reliably verify a kernel. Fortunately, the Power
Architecture Platform Reference (PAPR) defines a method we can call to ask
for more memory: the broad and powerful ibm,client-architecture-support
(CAS) method.
CAS can do an enormous amount of things on a PAPR platform: as well as
asking for memory, you can set the supported processor level, the interrupt
controller, hash vs radix mmu, and so on.
If:
- we are running under what we think is PowerVM (compatible property of /
begins with "IBM"), and
- the full amount of RMA is less than 512MB (as determined by the reg
property of /memory)
then call CAS as follows: (refer to the Linux on Power Architecture
Reference, LoPAR, which is public, at B.5.2.3):
- Use the "any" PVR value and supply 2 option vectors.
- Set option vector 1 (PowerPC Server Processor Architecture Level)
to "ignore".
- Set option vector 2 with default or Linux-like options, including a
min-rma-size of 512MB.
- Set option vector 3 to request Floating Point, VMX and Decimal Floating
point, but don't abort the boot if we can't get them.
- Set option vector 4 to request a minimum VP percentage to 1%, which is
what Linux requests, and is below the default of 10%. Without this,
some systems with very large or very small configurations fail to boot.
This will cause a CAS reboot and the partition will restart with 512MB
of RMA. Importantly, grub will notice the 512MB and not call CAS again.
Notes about the choices of parameters:
- A partition can be configured with only 256MB of memory, which would
mean this request couldn't be satisfied, but PFW refuses to load with
only 256MB of memory, so it's a bit moot. SLOF will run fine with 256MB,
but we will never call CAS under qemu/SLOF because /compatible won't
begin with "IBM".)
- unspecified CAS vectors take on default values. Some of these values
might restrict the ability of certain hardware configurations to boot.
This is why we need to specify the VP percentage in vector 4, which is
in turn why we need to specify vector 3.
Finally, we should have enough memory to verify a kernel, and we will
reach Linux. One of the first things Linux does while still running under
OpenFirmware is to call CAS with a much fuller set of options (including
asking for 512MB of memory). Linux includes a much more restrictive set of
PVR values and processor support levels, and this CAS invocation will likely
induce another reboot. On this reboot grub will again notice the higher RMA,
and not call CAS. We will get to Linux again, Linux will call CAS again, but
because the values are now set for Linux this will not induce another CAS
reboot and we will finally boot all the way to userspace.
On all subsequent boots, everything will be configured with 512MB of RMA,
so there will be no further CAS reboots from grub. (phyp is super sticky
with the RMA size - it persists even on cold boots. So if you've ever booted
Linux in a partition, you'll probably never have grub call CAS. It'll only
ever fire the first time a partition loads grub, or if you deliberately lower
the amount of memory your partition has below 512MB.)
Signed-off-by: Daniel Axtens <dja@axtens.net>
Signed-off-by: Stefan Berger <stefanb@linux.ibm.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Tested-by: Nageswara R Sastry <rnsastry@linux.ibm.com>
Reviewed-by: Robbie Harwood <rharwood@redhat.com>
Found during a test build on Debian/hurd-i386 with --disable-werror enabled:
In file included from grub-core/osdep/getroot.c:12:
grub-core/osdep/hurd/getroot.c: In function ‘grub_util_find_hurd_root_device’:
grub-core/osdep/hurd/getroot.c:126:13: error: unused variable ‘next’ [-Werror=unused-variable]
126 | char *next;
| ^~~~
grub-core/osdep/hurd/getroot.c:125:14: error: unused variable ‘size’ [-Werror=unused-variable]
125 | size_t size;
| ^~~~
Fixes: e981b0a24 (osdep/hurd/getroot: Use "part:" qualifier)
Signed-off-by: John Paul Adrian Glaubitz <glaubitz@physik.fu-berlin.de>
Reviewed-by: Samuel Thibault <samuel.thibault@ens-lyon.org>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
This prevents load_all_modules from failing when called before any
modules have been loaded. Failures in GDB user-defined functions cause
any function which called them to also fail.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
GDB logging is redirected to write .segments.tmp, which means that GDB
will wrap lines longer than what it thinks is the screen width
(typically 80 characters). When wrapping does occur it causes gmodule.pl
to misbehave. So disable line wrapping by using GDB's "with" command so
that its guaranteed to return the width to the previous value upon
command completion.
Also disable command tracing when dumping the module sections because that
output will go to .segments.tmp and thus cause gmodule.pl to misbehave.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
An error in any GDB command causes it to immediately abort with an error,
this includes any command that calls that command. This leads to an issue
in dump_module_sections where an error causes the command to exit without
turning off file redirection. The user then ends up with a GDB command
line where commands output nothing to the console.
Instead do the work of dump_module_sections in the command
dump_module_sections_helper and run the command using GDB's pipe command
which does the redirection and undoes the redirection when it finishes
regardless of any errors in the command.
Also, remove .segments.tmp file prior to loading modules in case one was
left from a previous run.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
On EFI platforms where EFI calls do not require a wrapper (notably i386-efi
and arm64-efi), the func argument needs to be wrapped in parenthesis to
allow valid syntax when func is an expression which evaluates to a function
pointer. On EFI platforms that do need a wrapper, this was never an issue
because func is passed to the C function wrapper as an argument and thus
does not need parenthesis to be evaluated.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The "addr" is used to request the memory with specific ranges but the real
loadable address come from the relocator. Thus, print the final retrieved
addresses, virtual and physical, for initrd.
On the occasion migrate to PRIxGRUB_ADDR and PRIxGRUB_SIZE format specifiers.
Signed-off-by: Jeremy Szu <jeremy.szu@canonical.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The functional test requires unicode.pf2 to run successfully, so
explicitly have the test return ERROR when its not found.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Tested-by: Thomas Schmitt <scdbackup@gmx.net>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Tests should be SKIP'd only when they do not apply to a particular target.
Hard errors are for when the test should run but can not be setup properly.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
These are not added to grub-fs-tester because they are not generated and
none of the filesystem tests are run on these ISOs. The test is to run the
command "ls /" on the ISO, and a failure is determined if the command
times out, has non-zero return value or has any output.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Tested-by: Thomas Schmitt <scdbackup@gmx.net>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The disk sector size provided by sysfs file system considers the sector
size of 512 irrespective of disk sector size, thus causing the read by
the GRUB to an incorrect offset from what was originally intended.
Considering the 512 sector size of sysfs data the actual sector needs to
be modified corresponding to disk sector size.
Signed-off-by: Mukesh Kumar Chaurasiya <mchauras@linux.vnet.ibm.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
By using a shell variable that is set once by the expansion of an autoconf
variable, the resulting script is more readable.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
In the function send_dhcp_packet(), added an error check for the return
value of grub_netbuff_push().
Fixes: CID 404614
Signed-off-by: Alec Brown <alec.r.brown@oracle.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
We do a lot of math about heap growth in hot path of grub_memalign().
However, the result is only used if out of memory is encountered, which
is seldom.
This patch moves these calculations away from hot path. These
calculations are now only done if out of memory is encountered. This
change can also help compiler to optimize integer overflow checks away.
Signed-off-by: Zhang Boyang <zhangboyang.id@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
When grub_memalign() encounters out-of-memory, it will try
grub_mm_add_region_fn() to request more memory from system firmware.
However, it doesn't preallocate memory space for future allocation
requests. In extreme cases, it requires one call to
grub_mm_add_region_fn() for each memory allocation request. This can
be very slow.
This patch introduces GRUB_MM_HEAP_GROW_EXTRA, the minimal heap growth
granularity. The new region size is now set to the bigger one of its
original value and GRUB_MM_HEAP_GROW_EXTRA. Thus, it will result in some
memory space preallocated if current allocations request is small.
The value of GRUB_MM_HEAP_GROW_EXTRA is set to 1MB. If this value is
smaller, the cost of small memory allocations will be higher. If this
value is larger, more memory will be wasted and it might cause
out-of-memory on machines with small amount of RAM.
Signed-off-by: Zhang Boyang <zhangboyang.id@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
When grub_memalign() encounters out-of-memory, it will try
grub_mm_add_region_fn() to request more memory from system firmware.
However, the size passed to it doesn't take region management overhead
into account. Adding a memory area of "size" bytes may result in a heap
region of less than "size" bytes really available. Thus, the new region
may not be adequate for current allocation request, confusing
out-of-memory handling code.
This patch introduces GRUB_MM_MGMT_OVERHEAD to address the region
management overhead (e.g. metadata, padding). The value of this new
constant must be large enough to make sure grub_memalign(align, size)
always succeeds after a successful call to
grub_mm_init_region(addr, size + align + GRUB_MM_MGMT_OVERHEAD),
for any given addr and size (assuming no integer overflow).
The size passed to grub_mm_add_region_fn() is now correctly adjusted,
thus if grub_mm_add_region_fn() succeeded, current allocation request
can always succeed.
Signed-off-by: Zhang Boyang <zhangboyang.id@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
When re-running a failed test, even the non-standard grub-shell QEMU
arguments should be preserved in the run.sh to more precisely replay
the failed test run.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Now it becomes trivial to re-run a test from the output in its working
directory. This also makes it easy to send a reproducible failing test to
the mailing list. This has allowed a refactor so that the duplicated code
to call QEMU has be condensed (e.g. the use of timeout and file descriptor
redirection). The run.sh script will pass any arguments given to QEMU.
This allows QEMU to be easily started in a state ready for GDB to be
attached.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
This allows turning on shell tracing for grub-shell and grub-fs-tester
when its not practical or not possible to use command line arguments
(e.g. from "make check"). Turn on tracing when the envvar is an integer
greater than 1, since these can generate a lot of output. Since this
change uses the environment variables to set the default value for debug
in grub-shell, this allows enabling grub-shell's debug mode which will
preserve various generated output files that are helpful for debugging
tests.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
If processing of a SUSP CE entry leads to a continuation area which
begins by entry CE or ST, then these entries were skipped without
interpretation. In case of CE this would lead to premature end of
processing the SUSP entries of the file. In case of ST this could
cause following non-SUSP bytes to be interpreted as SUSP entries.
Signed-off-by: Thomas Schmitt <scdbackup@gmx.net>
Tested-by: Lidong Chen <lidong.chen@oracle.com>
Reviewed-by: Thomas Schmitt <scdbackup@gmx.net>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
An SL entry consists of the entry info and the component area.
The entry info should take up 5 bytes instead of sizeof(*entry).
The area after the first 5 bytes is the component area. It is
incorrect to use the sizeof(*entry) to check the entry boundary.
Signed-off-by: Lidong Chen <lidong.chen@oracle.com>
Reviewed-by: Thomas Schmitt <scdbackup@gmx.net>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Added a check for the SP entry data boundary before reading it.
Signed-off-by: Lidong Chen <lidong.chen@oracle.com>
Reviewed-by: Thomas Schmitt <scdbackup@gmx.net>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
In the code, the for loop advanced the entry pointer to the next entry before
checking if the next entry is within the system use area boundary. Another
issue in the code was that there is no check for the size of system use area.
For a corrupted system, the size of system use area can be less than the size
of minimum SUSP entry size (4 bytes). These can cause buffer overrun. The fixes
added the checks to ensure the read is valid and within the boundary.
Signed-off-by: Lidong Chen <lidong.chen@oracle.com>
Reviewed-by: Thomas Schmitt <scdbackup@gmx.net>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
There is no check for the end of block when reading
directory extents. It resulted in read_node() always
read from the same offset in the while loop, thus
caused infinite loop. The fix added a check for the
end of the block and ensure the read is within directory
boundary.
Signed-off-by: Lidong Chen <lidong.chen@oracle.com>
Reviewed-by: Thomas Schmitt <scdbackup@gmx.net>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The logical sector size used by LUKS1 is 512 bytes and LUKS2 uses 512 to
4069 bytes. The default password used is "pass", but can be overridden
by setting the PASS environment variable. The device mapper name is set
to the name of the temp directory so that its easy to correlate device
mapper name with a particular test run. Also since this name is unique
per test run, multiple simultaneous test runs are allowed.
Note that cryptsetup is passing the --disable-locks parameter to allow
cryptsetup run successfully when /run/lock/cryptsetup is not accessible.
Since the device mapper name is unique per test run, there is no need to
worry about locking the device to serialize access.
Signed-off-by: Pierre-Louis Bonicoli <pierre-louis.bonicoli@libregerbil.fr>
Tested-by: Glenn Washburn <development@efficientek.com>
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Patrick Steinhardt <ps@pks.im>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
This lets a LUKS2 cryptodisk have its cipher and hash filled out,
otherwise they wouldn't be initialized if cheat mounted.
Signed-off-by: Josselin Poiret <dev@jpoiret.xyz>
Tested-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Patrick Steinhardt <ps@pks.im>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Changes UUID comparisons so that LUKS1 and LUKS2 are both recognized
as being LUKS cryptodisks.
Signed-off-by: Josselin Poiret <dev@jpoiret.xyz>
Tested-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Patrick Steinhardt <ps@pks.im>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
When using grub-probe with cryptodisk, the mapped block device from the host
is used directly instead of decrypting the source device in GRUB code.
In that case, the sector size and count of the host device needs to be used.
This is especially important when using LUKS2, which does not assign
total_sectors and log_sector_size when scanning, but only later when the
segments in the JSON area are evaluated. With an unset log_sector_size,
grub_device_open() complains.
This fixes grub-probe failing with
"error: sector sizes of 1 bytes aren't supported yet.".
Signed-off-by: Fabian Vogt <fvogt@suse.de>
Reviewed-by: Patrick Steinhardt <ps@pks.im>
Tested-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Patrick Steinhardt <ps@pks.im>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Oops. You're allowed to have up to n = NAT_JOURNAL_ENTRIES entries
_inclusive_, because the loop below uses i < n, not i <= n. D'oh.
Fixes: 4bd9877f62 (fs/f2fs: Do not read past the end of nat journal entries)
Reported-by: программист нект <programmer11180@programist.ru>
Tested-by: программист нект <programmer11180@programist.ru>
Signed-off-by: Daniel Axtens <dja@axtens.net>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
When building .img files, a .interp section from the .image files will
sometimes be copied into the .img file. This additional section pushes
the .img file beyond the 512-byte limit and causes grub-install to fail
to run for i386-pc platforms.
Signed-off-by: Nicholas Vinson <nvinson234@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The grub_cmd_cryptomount make check test performs some functional testing
of cryptomount and by extension the underlying cryptodisk infrastructure.
A utility test script named grub-shell-luks-tester is created to handle the
complexities of the testing, making it simpler to add new test cases in
grub_cmd_cryptomount.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
This allows test case scripts to use the appropriate halt command for
the built architecture to end execution early. Otherwise, test case
scripts have no way to know the appropriate mechanism for halting the
test case early.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
When turning on shell tracing the trim line will be output before we
actually want to start the trim. However, in this case the trim line never
starts from the beginning of the line. So start trimming from the correct
line by matching from the beginning of the line.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
This will be useful for tests that have unwanted output from setup. This is
not documented because its only intended to be internal at the moment. Also,
--no-trim is allowed to explicitly turn off trim.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
This keeps the generated files to aid in diagnosing the source of the failure.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
This allows us to test if unexpected output in test scripts is because of
a bug in GRUB, because there was an error in QEMU, or QEMU was killed due
to a timeout.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The "transparent" parameter to grub_gzio_open() was removed in 2010, fc2ef1172c
(* grub-core/io/gzio.c (grub_gzio_open): Removed "transparent" parameter.)
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Commit f5759a878 (normal/help: Add paging instructions to normal and help
prompts) changed the output of the help command, which broke the help
test. This change allows the test to pass.
On the occasion do s/outpu/output/.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
We currently rely on some pretty fragile comparison by name to
identify whether a serial port being configured is identical
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The various functions to add a port used to return port->name, and
the callers would immediately iterate all registered ports to "find"
the one just created by comparing that return value with ... port->name.
This is a waste of cycles and code. Instead, have those functions
return "port" directly.
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
We are comparing strings after all.
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
This adds the ability to explicitly add an MMIO based serial port
via the "serial" command. The syntax is:
serial --port=mmio,<hex_address>{.b,.w,.l,.q}
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
It is common for PCI based UARTs to use larger than one byte access
sizes. This adds support for this and uses the information present
in SPCR accordingly.
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
"serial auto" is now equivalent to just "serial" and will use the
SPCR to discover the port if present, otherwise defaults to "com0"
as before.
This allows to support MMIO ports specified by ACPI which is needed
on AWS EC2 "metal" instances, and will enable GRUB to pickup the
port configuration specified by ACPI in other cases.
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
This will allow ports to be added with a pre-set configuration.
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
And while at it, unify it as clock frequency in Hz, to match the value in
grub_serial_config struct and do the division by 16 in one common place.
This will simplify adding SPCR support.
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
This adds the ability for the driver to access UARTs via MMIO instead
of PIO selectively at runtime, and exposes a new function to add an
MMIO port.
In an ideal world, MMIO accessors would be generic and have architecture
specific memory barriers. However, existing drivers don't have them and
most of those "bare metal" drivers tend to be for x86 which doesn't need
them. If necessary, those can be added later.
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
This adds the definition of the two ACPI tables according to the spec.
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
And convert grub_acpi_find_fadt() to use it.
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The i386-pc mode supports MBR partition scheme where maximum partition
size is 2 TiB. In case of large partitions left shift expression with
unsigned long int "length" object may cause integer overflow making
calculated partition size less than true value. This issue is fixed by
increasing the size of "length" integer type.
Signed-off-by: Maxim Fomin <maxim@fomin.one>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
This allows the cmp command to be used in GRUB scripts to conditionally
run commands based on whether two files are the same.
The command is now quiet by default and the -v switch can be given to enable
verbose mode, the previous behavior.
Update documentation accordingly.
Suggested-by: Li Gen <ligenlive@gmail.com>
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
This was fixed here: 3cf2e848bc (disk/cryptodisk: Allows UUIDs to be compared
in a dash-insensitive manner).
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
This allows seeing full QEMU output of grub-shell, which can be invaluable
when debugging failing tests.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The xenpolicy variable was left set from previous function call. This
resulted in all-but-first menu entries including XSM policy, even if it
did not exist.
Fix this by initializing the xenpolicy variable.
Signed-off-by: Marek Marczykowski-Górecki <marmarek@invisiblethingslab.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
If max_char_width or max_char_height are negative wrong values can be propagated
by grub_font_get_max_char_width() or grub_font_get_max_char_height(). Prevent
this from happening.
Signed-off-by: Zhang Boyang <zhangboyang.id@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Like glyphs in ascii_font_glyph[], assign null_font to
unknown_glyph->font in order to prevent grub_font_get_*() from
dereferencing NULL pointer.
Signed-off-by: Zhang Boyang <zhangboyang.id@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
There is a problem in ascii_glyph_lookup(). It doesn't check the return
value of grub_malloc(). If memory can't be allocated, then NULL pointer
will be written to.
This patch fixes the problem by fallbacking to unknown_glyph when
grub_malloc() returns NULL.
Signed-off-by: Zhang Boyang <zhangboyang.id@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
This patch adds support for plain encryption mode, plain dm-crypt, via
new module/command named "plainmount".
Signed-off-by: Maxim Fomin <maxim@fomin.one>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Reviewed-by: Glenn Washburn <development@efficientek.com>
The final piece needed to add UEFI file system transposition support is to
ensure the boot media can be located regardless of how the boot partition
was instantiated. Especially, we do not want to be reliant on brittle
partition UUIDs, as these only work if a boot media is duplicated at the
block level and not at the file system level.
To accomplish this for EFI boot, we now create a UUID file in a .disk/
directory, that can then be searched for.
Note: The switch from make_image_fwdisk_abs() to make_image_abs() is
needed in order to use the search functionality.
Signed-off-by: Pete Batard <pete@akeo.ie>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
To enable file system transposition support for UEFI, we also must ensure that
there exists a copy of the EFI bootloaders, that are currently embedded in the
efi.img for xorriso, at their expected UEFI location on the ISO 9660 file system.
This is accomplished by removing the use of a temporary directory to create the
efi/ content, to instead place it at the root of the ISO 9660 content.
Signed-off-by: Pete Batard <pete@akeo.ie>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
In order to add file system transposition support for UEFI, i.e. the ability
to copy the content of an grub-mkrescue ISO 9660 image onto user-formatted
media, and have that boot on UEFI systems, the first thing we need to do is
add support for the file systems that are natively handled by UEFI. This
mandatorily includes FAT, but we also include NTFS as the latter is also
commonly supported on modern x64 platforms.
Signed-off-by: Pete Batard <pete@akeo.ie>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
SC2120 (warning): function references arguments, but none are ever passed.
In grub-completion.bash.in line 63:
__grub_get_options_from_help () {
^-- SC2120 (warning)
local prog
if [ $# -ge 1 ]; then
prog="$1"
The arg of __grub_get_options_from_help() is optional. So, the current
code meets the exception and does not need to be modified. Ignoring the
warning then.
More: https://github.com/koalaman/shellcheck/wiki/SC2120
Signed-off-by: t.feng <fengtao40@huawei.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
SC2155 (warning): Declare and assign separately to avoid masking return values.
The exit status of the command is overridden by the exit status of the
creation of the local variable.
In grub-completion.bash.in line 115:
local config_file=$(__grub_dir)/grub.cfg
^---------^ SC2155 (warning)
In grub-completion.bash.in line 126:
local grub_dir=$(__grub_dir)
^------^ SC2155 (warning)
More: https://github.com/koalaman/shellcheck/wiki/SC2155
Signed-off-by: t.feng <fengtao40@huawei.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
SC2207 (warning): Prefer mapfile or read -a to split
command output (or quote to avoid splitting).
In grub-completion.bash.in line 56:
COMPREPLY=($(compgen -P "${2-}" -W "${1-}" -S "${4-}" -- "$cur"))
^-- SC2207 (warning)
In grub-completion.bash.in line 119:
COMPREPLY=( $(compgen \
^-- SC2207 (warning)
In grub-completion.bash.in line 128:
COMPREPLY=( $( compgen -f -X '!*/*.mod' -- "${grub_dir}/$cur" | {
^-- SC2207 (warning)
COMPREPLY=($(command)) are doing unquoted command expansion in an array.
This will invoke the shell's sloppy word splitting and glob expansion.
If we want to split the output into lines or words, use read -r and
loops will be better. This prevents the shell from doing unwanted
splitting and glob expansion, and therefore avoiding problems with
output containing spaces or special characters.
More: https://github.com/koalaman/shellcheck/wiki/SC2207
Signed-off-by: t.feng <fengtao40@huawei.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
SC2070 (error): -n doesn't work with unquoted arguments.
Quote or use [[ ]].
In grub-completion.bash.in line 130:
[ -n $tmp ] && {
^--^ SC2070 (error)
More: https://github.com/koalaman/shellcheck/wiki/SC2070
Signed-off-by: t.feng <fengtao40@huawei.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
grub_file_open() calls grub_file_get_device_name(), but doesn't check
the return. Instead, it checks if grub_errno is set.
However, nothing initialises grub_errno here when grub_file_open()
starts. This means that trying to open one file that doesn't exist and
then trying to open another file that does will (incorrectly) also
fail to open that second file.
Let's fix that.
Signed-off-by: Steve McIntyre <steve@einval.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The current i386 initrd is limited under 1 GiB memory and it works with
most compressed initrds (also initrd_addr_max case reported by kernel).
addr = (addr_max - aligned_size) & ~0xFFF;
Above line is used to calculate the reasonable address to store the initrd.
However, if initrd size is greater than 1 GiB or initrd_addr_max, then it
will get overflow, especially on x86_64 arch.
Therefore, add a check point to prevent it overflows as well as having
a debug log for complex story of initrd addresses.
Signed-off-by: Jeremy Szu <jeremy.szu@canonical.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Only perform call to fwsetup if one is on EFI platform. On all other
platforms fwsetup command does not exists, and thus returns 0 and
a useless uefi-firmware menu entry gets generated.
Signed-off-by: Dimitri John Ledkov <dimitri.ledkov@canonical.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Fixes: 58ea11d5b (fs/hfsplus: Don't fetch a key beyond the end of the node)
Signed-off-by: t.feng <fengtao40@huawei.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The l1_entries and l2_entries were not freed at the end of file read.
Fixes: 5825b3794 (BFS implementation based on the specification)
Signed-off-by: t.feng <fengtao40@huawei.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The hashtable is not freed if GRUB_AFFS_FILETYPE_HARDLINK and
grub_disk_read() failed. If grub_affs_create_node() returns non-zero
the hashtable should be freed too.
By the way, the hashtable argument is unused in grub_affs_create_node().
So, we can remove the argument and free it in grub_affs_iterate_dir().
It allocates the memory and it should be responsible for releasing it.
This is why commit ebf32bc4e9 (fs/affs: Fix resource leaks) missed
this memory leak.
Fixes: ebf32bc4e9 (fs/affs: Fix resource leaks)
Signed-off-by: t.feng <fengtao40@huawei.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
An unchecked decrement operation in cl_print() would cause a few
integers to underflow. Where an output terminal's state is stored in
cl_term, the values cl_term->ystart and cl_term->pos.y both underflow.
This can be replicated with the following steps:
1. Get to the GRUB command line
2. Hold down the "d" key (or any key that enters a visible character)
until it fills the entire row
3. Press "HOME" and then press "CTRL-k". This will clear every
character entered in step 2
4. Continuously press "CTRL-y" until the terminal scrolls the original
prompt ("grub> ") passed the terminal's top row. Now, no prompt
should be visible. This step causes cl_term->ystart to underflow
5. Press "HOME" and then "d" (or any visible character). This can have
different visual effects for different systems, but it will always
cause cl_term->pos.y to underflow
On BIOS systems, these underflows cause the output terminal to
completely stop displaying anything. Characters can still be
entered and commands can be run, but nothing will display on the
terminal. From here, you can only get the display working by running
a command to switch the current output terminal to a different type:
terminal_output <OTHER_TERMINAL>
On UEFI systems, these replication steps do not break the output
terminal. Until you press "ENTER", the cursor stops responding to input,
but you can press "ENTER" after step 5 and the command line will
work properly again. This patch is mostly important for BIOS systems
where the output terminal is rendered unusable after the underflows
occur.
This patch adds two checks, one for each variable. It ensures that
cl_term->ystart does not decrement passed 0. It also ensures that
cl_term->pos.y does not get set passed the terminal's bottom row.
When the previously listed replication steps are followed with this
patch, the terminal's cursor will be set to the top row and the command
line is still usable, even on BIOS systems.
Signed-off-by: Ryan Cohen <rcohenprogramming@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Coordinates passed to screen_write_char() did not have any checks to
ensure they are not out-of-bounds. This adds an if statement to prevent
out-of-bounds writes to the VGA text buffer.
Signed-off-by: Ryan Cohen <rcohenprogramming@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Per "man 5 cpio", the namesize in the cpio header includes the trailing
NUL byte of the pathname and the pathname is followed by NUL bytes, but
the current implementation ignores the trailing NUL byte when making
the newc header. Although make_header() tries to pad the pathname string,
the padding won't happen when strlen(name) + sizeof(struct newc_head)
is a multiple of 4, and the non-NULL-terminated pathname may lead to
unexpected results.
Assume that a file is created with 'echo -n aaaa > /boot/test12' and
loaded by grub2:
linux /boot/vmlinuz
initrd newc:test12:/boot/test12 /boot/initrd
The initrd command eventually invoked grub_initrd_load() and sent
't''e''s''t''1''2' to make_header() to generate the header:
00000070 30 37 30 37 30 31 33 30 31 43 41 30 44 45 30 30 |070701301CA0DE00|
00000080 30 30 38 31 41 34 30 30 30 30 30 33 45 38 30 30 |0081A4000003E800|
00000090 30 30 30 30 36 34 30 30 30 30 30 30 30 31 36 33 |0000640000000163|
000000a0 37 36 45 34 35 32 30 30 30 30 30 30 30 34 30 30 |76E4520000000400|
000000b0 30 30 30 30 30 38 30 30 30 30 30 30 31 33 30 30 |0000080000001300|
000000c0 30 30 30 30 30 30 30 30 30 30 30 30 30 30 30 30 |0000000000000000|
000000d0 30 30 30 30 30 36 30 30 30 30 30 30 30 30 74 65 |00000600000000te|
^namesize
000000e0 73 74 31 32 61 61 61 61 30 37 30 37 30 31 30 30 |st12aaaa07070100|
^^ end of the pathname
Since strlen("test12") + sizeof(struct newc_head) is 116 = 29 * 4,
make_header() didn't pad the pathname, and the file content followed
"test12" immediately. This violates the cpio format and may trigger such
error during linux boot:
Initramfs unpacking failed: ZSTD-compressed data is trunc
To avoid the potential problems, this commit counts the trailing NUL byte
in when calling make_header() and adjusts the initrd size accordingly.
Now the header becomes
00000070 30 37 30 37 30 31 33 30 31 43 41 30 44 45 30 30 |070701301CA0DE00|
00000080 30 30 38 31 41 34 30 30 30 30 30 33 45 38 30 30 |0081A4000003E800|
00000090 30 30 30 30 36 34 30 30 30 30 30 30 30 31 36 33 |0000640000000163|
000000a0 37 36 45 34 35 32 30 30 30 30 30 30 30 34 30 30 |76E4520000000400|
000000b0 30 30 30 30 30 38 30 30 30 30 30 30 31 33 30 30 |0000080000001300|
000000c0 30 30 30 30 30 30 30 30 30 30 30 30 30 30 30 30 |0000000000000000|
000000d0 30 30 30 30 30 37 30 30 30 30 30 30 30 30 74 65 |00000700000000te|
^namesize
000000e0 73 74 31 32 00 00 00 00 61 61 61 61 30 37 30 37 |st12....aaaa0707|
^^ end of the pathname
Besides the trailing NUL byte, make_header() pads 3 more NUL bytes, and
the user can safely read the pathname without a further check.
To conform to the cpio format, the headers for "TRAILER!!!" are also
adjusted to include the trailing NUL byte, not ignore it.
Signed-off-by: Gary Lin <glin@suse.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Validate the length of Allocation Extent Descriptor in grub_udf_read_block(),
based on the details in UDF spec. v2.01 section 2.3.11.
Fixes: CID 314037
Signed-off-by: Jagannathan Raman <jag.raman@oracle.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
This enables an early failure; for i386-ieee1275 and powerpc-ieee1275 on
Linux, without /dev/nvram the system may be left in an unbootable state.
Signed-off-by: Ismael Luceno <iluceno@suse.de>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
When running tests there are many invocations of grub-shell, and because
the output files are all random names in the same tmp directory, it
becomes more work to figure out which files went with which grub-shell
invocations. So all generated files from one invocation of grub-shell
are put into a randomly named directory, so as not to collide with other
grub-shell invocations. And now that the generated files can be put in
a location where they will not get stepped on, and they can be named
sensible names.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The out->ncomb is a bit-field of 8 bits. So, the max possible value is 255.
However, code in grub_unicode_aglomerate_comb() doesn't check for an
overflow when incrementing out->ncomb. If out->ncomb is already 255,
after incrementing it will get 0 instead of 256, and cause illegal
memory access in subsequent processing.
This patch introduces GRUB_UNICODE_NCOMB_MAX to represent the max
acceptable value of ncomb. The code now checks for this limit and
ignores additional combining characters when limit is reached.
Reported-by: Daniel Axtens <dja@axtens.net>
Signed-off-by: Zhang Boyang <zhangboyang.id@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The calculations in blit_comb() need information from glyph's font, e.g.
grub_font_get_xheight(main_glyph->font). However, main_glyph->font is
NULL if main_glyph comes from ascii_font_glyph[]. Therefore
grub_font_get_*() crashes because of NULL pointer.
There is already a solution, the null_font. So, assign it to those glyphs
in ascii_font_glyph[].
Reported-by: Daniel Axtens <dja@axtens.net>
Signed-off-by: Zhang Boyang <zhangboyang.id@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
As a mitigation and hardening measure add sanity checks to
grub_font_blit_glyph() and grub_font_blit_glyph_mirror(). This patch
makes these two functions do nothing if target blitting area isn't fully
contained in target bitmap. Therefore, if complex calculations in caller
overflows and malicious coordinates are given, we are still safe because
any coordinates which result in out-of-bound-write are rejected. However,
this patch only checks for invalid coordinates, and doesn't provide any
protection against invalid source glyph or destination glyph, e.g.
mismatch between glyph size and buffer size.
This hardening measure is designed to mitigate possible overflows in
blit_comb(). If overflow occurs, it may return invalid bounding box
during dry run and call grub_font_blit_glyph() with malicious
coordinates during actual blitting. However, we are still safe because
the scratch glyph itself is valid, although its size makes no sense, and
any invalid coordinates are rejected.
It would be better to call grub_fatal() if illegal parameter is detected.
However, doing this may end up in a dangerous recursion because grub_fatal()
would print messages to the screen and we are in the progress of drawing
characters on the screen.
Reported-by: Daniel Axtens <dja@axtens.net>
Signed-off-by: Zhang Boyang <zhangboyang.id@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The expression (ctx.bounds.height - combining_glyphs[i]->height) / 2 may
evaluate to a very big invalid value even if both ctx.bounds.height and
combining_glyphs[i]->height are small integers. For example, if
ctx.bounds.height is 10 and combining_glyphs[i]->height is 12, this
expression evaluates to 2147483647 (expected -1). This is because
coordinates are allowed to be negative but ctx.bounds.height is an
unsigned int. So, the subtraction operates on unsigned ints and
underflows to a very big value. The division makes things even worse.
The quotient is still an invalid value even if converted back to int.
This patch fixes the problem by casting ctx.bounds.height to int. As
a result the subtraction will operate on int and grub_uint16_t which
will be promoted to an int. So, the underflow will no longer happen. Other
uses of ctx.bounds.height (and ctx.bounds.width) are also casted to int,
to ensure coordinates are always calculated on signed integers.
Fixes: CVE-2022-3775
Reported-by: Daniel Axtens <dja@axtens.net>
Signed-off-by: Zhang Boyang <zhangboyang.id@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Expressions like u64 = u32 * u32 are unsafe because their products are
truncated to u32 even if left hand side is u64. This patch fixes all
problems like that one in fbutil.
To get right result not only left hand side have to be u64 but it's also
necessary to cast at least one of the operands of all leaf operators of
right hand side to u64, e.g. u64 = u32 * u32 + u32 * u32 should be
u64 = (u64)u32 * u32 + (u64)u32 * u32.
For 1-bit bitmaps grub_uint64_t have to be used. It's safe because any
combination of values in (grub_uint64_t)u32 * u32 + u32 expression will
not overflow grub_uint64_t.
Other expressions like ptr + u32 * u32 + u32 * u32 are also vulnerable.
They should be ptr + (grub_addr_t)u32 * u32 + (grub_addr_t)u32 * u32.
This patch also adds a comment to grub_video_fb_get_video_ptr() which
says it's arguments must be valid and no sanity check is performed
(like its siblings in grub-core/video/fb/fbutil.c).
Signed-off-by: Zhang Boyang <zhangboyang.id@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
As a mitigation and hardening measure enforce verification of font
files. Then only trusted font files can be load. This will reduce the
attack surface at cost of losing the ability of end-users to customize
fonts if e.g. UEFI Secure Boot is enabled. Vendors can always customize
fonts because they have ability to pack fonts into their GRUB bundles.
This goal is achieved by:
* Removing GRUB_FILE_TYPE_FONT from shim lock verifier's
skip-verification list.
* Adding GRUB_FILE_TYPE_FONT to lockdown verifier's defer-auth list,
so font files must be verified by a verifier before they can be loaded.
Suggested-by: Daniel Kiper <daniel.kiper@oracle.com>
Signed-off-by: Zhang Boyang <zhangboyang.id@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
If search target is less than all entries in font->index then "hi"
variable is set to -1, which translates to SIZE_MAX and leads to errors.
This patch fixes the problem by replacing the entire binary search code
with the libstdc++'s std::lower_bound() implementation.
Signed-off-by: Zhang Boyang <zhangboyang.id@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The BMP index (font->bmp_idx) is designed as a reverse lookup table of
char entries (font->char_index), in order to speed up lookups for BMP
chars (i.e. code < 0x10000). The values in BMP index are the subscripts
of the corresponding char entries, stored in grub_uint16_t, while 0xffff
means not found.
This patch fixes the problem of large subscript truncated to grub_uint16_t,
leading BMP index to return wrong char entry or report false miss. The
code now checks for bounds and uses BMP index as a hint, and fallbacks
to binary-search if necessary.
On the occasion add a comment about BMP index is initialized to 0xffff.
Signed-off-by: Zhang Boyang <zhangboyang.id@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
In fact it can't overflow at all because glyph_id->ncomb is only 8-bit
wide. But let's keep safe if somebody changes the width of glyph_id->ncomb
in the future. This patch also fixes the inconsistency between
render_max_comb_glyphs and render_combining_glyphs when grub_malloc()
returns NULL.
Signed-off-by: Zhang Boyang <zhangboyang.id@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Remove grub_font_dup_glyph() since nobody is using it since 2013, and
I'm too lazy to fix the integer overflow problem in it.
Signed-off-by: Zhang Boyang <zhangboyang.id@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
This patch fixes several integer overflows in grub_font_construct_glyph().
Glyphs of invalid size, zero or leading to an overflow, are rejected.
The inconsistency between "glyph" and "max_glyph_size" when grub_malloc()
returns NULL is fixed too.
Fixes: CVE-2022-2601
Reported-by: Zhang Boyang <zhangboyang.id@gmail.com>
Signed-off-by: Zhang Boyang <zhangboyang.id@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The length of memory allocation and file read may overflow. This patch
fixes the problem by using safemath macros.
There is a lot of code repetition like "(x * y + 7) / 8". It is unsafe
if overflow happens. This patch introduces grub_video_bitmap_calc_1bpp_bufsz().
It is safe replacement for such code. It has safemath-like prototype.
This patch also introduces grub_cast(value, pointer), it casts value to
typeof(*pointer) then store the value to *pointer. It returns true when
overflow occurs or false if there is no overflow. The semantics of arguments
and return value are designed to be consistent with other safemath macros.
Signed-off-by: Zhang Boyang <zhangboyang.id@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Check glyph's width and height against limits specified in font's
metadata. Reject the glyph (and font) if such limits are exceeded.
Signed-off-by: Zhang Boyang <zhangboyang.id@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The commit eb33e61b3 (multiboot: fix memory leak) did not fix all
issues. Fix all of them right now.
Fixes: eb33e61b3 (multiboot: fix memory leak)
Signed-off-by: t.feng <fengtao40@huawei.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
zpool status by default prints basenames of VDEVs, which means that GRUB
would have to go around guessing to see whether a VDEV exists. Instead,
it'd be more robust to simply tell zpool to give us full paths to VDEVs
via -P.
Signed-off-by: Arsen Arsenović <arsen@aarsen.me>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
This is not an ideal solution, as interactive users must always run
a command in order to get the behavior they want, but it avoids
problematic interactions between prompting and sourcing files.
Signed-off-by: Robbie Harwood <rharwood@redhat.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Currently if an EFI firmware fails to do a TPM measurement for a file,
the error will be propagated to the verifiers framework which will
prevent it to be opened. This mean that buggy firmwares will lead to
the system not booting because files won't be allowed to be loaded. But
a failure to do a TPM measurement isn't expected to be a fatal error
that causes the system to be unbootable.
To avoid this, don't return errors from .write and .verify_string
callbacks and just print a debug message in the case of a TPM
measurement failure. Add an environment variable, tpm_fail_fatal, to
restore the previous behavior.
Also-authored-by: Javier Martinez Canillas <javierm@redhat.com>
Signed-off-by: Javier Martinez Canillas <javierm@redhat.com>
Signed-off-by: Robbie Harwood <rharwood@redhat.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Add an include on stdbool.h, making the bool type generally available
within the GRUB without needing to add a file-specific include every
time it would be used.
Signed-off-by: Robbie Harwood <rharwood@redhat.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The GRUB emulator is used as a debugging utility but it could also be
used as a user-space bootloader if there is support to boot an operating
system.
The Linux kernel is already able to (re)boot another kernel via the
kexec boot mechanism. So the grub-emu tool could rely on this feature
and have linux and initrd commands that are used to pass a kernel,
initramfs image and command line parameters to kexec for booting
a selected menu entry.
By default the systemctl kexec option is used so systemd can shutdown
all of the running services before doing a reboot using kexec. But if
this is not present, it can fall back to executing the kexec user-space
tool directly. The ability to force a kexec-reboot when systemctl kexec
fails must only be used in controlled environments to avoid possible
filesystem corruption and data loss.
Signed-off-by: Raymund Will <rw@suse.com>
Signed-off-by: John Jolly <jjolly@suse.com>
Signed-off-by: Javier Martinez Canillas <javierm@redhat.com>
Signed-off-by: Robbie Harwood <rharwood@redhat.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
A user may wish to use an image that is not sorted as the "latest"
version as the top-level entry. For example, in Arch Linux, if a user
has the LTS and regular kernels installed, "/boot/vmlinuz-linux-lts"
gets sorted as the "latest" compared to "/boot/vmlinuz-linux", meaning
the LTS kernel becomes the top-level entry. However, a user may wish to
use the regular kernel as the top-level default with the LTS only
existing as a backup.
This need can be seen in Arch Linux's AUR with two user-submitted
packages[0][1] providing an update hook which patches /etc/grub.d/10_linux
to move the desired kernel to the top-level. This patch serves to solve
this in a more generic way.
Introduce the GRUB_TOP_LEVEL, GRUB_TOP_LEVEL_XEN and GRUB_TOP_LEVEL_OS_PROBER
variables to allow users to specify the top-level entry.
Create grub_move_to_front() as a helper function which moves entries to
the front of a list. This function does the heavy lifting of moving
the menu entry to the front in each script.
In 10_netbsd, since there isn't an explicit list variable, extract the
items that are being iterated through into a list so that we can
optionally apply grub_move_to_front() to the list before the loop.
[0]: https://aur.archlinux.org/packages/grub-linux-default-hook
[1]: https://aur.archlinux.org/packages/grub-linux-rt-default-hook
Signed-off-by: Denton Liu <liu.denton@gmail.com>
Reviewed-by: Oskari Pirhonen <xxc3ncoredxx@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
In grub-core/video/readers/jpeg.c, the height and width of a JPEG image don't
have an upper limit for how big the JPEG image can be. In Coverity, this is
getting flagged as an untrusted loop bound. This issue can also seen in PNG and
TGA format images as well but Coverity isn't flagging it. To prevent this, the
constant IMAGE_HW_MAX_PX is being added to include/grub/bitmap.h, which has
a value of 16384, to act as an artificial limit and restrict the height and
width of images. This value was picked as it is double the current max
resolution size, which is 8K.
Fixes: CID 292450
Signed-off-by: Alec Brown <alec.r.brown@oracle.com>
Reviewed-by: Darren Kenny <darren.kenny@oracle.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
This is "belt and braces" with commit 12e20a6a69 (disk/diskfilter:
Check calloc() result for NULL): we end up trying to use too much memory
in situations like corrupted Linux software RAID setups purporting to
use a huge number of disks. Simply refuse to permit such configurations.
1024 is a bit arbitrary, yes, and I feel a bit like I'm tempting fate
here, but I think 1024 disks in an array (that GRUB has to read to boot!)
should be enough for anyone.
Signed-off-by: Daniel Axtens <dja@axtens.net>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Now that we implemented support for the LoadFile2 protocol for initrd
loading, there is no longer a need to pass the initrd parameters via
the device tree. This means that when the LoadFile2 protocol is being
used, there is no reason to update the device tree in the first place,
and so we can ignore it entirely.
The only remaining reason to deal with the devicetree is if we are
using the "devicetree" command to load one from disk, so tweak the
logic in grub_fdt_install() to take that into account.
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
Reviewed-by: Leif Lindholm <quic_llindhol@quicinc.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Recent Linux kernels will invoke the LoadFile2 protocol installed on
a well-known vendor media path to load the initrd if it is exposed by
the firmware. Using this method is preferred for two reasons:
- the Linux kernel is in charge of allocating the memory, and so it can
implement any placement policy it wants (given that these tend to
change between kernel versions),
- it is no longer necessary to modify the device tree provided by the
firmware.
So let's install this protocol when handling the "initrd" command if
such a recent kernel was detected (based on the PE/COFF image version),
and defer loading the initrd contents until the point where the kernel
invokes the LoadFile2 protocol.
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
Reviewed-by: Heinrich Schuchardt <heinrich.schuchardt@canonical.com>
Tested-by: Ilias Apalodimas <ilias.apalodimas@linaro.org>
Reviewed-by: Ilias Apalodimas <ilias.apalodimas@linaro.org>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
When GRUB runs on top of EFI firmware, it only has access to block and
network device abstractions exposed by the firmware, and it is up to the
firmware to quiesce the underlying hardware when exiting boot services
and handing over to the OS.
This is especially important for network devices, to prevent incoming
packets from being DMA'd straight into memory after the OS has taken
over but before it has managed to reconfigure the network hardware.
GRUB handles this by means of the grub_net_fini_hw() preboot hook, which
is executed before calling into the booted image. This means that all
network devices disappear or become inoperable before the EFI stub
executes on EFI targeted builds. This is problematic as it prevents the
EFI stub from calling back into GRUB provided protocols such as
LoadFile2 for the initrd, which we will provide in a subsequent patch.
So add a flag that indicates to the network core that EFI network
devices should not be closed when grub_net_fini_hw() is called.
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
Reviewed-by: Heinrich Schuchardt <heinrich.schuchardt@canonical.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The way we load the Linux and PE/COFF image headers depends on a fixed
placement of the COFF header at offset 0x40 into the file. This is
a reasonable default, given that this is where Linux emits it today.
However, in order to comply with the PE/COFF spec, which allows this
header to appear anywhere in the file, let's ensure that we read the
header from where it actually appears in the file if it is not located
at offset 0x40.
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Xen has its own version of the image header, to account for the
additional PE/COFF header fields. Since we are adding references to
those in the shared EFI loader code, update the common definitions
and drop the Xen specific one which no longer has a purpose.
Since in both cases, the call to grub_arch_efi_linux_check_image() is
preceded by a load of the image header, let's move the load into that
function, and rename it to grub_arch_efi_linux_load_image_header().
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The PE/COFF spec permits the COFF signature and file header to appear
anywhere in the file, and the actual offset is recorded in 4 byte
little endian field at offset 0x3c of the image.
When GRUB is emitted as a PE/COFF binary, we reuse the 128 byte MS-DOS
stub (even for non-x86 architectures), putting the COFF signature and
file header at offset 0x80. However, other PE/COFF images may use
different values, and non-x86 Linux kernels use an offset of 0x40
instead.
So let's get rid of the grub_pe32_header struct from pe32.h, given that
it does not represent anything defined by the PE/COFF spec. Instead,
introduce a minimal struct grub_msdos_image_header type based on the
PE/COFF spec's description of the image header, and use the offset
recorded at file position 0x3c to discover the actual location of the PE
signature and the COFF image header.
The remaining fields are moved into a struct grub_pe_image_header,
which we will use later to access COFF header fields of arbitrary
images (and which may therefore appear at different offsets)
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The grub_buffer_free() should handle NULL input pointer, similar to
grub_free(). If the pointer is not referencing any memory location,
grub_buffer_free() need not perform any function.
Fixes: CID 396931
Signed-off-by: Jagannathan Raman <jag.raman@oracle.com>
Reviewed-by: Ross Philipson <ross.philipson@oracle.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The dnode_get_path() traverses dnode structures to locate the dnode leaf
of a given path. When the leaf is a symlink to another path, it restarts
the traversal either from root or from a different path. In such cases,
dn_new must be re-initialized
Passes "make check".
Fixes: CID 86750
Signed-off-by: Jagannathan Raman <jag.raman@oracle.com>
Reviewed-by: Ross Philipson <ross.philipson@oracle.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
After doing some validation with clang from versions 3.8 and up, the
builds prior to version 8.0.0 fail due to the use of safemath functions
at link time.
Signed-off-by: Darren Kenny <darren.kenny@oracle.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Building the current code with clang and the latest gnulib fails due to
the use of a variable-length-array (vla) warning, which turns in to an
error due to the presence of the -Werror during the build.
The gnulib team stated that their code should not be built with -Werror.
At present, the only way to do this is for the complete code-base, by
using the --disable-werror option to configure.
Rather than doing this, and failing to gain any benefit that it provides,
instead, if building with clang, this patch makes it possible to specifically
not error on vlas, while retaining the -Werror functionality otherwise.
Signed-off-by: Darren Kenny <darren.kenny@oracle.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The recent gnulib updates require an implementation of abort(), but the
current macro provided by changeset:
cd37d3d391 gnulib: Drop no-abort.patch
to config.h.in does not work with the clang compiler since it doesn't
provide a __builtin_trap() implementation, so this element of the
changeset needs to be reverted, and replaced.
After some discussion with Vladimir 'phcoder' Serbinenko and Daniel Kiper
it was suggested to bring back in the change from the changeset:
db7337a3d3 * grub-core/gnulib/regcomp.c (regerror): ...
Which implements abort() as an inline call to grub_abort(), but since
that was made static by changeset:
a8f15bceea * grub-core/kern/misc.c (grub_abort): Make static
it is also necessary to revert the specific part that makes it a static
function too.
Another implementation of abort() was found in grub-core/kern/compiler-rt.c
which needs to also be removed to be consistent.
Signed-off-by: Darren Kenny <darren.kenny@oracle.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
In the function grub_cryptodisk_endecrypt(), a for loop is incrementing the
variable i by (1U << log_sector_size). The variable i is of type grub_size_t
which is a 64-bit unsigned integer on x86_64 architecture. On the other hand, 1U
is a 32-bit unsigned integer. By performing a left shift on a 32-bit value and
assigning it to a 64-bit variable, the 64-bit variable may have incorrect values
in the high 32-bits if the shift has an overflow. To avoid this, we replace 1U
with (grub_size_t)1.
Fixes: CID 307788
Signed-off-by: Alec Brown <alec.r.brown@oracle.com>
Reviewed-by: Darren Kenny <darren.kenny@oracle.com>
Reviewed-by: Patrick Steinhardt <ps@pks.im>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Every heap grow will cause all disk caches invalidated which decreases
performance severely. This patch moves disk cache invalidation code to
the last of memory squeezing measures. So, disk caches are released only
when there are no other ways to get free memory.
Signed-off-by: Zhang Boyang <zhangboyang.id@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Reviewed-by: Patrick Steinhardt <ps@pks.im>
For printf()/fprintf() functions, unsigned integers should use %u as the
valid conversion specifier instead of %d.
Signed-off-by: Qiumiao Zhang <zhangqiumiao1@huawei.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The stack check logs a console message on failure, and the EFI API expects
a NULL terminated UCS-2 string. In order to define a UCS-2 string literal,
kernel.img on amd64 and i386 EFI targets is built with -fshort-wchar.
Also compile kernel.img on other EFI targets with -fshort-wchar.
Fixes: 37ddd94 (kern/efi/init: Log a console error during a stack check failure)
Reported-by: Glenn Washburn <development@efficientek.com>
Signed-off-by: Chris Coulson <chris.coulson@canonical.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
This is useful on cloud instances with remote serial ports as it can be
difficult to connect "fast enough" to get the initial menu display
Signed-off-by: Benjamin Herrenschmidt <benh@kernel.crashing.org>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The point of no return is used to define a point where no change should
be reverted in a wake of fatal error that consequently aborts the
process. The powerpc-ieee1275 install apparently missed this point of no
return definition that newly installed modules could be inadvertently
reverted after successful image embedding so that boot failure is
incurred due to inconsistent state.
Signed-off-by: Michael Chang <mchang@suse.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
With wildly corrupt inputs, we can end up trying to calloc a very
large amount of memory, which will fail and give us a NULL pointer.
We need to check that to avoid a crash. (And, even if we blocked
such inputs, it is good practice to check the results of allocations
anyway.)
Signed-off-by: Daniel Axtens <dja@axtens.net>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
A user can now specify UUID strings with dashes, instead of having to remove
dashes. This is backwards-compatibility preserving and also fixes a source
of user confusion over the inconsistency with how UUIDs are specified
between file system UUIDs and cryptomount UUIDs. Since cryptsetup, the
reference implementation for LUKS, displays and generates UUIDs with dashes
there has been additional confusion when using the UUID strings from
cryptsetup as exact input into GRUB does not find the expected cryptodisk.
A new function grub_uuidcasecmp() is added that is general enough to be used
other places where UUIDs are being compared.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Patrick Steinhardt <ps@pks.im>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Variable values may contain spaces at the end or newlines. However, when
displayed without quotes this is not obvious and can lead to confusion as
to the actual contents of variables. Also for some variables grub_env_get()
returns a NULL pointer instead of a pointer to an empty string and
previously would be printed as "var=(null)". Now such variables will be
displayed as "var=''".
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
This adds acpi as bootstrap module whenever it is available. This opens the
path for proper IRQ routing for fully-userland disk drivers.
Signed-off-by: Samuel Thibault <samuel.thibault@ens-lyon.org>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Allow treating util/grub-module-verifierXX.c as a file you can build
directly so syntax checkers like vim's "syntastic" plugin, which uses
"gcc -x c -fsyntax-only" to build it, will work.
One still has to do whatever setup is required to make it pick the
right include dirs, which -I options we use, etc., but this makes
it so you can do the checking on the file you're editing, rather
than on a different file.
Signed-off-by: Peter Jones <pjones@redhat.com>
Signed-off-by: Robbie Harwood <rharwood@redhat.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Fix the incorrect return value of __clzsi2() function.
Fixes: e795b90 (RISC-V: Add libgcc helpers for clz)
Signed-off-by: Tuan Phan <tphan@ventanamicro.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
We have multiple reports of things being slower with a 1 MiB initial static
allocation, and a report (more difficult to nail down) of a boot failure
as a result of the smaller initial allocation.
Make the initial memory allocation 32 MiB.
Signed-off-by: Daniel Axtens <dja@axtens.net>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
In addition to C locale there is also C.UTF-8 locale now. Filter that as
well, by using ${grub_lang}, which contains a stripped value.
This fixes the following message and resulting boot failure:
error: file `/boot/grub/locale/C.gmo' not found.
Signed-off-by: Christian Hesse <mail@eworm.de>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
In some filesystem utils like mksquashfs, they will silently change
behaviour and cause timestamps to unexpectedly change. Build
environments like Debian's set SOURCE_DATE_EPOCH in the environment,
so remove it. Reproducible builds are good and useful for shipped
artifacts, but this causes build-time tests to fail.
Signed-off-by: Steve McIntyre <steve@einval.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The EFI_CONFORMANCE_PROFILES_TABLE_GUID is used for a table of GUIDs for conformance
profiles (cf. UEFI specification 2.10, 4.6.5 EFI_CONFORMANCE_PROFILE_TABLE).
The lsefisystab command is used to display installed EFI configuration tables.
Currently it only shows the GUID but not a short text for the table.
Provide a short text for the EFI_CONFORMANCE_PROFILES_TABLE_GUID.
Signed-off-by: Heinrich Schuchardt <heinrich.schuchardt@canonical.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Recently, ext4 added the large_dir feature, which adds support for
a 3 level htree directory support.
The GRUB supports existing file systems with htree directories by
ignoring their existence, and since the index nodes for the hash tree
look like deleted directory entries (by design), the GRUB can simply do
a brute force O(n) linear search of directories. The same is true for
3 level deep htrees indicated by large_dir feature flag.
Hence, it is safe for the GRUB to ignore the large_dir incompat feature.
Fixes: https://savannah.gnu.org/bugs/?61606
Signed-off-by: Theodore Ts'o <tytso@mit.edu>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
A new option is added to the loopback command, -D or --decompress, which
when specified transparently decompresses the backing file. This allows
compressed images to be used as if they were uncompressed.
Add documentation to support this change.
Suggested-by: Li Gen <ligenlive@gmail.com>
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The list of targets that support PCI is in gentpl.py. However, there is no
support for generating makefile script from a .def file that will apply
globally to the makefile, but on a per target basis. So instead, use
gentpl.py in configure to get the list of targets and check if the current
build target is one of them. If it is, set the automake conditional
COND_HAVE_PCI. Then in conf/Makefile.common add -DGRUB_HAS_PCI for the
platform if COND_HAVE_PCI is true.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Store returned value from grub_getkey() in int instead of char to
prevent throwing away the extended bits. This was a problem because,
for instance, the left arrow key press would return
(GRUB_TERM_EXTENDED | 0x4b), which would have the GRUB_TERM_EXTENDED
thrown away leaving 0x4b or 'K'. These extended keys should either
work as intended or do nothing. This change has them do nothing,
instead of inserting a key not pressed by the user.
Signed-off-by: Li Gen <ligenlive@gmail.com>
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The register_key_notify() method should have an output parameter which is
a pointer to the unique handle assigned to the registered notification.
Signed-off-by: Li Gen <ligenlive@gmail.com>
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The netmask configured in firmware is not respected on ppc64 (big endian).
When 255.255.252.0 is set as netmask in firmware, the following is the
value of bootpath string in grub_ieee1275_parse_bootpath():
/vdevice/l-lan@30000002:speed=auto,duplex=auto,192.168.88.10,,192.168.89.113,192.168.88.1,5,5,255.255.252.0,512
The netmask in this bootpath is not a problem, since it's a value specified
in firmware. But the value of subnet_mask.ipv4 was set with 0xfffffc00, and
__builtin_ctz(~grub_le_to_cpu32(subnet_mask.ipv4)) returned 16 (not 22).
As a result, 16 was used for netmask wrongly:
1111 1111 1111 1111 1111 1100 0000 0000 # subnet_mask.ipv4(=0xfffffc00)
0000 0000 1111 1100 1111 1111 1111 1111 # grub_le_to_cpu32(subnet_mask.ipv4)
1111 1111 0000 0011 0000 0000 0000 0000 # ~grub_le_to_cpu32(subnet_mask.ipv4)
and the count of zero with __builtin_ctz() can be 16. This patch changes
it as below:
1111 1111 1111 1111 1111 1100 0000 0000 # subnet_mask.ipv4(=0xfffffc00)
0000 0000 1111 1100 1111 1111 1111 1111 # grub_le_to_cpu32(subnet_mask.ipv4)
1111 1111 1111 1111 1111 1100 0000 0000 # grub_be_to_cpu32(subnet_mask.ipv4)
0000 0000 0000 0000 0000 0011 1111 1111 # ~grub_be_to_cpu32(subnet_mask.ipv4)
The count of zero with __builtin_clz() can be 22 (clz counts the number
of one bits preceding the most significant zero bit).
Signed-off-by: Masahiro Matsuya <mmatsuya@redhat.com>
Signed-off-by: Robbie Harwood <rharwood@redhat.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Numerous register fields in the relocator state are simply not
used depending on the relocator. This causes Coverity to flag
these fields but there is no real bug here. Simply initializing
the variable to {0} solves the issue. Fixed in the else case too
for consistency.
Fixes: CID 396932
Signed-off-by: Ross Philipson <ross.philipson@oracle.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
This is trivial, but it might save some time to beginners.
Signed-off-by: Andrea G. Monaco <andrea.monaco@autistici.org>
Reviewed-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Add a new --is-supported option to commands/efi/efifwsetup and
conditionalize display on it.
Signed-off-by: Robbie Harwood <rharwood@redhat.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The "fwsetup" command is only registered if the firmware supports booting
to the firmware setup UI. But it could be possible that the GRUB config
already contains a "fwsetup" entry, because it was generated in a machine
that has support for this feature.
To prevent users getting an error like:
error: ../../grub-core/script/function.c:109:can't find command `fwsetup'.
if it is not supported by the firmware, let's just always register the
command but print a more accurate message if the firmware doesn't
support this option.
Signed-off-by: Javier Martinez Canillas <javierm@redhat.com>
Signed-off-by: Robbie Harwood <rharwood@redhat.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The 30_uefi-firmware template checks if an OsIndicationsSupported UEFI var
exists and EFI_OS_INDICATIONS_BOOT_TO_FW_UI bit is set, to decide whether
a "fwsetup" menu entry would be added or not to the GRUB menu.
But this has the problem that it will only work if the configuration file
was created on an UEFI machine that supports booting to a firmware UI.
This for example doesn't support creating GRUB config files when executing
on systems that support both UEFI and legacy BIOS booting. Since creating
the config file from legacy BIOS wouldn't allow to access the firmware UI.
To prevent this, make the template to unconditionally create the grub.cfg
snippet but check at runtime if was booted through UEFI to decide if this
entry should be added. That way it won't be added when booting with BIOS.
There's no need to check if EFI_OS_INDICATIONS_BOOT_TO_FW_UI bit is set,
since that's already done by the "fwsetup" command when is executed.
Signed-off-by: Javier Martinez Canillas <javierm@redhat.com>
Signed-off-by: Robbie Harwood <rharwood@redhat.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Each call of grub_efi_get_variable() needs a grub_free().
Signed-off-by: Robbie Harwood <rharwood@redhat.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Coverity reports that dnode_end_t argument of fill_fs_info() is too
large to pass-by-value. Therefore, replace the argument with a pointer.
Fixes: CID 73631
Signed-off-by: Jagannathan Raman <jag.raman@oracle.com>
Reviewed-by: Darren Kenny <darren.kenny@oracle.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
It was reported in the #grub IRC channel on Libera that decryption of
LUKS2 partitions fails with errors about invalid digests and/or salts.
In all of these cases, what failed was decoding the Base64
representation of these, where the encoded data contained invalid
characters.
As it turns out, the root cause is that json-c, which is used by
cryptsetup to read and write the JSON header, will escape some
characters by prepending a backslash when writing JSON strings by
default. Most importantly, json-c also escapes the forward slash, which
is part of the Base64 alphabet. Because GRUB doesn't know to unescape
such characters, decoding this string will rightfully fail.
Interestingly, this issue has until now only been reported by users of
Ubuntu 18.04. And a bit of digging in fact reveals that cryptsetup has
changed the logic in a054206d (Suppress useless slash escaping in json
lib, 2018-04-20), which has been released with cryptsetup v2.0.3. Ubuntu
18.04 is still shipping with cryptsetup v2.0.2 though, which explains
why this is not a more frequent issue.
Fix the issue by using our new grub_json_unescape() helper function
that handles unescaping for us.
Reported-by: Afdal
Signed-off-by: Patrick Steinhardt <ps@pks.im>
Reviewed-by: Daniel Kiper <dkiper@net-space.pl>
JSON strings require certain characters to be encoded, either by using
a single reverse solidus character "\" for a set of popular characters,
or by using a Unicode representation of "\uXXXXX". The jsmn library
doesn't handle unescaping for us, so we must implement this functionality
for ourselves.
Add a new function grub_json_unescape() that takes a potentially
escaped JSON string as input and returns a new unescaped string.
Signed-off-by: Patrick Steinhardt <ps@pks.im>
Reviewed-by: Daniel Kiper <dkiper@net-space.pl>
In the case of an error grub_initrd_load() uses argv[] to print the
filename that caused the error. It is also possible to obtain the
filename from the file handles and there is no need to duplicate that
information in argv[], so let's drop it.
Signed-off-by: Nikita Ermakov <arei@altlinux.org>
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
In grub-core/loader/i386/bsdXX.c and grub-core/loader/multiboot_elfxx.c, error
conditionals are simplified to statements such as "if (err)". Even though the
assumption that non-zero values give errors is correct, it would be clearer and
more consistent to compare these conditionals to GRUB_ERR_NONE.
Signed-off-by: Alec Brown <alec.r.brown@oracle.com>
Reviewed-by: Darren Kenny <darren.kenny@oracle.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
In util/grub-module-verifierXX.c, the function get_shnum() returns the variable
shnum, which is of the type Elf_Word. In the function, shnum can be obtained by
the e_shnum member of an Elf_Ehdr or the sh_size member of an Elf_Shdr. The
sh_size member can either be grub_uint32_t or grub_uint64_t, depending on the
architecture, but Elf_Word is only grub_uint32_t. To account for when sh_size is
grub_uint64_t, we can set shnum to have type Elf_Shnum and have get_shnum()
return an Elf_Shnum.
Signed-off-by: Alec Brown <alec.r.brown@oracle.com>
Reviewed-by: Darren Kenny <darren.kenny@oracle.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
In bsdXX.c and multiboot_elfxx.c, e_phnum is used to obtain the number of
program header table entries, but it wasn't being checked if the value was
there.
According to the elf(5) manual page,
"If the number of entries in the program header table is larger than or equal to
PN_XNUM (0xffff), this member holds PN_XNUM (0xffff) and the real number of
entries in the program header table is held in the sh_info member of the
initial entry in section header table. Otherwise, the sh_info member of the
initial entry contains the value zero."
Since this check wasn't being made, grub_elfXX_get_phnum() is being added to
elfXX.c to make this check and use e_phnum if it doesn't have PN_XNUM as a
value, else use sh_info. We also need to make sure e_phnum isn't greater than
PN_XNUM and sh_info isn't less than PN_XNUM.
Note that even though elf.c and elfXX.c are located in grub-core/kern, they are
compiled as modules and don't need the EXPORT_FUNC() macro to define the functions
in elf.h.
Also, changed casts of phnum to match variables being set as well as dropped
casts when unnecessary.
Signed-off-by: Alec Brown <alec.r.brown@oracle.com>
Reviewed-by: Darren Kenny <darren.kenny@oracle.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
In multiboot_elfxx.c, e_shstrndx is used to obtain the section header table
index of the section name string table, but it wasn't being checked if the value
was there.
According to the elf(5) manual page,
"If the index of section name string table section is larger than or equal to
SHN_LORESERVE (0xff00), this member holds SHN_XINDEX (0xffff) and the real
index of the section name string table section is held in the sh_link member of
the initial entry in section header table. Otherwise, the sh_link member of the
initial entry in section header table contains the value zero."
Since this check wasn't being made, grub_elfXX_get_shstrndx() is being added to
elfXX.c to make this check and use e_shstrndx if it doesn't have SHN_XINDEX as a
value, else use sh_link. We also need to make sure e_shstrndx isn't greater than
or equal to SHN_LORESERVE and sh_link isn't less than SHN_LORESERVE.
Note that even though elf.c and elfXX.c are located in grub-core/kern, they are
compiled as modules and don't need the EXPORT_FUNC() macro to define the functions
in elf.h.
Signed-off-by: Alec Brown <alec.r.brown@oracle.com>
Reviewed-by: Darren Kenny <darren.kenny@oracle.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
In bsdXX.c and multiboot_elfxx.c, e_shnum is used to obtain the number of
section header table entries, but it wasn't being checked if the value was
there.
According to the elf(5) manual page,
"If the number of entries in the section header table is larger than or equal to
SHN_LORESERVE (0xff00), e_shnum holds the value zero and the real number of
entries in the section header table is held in the sh_size member of the initial
entry in section header table. Otherwise, the sh_size member of the initial
entry in the section header table holds the value zero."
Since this check wasn't being made, grub_elfXX_get_shnum() is being added to
elfXX.c to make this check and use whichever member doesn't have a value of
zero. If both are zero, then we must return an error. We also need to make sure
that e_shnum doesn't have a value greater than or equal to SHN_LORESERVE and
sh_size isn't less than SHN_LORESERVE.
In order to get this function to work, the type ElfXX_Shnum is being added where
Elf32_Shnum defines Elf32_Word and Elf64_Shnum defines Elf64_Xword. This new
type is needed because if shnum obtains a value from sh_size, sh_size could be
of type El32_Word for Elf32_Shdr structures or Elf64_Xword for Elf64_Shdr
structures.
Note that even though elf.c and elfXX.c are located in grub-core/kern, they are
compiled as modules and don't need the EXPORT_FUNC() macro to define the functions
in elf.h.
For a few smaller changes, changed casts of shnum to match variables being set
as well as dropped casts when unnecessary and fixed spacing errors in bsdXX.c.
Also, shnum is an unsigned integer and is compared to int i in multiboot_elfxx.c,
it should be unsigned to match shnum.
Signed-off-by: Alec Brown <alec.r.brown@oracle.com>
Reviewed-by: Darren Kenny <darren.kenny@oracle.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The linux_xen template orders the "early" initrd file(s) _first_
(i.e., before the "real" initrd files) and that seems reasonable,
as microcode updates usually come first.
However, this usually breaks Linux boot with initrd under Xen
because Xen assumes the real initrd is the first multiboot[2]
module after the kernel, passing its address over to Linux in
Xen's start_info struct.
So, if a microcode-only initrd (i.e., without init/userspace)
is found by grub-mkconfig, it ends up considered as a normal
initrd by the Linux kernel, which cannot do anything with it
(as it has no other files) and panic()s unable to mount root
if it depends on a initrd to do that (e.g., root=UUID=...).
...
Well, since Xen doesn't actually use the provided microcode
by default / unless the 'ucode=<module number|scan>' option
is enabled, this isn't used in the general case (and breaks).
Additionally, if an user enables the 'ucode=' option, that
either specifies which module is to be used for microcode,
or scans all modules (regardless of being first) for that.
Thus, for Xen:
- it is *not required* to have microcode first,
- but it is *required* to have real initrd first
So, fix it by ordering the real initrd before early initrd(s).
After:
# touch /boot/xen /boot/microcode.cpio
# grub-mkconfig 2>/dev/null | grep -P '^\t(multiboot|module)'
multiboot /boot/xen ...
module /boot/vmlinuz-5.4.0-122-generic ...
module --nounzip /boot/initrd.img-5.4.0-122-generic
module --nounzip /boot/microcode.cpio
...
Corner case specific to Xen implementation details:
It is actually _possible_ to have a microcode initrd first,
but that requires a non-default option (so can't rely on it),
and it turns out to be inconsistent with its counterpart
(really shouldn't rely on it, as it may get confusing; below).
'ucode=1' does manually specify the first module is microcode
_AND_ clears its bit in the module bitmap. The next module is
now the 'new first', and gets passed to Linux as initrd. Good.
'ucode=scan' checks all modules for microcode, but does _NOT_
clear a bit if it finds one (reasonable, as it can find that
prepended in a "real" initrd anyway, which needs to be used).
The first module still gets passed to Linux as initrd. Bad.
Fixes: e86f6aafb8 (grub-mkconfig/20_linux_xen: Support multiple early initrd images)
Signed-off-by: Mauricio Faria de Oliveira <mfo@canonical.com>
Acked-by: Juergen Gross <jgross@suse.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The linux_xen template can put multiple initrd files in the
same multiboot[2] module[2] command, which is against specs.
This causes ONLY the _first_ initrd file to be loaded; other
files just have filenames in a "cmdline" string of the first
initrd file and are NOT loaded.
Fix this by inserting a module[2] command per initrd file.
Before:
# touch /boot/xen /boot/microcode.cpio
# grub-mkconfig 2>/dev/null | grep -P '^\t(multiboot|module)'
multiboot /boot/xen ...
module /boot/vmlinuz-5.4.0-122-generic ...
module --nounzip /boot/microcode.cpio /boot/initrd.img-5.4.0-122-generic
After:
# touch /boot/xen /boot/microcode.cpio
# grub-mkconfig 2>/dev/null | grep -P '^\t(multiboot|module)'
multiboot /boot/xen ...
module /boot/vmlinuz-5.4.0-122-generic ...
module --nounzip /boot/microcode.cpio
module --nounzip /boot/initrd.img-5.4.0-122-generic
Cause:
The code was copied from the linux template, which is *apparently*
equivalent.. but its initrd command grub_cmd_initrd() *supports*
multiple files (see grub_initrd_init()), while module/module2 in
grub_cmd_module() *does not* (see grub_multiboot[2]_add_module()).
See commit e86f6aafb8 (grub-mkconfig/20_linux_xen: Support multiple early initrd images):
'This is basically a copy of a698240d "grub-mkconfig/10_linux:
Support multiple early initrd images" ...'
Specs:
Both multiboot and multiboot2 specifications mention support for
'multiple boot modules' (struct/tag used for kernel/initrd files):
"Boot loaders don’t have to support multiple boot modules,
but they are strongly encouraged to" [1,2]
However, there is a 1:1 relationship between boot modules and files,
more or less clearly; note the usage of singular/plural "module(s)".
(Multiboot2, clearly: "One tag appears per module".)
Multiboot [1]:
"the ‘mods’ fields indicate ... what boot modules
were loaded ..., and where they can be found.
‘mods_count’ contains the number of modules loaded"
"The first two fields contain the start and end addresses
of the boot module itself."
Multiboot2 [2]:
"This tag indicates ... what boot module was loaded ...,
and where it can be found."
"The ‘mod_start’ and ‘mod_end’ contain the start and end
physical addresses of the boot module itself."
"One tag appears per module.
This tag type may appear multiple times."
And both clearly mention the 'string' field of a boot module,
which is to be used by the operating system, not boot loader:
"The ‘string’ field provides an arbitrary string to be
associated with that particular boot module ...
its exact use is specific to the operating system."
Links:
[1] https://www.gnu.org/software/grub/manual/multiboot/multiboot.html
3.3 Boot information format
[2] https://www.gnu.org/software/grub/manual/multiboot2/multiboot.html
3.6.6 Modules
Fixes: e86f6aafb8 (grub-mkconfig/20_linux_xen: Support multiple early initrd images)
Signed-off-by: Mauricio Faria de Oliveira <mfo@canonical.com>
Acked-by: Juergen Gross <jgross@suse.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Note this cast was fixed in grub_strcasecmp() in commit ce41ab7aab
(* grub-core/kern/misc.c (grub_strcmp): Use unsigned comparison as per
common usage and preffered in several parts of code.), but this commit
omitted fixing it in grub_strncasecmp().
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The previous behavior ignored an error and the output from grub-mkrescue.
This made it difficult to discover that grub-mkrescue was the reason that
tests which rely on grub-shell were failing. Even after discovering
grub-mkrescue was the culprit, there was no output to indicate why it was
failing. It turns out that grub-mkrescue is a thin wrapper around xorriso.
So if you do not have xorriso installed it will fail with an error message
about not being able to find xorriso.
This change will allow grub-mkrescue output to be written to stderr, only
if grub-mkrescue fails. If grub-mkrescue succeeds, there will be no output
from grub-mkrescue so as not to interfere with the functioning of tests.
This change should have no effect on the running of tests or other uses of
grub-shell as it only modifies the error path.
Also, if grub-mkrescue fails, the script exits early. Since grub-shell
needs the ISO image created by grub-mkresue to boot the QEMU instance,
a failure here should be considered fatal.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The "ARM\x64" magic number in the file header identifies an image as one
that implements the bare metal boot protocol, allowing the loader to
simply move the file to a suitably aligned address in memory, with
sufficient headroom for the trailing .bss segment (the required memory
size is described in the header as well).
Note of this matters for GRUB, as it only supports EFI boot. EFI does
not care about this magic number, and nor should GRUB: this prevents us
from booting other PE linux images, such as the generic EFI zboot
decompressor, which is a pure PE/COFF image, and does not implement the
bare metal boot protocol.
So drop the magic number check.
Signed-off-by: Ard Biesheuvel <ardb@kernel.org>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Because grub_util_mkdir() is implemented to not return a value on any
platform, grub_instal_mkdir_p() can test for success by confirming that
the directory requested exists after attempting to create it, otherwise
it should fail with an error and exit.
While fixing this, a flaw in the logic was shown, where the first match
of the path separator, which almost always was the first character in
the path (e.g. /boot/grub2) would result in creating a directory with an
empty name (i.e. ""). To avoid that, it should skip the handling of the
path separator where p is pointing to the first character.
Signed-off-by: Darren Kenny <darren.kenny@oracle.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Coverity signaled 2 issues where the return value of grub_util_mkdir()
was not being tested.
The Windows variant of this code defines the function as having no
return value (void), but the UNIX variants all are mapped using a macro
to the libc mkdir() function, which returns an int value.
To be consistent, the mapping should cast to void to for these too.
Fixes: CID 73583
Fixes: CID 73617
Signed-off-by: Darren Kenny <darren.kenny@oracle.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Update the read hook to take into account encrypted volumes on a partition.
GRUB disk read hooks supply an absolute sector number at which the read is
started from. If the encrypted volume is in a partition, the sector number
given to the read hook will be offset by the number of the sector at the
start of the partition. The read hook then needs to subtract the partition
start from the supplied sector to get the correct start sector for the read
into the detached header file.
Reported-by: brutser <brutser@perso.be>
Signed-off-by: Glenn Washburn <development@efficientek.com>
Tested-by: brutser <brutser@perso.be>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
By using shell variable that are set once by the expansion of an autoconf
variable, the resulting shell script is more easily moved and modified
from the build/install directory it was generated for. The resulting
script is more readable as well.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
On my laptop running at 2.4GHz, if I run a VM where tsc calibration
using pmtimer will fail presuming a broken pmtimer, it takes ~51 seconds
to do so (as measured with the stopwatch on my phone), with a tsc delta
of 0x1cd1c85300, or around 125 billion cycles.
If instead of trying to wait for 5-200ms to show up on the pmtimer, we
try to wait for 5-200us, it decides it's broken in ~0x2626aa0 TSCs, aka
~2.4 million cycles, or more or less instantly.
Additionally, this reading the pmtimer was returning 0xffffffff anyway,
and that's obviously an invalid return. I've added a check for that and
0 so we don't bother waiting for the test if what we're seeing is dead
pins with no response at all.
If "debug" includes "pmtimer", you will see one of the following three
outcomes. If pmtimer gives all 0 or all 1 bits, you will see:
pmtimer: 0xffffff bad_reads: 1
pmtimer: 0xffffff bad_reads: 2
pmtimer: 0xffffff bad_reads: 3
pmtimer: 0xffffff bad_reads: 4
pmtimer: 0xffffff bad_reads: 5
pmtimer: 0xffffff bad_reads: 6
pmtimer: 0xffffff bad_reads: 7
pmtimer: 0xffffff bad_reads: 8
pmtimer: 0xffffff bad_reads: 9
pmtimer: 0xffffff bad_reads: 10
timer is broken; giving up.
This outcome was tested using qemu+kvm with UEFI (OVMF) firmware and
these options: -machine pc-q35-2.10 -cpu Broadwell-noTSX
If pmtimer gives any other bit patterns but is not actually marching
forward fast enough to use for clock calibration, you will see:
pmtimer delta is 0x0 (1904 iterations)
tsc delta is implausible: 0x2626aa0
This outcome was tested using GRUB patched to not ignore bad reads using
qemu+kvm with UEFI (OVMF) firmware, and these options:
-machine pc-q35-2.10 -cpu Broadwell-noTSX
If pmtimer actually works, you'll see something like:
pmtimer delta is 0xdff
tsc delta is 0x278756
This outcome was tested using qemu+kvm with UEFI (OVMF) firmware, and
these options: -machine pc-i440fx-2.4 -cpu Broadwell-noTSX
I've also tested this outcome on a real Intel Xeon E3-1275v3 on an Intel
Server Board S1200V3RPS using the SDV.RP.B8 "Release" build here:
https://www.intel.com/content/www/us/en/download/674448/firmware-update-for-the-intel-server-board-s1200rp-uefi-development-kit-release-vb8.html
Signed-off-by: Peter Jones <pjones@redhat.com>
Signed-off-by: Javier Martinez Canillas <javierm@redhat.com>
Signed-off-by: Robbie Harwood <rharwood@redhat.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
luks2_get_keyslot() can fail for a variety of reasons that do not necessarily
mean the next keyslot should not be tried (e.g. a new kdf type). So always
try the next slot. This will make GRUB more resilient to non-spec json data
that 3rd party systems may add. We do not care if some of the keyslots are
unusable, only if there is at least one that is.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
This command is meant to behave similarly to the "mode" command of the EFI
Shell application. In addition to allowing mode selection by giving the
number of columns and rows as arguments, the command allows specifying the
mode number to select the mode. Also supported are the arguments "min" and
"max", which set the mode to the minimum and maximum mode respectively as
calculated by the columns * rows of that mode.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
When filesystem detection fails, all that's currently debug-logged is
a series of messages like:
grub-core/kern/fs.c:56:fs: Detecting ntfs...
grub-core/kern/fs.c:76:fs: ntfs detection failed.
repeated for each filesystem. Any messages provided to grub_error() by
the filesystem are lost, and one has to break out gdb to figure out what
went wrong.
With this change, one instead sees:
grub-core/kern/fs.c:56:fs: Detecting fat...
grub-core/osdep/hostdisk.c:357:hostdisk: reusing open device
`/path/to/device'
grub-core/kern/fs.c:77:fs: error: invalid modification timestamp for /.
grub-core/kern/fs.c:79:fs: fat detection failed.
in the debug prints.
Signed-off-by: Robbie Harwood <rharwood@redhat.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The way the code is written the tofree variable would never be passed to
the free_subchunk() function uninitialized. Coverity cannot determine
this and flags the situation as "Using uninitialized value...". The fix
is just to initialize the local struct.
Fixes: CID 314016
Signed-off-by: Ross Philipson <ross.philipson@oracle.com>
Reviewed-by: Darren Kenny <darren.kenny@oracle.com>
Tested-by: Alec Brown <alec.r.brown@oracle.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The EFI_CC_MEASUREMENT_PROTOCOL abstracts the measurement for virtual firmware
in confidential computing environment. It is similar to the EFI_TCG2_PROTOCOL.
It was proposed by Intel and ARM and approved by UEFI organization.
It is defined in Intel GHCI specification: https://cdrdv2.intel.com/v1/dl/getContent/726790 .
The EDKII header file is available at https://github.com/tianocore/edk2/blob/master/MdePkg/Include/Protocol/CcMeasurement.h .
Signed-off-by: Lu Ken <ken.lu@intel.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The event description is a string, so using grub_strcpy() is cleaner than
using grub_memcpy().
Signed-off-by: Lu Ken <ken.lu@intel.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
1. Use macro GRUB_ERR_NONE instead of hard code 0.
2. Keep lowercase of the first char for the status string of log event.
Signed-off-by: Lu Ken <ken.lu@intel.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Introduce ERROR_PLATFORM_NOT_SUPPORT_SSP environment variable to treat
the "--enable-stack-protector is only supported on EFI platforms" message
as a warning instead of an error. If ERROR_PLATFORM_NOT_SUPPORT_SSP is
set to "no" (case-insensitive), then the message will be printed as
a warning. Otherwise, it prints as an error. The default behavior is to
print the message as an error.
For any wrapper build script that has some variation of:
for p in SELECTED_GRUB_PLATFORMS; do \
configure --enable-stack-protector \
--with-platform${P} ... || die; \
done
make
The GRUB will fail to build if SELECTED_GRUB_PLATFORMS contains a platform
that does not support SSP.
Such wrapper scripts need to work-around this issue by modifying the
above for-loop, so it conditionally passes --enable-stack-protector to
configure for the proper GRUB platform(s).
However, if the above example is modified to have to conditionally pass
in --enable-stack-protector, its behavior is effectively the same as the
proposed change. Additionally, The list of SSP supported platforms is
now in 2 places. One in the configure script and one in the build wrapper
script. If the second list is not properly maintained it could mistakenly
disable SSP for a platform that later gained support for it.
Signed-off-by: Nicholas Vinson <nvinson234@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
With gsub substitutions the offsets should be validated against the
number of glyphs in a font face and the memory allocated for the gsub
substitution data.
Both the number of glyphs and the last address in the allocated data are
passed in to process_cursive(), where the number of glyphs validates the end
of the range.
Enabling memory allocation validation uses two macros, one to simply check the
address against the allocated space, and the other to check that the number of
items of a given size doesn't extend outside of the allocated space.
Fixes: CID 73770
Fixes: CID 314040
Signed-off-by: Darren Kenny <darren.kenny@oracle.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
There are no users left of version_find_latest(), version_test_gt(), and
version_test_numeric(). Remove those unused helper functions. Using
those helper functions is what caused the quadratic sorting performance
issues in the first place, so removing them is a net win.
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Reviewed-by: Robbie Harwood <rharwood@redhat.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The current implementation of the 10_kfreebsd script implements its menu
items sorting in bash with a quadratic algorithm, calling "sed", "sort",
"head", and "grep" to compare versions between individual lines, which
is annoyingly slow for kernel developers who can easily end up with
50-100 kernels in their boot partition.
This fix is ported from the 10_linux script, which has a similar
quadratic code pattern.
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Cc: debian-bsd@lists.debian.org
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The current implementation of the 10_hurd script implements its menu
items sorting in bash with a quadratic algorithm, calling "sed", "sort",
"head", and "grep" to compare versions between individual lines, which
is annoyingly slow for kernel developers who can easily end up with
50-100 kernels in their boot partition.
This fix is ported from the 10_linux script, which has a similar
quadratic code pattern.
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Cc: Samuel Thibault <samuel.thibault@ens-lyon.org>
Tested-by: Samuel Thibault <samuel.thibault@ens-lyon.org>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The current implementation of the 20_linux_xen script implements its
menu items sorting in bash with a quadratic algorithm, calling "sed",
"sort", "head", and "grep" to compare versions between individual lines,
which is annoyingly slow for kernel developers who can easily end up
with 50-100 kernels in their boot partition.
This fix is ported from the 10_linux script, which has a similar
quadratic code pattern.
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Cc: xen-devel@lists.xenproject.org
Tested-by: Jason Andryuk <jandryuk@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The current implementation of the 10_linux script implements its menu
items sorting in bash with a quadratic algorithm, calling "sed", "sort",
"head", and "grep" to compare versions between individual lines, which
is annoyingly slow for kernel developers who can easily end up with
50-100 kernels in /boot.
As an example, on a Intel(R) Core(TM) i7-8650U CPU @ 1.90GHz, running:
/usr/sbin/grub-mkconfig > /dev/null
With 44 kernels in /boot, this command takes 10-15 seconds to complete.
After this fix, the same command runs in 5 seconds.
With 116 kernels in /boot, this command takes 40 seconds to complete.
After this fix, the same command runs in 8 seconds.
For reference, the quadratic algorithm here is:
while [ "x$list" != "x" ] ; do <--- outer loop
linux=`version_find_latest $list`
version_find_latest()
for i in "$@" ; do <--- inner loop
version_test_gt()
fork+exec sed
version_test_numeric()
version_sort
fork+exec sort
fork+exec head -n 1
fork+exec grep
list=`echo $list | tr ' ' '\n' | fgrep -vx "$linux" | tr '\n' ' '`
tr
fgrep
tr
So all commands executed under version_test_gt() are executed
O(n^2) times where n is the number of kernel images in /boot.
Here is the improved algorithm proposed:
- Prepare a list with all the relevant information for ordering by a single
sort(1) execution. This is done by renaming ".old" suffixes by " 1" and
by suffixing all other files with " 2", thus making sure the ".old" entries
will follow the non-old entries in reverse-sorted-order.
- Call version_reverse_sort on the list (sort -r -V): A single execution of
sort(1). For instance, GNU coreutils' sort will reverse-sort the list in
O(n*log(n)) with a merge sort.
- Replace the " 1" suffixes by ".old", and remove the " 2" suffixes.
- Iterate on the reverse-sorted list to output each menu entry item.
Therefore, the algorithm proposed has O(n*log(n)) complexity with GNU
coreutils' sort compared to the prior O(n^2) complexity. Moreover, the
constant time required for each list entry is much less because sorting
is done within a single execution of sort(1) rather than requiring
O(n^2) executions of sed(1), sort(1), head(1), and grep(1) in
sub-shells.
Signed-off-by: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Reviewed-by: Robbie Harwood <rharwood@redhat.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Using the disk read hook mechanism, setup a read hook on the source disk
which will read from the given header file during the scan and recovery
cryptodisk backend functions. Disk read hooks are executed after the data
has been read from the disk. This is okay, because the read hook is given
the read buffer before its sent back to the caller. In this case, the hook
can then overwrite the data read from the disk device with data from the
header file sent in as the read hook data. This is transparent to the
read caller. Since the callers of this function have just opened the
source disk, there are no current read hooks, so there's no need to
save/restore them nor consider if they should be called or not.
This hook assumes that the header is at the start of the volume, which
is not the case for some formats (e.g. GELI). So GELI will return an
error if a detached header is specified. It also can only be used
with formats where the detached header file can be written to the
first blocks of the volume and the volume could still be unlocked.
So the header file can not be formatted differently from the on-disk
header. If these assumpts are not met, detached header file processing
must be specially handled in the cryptodisk backend module.
The hook will be called potentially many times by a backend. This is fine
because of the assumptions mentioned and the read hook reads from absolute
offsets and is stateless.
Also add a --header (short -H) option to cryptomount which takes a file
argument.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Patrick Steinhardt <ps@pks.im>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
It will be desirable in the future to allow having the read hook modify the
data passed back from a read function call on a disk or file. This adds that
infrastructure and has no impact on code flow for existing uses of the read
hook. Also changed is that now when the read hook callback is called it can
also indicate what error code should be sent back to the read caller.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Patrick Steinhardt <ps@pks.im>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Document the variables net_<interface>_clientid, net_<interface>_clientuuid,
lockdown, and shim_lock in the list of special environment variables.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Adjust the interface of grub_efi_mm_add_regions() to take a set of
GRUB_MM_ADD_REGION_* flags, which most notably is currently only the
GRUB_MM_ADD_REGION_CONSECUTIVE flag. This allows us to set the function
up as callback for the memory subsystem and have it call out to us in
case there's not enough pages available in the current heap.
Signed-off-by: Patrick Steinhardt <ps@pks.im>
Signed-off-by: Daniel Axtens <dja@axtens.net>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Tested-by: Patrick Steinhardt <ps@pks.im>
The function add_memory_regions() is currently only called on system
initialization to allocate a fixed amount of pages. As such, it didn't
need to return any errors: in case it failed, we cannot proceed anyway.
This will change with the upcoming support for requesting more memory
from the firmware at runtime, where it doesn't make sense anymore to
fail hard.
Refactor the function to return an error to prepare for this. Note that
this does not change the behaviour when initializing the memory system
because grub_efi_mm_init() knows to call grub_fatal() in case
grub_efi_mm_add_regions() returns an error.
Signed-off-by: Patrick Steinhardt <ps@pks.im>
Signed-off-by: Daniel Axtens <dja@axtens.net>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Tested-by: Patrick Steinhardt <ps@pks.im>
In preparation of support for runtime-allocating additional memory
region, this patch extracts the function to retrieve the EFI memory
map and add a subset of it to GRUB's own memory regions.
Signed-off-by: Patrick Steinhardt <ps@pks.im>
Signed-off-by: Daniel Axtens <dja@axtens.net>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Tested-by: Patrick Steinhardt <ps@pks.im>
When initializing the EFI memory subsystem, we will by default request
a quarter of the available memory, bounded by a minimum/maximum value.
Given that we're about to extend the EFI memory system to dynamically
request additional pages from the firmware as required, this scaling of
requested memory based on available memory will not make a lot of sense
anymore.
Remove this logic as a preparatory patch such that we'll instead defer
to the runtime memory allocator. Note that ideally, we'd want to change
this after dynamic requesting of pages has been implemented for the EFI
platform. But because we'll need to split up initialization of the
memory subsystem and the request of pages from the firmware, we'd have
to duplicate quite some logic at first only to remove it afterwards
again. This seems quite pointless, so we instead have patches slightly
out of order.
Signed-off-by: Patrick Steinhardt <ps@pks.im>
Signed-off-by: Daniel Axtens <dja@axtens.net>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Tested-by: Patrick Steinhardt <ps@pks.im>
Currently, all platforms will set up their heap on initialization of the
platform code. While this works mostly fine, it poses some limitations
on memory management on us. Most notably, allocating big chunks of
memory in the gigabyte range would require us to pre-request this many
bytes from the firmware and add it to the heap from the beginning on
some platforms like EFI. As this isn't needed for most configurations,
it is inefficient and may even negatively impact some usecases when,
e.g., chainloading. Nonetheless, allocating big chunks of memory is
required sometimes, where one example is the upcoming support for the
Argon2 key derival function in LUKS2.
In order to avoid pre-allocating big chunks of memory, this commit
implements a runtime mechanism to add more pages to the system. When
a given allocation cannot be currently satisfied, we'll call a given
callback set up by the platform's own memory management subsystem,
asking it to add a memory area with at least "n" bytes. If this
succeeds, we retry searching for a valid memory region, which should
now succeed.
If this fails, we try asking for "n" bytes, possibly spread across
multiple regions, in hopes that region merging means that we end up
with enough memory for things to work out.
Signed-off-by: Patrick Steinhardt <ps@pks.im>
Signed-off-by: Daniel Axtens <dja@axtens.net>
Tested-by: Stefan Berger <stefanb@linux.ibm.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Tested-by: Patrick Steinhardt <ps@pks.im>
In grub_memalign(), there's a commented section which would allow for
unloading of unneeded modules in case where there is not enough free
memory available to satisfy a request. Given that this code is never
compiled in, let's remove it together with grub_dl_unload_unneeded().
Signed-off-by: Patrick Steinhardt <ps@pks.im>
Signed-off-by: Daniel Axtens <dja@axtens.net>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Tested-by: Patrick Steinhardt <ps@pks.im>
This is handy for debugging. Enable with "set debug=regions".
Signed-off-by: Daniel Axtens <dja@axtens.net>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Tested-by: Patrick Steinhardt <ps@pks.im>
On x86_64-efi (at least) regions seem to be added from top down. The mm
code will merge a new region with an existing region that comes
immediately before the new region. This allows larger allocations to be
satisfied that would otherwise be the case.
On powerpc-ieee1275, however, regions are added from bottom up. So if
we add 3x 32MB regions, we can still only satisfy a 32MB allocation,
rather than the 96MB allocation we might otherwise be able to satisfy.
* Define 'post_size' as being bytes lost to the end of an allocation
due to being given weird sizes from firmware that are not multiples
of GRUB_MM_ALIGN.
* Allow merging of regions immediately _after_ existing regions, not
just before. As with the other approach, we create an allocated
block to represent the new space and the pass it to grub_free() to
get the metadata right.
Signed-off-by: Daniel Axtens <dja@axtens.net>
Tested-by: Stefan Berger <stefanb@linux.ibm.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Tested-by: Patrick Steinhardt <ps@pks.im>
grub_mm_region_init() does:
h = (grub_mm_header_t) (r + 1);
where h is a grub_mm_header_t and r is a grub_mm_region_t.
Cells are supposed to be GRUB_MM_ALIGN aligned, but while grub_mm_dump
ensures this vs the region header, grub_mm_region_init() does not.
It's better to be explicit than implicit here: rather than changing
grub_mm_region_init() to ALIGN_UP(), require that the struct is
explicitly a multiple of the header size.
Signed-off-by: Daniel Axtens <dja@axtens.net>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Tested-by: Patrick Steinhardt <ps@pks.im>
This breaks the tests on pseries - just restrict it to x86 platforms
that don't specify an EFI.
Signed-off-by: Daniel Axtens <dja@axtens.net>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The corpus was generating issues in grub_btrfs_read_logical() when
attempting to iterate over stripe entries in the superblock's
bootmapping.
In most cases the reason for the failure was that the number of stripes
in chunk->nstripes exceeded the possible space statically allocated in
superblock bootmapping space. Each stripe entry in the bootmapping block
consists of a grub_btrfs_key followed by a grub_btrfs_chunk_stripe.
Another issue that came up was that while calculating the chunk size,
in an earlier piece of code in that function, depending on the data
provided in the btrfs file system, it would end up calculating a size
that was too small to contain even 1 grub_btrfs_chunk_item, which is
obviously invalid too.
Signed-off-by: Darren Kenny <darren.kenny@oracle.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The fuzzer is generating btrfs file systems that have chunks with
invalid combinations of stripes and substripes for the given RAID
configurations.
After examining the Linux kernel fs/btrfs/tree-checker.c code, it
appears that sub-stripes should only be applied to RAID10, and in that
case there should only ever be 2 of them.
Similarly, RAID single should only have 1 stripe, and RAID1/1C3/1C4
should have 2. 3 or 4 stripes respectively, which is what redundancy
corresponds.
Some of the chunks ended up with a size of 0, which grub_malloc() still
returned memory for and in turn generated ASAN errors later when
accessed.
While it would be possible to specifically limit the number of stripes,
a more correct test was on the combination of the chunk item, and the
number of stripes by the size of the chunk stripe structure in
comparison to the size of the chunk itself.
Signed-off-by: Darren Kenny <darren.kenny@oracle.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
According to the btrfs code in Linux, the structure of a directory item
leaf should be of the form:
|struct btrfs_dir_item|name|data|
in GRUB the name len and data len are in the grub_btrfs_dir_item
structure's n and m fields respectively.
The combined size of the structure, name and data should be less than
the allocated memory, a difference to the Linux kernel's struct
btrfs_dir_item is that the grub_btrfs_dir_item has an extra field for
where the name is stored, so we adjust for that too.
Signed-off-by: Darren Kenny <darren.kenny@oracle.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
A corrupt f2fs file system might specify a name length which is greater
than the maximum name length supported by the GRUB f2fs driver.
We will allocate enough memory to store the overly long name, but there
are only F2FS_NAME_LEN bytes in the source, so we would read past the end
of the source.
While checking directory entries, do not copy a file name with an invalid
length.
Signed-off-by: Sudhakar Kuppusamy <sudhakar@linux.ibm.com>
Signed-off-by: Daniel Axtens <dja@axtens.net>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
A corrupt f2fs filesystem could have a block offset or a bitmap
offset that would cause us to read beyond the bounds of the nat
bitmap.
Introduce the nat_bitmap_size member in grub_f2fs_data which holds
the size of nat bitmap.
Set the size when loading the nat bitmap in nat_bitmap_ptr(), and
catch when an invalid offset would create a pointer past the end of
the allocated space.
Check against the bitmap size in grub_f2fs_test_bit() test bit to avoid
reading past the end of the nat bitmap.
Signed-off-by: Sudhakar Kuppusamy <sudhakar@linux.ibm.com>
Signed-off-by: Daniel Axtens <dja@axtens.net>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
A corrupt f2fs file system could specify a nat journal entry count
that is beyond the maximum NAT_JOURNAL_ENTRIES.
Check if the specified nat journal entry count before accessing the
array, and throw an error if it is too large.
Signed-off-by: Sudhakar Kuppusamy <sudhakar@linux.ibm.com>
Signed-off-by: Daniel Axtens <dja@axtens.net>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
In a similar vein to the previous patch, parse_line() would write
a NUL byte past the end of the buffer if there was an HTTP header
with a LF rather than a CRLF.
RFC-2616 says:
Many HTTP/1.1 header field values consist of words separated by LWS
or special characters. These special characters MUST be in a quoted
string to be used within a parameter value (as defined in section 3.6).
We don't support quoted sections or continuation lines, etc.
If we see an LF that's not part of a CRLF, bail out.
Fixes: CVE-2022-28734
Signed-off-by: Daniel Axtens <dja@axtens.net>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
GRUB has special code for handling an http header that is split
across two packets.
The code tracks the end of line by looking for a "\n" byte. The
code for split headers has always advanced the pointer just past the
end of the line, whereas the code that handles unsplit headers does
not advance the pointer. This extra advance causes the length to be
one greater, which breaks an assumption in parse_line(), leading to
it writing a NUL byte one byte past the end of the buffer where we
reconstruct the line from the two packets.
It's conceivable that an attacker controlled set of packets could
cause this to zero out the first byte of the "next" pointer of the
grub_mm_region structure following the current_line buffer.
Do not advance the pointer in the split header case.
Fixes: CVE-2022-28734
Signed-off-by: Daniel Axtens <dja@axtens.net>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
It's possible for data->sock to get torn down in tcp error handling.
If we unconditionally tear it down again we will end up doing writes
to an offset of the NULL pointer when we go to tear it down again.
Detect if it has been torn down and don't do it again.
Signed-off-by: Daniel Axtens <dja@axtens.net>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Under tftp errors, we print a tftp error message from the tftp header.
However, the tftph pointer is a pointer inside nb, the netbuff. Previously,
we were freeing the nb and then dereferencing it. Don't do that, use it
and then free it later.
This isn't really _bad_ per se, especially as we're single-threaded, but
it trips up fuzzers.
Signed-off-by: Daniel Axtens <dja@axtens.net>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
A malicious tftp server can cause UAFs and a double free.
An attempt to read from a network file is handled by grub_net_fs_read(). If
the read is at an offset other than the current offset, grub_net_seek_real()
is invoked.
In grub_net_seek_real(), if a backwards seek cannot be satisfied from the
currently received packets, and the underlying transport does not provide
a seek method, then grub_net_seek_real() will close and reopen the network
protocol layer.
For tftp, the ->close() call goes to tftp_close() and frees the tftp_data_t
file->data. The file->data pointer is not nulled out after the free.
If the ->open() call fails, the file->data will not be reallocated and will
continue point to a freed memory block. This could happen from a server
refusing to send the requisite ack to the new tftp request, for example.
The seek and the read will then fail, but the grub_file continues to exist:
the failed seek does not necessarily cause the entire file to be thrown
away (e.g. where the file is checked to see if it is gzipped/lzio/xz/etc.,
a read failure is interpreted as a decompressor passing on the file, not as
an invalidation of the entire grub_file_t structure).
This means subsequent attempts to read or seek the file will use the old
file->data after free. Eventually, the file will be close()d again and
file->data will be freed again.
Mark a net_fs file that doesn't reopen as broken. Do not permit read() or
close() on a broken file (seek is not exposed directly to the file API -
it is only called as part of read, so this blocks seeks as well).
As an additional defence, null out the ->data pointer if tftp_open() fails.
That would have lead to a simple null pointer dereference rather than
a mess of UAFs.
This may affect other protocols, I haven't checked.
Signed-off-by: Daniel Axtens <dja@axtens.net>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
I don't really understand what's going on here but fuzzing found
a bug where we read past the end of check_with. That's a C string,
so use grub_strlen() to make sure we don't overread it.
Signed-off-by: Daniel Axtens <dja@axtens.net>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
grub_net_dns_lookup() takes as inputs a pointer to an array of addresses
("addresses") for the given name, and pointer to a number of addresses
("naddresses"). grub_net_dns_lookup() is responsible for allocating
"addresses", and the caller is responsible for freeing it if
"naddresses" > 0.
The DNS recv_hook will sometimes set and free the addresses array,
for example if the packet is too short:
if (ptr + 10 >= nb->tail)
{
if (!*data->naddresses)
grub_free (*data->addresses);
grub_netbuff_free (nb);
return GRUB_ERR_NONE;
}
Later on the nslookup command code unconditionally frees the "addresses"
array. Normally this is fine: the array is either populated with valid
data or is NULL. But in these sorts of error cases it is neither NULL
nor valid and we get a double-free.
Only free "addresses" if "naddresses" > 0.
It looks like the other use of grub_net_dns_lookup() is not affected.
Signed-off-by: Daniel Axtens <dja@axtens.net>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
A netbuff shouldn't be too huge. It's bounded by MTU and TCP segment
reassembly. If we are asked to create one that is unreasonably big, refuse.
This is a hardening measure: if we hit this code, there's a bug somewhere
else that we should catch and fix.
This commit:
- stops the bug propagating any further.
- provides a spot to instrument in e.g. fuzzing to try to catch these bugs.
I have put instrumentation (e.g. __builtin_trap() to force a crash) here and
have not been able to find any more crashes.
Signed-off-by: Daniel Axtens <dja@axtens.net>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
We can receive packets with invalid IP fragmentation information. This
can lead to rsm->total_len underflowing and becoming very large.
Then, in grub_netbuff_alloc(), we add to this very large number, which can
cause it to overflow and wrap back around to a small positive number.
The allocation then succeeds, but the resulting buffer is too small and
subsequent operations can write past the end of the buffer.
Catch the underflow here.
Fixes: CVE-2022-28733
Signed-off-by: Daniel Axtens <dja@axtens.net>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
In some cases attempting to display arbitrary binary strings leads
to ASAN splats reading the widthspec array out of bounds.
Check the index. If it would be out of bounds, return a width of 1.
I don't know if that's strictly correct, but we're not really expecting
great display of arbitrary binary data, and it's certainly not worse than
an OOB read.
Signed-off-by: Daniel Axtens <dja@axtens.net>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Certain 1 px wide images caused a wild pointer write in
grub_jpeg_ycrcb_to_rgb(). This was caused because in grub_jpeg_decode_data(),
we have the following loop:
for (; data->r1 < nr1 && (!data->dri || rst);
data->r1++, data->bitmap_ptr += (vb * data->image_width - hb * nc1) * 3)
We did not check if vb * width >= hb * nc1.
On a 64-bit platform, if that turns out to be negative, it will underflow,
be interpreted as unsigned 64-bit, then be added to the 64-bit pointer, so
we see data->bitmap_ptr jump, e.g.:
0x6180_0000_0480 to
0x6181_0000_0498
^
~--- carry has occurred and this pointer is now far away from
any object.
On a 32-bit platform, it will decrement the pointer, creating a pointer
that won't crash but will overwrite random data.
Catch the underflow and error out.
Fixes: CVE-2021-3697
Signed-off-by: Daniel Axtens <dja@axtens.net>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
An invalid file could contain multiple start of stream blocks, which
would cause us to reallocate and leak our bitmap. Refuse to handle
multiple start of streams.
Additionally, fix a grub_error() call formatting.
Signed-off-by: Daniel Axtens <dja@axtens.net>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Fix a memory leak where an invalid file could cause us to reallocate
memory for a huffman table we had already allocated memory for.
Signed-off-by: Daniel Axtens <dja@axtens.net>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Fuzzing revealed some inputs that were taking a long time, potentially
forever, because they did not bail quickly upon encountering an I/O error.
Try to catch I/O errors sooner and bail out.
Signed-off-by: Daniel Axtens <dja@axtens.net>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
ASAN picked up two OOB global reads: we weren't checking if some code
values fit within the cplens or cpdext arrays. Check and throw an error
if not.
Signed-off-by: Daniel Axtens <dja@axtens.net>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
In fuzzing we observed crashes where a code would attempt to be inserted
into a huffman table before the start, leading to a set of heap OOB reads
and writes as table entries with negative indices were shifted around and
the new code written in.
Catch the case where we would underflow the array and bail.
Fixes: CVE-2021-3696
Signed-off-by: Daniel Axtens <dja@axtens.net>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
A 16-bit greyscale PNG without alpha is processed in the following loop:
for (i = 0; i < (data->image_width * data->image_height);
i++, d1 += 4, d2 += 2)
{
d1[R3] = d2[1];
d1[G3] = d2[1];
d1[B3] = d2[1];
}
The increment of d1 is wrong. d1 is incremented by 4 bytes per iteration,
but there are only 3 bytes allocated for storage. This means that image
data will overwrite somewhat-attacker-controlled parts of memory - 3 bytes
out of every 4 following the end of the image.
This has existed since greyscale support was added in 2013 in commit
3ccf16dff9 (grub-core/video/readers/png.c: Support grayscale).
Saving starfield.png as a 16-bit greyscale image without alpha in the gimp
and attempting to load it causes grub-emu to crash - I don't think this code
has ever worked.
Delete all PNG greyscale support.
Fixes: CVE-2021-3695
Signed-off-by: Daniel Axtens <dja@axtens.net>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
This causes the bitmap to be leaked. Do not permit multiple image headers.
Signed-off-by: Daniel Axtens <dja@axtens.net>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Fuzzing revealed some inputs that were taking a long time, potentially
forever, because they did not bail quickly upon encountering an I/O error.
Try to catch I/O errors sooner and bail out.
Signed-off-by: Daniel Axtens <dja@axtens.net>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
If we have an error in grub_file_open() before we free device_name, we
will leak it.
Free device_name in the error path and null out the pointer in the good
path once we free it there.
Signed-off-by: Daniel Axtens <dja@axtens.net>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
We must not allow other verifiers to pass things like the GRUB modules.
Instead of maintaining a blocklist, maintain an allowlist of things
that we do not care about.
This allowlist really should be made reusable, and shared by the
lockdown verifier, but this is the minimal patch addressing
security concerns where the TPM verifier was able to mark modules
as verified (or the OpenPGP verifier for that matter), when it
should not do so on shim-powered secure boot systems.
Fixes: CVE-2022-28735
Signed-off-by: Julian Andres Klode <julian.klode@canonical.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
This ports the EFI chainloader to use grub_loader_set_ex() in order to fix
a use-after-free bug that occurs when grub_cmd_chainloader() is executed
more than once before a boot attempt is performed.
Fixes: CVE-2022-28736
Signed-off-by: Chris Coulson <chris.coulson@canonical.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Loaders rely on global variables for saving context which is consumed
in the boot hook and freed in the unload hook. In the case where a loader
command is executed twice, calling grub_loader_set() a second time executes
the unload hook, but in some cases this runs when the loader's global
context has already been updated, resulting in the updated context being
freed and potential use-after-free bugs when the boot hook is subsequently
called.
This adds a new API, grub_loader_set_ex(), which allows a loader to specify
context that is passed to its boot and unload hooks. This is an alternative
to requiring that loaders call grub_loader_unset() before mutating their
global context.
Signed-off-by: Chris Coulson <chris.coulson@canonical.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The chainloader command retains the source buffer and device path passed
to LoadImage(), requiring the unload hook passed to grub_loader_set() to
free them. It isn't required to retain this state though - they aren't
required by StartImage() or anything else in the boot hook, so clean them
up before grub_cmd_chainloader() finishes.
Signed-off-by: Chris Coulson <chris.coulson@canonical.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Coverity reports that while loopis in the following functions uses
tainted data as boundary:
zfs_mount() -> check_mos_features() -> dnode_get() -> zfs_log2()
zfs_mount() -> grub_memmove()
The defect type is "Untrusted loop bound" caused as a result of
"tainted_data_downcast". Coverity does not like the pointer downcast
here and we need to address it.
We believe Coverity flags pointer downcast for the following two
reasons:
1. External data: The pointer downcast could indicate that the source is
external data, which we need to further sanitize - such as verifying its
limits. In this case, the data is read from an external source, which is
a disk. But, zio_read(), which reads the data from the disk, sanitizes it
using a checksum. checksum is the best facility that ZFS offers to verify
external data, and we don't believe a better way exists. Therefore, no
further action is possible for this.
2. Corruption due to alignment: downcasting a pointer from a strict type
to less strict type could result in data corruption. For example, the
following cast would corrupt because uint32_t is 4-byte aligned, and
won't be able to point to 0x1003 which is not 4-byte aligned.
uint8_t *ptr = 0x1003;
uint32_t *word = ptr; (incorrect, alignment issues)
This patch converts the "osp" pointer in zfs_mount() from a "void" type
to "objset_phys_t" type to address this issue.
We are not sure if there are any other reasons why Coverity flags the
downcast. However, the fix for alignment issue masks/suppresses any
other issues from showing up.
Fixes: CID 314023
Signed-off-by: Jagannathan Raman <jag.raman@oracle.com>
Reviewed-by: Darren Kenny <darren.kenny@oracle.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Coverity reports that the while loop in the following function uses
tainted data as boundary:
fill_fs_info() -> dnode_get() -> zfs_log2()
The tainted originated from:
fill_fs_info() -> make_mdn()
The defect type is "Untrusted loop bound" caused as a result of
"tainted_data_downcast". Coverity does not like the pointer downcast
here and we need to address it.
We believe Coverity flags pointer downcast for the following two
reasons:
1. External data: The pointer downcast could indicate that the source is
external data, which we need to further sanitize - such as verifying its
limits. In this case, the data is read from an external source, which is
a disk. But, zio_read(), which reads the data from the disk, sanitizes it
using a checksum. checksum is the best facility that ZFS offers to verify
external data, and we don't believe a better way exists. Therefore, no
further action is possible for this.
2. Corruption due to alignment: downcasting a pointer from a strict type
to less strict type could result in data corruption. For example, the
following cast would corrupt because uint32_t is 4-byte aligned, and
won't be able to point to 0x1003 which is not 4-byte aligned.
uint8_t *ptr = 0x1003;
uint32_t *word = ptr; (incorrect, alignment issues)
This patch converts the "osp" pointer in make_mdn() from a "void" type
to "objset_phys_t" type to address the issue.
We are not sure if there are any other reasons why Coverity flags the
downcast. However, the fix for alignment issue masks/suppresses any
other issues from showing up.
Fixes: CID 314020
Signed-off-by: Jagannathan Raman <jag.raman@oracle.com>
Reviewed-by: Darren Kenny <darren.kenny@oracle.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
In util/grub-module-verifierXX.c, the function get_shdr() is used to obtain the
section header at a given index but isn't checking that there is an offset for
the section header table. To validate that there is, we can check that e_shoff
isn't 0.
Signed-off-by: Alec Brown <alec.r.brown@oracle.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Reviewed-by: Darren Kenny <darren.kenny@oracle.com>
In bsdXX.c, a couple of untrusted loop bound and untrusted allocation size bugs
were flagged by Coverity in the functions grub_openbsd_find_ramdisk() and
grub_freebsd_load_elfmodule(). These bugs were flagged by coverity because the
variable shdr was downcasting from a char pointer to an Elf_Shdr pointer
whenever it was used to set the base value in for loops. To avoid this, we need
to set shdr as an Elf_Shdr pointer where it is initialized.
In the function read_headers(), the function is reading elf section header data
from a file and passing it to the variable shdr as data for a char pointer. If
we switch the type of shdr to an Elf_Shdr pointer in read_headers() as well as
other functions, then we won't need to downcast to an Elf_Shdr pointer. By doing
this, the issue becomes masked from Coverity's view. In the following patches,
we check limits to ensure the data isn't tainted.
Also, switched use of (char *) to (grub_uint8_t *) to give a better indication
of pointer arithmetic and not suggest use of a C string.
Fixes: CID 314018
Fixes: CID 314030
Fixes: CID 314031
Fixes: CID 314039
Signed-off-by: Alec Brown <alec.r.brown@oracle.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Reviewed-by: Darren Kenny <darren.kenny@oracle.com>
Some devices report IoAlign values but seem to require buffers with
higher alignment.
The UEFI specification is saying: "IoAlign values of 0 and 1 mean that
the buffer can be placed anywhere in memory. Otherwise, IoAlign must
be a power of 2, and the requirement is that the start address of
a buffer must be evenly divisible by IoAlign with no remainder."
Some devices report IoAlign of 2, however seem to require 4 bytes
aligned buffers. It seems that this got misinterpreted by some vendors
assuming IoAlign is 2^IoAlign. There is also such a hint in an example
in earlier versions of the Driver Writer's Guide:
ScsiPassThruMode.IoAlign = 2; // Data must be alligned on 4-byte boundary
Some devices report no alignment requirements at all but seem to read
corrupted data or report read errors when passing unaligned buffers.
Work around by using an alignment of at least BlockSize (typically 512
bytes) in any case. If IoAlign (interpreted as per UEFI specification)
requests a higher alignment than BlockSize, follow IoAlign still.
Note: The problem has only noticed with compressed squashfs. It seems
that ext4 (and presumably other file system drivers) pass buffers with
a higher alignment already.
Signed-off-by: Stefan Agner <stefan@agner.ch>
Acked-by: Heinrich Schuchardt <heinrich.schuchardt@canaonical.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
When using userland drivers such as rumpdisk, we'd rather make ext2fs use
parted-based libstore partitioning support. That can be used for kernelland
drivers as well, so we can just make GRUB always use the "part:" qualifier
to switch ext2fs to it.
grub_util_find_hurd_root_device() then has to understand this syntax and
translate it into the /dev/ entry name.
Signed-off-by: Samuel Thibault <samuel.thibault@ens-lyon.org>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Add the options --key-file, --keyfile-offset, and --keyfile-size to
cryptomount and code to put read the requested key file data and pass
via the cargs struct. Note, key file data is for all intents and purposes
equivalent to a password given to cryptomount. So there is no need to
enable support for key files in the various crypto backends (e.g. LUKS1)
because the key data is passed just as if it were a password.
Signed-off-by: John Lane <john@lane.uk.net>
Signed-off-by: Denis 'GNUtoo' Carikli <GNUtoo@cyberdimension.org>
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The partmap test requires no elevated privileges. However, it uses parted
which can be used as a normal user, but is usually located in /sbin or
/usr/bin (eg. on Debian systems). Whereas the normal user does not usually
have /sbin or /usr/sbin added to their path, thus parted will not be found
causing the test to abort. Add /sbin and /usr/sbin to the path for the
partmap test so that the test can run successfully as an unprivileged user.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
On failure, the hfs test should show both the host and GRUB determined fs
UUID. Prior to this change, both outputs where generated by GRUB, which is
less helpful in determining the cause of failure.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The section is an itemized list of commands that are not listed else where
in the command sections.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Move loader commands documented in the general commands list into the
loader command section.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Also, add period to terminate sentence.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The --no-apm option is only available on the i396-pc target.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The commit 848724273e (net/net: Avoid unnecessary calls to
grub_net_tcp_retransmit()) needs to have its condition inverted to avoid
unnecessary calls to grub_net_tcp_retransmit(). As it is, it creates many
unnecessary calls and does not call grub_net_tcp_retransmit() when needed.
The call to grub_net_tcp_retransmit() should only be made when
grub_net_cards does _not_ equal NULL, meaning that there are potentially
network cards that need TCP retransmission.
Fixes: 848724273e (net/net: Avoid unnecessary calls to grub_net_tcp_retransmit())
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Add detection for initramfs of the form *.img.old. For example, Gentoo's
sys-kernel/genkernel installs it as initramfs-*.img and moves any existing
one to initramfs-*.img.old.
Apply the same scheme to initrd-*.img and initrd-*.gz files for consistency.
Signed-off-by: Oskari Pirhonen <xxc3ncoredxx@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Those are used with non-bootstrap disk drivers, for which libstore has to
open /dev/disk before calling device_open on it instead of on the device
master port. Normally in that case all /dev/ entries also have the @/dev/disk:
qualifier, so we can just drop it.
Signed-off-by: Samuel Thibault <samuel.thibault@ens-lyon.org>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
A recent fix that made appears to have broken the ability to create an
aarch64 boot image on a x86-based host.
This was due to an overzealous testing of the architecture when building
grub-mkimage and removing the code that build an ARM image when not built
on ARM.
On the occasion remove redundant break.
Fixes: 8541f319 (grub-mkimage: Only check aarch64 relocations when built for aarch64)
Signed-off-by: Darren Kenny <darren.kenny@oracle.com>
Tested-by: Selva Ganesan <selvaganesan89@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Although the EFI specification enforces support for FAT ESP, it's free
for EFI implementations to implement support for ESPs with other formats
(e.g. ext4, ntfs, etc), and at least U-Boot EFI will support ext4 ESP if
U-Boot is built with ext4 support. In some situations a GRUB installation
on such a non-FAT ESP could be useful (e.g. a NTFS-based USB disk that
can dual boot a Windows installation media and a Linux LiveCD).
As this is advanced and implementation-dependent behavior, let grub-install
allow this kind of installation, but only when --force is specified.
Signed-off-by: Icenowy Zheng <uwu@icenowy.me>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
During UEFI PXE boot in IPv6 network, if the DHCP server adopts stateful
automatic configuration, then the client receives a ICMP6_ROUTER_ADVERTISE
multicast message from the server. This may be received without the interface
having a configured network address, so orig_inf will be NULL, which can lead
to a NULL dereference when creating the default route. Actually, in this case,
the client obtains the default route through DHCPv6 instead of RA messages.
So if orig_inf == NULL and route_inf == NULL, we should not set the
default route.
Fixes: https://savannah.gnu.org/bugs/?62072
Signed-off-by: Qiumiao Zhang <zhangqiumiao1@huawei.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
ZFS file systems are not unmounted using umount, but instead by exporting
them. So export the ZFS file system that has the same label as the one that
was created during the test, if such one exists. This is required to delete
the loopback device that uses the ZFS image file. Otherwise the added code
to delete all loopback devices setup during the test run will never be able
to finish because the loopback device can not be deleted while in use.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
When all tests complete successfully, filesystems mounted by grub-fs-tester
will be unmounted before exiting. However, on certain test failures the
tester will exit with a failure code and not unmount previously mounted
filesystems. Now keep track of mounts and umounts and run an exit handler
on exit or process interruption that will umount all mounts that haven't
already been unmounted.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Using "*" to prefix list items leads to undesirable display output for
at least the generation of the html documentation. Use the @itemize and
@item directives to get itemized list output.
Also fix some wording and punctuation issues.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
It is not clear from the documentation what a "list" is in the context
of the "if" command. Note that its a list of simple commands separated
by a ";" and that only the exit status of the last command matters.
The same is true for the "cond" parameter to the "while" command.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Instead of "tmp" the name is prefixed by the name of the scripts (e.g.
grub-fs-tester). A timestamp is added in the name to allow for easily
seeing a chronological sorting of runs and the name of the filesystem
being tested. The random component is set to the minimal possible,
3 characters, because the timestamp should provide enough uniqueness.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Using the blkid cache can cause issues when running many file system tests
in parallel. We do not need it, as its only there to improve performance,
and using the cache does not provide significant performance improvements.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Autoconf will set a default CFLAGS of "-g -O2" if CFLAGS is not set.
CFLAGS was defaulted to "" early in configure to prevent this. A recent
commit ad9ccf660 (configure: Fix various new autotools warnings) added
AC_USE_SYSTEM_EXTENSIONS, which pulls in the autoconf CFLAGS check,
before we default CFLAGS and thus setting the autoconf default for
CFLAGS. Move the default setting of CFLAGS to before AC_USE_SYSTEM_EXTENSIONS
so that autoconf will see CFLAGS as set and not give it a default.
CFLAGS is also moved above AC_CONFIG_AUX_DIR, because CFLAGS should be
defaulted to "" as soon as possible to catch any autoconf macros that try
to use some other default. Regardless, this currently has no effect as that
macro does not consider the CFLAGS variable.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Robbie Harwood <rharwood@redhat.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The value of next_marker is adjusted based on the word sized value
read from data->file.
The updated next_marker value should reference a location in the file
just beyond the huffman table, and as such should not have a value
larger than the size of the file.
Fixes: CID 73657
Signed-off-by: Darren Kenny <darren.kenny@oracle.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The grub_absolute_pointer() is a compound expression that can only work
within a function. We are out of luck here when the pointer variables
require global definition due to ATTRIBUTE_TEXT that have to use fully
initialized global definition because of the way linkers work.
static gf_single_t * const gf_powx ATTRIBUTE_TEXT = (void *) 0x100000;
For the reason given above, use GCC diagnostic pragmas to suppress the
array-bounds warning.
Signed-off-by: Michael Chang <mchang@suse.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The GRUB is failing to build with GCC-12 in many places like this:
In function 'init_cbfsdisk',
inlined from 'grub_mod_init' at ../../grub-core/fs/cbfs.c:391:3:
../../grub-core/fs/cbfs.c:345:7: error: array subscript 0 is outside array bounds of 'grub_uint32_t[0]' {aka 'unsigned int[]'} [-Werror=array-bounds]
345 | ptr = *(grub_uint32_t *) 0xfffffffc;
| ~~~~^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
This is caused by GCC regression in 11/12 [1]. In a nut shell, the
warning is about detected invalid accesses at non-zero offsets to NULL
pointers. Since hardwired constant address is treated as NULL plus an
offset in the same underlying code, the warning is therefore triggered.
Instead of inserting #pragma all over the places where literal pointers
are accessed to avoid diagnosing array-bounds, we can try to borrow the
idea from Linux kernel that the absolute_pointer() macro [2][3] is used
to disconnect a pointer using literal address from it's original object,
hence GCC won't be able to make assumptions on the boundary while doing
pointer arithmetic. With that we can greatly reduce the code we have to
cover up by making initial literal pointer assignment to use the new
wrapper but not having to track everywhere literal pointers are
accessed. This also makes code looks cleaner.
Please note the grub_absolute_pointer() macro requires to be invoked in
a function as long as it is compound expression. Some global variables
with literal pointers has been changed to local ones in order to use
grub_absolute_pointer() to initialize it. The shuffling is basically done
in a selective and careful way that the variable's scope doesn't matter
being local or global, for example, the global variable must not get
modified at run time throughout. For the record, here's the list of
global variables got shuffled in this patch:
grub-core/commands/i386/pc/drivemap.c:int13slot
grub-core/term/i386/pc/console.c:bios_data_area
grub-core/term/ns8250.c:serial_hw_io_addr
[1] https://gcc.gnu.org/bugzilla/show_bug.cgi?id=99578
[2] https://elixir.bootlin.com/linux/v5.16.14/source/include/linux/compiler.h#L180
[3] https://elixir.bootlin.com/linux/v5.16.14/source/include/linux/compiler-gcc.h#L31
Signed-off-by: Michael Chang <mchang@suse.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The warning is real as long as dangling pointer to tmp_ may be used if
o32 and o64 are both NULL. However that is not going to happen and can
be ignored safely because the PE_OHDR is being used in a context that
either o32 or o64 must have been properly initialized. Sadly compiler
seems not to always optimize that unused tmp_ away so explicit
suppression remain needed here.
../util/mkimage.c: In function 'grub_install_generate_image':
../util/mkimage.c:1422:41: error: dangling pointer to 'tmp_' may be used [-Werror=dangling-pointer=]
1422 | PE_OHDR (o32, o64, header_size) = grub_host_to_target32 (header_size);
../util/mkimage.c:857:28: note: 'tmp_' declared here
857 | __typeof__((o64)->field) tmp_; \
| ^~~~
Signed-off-by: Michael Chang <mchang@suse.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
This patch handles automatic configuration of VLAN when booting from PXE
on UEFI hardware.
Signed-off-by: Chad Kimes <chkimes@github.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Previously there was no way to set the 802.1Q VLAN identifier, despite
support for vlantag in the net module. The only location vlantag was
being populated was from PXE boot and only for Open Firmware hardware.
This commit allows users to manually configure VLAN information for any
interface.
Example usage:
grub> net_ls_addr
efinet1 00:11:22:33:44:55 192.0.2.100
grub> net_set_vlan efinet1 100
grub> net_ls_addr
efinet1 00:11:22:33:44:55 192.0.2.100 vlan100
grub> net_set_vlan efinet1 0
efinet1 00:11:22:33:44:55 192.0.2.100
Signed-off-by: Chad Kimes <chkimes@github.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The initial implementation of the stack protector just busy looped
in __stack_chk_fail in order to reduce the amount of code being
executed after the stack has been compromised because of a lack of
firmware memory protections. With future firmware implementations
incorporating memory protections such as W^X, call in to boot services
when an error occurs in order to log a message to the console before
automatically rebooting the machine.
Signed-off-by: Chris Coulson <chris.coulson@canonical.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
In the function grub_xnu_boot(), struct grub_relocator32_state state is called
but isn't being initialized. This results in the members grub_uint32_t ebx,
grub_uint32_t ecx, grub_uint32_t edx, grub_uint32_t edi, and grub_uint32_t esi
being filled with junk data from the stack since none of them are being set to
any values. We can prevent this by setting state to {0}.
Fixes: CID 375035
Signed-off-by: Alec Brown <alec.r.brown@oracle.com>
Reviewed-by: Darren Kenny <darren.kenny@oracle.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
In the function grub_xnu_boot_resume(), struct grub_relocator32_state state is
called but isn't being initialized. This results in the members grub_uint32_t
ebx, grub_uint32_t ecx, grub_uint32_t edx, grub_uint32_t esi, and grub_uint32_t
edi being filled with junk data from the stack since none of them are being set
to any values. We can prevent this by setting state to {0}.
Fixes: CID 375031
Signed-off-by: Alec Brown <alec.r.brown@oracle.com>
Reviewed-by: Darren Kenny <darren.kenny@oracle.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
In the function grub_linux16_boot(), struct grub_relocator16_state state is
called but isn't being initialized. This results in the members grub_uint32_t
ebx, grub_uint32_t edx, grub_uint32_t esi, and grub_uint32_t ebp being filled
with junk data from the stack since none of them are being set to any values.
We can prevent this by setting state to {0}.
Fixes: CID 375028
Signed-off-by: Alec Brown <alec.r.brown@oracle.com>
Reviewed-by: Darren Kenny <darren.kenny@oracle.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
In the function grub_netbsd_setup_video(), struct grub_netbsd_btinfo_framebuf
params is called but isn't being initialized. The member grub_uint8_t
reserved[16] isn't set to any values and is instead filled with junk data from
the stack. We can prevent this by setting params to {0}.
Fixes: CID 375026
Signed-off-by: Alec Brown <alec.r.brown@oracle.com>
Reviewed-by: Darren Kenny <darren.kenny@oracle.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
In the function grub_net_ipv6_get_link_local(), grub_net_network_level_address_t
addr is called but isn't being initialized. This results in the member
grub_dns_option_t option being filled with junk data from the stack. We can
prevent this by setting the option member in addr to 0.
Fixes: CID 375033
Signed-off-by: Alec Brown <alec.r.brown@oracle.com>
Reviewed-by: Darren Kenny <darren.kenny@oracle.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
In the function grub_net_configure_by_dhcp_ack(),
grub_net_network_level_address_t addr is called but isn't being initialized.
This results in the member grub_dns_option_t option being filled with junk data
from the stack. To prevent this, we can set the option member in addr to 0.
Fixes: CID 375036
Signed-off-by: Alec Brown <alec.r.brown@oracle.com>
Reviewed-by: Darren Kenny <darren.kenny@oracle.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
In the function grub_net_arp_receive(), grub_net_network_level_address_t
sender_addr and target_addr are being called but aren't being initialized.
In both of these structs, each member is being set to a value except for
grub_dns_option_t option. This results in this member being filled with junk
data from the stack. To prevent this, we can set the option member in both
structs to 0.
Fixes: CID 375030
Signed-off-by: Alec Brown <alec.r.brown@oracle.com>
Reviewed-by: Darren Kenny <darren.kenny@oracle.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
If the machine has network cards found, but there are no tcp open sockets
(because the user doesn't use the network to boot), then grub_net_tcp_retransmit()
should be a noop. Thus GRUB doesn't need to call grub_get_time_ms(), which
does a call into firmware on powerpc-ieee1275, and probably other targets.
So only call grub_get_time_ms() if there are tcp sockets.
Aside from improving performance, its also useful to stay out of the firmware
as much as possible when debugging via QEMU because its a pain to get back
in to GRUB execution. grub_net_tcp_retransmit() can get called very frequently
via grub_net_poll_cards_idle() when GRUB is waiting for a keypress
(grub_getkey_noblock() calls grub_net_poll_cards_idle()). This can be annoying
when debugging an issue in GRUB on PowerPC in QEMU with GDB when GRUB is waiting
for a keypress because interrupting via GDB nearly always lands in the OpenBIOS
firmware's milliseconds call.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
In grub_net_poll_cards_idle_real(), only call grub_net_tcp_retransmit() if there
are network cards found. If there are no network card found, there can be no
tcp sockets to transmit on. So no need to go through that logic.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
This looks like it was a copy/paste error. If the net module is unloaded,
grub_net_poll_cards_idle should be NULL so that GRUB does not try to call
a function which now doesn't exist.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
These CFLAGS definitions are reset below them before they have a change to
affect anything. The exception is the *-emu case, which is put in the next
if block, which is the only place its used before getting reset.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
It appears as though the intent of this code is to define abort() and main()
symbols for some configure tests. However, it never gets used because the if
is only entered when not building for *-emu, but the next if block only runs
when building for *-emu. And the if block after that unconditionally resets
CFLAGS. So this code can have no effect.
Additionally, s/aclocal.m4/acinclude.m4/ and move grub_ASM_USCORE to put
with other marcos defined in acinclude.m4.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
According to the INSTALL, "The HOST_* variables override not prefixed
variables". This change makes it so, instead of previous behavior, which
was to ignore the HOST_CC environment variable.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Add linker flags when linking kernel.exec to have malloc and free point to
grub_malloc() and grub_free() respectively. Some gdb functionality depends on
gdb locating the symbols "malloc" and "free", such as dynamically creating
strings for arguments to injected function calls. A trivial example would
the gdb command 'p strlen("astring")'. Make sure not to do this on emu
platforms, or an infinite loop occurs because emu has a special
grub_malloc() that calls malloc().
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
When using "search" on EFI systems, we sometimes want to exclude devices
that are not EFI disks, e.g. md, lvm. This is typically used when
wanting to chainload when having a software raid (md) for EFI partition:
with no option, "search --file /EFI/redhat/shimx64.efi" sets root envvar
to "md/boot_efi" which cannot be used for chainloading since there is no
effective EFI device behind.
Signed-off-by: Renaud Métrich <rmetrich@redhat.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
GRUB EFI builds are now often used in combination with flicker-free
boot, but this breaks with upstream GRUB because the "Welcome to GRUB!"
message will kick the EFI fb into text mode and show the msg, breaking
the flicker-free experience.
EFI systems are so fast, that when the menu or the countdown are
enabled the message will be immediately overwritten, so in these cases
not printing the message does not matter.
And in case when the timeout_style is set to TIMEOUT_STYLE_HIDDEN,
the user has asked GRUB to be quiet (for example to allow flickfree
boot) and thus the message should not be printed.
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Reviewed-by: Robbie Harwood <rharwood@redhat.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
When the user has asked the menu code to be hidden/quiet and the current
entry is being autobooted because the timeout has expired don't show
the "Booting `%s'" msg.
This is necessary to let flicker-free boots really be flicker free,
otherwise the "Booting `%s'" msg will kick the EFI fb into text mode
and show the msg, breaking the flicker-free experience.
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Reviewed-by: Robbie Harwood <rharwood@redhat.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
To allow flickerfree boot the EFI console code does not call
grub_efi_set_text_mode(1) until some text is actually output. Depending
on if the output text is because of an error loading, e.g. the .cfg
file, or because of showing the menu the cursor needs to be on or off
when the first text is shown. So far the cursor was hardcoded to being
on, but this is causing drawing artifacts + slow drawing of the menu as
reported here: https://bugzilla.redhat.com/show_bug.cgi?id=1946969
Handle the cursorstate in the same way as the colorstate to fix this,
when no text has been output yet, just cache the cursorstate and then
use the last set value when the first text is output.
Fixes: 2d7c3abd87 (efi/console: Do not set text-mode until we actually need it)
Fixes: https://bugzilla.redhat.com/show_bug.cgi?id=1946969
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
GRUB_MOD_INIT(normal) does an unconditional:
grub_env_set ("color_normal", "light-gray/black");
which triggers a grub_term_setcolorstate() call. The original version
of the "efi/console: Do not set text-mode until we actually need it" patch,
https://lists.gnu.org/archive/html/grub-devel/2018-03/msg00125.html,
protected against this by caching the requested state in
grub_console_setcolorstate() and then only applying it when the first
text output actually happens. During refactoring to move the
grub_console_setcolorstate() up higher in the grub-core/term/efi/console.c
file the code to cache the color-state + bail early was accidentally dropped.
Restore the cache the color-state + bail early behavior from the original.
Fixes: 2d7c3abd87 (efi/console: Do not set text-mode until we actually need it)
Cc: Javier Martinez Canillas <javierm@redhat.com>
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
While it would appear unlikely that the memory allocated in *argv in
grub_parser_split_cmdline() would be leaked, we should try ensure that
it doesn't leak by calling grub_free() before we return from
grub_rescue_parse_line().
To avoid a possible double-free, grub_parser_split_cmdline() is being
changed to assign *argv = NULL when we've called grub_free() in the fail
section.
Fixes: CID 96680
Signed-off-by: Darren Kenny <darren.kenny@oracle.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Coverity flagged the switch checks for R_AARCH64_* as being logically
dead code, since it could never happen on x86 due to the masking of the
values earlier in the code.
A check for building on __arm__ (which gcc and clang define) and for
MKIMAGE_ELF64 (which GRUB defines) has been added to avoid this dead
code being built in.
Fixes: CID 158599
Signed-off-by: Darren Kenny <darren.kenny@oracle.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
$ ./configure --target=x86_64-w64-mingw32 --with-platform=efi --host=x86_64-w64-mingw32
$ make
[...]
cat syminfo.lst | sort | gawk -f ./genmoddep.awk > moddep.lst || (rm -f moddep.lst; exit 1)
__imp__errno in regexp is not defined
This happens because grub-core/lib/gnulib/malloc/dynarray_resize.c and
grub-core/lib/gnulib/malloc/dynarray_emplace_enlarge.c (both are used by
regexp module) from the latest Gnulib call __set_errno() which originally
sets errno variable (Windows builds add __imp__ prefix). Of course it is
not defined and grub_errno should be used instead.
Signed-off-by: Daniel Kiper <daniel.kiper@oracle.com>
Signed-off-by: Robbie Harwood <rharwood@redhat.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
- Fix type of size variable in luks2_verify_key()
- Avoid redefinition of SIZE_MAX and ATTRIBUTE_ERROR
- Work around gnulib's int types on older compilers
Signed-off-by: Robbie Harwood <rharwood@redhat.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
In addition to the changes carried in our gnulib patches, several
Coverity and code hygiene fixes that were previously downstream are also
included in this 3-year gnulib increment.
Unfortunately, fix-width.patch is retained.
Bump minimum autoconf version from 2.63 to 2.64 and automake from 1.11
to 1.14, as required by gnulib.
Sync bootstrap script itself with gnulib.
Update regexp module for new dynarray dependency.
Signed-off-by: Robbie Harwood <rharwood@redhat.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Originally added in commit db7337a3d (grub-core/lib/posix_wrap/stdlib.h
(abort): Removed), this patched out all relevant invocations of abort()
in gnulib. While it was not documented why at the time, testing suggests
that there's no abort() implementation available for gnulib to use.
gnulib's position is that the use of abort() is correct here, since it
happens when input violates a "shall" from POSIX. Additionally, the
code in question is probably not reachable. Since abort() is more
friendly to user-space, they prefer to make no change, so we can just
carry a define instead (suggested by Paul Eggert).
Signed-off-by: Robbie Harwood <rharwood@redhat.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Originally added in commit 9fbdec2f (bootstrap: Add gnulib's base64
module) and subsequently modified in commit 552c9fd08 (gnulib: Fix build
of base64 when compiling with memory debugging), fix-base64.patch
handled two problems we have using gnulib, which are exercised by the
base64 module but not directly caused by it.
First, GRUB defines its own bool type, while gnulib expects the
equivalent of stdbool.h to be present. Rather than patching gnulib,
instead use gnulib's stdbool module to provide a bool type if needed
(suggested by Simon Josefsson).
Second, our config.h doesn't always inherit config-util.h, which is
where gnulib-related options like _GL_ATTRIBUTE_CONST end up.
fix-base64.h worked around this by defining the attribute away, but this
workaround is better placed in config.h itself, not a gnulib patch.
Signed-off-by: Robbie Harwood <rharwood@redhat.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
gnulib defines go in config-util.h, and we need to know whether to
provide duplicates in config.h or not.
Signed-off-by: Robbie Harwood <rharwood@redhat.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The commit 9d25b0da9 (Remove emu libusb support.) dropped use of libusb,
but did not remove mention of it from INSTALL file.
Signed-off-by: Robbie Harwood <rharwood@redhat.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The mingw-w64-tools is especially important because with out it some
Windows builds may fail due to lack of proper pkg-config.
Signed-off-by: Daniel Kiper <daniel.kiper@oracle.com>
Reviewed-by: Robbie Harwood <rharwood@redhat.com>
These are probably stray references left after earlier removals.
Signed-off-by: Daniel Kiper <daniel.kiper@oracle.com>
Reviewed-by: Robbie Harwood <rharwood@redhat.com>
$ ./configure --target=i686-w64-mingw32 --with-platform=efi --host=i686-w64-mingw32
[...]
checking if __bss_start is defined by the compiler... no
checking if edata is defined by the compiler... no
checking if _edata is defined by the compiler... no
configure: error: none of __bss_start, edata or _edata is defined
This happens on machines with quite recent ld due to an error:
`edata' referenced in section `.text' of /tmp/cc72w9E4.o: defined in discarded section `.data' of conftest.exe
collect2: error: ld returned 1 exit status
So, we have to tell linker to not discard .data and .edata sections.
The trick comes from ld documentation:
3.6.7 Output Section Discarding
The linker will not normally create output sections with no contents.
This is for convenience when referring to input sections that may or may
not be present in any of the input files. For example:
.foo : { *(.foo) }
will only create a ‘.foo’ section in the output file if there is a
‘.foo’ section in at least one input file, and if the input sections are
not all empty. Other link script directives that allocate space in an
output section will also create the output section. So too will
assignments to dot even if the assignment does not create space, except
for ‘. = 0’, ‘. = . + 0’, ‘. = sym’, ‘. = . + sym’ and ‘. = ALIGN (. !=
0, expr, 1)’ when ‘sym’ is an absolute symbol of value 0 defined in the
script. This allows you to force output of an empty section with ‘. = .’.
This change does not impact generated binaries because the
conf/i386-cygwin-img-ld.sc linker script is used only when
you run configure.
Signed-off-by: Daniel Kiper <daniel.kiper@oracle.com>
Reviewed-by: Robbie Harwood <rharwood@redhat.com>
Latest GCC may complain in that way:
commands/i386/pc/sendkey.c: In function ‘grub_sendkey_postboot’:
commands/i386/pc/sendkey.c:223:21: error: writing 1 byte into a region of size 0 [-Werror=stringop-overflow=]
223 | *((char *) 0x41a) = 0x1e;
| ~~~~~~~~~~~~~~~~~~^~~~~~
The volatile keyword addition helps and additionally assures us the
compiler will not optimize out fixed assignments.
Signed-off-by: Daniel Kiper <daniel.kiper@oracle.com>
Reviewed-by: Robbie Harwood <rharwood@redhat.com>
Latest GCC may complain in that way:
In file included from ../include/grub/disk.h:31,
from ../include/grub/file.h:26,
from ../include/grub/loader.h:23,
from loader/i386/bsd.c:19:
loader/i386/bsd.c: In function ‘grub_cmd_openbsd’:
../include/grub/misc.h:71:10: error: ‘ptr’ may be used uninitialized in this function [-Werror=maybe-uninitialized]
71 | return grub_memmove (dest, src, n);
| ^~~~~~~~~~~~~~~~~~~~~~~~~~~
loader/i386/bsd.c:266:9: note: ‘ptr’ was declared here
266 | void *ptr;
| ^~~
So, let's fix it by assigning NULL to ptr in grub_bsd_add_meta().
Signed-off-by: Daniel Kiper <daniel.kiper@oracle.com>
Reviewed-by: Robbie Harwood <rharwood@redhat.com>
$ ./configure --target=x86_64-w64-mingw32 --with-platform=efi --host=x86_64-w64-mingw32
$ make
[...]
In file included from grub-core/osdep/platform.c:4:
grub-core/osdep/windows/platform.c: In function ‘grub_install_register_efi’:
grub-core/osdep/windows/platform.c:382:41: error: taking address of packed member of ‘struct grub_efi_file_path_device_path’ may result in an unaligned pointer value [-Werror=address-of-packed-member]
382 | path16_len = grub_utf8_to_utf16 (filep->path_name,
| ~~~~~^~~~~~~~~~~
Disable the -Wadress-of-packaed-member diagnostic for grub_utf8_to_utf16()
call which contains filep->path_name reference. It seems safe because the
structure is defined according to the UEFI spec and we hope authors did not
make any mistake... :-)
This fix is similar to the fix in the commit 8e8723a6b
(f2fs: Disable gcc9 -Waddress-of-packed-member).
Signed-off-by: Daniel Kiper <daniel.kiper@oracle.com>
Reviewed-by: Robbie Harwood <rharwood@redhat.com>
Commit 45bffae13 (util/resolve: Bail with error if moddep.lst file line is
too long) uses the %zu format specifier which has not been used in
any translated strings yet. So the sed scripts used for transliterating
certain languages need to be updated otherwise creation of the message
indexes will fail on an unknown format code. This is essentially the same
issue fixed for the %m format code in commit 2e246b6f (po: Fix replacement
of %m in sed programs).
Also reorder transliteration lines to be more lexicographically ordered.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
GRUB netbuff structure members track 2 different things: the extent of memory
allocated for the packet, and the extent of memory currently being worked on.
This works out in the structure as follows:
nb->head: beginning of the allocation
nb->data: beginning of the working data
nb->tail: end of the working data
nb->end: end of the allocation
The head and end pointers are set in grub_netbuff_alloc() and do not change.
The data and tail pointers are initialised to point at start of the
allocation (that is, head == data == tail initially), and are then
manipulated by grub_netbuff_*() functions. Key functions are as follows:
- grub_netbuff_put(): "put" more data into the packet - advance nb->tail
- grub_netbuff_unput(): trim the tail of the packet - retract nb->tail
- grub_netbuff_pull(): "consume" some packet data - advance nb->data
- grub_netbuff_reserve(): reserve space for future headers - advance nb->data and nb->tail
- grub_netbuff_push(): "un-consume" data to allow headers to be written - retract nb->data
Each of those functions does some form of error checking. For example,
grub_netbuff_put() does not allow nb->tail to exceed nb->end, and
grub_netbuff_push() does not allow nb->data to be before nb->head.
However, grub_netbuff_pull()'s error checking is a bit weird. It advances nb->data
and checks that it does not exceed nb->end. That allows you to get into the
situation where nb->data > nb->tail, which should not be.
Make grub_netbuff_pull() check against both nb->tail and nb->end. In theory just
checking against ->tail should be sufficient but the extra check should be
cheap and seems like good defensive practice.
Signed-off-by: Daniel Axtens <dja@axtens.net>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The libfuse 3.0.0 got released in 2016, with some API changes compared to 2.x.
This commit introduces support for 3.x while keeping it compatible with 2.6
as a fallback still.
To detect fuse3, switch configure over to use pkg-config, which is simpler yet
more reliable than looking for library and header manually. Also set
FUSE_USE_VERSION that way, as it depends on the used libfuse version.
Now that the CFLAGS are read from pkg-config, use just <fuse.h>, which works
with 2.x as well as 3.x and is recommended by libfuse upstream.
One behavior change of libfuse3 is that FUSE_ATOMIC_O_TRUNC is set by default,
which means that open with O_TRUNC is passed as-is instead of calling the
truncate operation. With libfuse2, truncate failed with -ENOSYS and that was
returned to the application. To make O_TRUNC fail with libfuse3, return -EROFS
explicitly if writing was requested.
Signed-off-by: Fabian Vogt <fvogt@suse.de>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
VLAN configuration seems to have never worked on little-endian systems.
This is likely because VLANTAG_IDENTIFIER is not byte-swapped before
copying into the net buffer, nor is inf->vlantag. We can resolve this by
using grub_cpu_to_be16{_compile_time}() and its inverse when copying
VLAN info to/from the net buffer.
Signed-off-by: Chad Kimes <chkimes@github.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The EFI_IMAGE_SECURITY_DATABASE_GUID is used for the image execution
information table (cf. UEFI specification 2.9, 32.5.3.1 Using The Image
Execution Information Table).
The lsefisystab command is used to display installed EFI configuration
tables. Currently it only shows the GUID but not a short text for the
table.
Provide a short text for the EFI_IMAGE_SECURITY_DATABASE_GUID.
Signed-off-by: Heinrich Schuchardt <heinrich.schuchardt@canonical.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
ChangeLog-2015 has been untouched for over 7 years now, and any
information in it is purely for historical purposes. At the same time,
grepping for code winds up matching this file quite a bit, almost never
accomplishing anything other than cluttering up your grep results. We
don't need this in the main repo, and "git show" will find it if you're
looking at the old history of commits on some file.
This patch deletes it and the Makefile.am rule to distribute it.
Signed-off-by: Peter Jones <pjones@redhat.com>
Reviewed-by: Daniel Axtens <dja@axtens.net>
Reviewed-by: Robbie Harwood <rharwood@redhat.com>
Reviewed-by: Javier Martinez Canillas <javierm@redhat.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
os-prober now effectively handles multiple paths passed to initrd, but
grub-mkconfig still truncates off any subsequent space-delimited paths.
Support proper parsing of space-delimited initrd paths passed from
os-prober for distributions, like Manjaro, that require it.
Fixes: https://savannah.gnu.org/bugs/?47681
Signed-off-by: Peter Levine <plevine457@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
This adds pci-arbiter and rumpdisk as bootstrap modules whenever they are
available. This opens the path for fully-userland disk support.
Signed-off-by: Samuel Thibault <samuel.thibault@ens-lyon.org>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
To prevent infinite recursion when grub_mm_debug is on, disable it when
calling grub_vprintf(). One such call loop is:
grub_vprintf() -> parse_printf_args() -> parse_printf_arg_fmt() ->
grub_debug_calloc() -> grub_printf() -> grub_vprintf().
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
These functions may be useful within modules as well. Export them so that
modules can use them.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Define MM_DEBUG in config.h when --enable-mm-debug is passed to configure.
It was being defined in config-util.h which only gets used when building
GRUB utilities for the host side. The enabling of debugging for memory
management in include/grub/mm.h explicitly does not happen when compiling
for the GRUB utilities. So this debugging code effectively could never be
enabled. Note, that MM_DEBUG is defined in an #if directive because the
enabling of debugging checks if MM_DEBUG is defined, not what its value is.
So even if MM_DEBUG were defined to nothing, the debugging code would
still be enabled.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
In GNU ld and ld.lld, -d is used with -r to allocate space to COMMON symbols.
This behavior is presumably to work around legacy projects which inspect
relocatable output by themselves and do not handle COMMON symbols. The GRUB
does not do this.
See https://github.com/llvm/llvm-project/issues/53660
-d is quite useless and ld.lld 15.0.0 will make -d no-op.
COMMON symbols have special symbol resolution semantics which can cause surprise
(see https://maskray.me/blog/2022-02-06-all-about-common-symbols). GCC<10 and
Clang<11 defaulted to -fcommon. Just use -fno-common to avoid COMMON symbols.
Signed-off-by: Fangrui Song <maskray@google.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
This allows for testing only tests that run directly on the build machine or
only tests that run in a virtualized environment. When testing multiple
targets on the same build machine the native tests only need to be run once
for all targets. Whereas, the nonnative tests must be run for each target
because the test is potentially compiled differently for each target.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
When using --no-floppy and a floppy was encountered, iterate_device()
was returning 1, causing the iteration to stop instead of continuing.
Signed-off-by: Renaud Métrich <rmetrich@redhat.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
This is causing the test grub_cmd_date() to fail because the returned
date is one day more than it should be.
This reverts commit 607d66116 (iee1275/datetime: Fix off-by-1 error.).
Signed-off-by: Glenn Washburn <development@efficientek.com>
Tested-by: Daniel Axtens <dja@axtens.net>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
This bashism allows converting NUM in base BASE to decimal. Its not needed
because the only place its used is to convert from hexidecimal and this can
also be done with the more portable $((0xHEXNUM)) syntax.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
In comparison to other i386 targets, on i386-efi the Q35 QEMU machine type
is used to do the testing to be able to make use of the EFI firmware in
QEMU. On the Q35 machine type there is no way to use ATA to communicate with
an IDE, only AHCI.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The image file can be useful in debugging an issue when the test fails.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
In grub-module-verifierXX.c, the function find_section() uses the value from
grub_target_to_host16 (e->e_shstrndx) to obtain the section header table index
of the section name string table, but it wasn't being checked if the value was
there.
According to the elf(5) manual page,
"If the index of section name string table section is larger than or equal
to SHN_LORESERVE (0xff00), this member holds SHN_XINDEX (0xffff) and the real
index of the section name string table section is held in the sh_link member of
the initial entry in section header table. Otherwise, the sh_link member of the
initial entry in section header table contains the value zero."
Since this check wasn't being made, the function get_shstrndx() is being added
to make this check and use e_shstrndx if it doesn't have SHN_XINDEX as a value,
else use sh_link. We also need to make sure e_shstrndx isn't greater than or
equal to SHN_LORESERVE and sh_link isn't less than SHN_LORESERVE.
Note that it may look as though the argument *arch isn't being used, it's
actually required in order to use the macros grub_target_to_host*(x) which are
unwinded to grub_target_to_host*_real(arch, (x)) based on defines earlier in
the file.
Signed-off-by: Alec Brown <alec.r.brown@oracle.com>
Reviewed-by: Darren Kenny <darren.kenny@oracle.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
In grub-module-verifierXX.c, grub_target_to_host16 (e->e_shnum) is used to
obtain the number of section header table entries, but it wasn't being
checked if the value was there.
According to the elf(5) manual page,
"If the number of entries in the section header table is larger than or equal
to SHN_LORESERVE (0xff00), e_shnum holds the value zero and the real number of
entries in the section header table is held in the sh_size member of the intial
entry in section header table. Otherwise, the sh_size member of the initial
entry in the section header table holds the value zero."
Since this check wasn't being made, the function get_shnum() is being added to
make this check and use whichever member doesn't have a value of zero. If both
are zero, then we must return an error. We also need to make sure that e_shnum
doesn't have a value greater than or equal to SHN_LORESERVE and sh_size isn't
less than SHN_LORESERVE.
Note that it may look as though the argument *arch isn't being used, it's
actually required in order to use the macros grub_target_to_host*(x) which are
unwinded to grub_target_to_host*_real(arch, (x)) based on defines earlier in
the file.
Fixes: CID 314021
Fixes: CID 314027
Fixes: CID 314033
Signed-off-by: Alec Brown <alec.r.brown@oracle.com>
Reviewed-by: Darren Kenny <darren.kenny@oracle.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Added the function get_shdr() which returns the section header at a given index
parameter passed into this function. This helps traverse the section header
table and reduces repeated calls to lengthy equations used to obtain section
headers.
Note that it may look as though the argument *arch isn't being used, it's
actually required in order to use the macros grub_target_to_host*(x) which are
unwinded to grub_target_to_host*_real(arch, (x)) based on defines earlier in the
file.
Signed-off-by: Alec Brown <alec.r.brown@oracle.com>
Reviewed-by: Darren Kenny <darren.kenny@oracle.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
In commit 178ac51073 (affs: Fix memory leaks), fixes were made to
grub_affs_iterate_dir() to prevent memory leaks from occurring after it
returns without freeing node. However, there were still some instances
where node was causing a memory leak when the function returns after
calling grub_affs_create_node(). In this function, new memory is
allocated to node but doesn't get freed until the hook() function is
called near the end. Before hook() is called, node should be freed in
grub_affs_create_node() before returning out of it.
Fixes: 178ac51073 (affs: Fix memory leaks)
Fixes: CID 73759
Signed-off-by: Alec Brown <alec.r.brown@oracle.com>
Reviewed-by: Darren Kenny <darren.kenny@oracle.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
As of version 2.38 binutils defaults to ISA specification version
2019-12-13. This version of the specification has has separated the
the csr read/write (csrr*/csrw*) instructions and the fence.i from
the I extension and put them into separate Zicsr and Zifencei
extensions.
This implies that we have to adjust the -march flag passed to the
compiler accordingly.
Signed-off-by: Heinrich Schuchardt <heinrich.schuchardt@canonical.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The UEFI specification defines that the EFI_BOOT_SERVICES.Exit(() service may return
EFI_SUCCESS or EFI_INVALID_PARAMETER. So it cannot be __attribute__((noreturn)).
Signed-off-by: Heinrich Schuchardt <heinrich.schuchardt@canonical.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Allow the use of HTTP servers listening on ports other 80. This is done
with an extension to the http notation:
(http[,server[,port]])
- or -
(http[,server[:port]])
Signed-off-by: Stephen Balousek <sbalousek@wickedloop.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Under certain conditions libgrub.pp gets generated with a such that it
contains a bunch of CPP defines, at least one of which contains "@MARKER@".
This line should not be used when generating libgrub_a_init.lst, otherwise
we get compiler errors like:
libgrub_a_init.c:22:18: error: stray ‘#’ in program
22 | extern void grub_#define_init (void);
| ^
libgrub_a_init.c:22:19: error: expected ‘=’, ‘,’, ‘;’, ‘asm’ or ‘__attribute__’ before ‘define_init’
22 | extern void grub_#define_init (void);
| ^~~~~~~~~~~
libgrub_a_init.c:23:18: error: stray ‘#’ in program
23 | extern void grub_#define_fini (void);
| ^
libgrub_a_init.c:23:19: error: expected ‘=’, ‘,’, ‘;’, ‘asm’ or ‘__attribute__’ before ‘define_fini’
23 | extern void grub_#define_fini (void);
| ^~~~~~~~~~~
...
When generating libgrub_a_init.lst only lines starting with "@MARKER@"
are desired.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
When generating video.lst, modules whose marker file contains the string
VIDEO_LIST_MARKER are selected. But when the marker file contains the CPP
defines, one of the defines is VIDEO_LIST_MARKER and is present in all
marker files, so they are all selected. By removing the defines, the correct
modules are selected.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The code reads each line into a buffer of size 1024 and does not check if
the line is longer. So a line longer than 1024 will be read as a valid line
followed by an invalid line. Then an error confusing to the user is sent
with the test "invalid line format". But the line format is perfectly fine,
the problem is in GRUB's parser. Check if we've hit a line longer than the
size of the buffer, and if so send a more correct and reasonable error.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
If the last non-NULL byte of "buf" is not a white-space character (such as
when a read line is longer than the size of "buf"), then "p" will eventually
point to the byte after the last byte in "buf". After which "p" will be
dereferenced in the while conditional leading to an out of bounds read. Make
sure that "p" is inside "buf" before dereferencing it.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Sometimes you only know which debug logging facility names you want to
turn off, not necessarily all the ones you want enabled. This patch allows
the debug string to contain facility names in the $debug variable which are
prefixed with a "-" to disable debug log messages for that conditional. Say
you want all debug logging on except for btrfs and scripting, then do:
"set debug=all,-btrfs,-scripting"
Note, that only the last occurrence of the facility name with or without a
leading "-" is considered. So simply appending ",-facilityname" to the
$debug variable will disable that conditional. To illustrate, the command
"set debug=all,-btrfs,-scripting,btrfs" will enable btrfs.
Also, add documentation explaining this new behavior.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The Coverity output is:
*** CID 366905: Memory - illegal accesses (USE_AFTER_FREE)
/grub-core/disk/cryptodisk.c: 1064 in grub_cryptodisk_scan_device_real()
1058 cleanup:
1059 if (askpass)
1060 {
1061 cargs->key_len = 0;
1062 grub_free (cargs->key_data);
1063 }
>>> CID 366905: Memory - illegal accesses (USE_AFTER_FREE)
>>> Using freed pointer "dev".
1064 return dev;
1065 }
1066
1067 #ifdef GRUB_UTIL
1068 #include <grub/util/misc.h>
1069 grub_err_t
Here the "dev" variable can point to a freed cryptodisk device if the
function grub_cryptodisk_insert() fails. This can happen only on a OOM
condition, but when this happens grub_cryptodisk_insert() calls grub_free on
the passed device. Since grub_cryptodisk_scan_device_real() assumes that
grub_cryptodisk_insert() is always successful, it will return the device,
though the device was freed.
Change grub_cryptodisk_insert() to not free the passed device on failure.
Then on grub_cryptodisk_insert() failure, free the device pointer. This is
done by going to the label "error", which will call cryptodisk_close() to
free the device and set the device pointer to NULL, so that a pointer to
freed memory is not returned.
Fixes: CID 366905
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The grub_mm_init_region() does some things that seem magical, especially
around region merging. Make it a bit clearer.
Signed-off-by: Daniel Axtens <dja@axtens.net>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The grub_free() possesses a surprising number of quirks, and also
uses single-letter variable names confusingly to iterate through
the free list.
Document what's going on.
Use prev and cur to iterate over the free list.
Signed-off-by: Daniel Axtens <dja@axtens.net>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Small allocations move the region's *first pointer. The comment
says that this happens for allocations under 64K. The code says
it's for allocations under 32K. Commit 45bf8b3a75 changed the
code intentionally: make the comment match.
Fixes: 45bf8b3a75 (* grub-core/kern/mm.c (grub_real_malloc): Decrease cut-off of moving the)
Signed-off-by: Daniel Axtens <dja@axtens.net>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
When iterating through the singly linked list of free blocks,
grub_real_malloc() uses p and q for the current and previous blocks
respectively. This isn't super clear, so swap to using prev and cur.
This makes another quirk more obvious. The comment at the top of
grub_real_malloc() might lead you to believe that the function will
allocate from *first if there is space in that block.
It actually doesn't do that, and it can't do that with the current
data structures. If we used up all of *first, we would need to change
the ->next of the previous block to point to *first->next, but we
can't do that because it's a singly linked list and we don't have
access to *first's previous block.
What grub_real_malloc() actually does is set *first to the initial
previous block, and *first->next is the block we try to allocate
from. That allows us to keep all the data structures consistent.
Document that.
Signed-off-by: Daniel Axtens <dja@axtens.net>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
I spent more than a trivial quantity of time figuring out pre_size and
whether a memory region's size contains the header cell or not.
Document the meanings of all the properties. Hopefully now no-one else
has to figure it out!
Signed-off-by: Daniel Axtens <dja@axtens.net>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Up to now GRUB can only embed to the first 64 KiB before primary
superblock of btrfs, effectively limiting the GRUB core size. That
could consequently pose restrictions to feature enablement like
advanced zstd compression.
This patch attempts to utilize full unused area reserved by btrfs for
the bootloader outlined in the document [1]:
The first 1MiB on each device is unused with the exception of primary
superblock that is on the offset 64KiB and spans 4KiB.
Apart from that, adjacent sectors to superblock and first block group
are not used for embedding in case of overflow and logged access to
adjacent sectors could be useful for tracing it up.
This patch has been tested to provide out of the box support for btrfs
zstd compression with which GRUB has been installed to the partition.
[1] https://btrfs.wiki.kernel.org/index.php/Manpage/btrfs(5)#BOOTLOADER_SUPPORT
Signed-off-by: Michael Chang <mchang@suse.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The iso9660 tests test creating isos with different combinations of
Joliet, Rock Ridge, and ISO 9660 conformance level. Refactor xorriso
argument generation for more readability and extensibility.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Thomas Schmitt <scdbackup@gmx.net>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Call grub_partition_get_name() unconditionally to initialize the part
variable. Then part will only be NULL when grub_partition_get_name() errors.
Note that when source->partition is NULL, then grub_partition_get_name()
returns an allocated empty string. So no comma or partition will be printed,
as desired.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Note that cargs.search_uuid does not need to be initialized in various parts
of the cryptomount argument parsing, just once when cargs is declared with
a struct initializer. The previous code used a global variable which would
retain the value across cryptomount invocations.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The crypto device modules should only be setting up the crypto devices and
not getting user input. This has the added benefit of simplifying the code
such that three essentially duplicate pieces of code are merged into one.
Add documentation of passphrase option for cryptomount as it is now usable.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Previously, the cryptomount arguments were passed by global variable and
function call argument, neither of which are ideal. This change passes data
via a grub_cryptomount_args struct, which can be added to over time as
opposed to continually adding arguments to the cryptodisk scan and
recover_key.
As an example, passing a password as a cryptomount argument is implemented.
However, the backends are not implemented, so testing this will return a not
implemented error.
Also, add comments to cryptomount argument parsing to make it more obvious
which argument states are being handled.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
When a cryptmount is specified with a UUID, but no cryptodisk backends find
a disk with that UUID, return a more detailed message giving telling the
user that they might not have a needed cryptobackend module loaded.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Update such that "cryptomount -u UUID" will not print two error messages
when an invalid passphrase is given and the most relevant error message
will be displayed.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
This displays an error notifying the user that they'll want to load
a backend module to make cryptomount useful.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The global "have_it" was never used by the crypto-backends, but was used to
determine if a crypto-backend successfully mounted a cryptodisk with a given
UUID. This is not needed however, because grub_device_iterate() will return
1 if and only if grub_cryptodisk_scan_device() returns 1. And
grub_cryptodisk_scan_device() will now only return 1 if a search_uuid has
been specified and a cryptodisk was successfully setup by a crypto-backend or
a cryptodisk of the requested UUID is already open.
To implement this grub_cryptodisk_scan_device_real() is modified to return
a cryptodisk or NULL on failure and having the appropriate grub_errno set to
indicated failure. Note that grub_cryptodisk_scan_device_real() will fail now
with a new errno GRUB_ERR_BAD_MODULE when none of the cryptodisk backend
modules succeed in identifying the source disk.
With this change grub_device_iterate() will return 1 when a crypto device is
successfully decrypted or when the source device has already been successfully
opened. Prior to this change, trying to mount an already successfully opened
device would trigger an error with the message "no such cryptodisk found",
which is at best misleading. The mount should silently succeed in this case,
which is what happens with this patch.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The commit ab2e53c8a (grub-mkconfig: Honor a symlink when generating
configuration by grub-mkconfig) has inadvertently discarded umask for
creating grub.cfg in the process of running grub-mkconfig. The resulting
wrong permission (0644) would allow unprivileged users to read GRUB
configuration file content. This presents a low confidentiality risk
as grub.cfg may contain non-secured plain-text passwords.
This patch restores the missing umask and sets the creation file mode
to 0600 preventing unprivileged access.
Fixes: CVE-2021-3981
Signed-off-by: Michael Chang <mchang@suse.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Create a library function for CloseProtocol() and use it for the SNP driver.
Signed-off-by: Heinrich Schuchardt <heinrich.schuchardt@canonical.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
In the context of the implementation of the EFI_LOAD_FILE2_PROTOCOL for
the initial ramdisk it was observed that opening the SNP protocol failed.
https://lists.gnu.org/archive/html/grub-devel/2021-10/msg00020.html
This is due to an incorrect call to CloseProtocol().
The first parameter of CloseProtocol() is the handle, not the interface.
We call OpenProtocol() with ControllerHandle == NULL. Hence we must also
call CloseProtcol() with ControllerHandel == NULL.
Each call of OpenProtocol() for the same network card handle is expected to
return the same interface pointer. If we want to close the protocol which
we opened non-exclusively when searching for a card, we have to do this
before opening the protocol exclusively.
As there is no guarantee that we successfully open the protocol add checks
in the transmit and receive functions.
Reported-by: Andreas Schwab <schwab@linux-m68k.org>
Signed-off-by: Heinrich Schuchardt <heinrich.schuchardt@canonical.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
minilzo fails to build on a number of Debian release architectures
(armel, mips64el, mipsel, ppc64el) with errors such as:
../../grub-core/lib/minilzo/minilzo.c: In function 'lzo_memops_get_le16':
../../grub-core/lib/minilzo/minilzo.c:3479:11: error: dereferencing type-punned pointer will break strict-aliasing rules [-Werror=strict-aliasing]
3479 | * (lzo_memops_TU2p) (lzo_memops_TU0p) (dd) = * (const lzo_memops_TU2p) (const lzo_memops_TU0p) (ss); \
| ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
../../grub-core/lib/minilzo/minilzo.c:3530:5: note: in expansion of macro 'LZO_MEMOPS_COPY2'
3530 | LZO_MEMOPS_COPY2(&v, ss);
| ^~~~~~~~~~~~~~~~
The latest upstream version is 2.10, so updating to it seems like a good
idea on general principles, and it fixes builds on all the above
architectures.
The update procedure documented in the GRUB Developers Manual worked; I
just updated the version numbers to make it clear that it's been
executed recently.
Signed-off-by: Colin Watson <cjwatson@debian.org>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Use the Git Book as a reference for documentation on Git as no other link
was provided. Other links were broken because they used @url instead of
@uref and needed a comma separator between link and link text.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Add a section with minimal description on setting up and running the test
suite with a link to the INSTALL documentation which is a little more
detailed in terms of package requirements.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The GRUB btrfs implementation can't handle two very basic btrfs
file layouts:
1. Mixed inline/regualr extents
# mkfs.btrfs -f test.img
# mount test.img /mnt/btrfs
# xfs_io -f -c "pwrite 0 1k" -c "sync" -c "falloc 0 4k" \
-c "pwrite 4k 4k" /mnt/btrfs/file
# umount /mnt/btrfs
# ./grub-fstest ./grub-fstest --debug=btrfs ~/test.img hex "/file"
Such mixed inline/regular extents case is not recommended layout,
but all existing tools and kernel can handle it without problem.
2. NO_HOLES feature
# mkfs.btrfs -f test.img -O no_holes
# mount test.img /mnt/btrfs
# xfs_io -f -c "pwrite 0 4k" -c "pwrite 8k 4k" /mnt/btrfs/file
# umount /mnt/btrfs
# ./grub-fstest ./grub-fstest --debug=btrfs ~/test.img hex "/file"
NO_HOLES feature is going to be the default mkfs feature in the incoming
v5.15 release, and kernel has support for it since v4.0.
The way GRUB btrfs code iterates through file extents relies on no gap
between extents.
If any gap is hit, then GRUB btrfs will error out, without any proper
reason to help debug the bug.
This is a bad assumption, since a long long time ago btrfs has a new
feature called NO_HOLES to allow btrfs to skip the padding hole extent
to reduce metadata usage.
The NO_HOLES feature is already stable since kernel v4.0 and is going to
be the default mkfs feature in the incoming v5.15 btrfs-progs release.
When there is a extent gap, instead of error out, just try next item.
This is still not ideal, as kernel/progs/U-boot all do the iteration
item by item, not relying on the file offset continuity.
But it will be way more time consuming to correct the whole behavior than
starting from scratch to build a proper designed btrfs module for GRUB.
Signed-off-by: Qu Wenruo <wqu@suse.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Commit 23e39f50ca (disk/ldm: Make sure comp data is freed before exiting from
make_vg()) fixed several spots in make_vg() where comp data was leaking memory
when an error was being handled but missed one. To avoid leaking memory, comp
should be freed when an error is being handled after comp has been successfully
allocated memory in the for loop.
Fixes: 23e39f50ca (disk/ldm: Make sure comp data is freed before exiting from make_vg())
Fixes: CID 73804
Signed-off-by: Alec Brown <alec.r.brown@oracle.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Commit 1fc860bb76 (commands/probe: Fix a resource leak when probing disks),
missed other cases where grub_device_close() should be called before a return
statement is called. Also found that grub_disk_close() wasn't being called when
an error is being returned. To avoid conflict with grub_errno, grub_error_push()
should be called before either grub_device_close() or grub_disk_close() is
called and grub_error_pop() should be called before grub_errno is returned.
Fixes: 1fc860bb76 (commands/probe: Fix a resource leak when probing disks)
Fixes: CID 292443
Signed-off-by: Alec Brown <alec.r.brown@oracle.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The POSIX locale is default or native operating system's locale
identical to the C locale, so no translation to human speaking languages
are provided. For this reason we should filter out LANG=POSIX as well as
LANG=C upon generating grub.cfg to avoid looking up for it's gettext's
message catalogs that will consequently result in an unpleasant message:
error: file `/boot/grub/locale/POSIX.gmo' not found
Signed-off-by: Michael Chang <mchang@suse.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
In huft_build() it is possible to reach the for loop where "r" is being
assigned to "q[j]" without "r.v" ever being initialized.
Fixes: CID 314024
Signed-off-by: Darren Kenny <darren.kenny@oracle.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
In zap_leaf_array_get() the chunk size passed in is considered tainted
by Coverity, and is being used before it is tested for validity. To fix
this the assignment of "la" is moved until after the test of the value
of "chunk".
Fixes: CID 314014
Signed-off-by: Darren Kenny <darren.kenny@oracle.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
In the function write_font_pf2() memory is allocated for font_name to
construct a new name, but it is not released before returning from the
function, leaking the allocated memory.
Fixes: CID 314015
Signed-off-by: Darren Kenny <darren.kenny@oracle.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
In the function cmd_cmp() within the while loop, srcnew and destnew are
being allocated but are never freed either before leaving scope or in
the recursive calls being made to cmd_cmp().
Fixes: CID 314032
Fixes: CID 314045
Signed-off-by: Darren Kenny <darren.kenny@oracle.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
In the function write_part(), the value of inname is not used beyond
the grub_util_fopen() call, so it should be freed to avoid leakage.
Fixes: CID 314028
Signed-off-by: Darren Kenny <darren.kenny@oracle.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The copy_all() function skips a section of code using continue, but
fails to free the memory in srcf first, leaking it.
Fixes: CID 314026
Signed-off-by: Darren Kenny <darren.kenny@oracle.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Prior to this change, the GRUB would only indicate that the check had
been failed, but not by what module. This made it difficult to track
down either the problem module, or debug the false positive further.
Before performing the license check, resolve the module name so that
it can be printed if the license check fails.
Signed-off-by: Robbie Harwood <rharwood@redhat.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Adding the conditional to debug log messages allows the GRUB user to
construct the $debug variable without needing to consult the source to
find the conditional (especially useful for situations where the source
is not readily available).
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Many tests abort due to not being root or missing tools, for instance mkfs
commands for file system tests. The tests are exited with code 77, which
means they were skipped. A skipped test is a test that should not be run,
e.g. a test specific to ARM64 should not be run on an x86 build. These aborts
are actually a hard error, code 99. That means that the test could not be
completed, but not because what was supposed to be tested failed, e.g. in
these cases where a missing tool prevents the running of a test.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
A recent refactoring of CUDA command code has exposed a bug in OpenBIOS [1]
which was causing system powerdown and system reset to fail, thus causing
the QEMU instance to hang. This in turn caused the grub-shell command to
timeout causing it to return an error code when the test actually completed
successfully.
Since it could be a while before the patch fixing this issue in OpenBIOS
filters down to the average distro, switch to PMU to allow powerdowns and
reboots to work as expected.
[1] https://gitlab.com/qemu-project/qemu/-/issues/624
Signed-off-by: Glenn Washburn <development@efficientek.com>
Signed-off-by: Daniel Axtens <dja@axtens.net>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
GET_ARRAY_INFO's info.nr_disks does not map to GET_DISK_INFO's
disk.number, which is an internal kernel index. If an array has had drives
added, removed, etc., there may be gaps in GET_DISK_INFO's results. But
since the consumer of devicelist cannot tolerate gaps (it expects to walk
a NULL-terminated list of device name strings), the devicelist index (j)
must be tracked separately from the disk.number index (i).
As part of this, since GRUB wants to only examine active (i.e. present
and non-failed) disks, the count of remaining disks (remaining) must be
tracked separately from the devicelist index (j).
Additionally, drop a line with empty spaces only.
Fixes: 49de079bbe (... (grub_util_raid_getmembers): Handle "removed" disks)
Fixes: 2b00217369 (... Added support for RAID and LVM)
Fixes: 1912043
Fixes: https://savannah.gnu.org/bugs/index.php?59887
Signed-off-by: Kees Cook <kees@ubuntu.com>
Reviewed-by: Petr Vorel <pvorel@suse.cz>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Many of the prerequisites for exercising the full "make check" test suite
have not been documented. This adds them along with a note that some tests
require elevated privileges to run.
Add an incomplete list of cross compiling toolchain packages for Debian
and trusted sources for other distros.
Add statement at the start of the document to clarify that package names
are from Debian 11.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The filesystem images created for the filesystem test can be useful when
debugging why a filesystem test failed. So, keep them around and let the
user clean them up.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Allow the HFS tests to not be skipped if the mac_roman modules is loaded in
the kernel, but not accessible to modprobe.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The ";", semi-colon, character is not a valid character for a FAT filesystem
label. This test used to succeed because prior to v4.2 of dosfstools
mkfs.vfat did not enforce the character restrictions for volume labels. So,
change the volume label string to be valid but contain symbol characters to
test odd volume labels.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Apparently there used to be a -B option for mkfs.minix to create a volume
with a specified block size. This version is hard to come by and does not
appear to be available in Debian distributions. So, remove support for
testing a variety of blocks sizes for MINIX3. This allows the MINIX tests
to run because they were being skipped due to not finding a mkfs.minix with
the -B option.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
This bring this test in line with the rest of the test scripts.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
These tests were not performed and therefore did not pass, nor fail. This
fixes misleading test exit code where, for instance, the pseries_test will
pass on i386-pc, which is not a pseries architecture.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
A test exiting with code 99 means that there was an error in the test itself
and not a failure in the thing being tested (also known as a hard error).
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
When a test program fails because it failed to setup the test properly, this
does not indicate a failure in what is attempting to be tested because the
test is never run. So exit with a hard error exit status to note this
difference. This will allow easier detection of tests that are not actually
being run and those that are really failing.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The script grub-shell does the bulk of the testing. If it returns an error
code, that means that the test failed and the test should immediately exit
with that error code. When grub-shell is used as a non-terminating command
in a pipeline, e.g. when data needs to be extracted from its output, its
error code will be occluded by the last command in the pipeline. Refactor
tests so that the shell will error with the exit code of grub-shell by
breaking up pipelines such that grub-shell is always the last command in
the pipeline that it is used in.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
When using the output of a subshell as input, its error code is ignored in
the context of "set -e". Many test scripts use grub-shell in a subshell with
output used as an argument to the test command to test for expected output.
Refactor these tests so that the subshell output goes to a shell variable,
so that if the subshell errors the script will immediately exit with an
error code.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
This helps to ensure that error codes do not get ignored.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Currently, the filesystem timestamp check in grub-fs-tester uses the
squashfs image file's last modified timestamp and checks to see if that
time stamp is within 3 seconds of the superblock timestamp as determined by
grub. The image file's timestamp could be more than 3 seconds off if
mksquashfs takes more than 3 seconds to generate the image, as is the case
on a virtual machine. Instead use squashfs tools to get the filesystem
timestamp directly.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Perhaps using a newer UEFI firmware is the reason for the created test disk
showing up as hd2 instead of hd3.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The following procedure to build xen/pvgrub is broken.
git clone https://git.savannah.gnu.org/git/grub.git
cd grub
./bootstrap
mkdir build-xen
cd build-xen
../configure --with-platform=xen
make
It fails with the message:
/usr/lib64/gcc/x86_64-suse-linux/10/../../../../x86_64-suse-linux/bin/ld:
section .note.gnu.property VMA [0000000000400158,0000000000400187]
overlaps section .bss VMA [000000000000f000,000000000041e1af]
The most significant factor is that new assembler (GNU as) generates the
.note.gnu.property section as default. This note section overlaps with
.bss because it doesn't reposition with -Wl,-Ttext,0 with which the base
address of .text section is set, rather the address of .note.gnu.property
is calculated for some reason from 0x400000 where the ELF executable
defaults to start.
Using -Ttext-segment doesn't help either, though it is said to set the
address of the first byte of the text segment according to "man ld".
What it actually does is to override the default 0x400000, aka the image
base address, to something else. The entire process can be observed in
the default linker script used by gcc [1]. Therefore we can't expect it
to achieve the same thing as -Ttext given that the first segment where
.text resides is offset by SIZEOF_HEADERS plus some sections may be
preceding it within the first segment. The end result is .text always
has to start with non-zero address with -Wl,-Ttext-segment,0 if using
default linker script.
It is also worth mentioning that binutils upstream apparently doesn't
seem to consider this as a bug [2] and proposed to use -Wl,-Ttext-segment,0
which is not fruitful as what has been tested by Gentoo [3].
As long as GRUB didn't use ISA information encoded in .note.gnu.property,
we can safely drop it via -Wa,-mx86-used-note=no assembler option to
fix the linker error above.
This is considered a better approach than using custom linker script to
drop the .note.gnu.property section because object file manipulation can
also be hampered one way or the other in that linker script may not be
helpful. See also this commit removing the section in the process of objcopy.
6643507ce build: Fix GRUB i386-pc build with Ubuntu gcc
[1] In /usr/lib64/ldscripts/elf_x86_64.x or use 'gcc -Wl,--verbose ...'
PROVIDE (__executable_start = SEGMENT_START("text-segment", 0x400000));
. = SEGMENT_START("text-segment", 0x400000) + SIZEOF_HEADERS;
[2] https://sourceware.org/bugzilla/show_bug.cgi?id=27377
[3] https://bugs.gentoo.org/787221
Signed-off-by: Michael Chang <mchang@suse.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Currently the grub_diskfilter_memberlist() function returns all physical
volumes added to a volume group to which a logical volume (LV) belongs.
However, this is suboptimal as it doesn't fit the intended behavior of
returning underlying devices that make up the LV. To give a clear
picture, the result should be identical to running commands below to
display the logical volumes with underlying physical volumes in use.
localhost:~ # lvs -o lv_name,vg_name,devices /dev/system/root
LV VG Devices
root system /dev/vda2(512)
localhost:~ # lvdisplay --maps /dev/system/root
--- Logical volume ---
...
--- Segments ---
Logical extents 0 to 4604:
Type linear
Physical volume /dev/vda2
Physical extents 512 to 5116
As shown above, we can know system-root LV uses only /dev/vda2 to
allocate it's extents, or we can say that /dev/vda2 is the member device
comprising the system-root LV.
It is important to be precise on the member devices, because that helps
to avoid pulling in excessive dependency. Let's use an example to
demonstrate why it is needed.
localhost:~ # findmnt /
TARGET SOURCE FSTYPE OPTIONS
/ /dev/mapper/system-root ext4 rw,relatime
localhost:~ # pvs
PV VG Fmt Attr PSize PFree
/dev/mapper/data system lvm2 a-- 1020.00m 0
/dev/vda2 system lvm2 a-- 19.99g 0
localhost:~ # cryptsetup status /dev/mapper/data
/dev/mapper/data is active and is in use.
type: LUKS1
cipher: aes-xts-plain64
keysize: 512 bits
key location: dm-crypt
device: /dev/vdb
sector size: 512
offset: 4096 sectors
size: 2093056 sectors
mode: read/write
localhost:~ # vgs
VG #PV #LV #SN Attr VSize VFree
system 2 3 0 wz--n- 20.98g 0
localhost:~ # lvs -o lv_name,vg_name,devices
LV VG Devices
data system /dev/mapper/data(0)
root system /dev/vda2(512)
swap system /dev/vda2(0)
We can learn from above that /dev/mapper/data is an encrypted volume and
also gets assigned to volume group "system" as one of it's physical
volumes. And also it is not used by root device, /dev/mapper/system-root,
for allocating extents, so it shouldn't be taking part in the process of
setting up GRUB to access root device.
However, running grub-install reports error as volume group "system"
contains encrypted volume.
error: attempt to install to encrypted disk without cryptodisk
enabled. Set `GRUB_ENABLE_CRYPTODISK=y' in file `/etc/default/grub'.
Certainly we can enable GRUB_ENABLE_CRYPTODISK=y and move on, but that
is not always acceptable since the server may need to be booted unattended.
Additionally, typing passphrase for every system startup can be a big
hassle of which most users would like to avoid.
This patch solves the problem by returning exact physical volume, /dev/vda2,
rightly used by system-root from the example above, thus grub-install will
not error out because the excessive encrypted device to boot the root device
is not configured.
Signed-off-by: Michael Chang <mchang@suse.com>
Tested-by: Olav Reinert <seroton10@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
When a file on ext4 is stored as sparse the data belonging to
zero-filled blocks is not written to storage and the extent map is
missing entries for these blocks. Such case can happen both for depth
0 extents (leafs) as well as higher-level tables.
Consider a scenario of a file which has a zero-filled beginning (e.g.
ISO image). In such case real data starts at block 8. If such a file is
stored using 2-level extent structure the extent list in the inode will
be depth 1 and will have an entry to a depth 0 (leaf) extent header for
blocks 8-n.
Unfortunately existing GRUB2 ext2 driver is only able to handle missing
entries in leaf extent tables, for which the grub_ext2_read_block()
function returns 0. In case the whole leaf extent list is missing for
a block the function fails with "invalid extent" error.
The fix for this problem relies on the grub_ext4_find_leaf() helper
function to distinguish two error cases: missing extent and error
walking through the extent tree. The existing error message is raised
only for the latter case, while for the missing leaf extent zero is
returned from grub_ext2_read_block() indicating a sparse block.
Signed-off-by: Krzysztof Nowicki <krzysztof.nowicki@nokia.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Open Hack'Ware was the only user. It added a lot of complexity.
Signed-off-by: Daniel Axtens <dja@axtens.net>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Open Hack'Ware was an alternative firmware of powerpc under QEMU.
The last commit to any Open Hack'Ware repo I can find is from 2014 [1].
Open Hack'Ware was used for the QEMU "prep" machine type, which was
deprecated in QEMU in commit 54c86f5a4844 (hw/ppc: deprecate the
machine type 'prep', replaced by '40p') in QEMU v3.1, and had reportedly
been broken for years before without anyone noticing. Support was removed
in February 2020 by commit b2ce76a0730e (hw/ppc/prep: Remove the
deprecated "prep" machine and the OpenHackware BIOS).
Open Hack'Ware's limitations require some messy code in GRUB. This
complexity is not worth carrying any more.
Remove detection of Open Hack'Ware. We will clean up the feature flags
in following commits.
[1]: https://github.com/qemu/openhackware and
https://repo.or.cz/w/openhackware.git are QEMU submodules. They have
only small changes on top of OHW v0.4.1, which was imported into
QEMU SCM in 2010. I can't find anything resembling an official repo
any more.
Signed-off-by: Daniel Axtens <dja@axtens.net>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
This code was broken by commit 3f05d693 (malloc: Use overflow checking
primitives where we do complex allocations), which added overflow
checking in many areas. The problem here is that the changes update the
local variable sz, which was already in use and which was not updated
before the change. So the code using sz was getting a different value of
than it would have previously for the same UDF image. This causes the
logic getting the destination of the symlink to not realize that its
gotten the full destination, but keeps trying to read past the end of
the destination. The bytes after the end are generally NULL padding
bytes, but that's not a valid component type (ECMA-167 14.16.1.1). So
grub_udf_read_symlink() branches to error logic, returning NULL, instead
of the symlink destination path.
The result of this bug is that the UDF filesystem tests were failing in
the symlink test with the grub-fstest error message:
grub-fstest: error: cannot open `(loop0)/sym': invalid symlink.
This change stores the result of doubling sz in another local variable s,
so as not to modify sz. Also remove unnecessary grub_add(), which increased
the output by 1, presumably to account for a NULL byte. This isn't needed
because an output buffer of size twice sz is already guaranteed to be more
than enough to contain the path components converted to UTF-8. The value of
sz contains at least 4 bytes for the path component header (ECMA-167 14.16.1),
which means that 2 * 4 bytes are allocated but will not be used for UTF-8
characters, so the NULL byte is accounted for.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
When generating grub.cfg using grub-mkconfig and the scripts 10_linux and
20_linux_xen there is no way to add kernel command line parameters _only_ to
the recovery entries generated.
This is needed to e.g. start a debug shell in installations using systemd
using the kernel command line parameter "systemd.debug-shell" or to recover
in a system with encrypted root in situations where the decryption of the
root filesystem per crypttab in the intiramfs image is broken and the recovery
entry should contain information how to decrypt the rootfs (cryptopts=).
This patch does not change the default behaviour of the GRUB if
GRUB_CMDLINE_LINUX_RECOVERY is not set.
If GRUB_CMDLINE_LINUX_RECOVERY is set and the generated recovery entry should
include the kernel parameter "single" the parameter must be explicitly included
in GRUB_CMDLINE_LINUX_RECOVERY.
As far as I know all credits for the idea and the initial implementation go to
Kyle Ranking of Purism.
Signed-off-by: Kyle Rankin <kyle.rankin@puri.sm>
Signed-off-by: Chris Vogel <chris@z9.de>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The gcc by default assumes executable stack is required if the source
object file doesn't have .note.GNU-stack section in place. If any of the
source objects doesn't incorporate the GNU-stack note, the resulting
program will have executable stack flag set in PT_GNU_STACK program
header to instruct program loader or kernel to set up the executable
stack when program loads to memory.
Usually the .note.GNU-stack section will be generated by gcc
automatically if it finds that executable stack is not required. However
it doesn't take care of generating .note.GNU-stack section for those
object files built from assembler sources. This leads to unnecessary
risk of security of exploiting the executable stack because those
assembler sources don't actually require stack to be executable to work.
The grub-emu and grub-emu-lite are found to flag stack as executable
revealed by execstack tool.
$ mkdir -p build-emu && cd build-emu
$ ../configure --with-platform=emu && make
$ execstack -q grub-core/grub-emu grub-core/grub-emu-lite
X grub-core/grub-emu
X grub-core/grub-emu-lite
This patch will add the missing GNU-stack note to the assembler source
used by both utilities, therefore the result doesn't count on gcc
default behavior and the executable stack is disabled.
$ execstack -q grub-core/grub-emu grub-core/grub-emu-lite
- grub-core/grub-emu
- grub-core/grub-emu-lite
Signed-off-by: Michael Chang <mchang@suse.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The test for the ability to decompress zisofs encoded files is supposed
to fail due to the lack of this ability in GRUB. But it fails early with
xorriso : FAILURE : -volid: Text too long (1650 > 32)
because "ziso9660" is not in the list of filesystems which accept at most
32 bytes in their FSLABEL. If this is fixed, the test returns false
success because the xorriso run does not produce any zisofs compressed
files. The problem is in the sequence of native xorriso commands used.
The command -set_filter_r applies only to the files which are already
inserted into the emerging ISO filesystem. In the current sequence no
files have been inserted yet by command -add when the last of two
-set_filter_r commands is executed. After this is corrected, xorriso
refuses to work because the global settings of command -zisofs can be
made only before command -set_filter_r has attached zisofs filters to
the data files in the emerging ISO. Further: A bug in xorriso causes
a false warning about FSLABEL being too long for Joliet. Shortcomings
of Joliet cause warnings about symbolic links. Such warnings might
distract from the actual reason why the test is expected to fail.
So, add "ziso9660" to the 32-byte FSLABEL list.
Fix the xorriso run to produce compressed files which for now cause
righteous failure of the test. Do this by removing a surplus group of
-set_filter_r and -zisofs commands, by moving the other such group
behind -add, and by swapping -set_filter_r and -zisofs.
Remove the -as mkisofs options which produce a Joliet filesystem tree.
Signed-off-by: Thomas Schmitt <scdbackup@gmx.net>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
This conforms to the behavior of the -s option of the Bash read command.
docs/grub: Document the -s option for the read command.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
This is primarily useful to do something like "loopback newdev (dev)8+" to
create a device that skips the first 4 KiB, which may contain a container
header, e.g. a non-standard RAID1 header, that GRUB does not recognize. This
would allow that container data to be potentially accessed up to the end of
container, which may be necessary for some layouts that store data at the
end. There is currently not a good way to programmatically get the number
of sectors on a disk to set the appropriate length of the blocklist.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
It helps to avoid an error on distros which has only python3 binary:
./autogen.sh: line 20: python: command not found
Use python3 as the default as python2 is EOL since Jan 2020. However,
check also for python which is on most distros, if not all, python2
because code still works on python2.
Although it should not be needed keep the possibility to define PYTHON
variable.
For detection use "command -v" which is POSIX and supported on all
common shells (bash, zsh, dash, busybox sh, mksh) instead requiring
"which" as an extra dependency (usable on containers).
Update the INSTALL file too.
Signed-off-by: Petr Vorel <pvorel@suse.cz>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The bootstrap.conf uses patch, let's require it.
Better than multiple messages:
./bootstrap.conf: line 84: patch: command not found
Mention it also in the INSTALL file.
Signed-off-by: Petr Vorel <pvorel@suse.cz>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The iso9660_test fails if the effective locale is not UTF-8. This happens
because xorriso needs to convert file names and FSLABEL to UCS-2 when
preparing a Joliet tree. The grub-fs-tester obviously intends to use UTF-8
as character set, but xorriso assumes by default the result of nl_langinfo(3)
with item CODESET. So, override the result of nl_langinfo(CODESET) by options
of xorriso -as mkisofs.
Signed-off-by: Thomas Schmitt <scdbackup@gmx.net>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The Clang does not support -falign-jumps and only recently gained support
for -falign-loops. The -falign-jumps=1 should be tested beside
-fliang-loops=1 to avoid passing unrecognized options to the Clang:
clang-14: error: optimization flag '-falign-jumps=1' is not supported [-Werror,-Wignored-optimization-argument]
The -falign-functions=1 is supported by GCC 5.1.0/Clang 3.8.0. So, just
add the option unconditionally.
Signed-off-by: Fangrui Song <maskray@google.com>
Acked-by: Paul Menzel <pmenzel@molgen.mpg.de>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The GCC warns "cc1: warning: ‘-malign-loops’ is obsolete, use ‘-falign-loops’".
The Clang silently ignores -malign-{jumps,loops,functions}.
The preferred -falign-* forms have been supported since GCC 3.2. So, just
remove -malign-{jumps,loops,functions}.
Signed-off-by: Fangrui Song <maskray@google.com>
Acked-by: Paul Menzel <pmenzel@molgen.mpg.de>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The commit 8b1e5d193 (fs/xfs: Add bigtime incompat feature support)
introduced the bigtime support by adding some features in v3 inodes.
This change extended grub_xfs_inode struct by 76 bytes but also changed
the computation of XFS_V2_INODE_SIZE and XFS_V3_INODE_SIZE. Prior this
commit, XFS_V2_INODE_SIZE was 100 bytes. After the commit it's 84 bytes
XFS_V2_INODE_SIZE becomes 16 bytes too small.
As a result, the data structures aren't properly aligned and the GRUB
generates "attempt to read or write outside of partition" errors when
trying to read the XFS filesystem:
GNU GRUB version 2.11
....
grub> set debug=efi,gpt,xfs
grub> insmod part_gpt
grub> ls (hd0,gpt1)/
partmap/gpt.c:93: Read a valid GPT header
partmap/gpt.c:115: GPT entry 0: start=4096, length=1953125
fs/xfs.c:931: Reading sb
fs/xfs.c:270: Validating superblock
fs/xfs.c:295: XFS v4 superblock detected
fs/xfs.c:962: Reading root ino 128
fs/xfs.c:515: Reading inode (128) - 64, 0
fs/xfs.c:515: Reading inode (739521961424144223) - 344365866970255880, 3840
error: attempt to read or write outside of partition.
This commit change the XFS_V2_INODE_SIZE computation by subtracting 76
bytes instead of 92 bytes from the actual size of grub_xfs_inode struct.
This 76 bytes value comes from added members:
20 grub_uint8_t unused5
1 grub_uint64_t flags2
48 grub_uint8_t unused6
This patch explicitly splits the v2 and v3 parts of the structure.
The unused4 is still ending of the v2 structures and the v3 starts
at unused5. Thanks to this we will avoid future corruptions of v2
or v3 inodes.
The XFS_V2_INODE_SIZE is returning to its expected size and the
filesystem is back to a readable state:
GNU GRUB version 2.11
....
grub> set debug=efi,gpt,xfs
grub> insmod part_gpt
grub> ls (hd0,gpt1)/
partmap/gpt.c:93: Read a valid GPT header
partmap/gpt.c:115: GPT entry 0: start=4096, length=1953125
fs/xfs.c:931: Reading sb
fs/xfs.c:270: Validating superblock
fs/xfs.c:295: XFS v4 superblock detected
fs/xfs.c:962: Reading root ino 128
fs/xfs.c:515: Reading inode (128) - 64, 0
fs/xfs.c:515: Reading inode (128) - 64, 0
fs/xfs.c:931: Reading sb
fs/xfs.c:270: Validating superblock
fs/xfs.c:295: XFS v4 superblock detected
fs/xfs.c:962: Reading root ino 128
fs/xfs.c:515: Reading inode (128) - 64, 0
fs/xfs.c:515: Reading inode (128) - 64, 0
fs/xfs.c:515: Reading inode (128) - 64, 0
fs/xfs.c:515: Reading inode (131) - 64, 768
efi/ fs/xfs.c:515: Reading inode (3145856) - 1464904, 0
grub2/ fs/xfs.c:515: Reading inode (132) - 64, 1024
grub/ fs/xfs.c:515: Reading inode (139) - 64, 2816
grub>
Fixes: 8b1e5d193 (fs/xfs: Add bigtime incompat feature support)
Signed-off-by: Erwan Velu <e.velu@criteo.com>
Tested-by: Carlos Maiolino <cmaiolino@redhat.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Avoid a warning
lib/libgcrypt-grub/cipher/rijndael.c:229:9:
warning: suggest braces around empty body in an ‘if’ statement [-Wempty-body]
229 | ;
| ^
Signed-off-by: Heinrich Schuchardt <heinrich.schuchardt@canonical.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Avoid a warning
lib/libgcrypt-grub/cipher/rijndael.c:352:21: warning:
comparison of integer expressions of different signedness:
‘int’ and ‘unsigned int’ [-Wsign-compare]
352 | for (i = 0; i < keylen; i++)
|
Signed-off-by: Heinrich Schuchardt <heinrich.schuchardt@canonical.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
In the case that one passes a write mask with ":" the write_mask is
obtained from grub_strtoul() and then promptly overwritten by 0xffffffff
three lines later.
This appears to have been so since the initial version of setpci in 2009.
I'm surprised no one else has hit this issue in the past 12 years...
Signed-off-by: Wouter van Kesteren <woutershep@gmail.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The sysfs_partition_path() calls udevadm to resolve the sysfs path for
a block device. That can be accomplished by stating the device node
and using the major/minor to follow the symlinks in /sys/dev/block/.
This cuts the execution time of grub-mkconfig to somewhere near 55% on
system without LVM (which uses libdevmapper instead sysfs_partition_path()).
Remove udevadm call as it does not help us more than calling stat() directly.
Signed-off-by: Jeff Mahoney <jeffm@suse.com>
Signed-off-by: Petr Vorel <pvorel@suse.cz>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
... to factor out fix for glibc 2.25 introduced in 7a5b301e3 (build: Use
AC_HEADER_MAJOR to find device macros).
Note: Once glibc 2.25 is old enough and this fix is not needed also
AC_HEADER_MAJOR in configure.ac should be removed.
Signed-off-by: Petr Vorel <pvorel@suse.cz>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The HEAP_MAX_ADDR is confusing. Currently it is set to 32MB, except on
ieee1275 on x86, where it is 64MB.
There is a comment which purports to explain it:
/* If possible, we will avoid claiming heap above this address, because it
seems to cause relocation problems with OSes that link at 4 MiB */
This doesn't make a lot of sense when the constants are well above 4MB
already. It was not always this way. Prior to commit 7b5d0fe444
(Increase heap limit) in 2010, HEAP_MAX_SIZE and HEAP_MAX_ADDR were
indeed 4MB. However, when the constants were increased the comment was
left unchanged.
It's been over a decade. It doesn't seem like we have problems with
claims over 4MB on powerpc or x86 ieee1275. The SPARC does things
completely differently and never used the constant.
Drop the constant and the check.
The only use of HEAP_MIN_SIZE was to potentially override the
HEAP_MAX_ADDR check. It is now unused. Remove it too.
Signed-off-by: Daniel Axtens <dja@axtens.net>
Signed-off-by: Stefan Berger <stefanb@linux.ibm.com>
Tested-by: Stefan Berger <stefanb@linux.ibm.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The "ide-drive" device was removed in QEMU 6.0. The "ide-hd" has been
available for more than 10 years now in QEMU. Thus there shouldn't be
any need for backwards compatible names.
Signed-off-by: Marius Bakke <marius@gnu.org>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
This incompat feature is used to denote that the filesystem stored its
metadata checksum seed in the superblock. This is used to allow tune2fs
changing the UUID on a mounted metdata_csum filesystem without having
to rewrite all the disk metadata. However, the GRUB doesn't use the
metadata checksum at all. So, it can just ignore this feature if it
is enabled. This is consistent with the GRUB filesystem code in general
which just does a best effort to access the filesystem's data.
The checksum seed incompat feature has to be removed from the ignore
list if the support for metadata checksum verification is added to the
GRUB ext2 driver later.
Suggested-by: Eric Sandeen <esandeen@redhat.com>
Suggested-by: Lukas Czerner <lczerner@redhat.com>
Signed-off-by: Javier Martinez Canillas <javierm@redhat.com>
Reviewed-by: Lukas Czerner <lczerner@redhat.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
The underlying type of 1ULL does not change across architectures but
grub_uint64_t does. This allows using the BF64_*CODE() macros as
arguments to format string functions that use the PRI* format string
macros that also vary with architecture.
Change the grub_error() call, where this was previously an issue and
temporarily fixed by casting and using a format string literal code,
to now use PRI* macros and remove casting.
Signed-off-by: Glenn Washburn <development@efficientek.com>
Reviewed-by: Daniel Kiper <daniel.kiper@oracle.com>
Skip versions between 2.07 and 2.10 to avoid leading zeros in minor
version number. This way version parsing in scripts should be easier.
Signed-off-by: Daniel Kiper <daniel.kiper@oracle.com>
2021-06-08 17:13:50 +02:00
778 changed files with 47466 additions and 63417 deletions
@ -1705,19 +1876,28 @@ if test x"$starfield_excuse" = x && test x"$enable_build_grub_mkfont" = xno ; th
starfield_excuse="No build-time grub-mkfont"
fi
if test x"$starfield_excuse" = x; then
for ext in pcf pcf.gz bdf bdf.gz ttf ttf.gz; do
for dir in . /usr/src /usr/share/fonts/X11/misc /usr/share/fonts/truetype/ttf-dejavu /usr/share/fonts/dejavu /usr/share/fonts/truetype; do
if test -f "$dir/DejaVuSans.$ext"; then
DJVU_FONT_SOURCE="$dir/DejaVuSans.$ext"
break 2
fi
done
done
AC_ARG_WITH([dejavufont],
AS_HELP_STRING([--with-dejavufont=FILE],
[set the DejeVu source [[guessed]]]))
if test "x$DJVU_FONT_SOURCE" = x; then
starfield_excuse="No DejaVu found"
fi
if test "x$with_dejavufont" = x; then
# search in well-known directories
if test x"$starfield_excuse" = x; then
for ext in pcf pcf.gz bdf bdf.gz ttf ttf.gz; do
for dir in . /usr/src /usr/share/fonts/X11/misc /usr/share/fonts/truetype/ttf-dejavu /usr/share/fonts/dejavu /usr/share/fonts/truetype /usr/pkg/share/fonts/X11/TTF /usr/local/share/fonts/dejavu /usr/X11R6/lib/X11/fonts/TTF /usr/share/fonts/dejavu-sans-fonts /usr/share/fonts/truetype/dejavu; do
if test -f "$dir/DejaVuSans.$ext"; then
DJVU_FONT_SOURCE="$dir/DejaVuSans.$ext"
break 2
fi
done
done
if test "x$DJVU_FONT_SOURCE" = x; then
starfield_excuse="No DejaVu found"
fi
fi
else
DJVU_FONT_SOURCE="$with_dejavufont"
fi
if test x"$enable_grub_themes" = xyes && test x"$starfield_excuse" != x; then
@ -1726,21 +1906,28 @@ fi
AC_SUBST([DJVU_FONT_SOURCE])
FONT_SOURCE=
AC_ARG_WITH([unifont],
AS_HELP_STRING([--with-unifont=FILE],
[set the unifont source [[guessed]]]))
for ext in pcf pcf.gz bdf bdf.gz ttf ttf.gz; do
for dir in . /usr/src /usr/share/fonts/X11/misc /usr/share/fonts/unifont /usr/share/fonts/uni /usr/share/fonts/truetype/unifont /usr/share/fonts/misc; do
# PCF and BDF from version 6.3 isn't hanled properly by libfreetype.
if test "$md5" = 0a54834d2788c83886a3e1785a6a1e61 || test "$md5" = 28f2565c7a41d8d407e2551159385edb || test "$md5" = dae5e588461b3b92b87b6ffee734f936 || test "$md5" = 4a3d687aa5bb329ed05f4263a1016791 ; then
continue
if test "x$with_unifont" = x; then
# search in well-known directories
for ext in pcf pcf.gz bdf bdf.gz ttf ttf.gz otf otf.gz; do
for dir in . /usr/src /usr/share/fonts/X11/misc /usr/share/fonts/unifont /usr/share/fonts/uni /usr/share/fonts/truetype/unifont /usr/share/fonts/misc /usr/pkg/share/fonts/X11/misc /usr/local/share/fonts/gnu-unifont /usr/local/share/fonts/unifont; do
# PCF and BDF from version 6.3 isn't hanled properly by libfreetype.
if test "$md5" = 0a54834d2788c83886a3e1785a6a1e61 || test "$md5" = 28f2565c7a41d8d407e2551159385edb || test "$md5" = dae5e588461b3b92b87b6ffee734f936 || test "$md5" = 4a3d687aa5bb329ed05f4263a1016791 ; then
continue
fi
FONT_SOURCE="$dir/unifont.$ext"
break 2
fi
FONT_SOURCE="$dir/unifont.$ext"
break 2
fi
done
done
done
else
FONT_SOURCE="$with_unifont"
fi
if test x"$enable_build_grub_mkfont" = xno ; then
FONT_SOURCE=
@ -1769,17 +1956,11 @@ if test x"$enable_grub_mount" = xno ; then
returngrub_error(GRUB_ERR_BAD_ARGUMENT,N_("empty entry in PCR list"));
if(next_pcr!=NULL)
*next_pcr='\0';
pcr=grub_strtoul(current_pcr,&pcr_end,10);
if(*current_pcr=='\0'||*pcr_end!='\0')
returngrub_error(GRUB_ERR_BAD_NUMBER,N_("entry '%s' in PCR list is not a number"),current_pcr);
if(pcr>TPM_MAX_PCRS-1)
returngrub_error(GRUB_ERR_OUT_OF_RANGE,N_("entry %llu in PCR list is too large to be a PCR number, PCR numbers range from 0 to %u"),(unsignedlonglong)pcr,TPM_MAX_PCRS-1);
pcrs[i]=(grub_uint8_t)pcr;
++(*pcr_count);
if(next_pcr==NULL)
break;
current_pcr=next_pcr+1;
if(*current_pcr=='\0')
returngrub_error(GRUB_ERR_BAD_ARGUMENT,N_("trailing comma at the end of PCR list"));
}
if(i==TPM_MAX_PCRS)
returngrub_error(GRUB_ERR_OUT_OF_RANGE,N_("too many PCRs in PCR list, the maximum number of PCRs is %u"),TPM_MAX_PCRS);
returngrub_error(GRUB_ERR_BAD_NUMBER,N_("TPM handle value '%s' is not a number"),value);
if(num>GRUB_UINT_MAX)
returngrub_error(GRUB_ERR_OUT_OF_RANGE,N_("value %llu is too large to be a TPM handle, TPM handles are unsigned 32-bit integers"),(unsignedlonglong)num);
