Skip to content

PCI: Fix crash when access broken ROM

Guixin Liu linux-pci 6 versions – Applied

Status from this message: Applied to pci/rom for v7.2, thanks for all your work on this!

Get series with b4
b4 am [email protected]

Versions newest first · linked by author and title

  1. v12 Latest2 patches 1 replyApplied
  2. v112 patches 9 repliesSuperseded
  3. v102 patches 4 repliesSuperseded
Earlier versions · 3
VersionPostedPatchesRepliesStatus
v92 patches4 repliesSuperseded
v82 patches4 repliesSuperseded
v72 patches5 repliesSuperseded

Changes in v12 from the cover letter

- Add rb tag from Krzysztof Wilczyński in the first patch, thanks.
- Change "get" to "Get".
- Renamed parameter last_image → expect_valid in pci_rom_is_header_valid() to better reflect its semantics: it indicates whether the caller expects the image to be valid (and thus whether a missing/invalid signature should be reported as an error or as normal end-of-chain).
- Tightened image alignment check: replaced the 2-byte alignment check with a 512-byte (PCI_ROM_IMAGE_SECTOR_SIZE) alignment check on image, per PCI Firmware Specification r3.3, sec 5.1, which mandates that each ROM image starts on a 512-byte boundary. This also satisfies the natural-alignment requirement of readw() on architectures such as arm64.
- Updated comment to cite the PCI Firmware Spec r3.3 sec 5.1 as the authoritative source for the alignment requirement, and to explain the relationship between page-aligned rom, sector-aligned image, and the IOMEM access constraint.
- Fixed off-by-one in overflow checks: check_add_overflow() now uses PCI_ROM_HEADER_SIZE - 1 and data_len - 1 so that header_end / end represent the inclusive last byte of the region, matching the subsequent > rom_end comparison.
- Refactored signature-check log flow: collapsed the dual-return branches into a single if (signature != PCI_ROM_IMAGE_SIGNATURE) block, emitting the appropriate pci_info() based on expect_valid, then returning false; success path returns true at the end.
- Reorder pci_rom_is_data_struct_valid() to check the "PCIR" signature before reading data_len, so bad signatures are still logged.
- Collapse the signature branch to early-return on failure, matching the style of pci_rom_is_header_valid().
- Add PCI_ROM_DATA_STRUCT_MIN_LEN (0x18), the PCI 2.x baseline PCI Data Structure length.
- Reject data_len < PCI_ROM_DATA_STRUCT_MIN_LEN to keep the fixed-offset reads (PCI_ROM_IMAGE_LEN @0x10, PCI_ROM_LAST_IMAGE_INDICATOR @0x15) in pci_get_rom_size() inside the mapped ROM window.
- Cite PCI Firmware Spec r3.3 sec 5.1.3 Table 5-2 in the new macro's comment.

Patches in v12 Applied

#Subject and filesSizeRepliesApplied
1/2PCI: Introduce named defines for PCI ROMdrivers/pci/rom.c+28 -120 replies
2/2PCI: Check ROM header and data structure addr before accessingdrivers/pci/rom.c+110 -180 replies

Reviews trailers per patch

All 2 patches have a Reviewed-by or Acked-by.

PatchAndy ShevchenkoKrzysztof Wilczyński
1/2R-b R-b
2/2R-b

Trailers given in replies are filled; those carried in the patches from earlier versions are outlined with dashes.

Read the v12 thread

Cover letter [PATCH v12 0/2] PCI: Fix crash when access broken ROM

v11 -> v12:
- Add rb tag from Krzysztof Wilczyński in the first patch, thanks.
- Change "get" to "Get".
- Renamed parameter last_image → expect_valid in pci_rom_is_header_valid() to better reflect its semantics: it indicates whether the caller expects the image to be valid (and thus whether a missing/invalid signature should be reported as an error or as normal end-of-chain).
- Tightened image alignment check: replaced the 2-byte alignment check with a 512-byte (PCI_ROM_IMAGE_SECTOR_SIZE) alignment check on image, per PCI Firmware Specification r3.3, sec 5.1, which mandates that each ROM image starts on a 512-byte boundary. This also satisfies the natural-alignment requirement of readw() on architectures such as arm64.
- Updated comment to cite the PCI Firmware Spec r3.3 sec 5.1 as the authoritative source for the alignment requirement, and to explain the relationship between page-aligned rom, sector-aligned image, and the IOMEM access constraint.
- Fixed off-by-one in overflow checks: check_add_overflow() now uses PCI_ROM_HEADER_SIZE - 1 and data_len - 1 so that header_end / end represent the inclusive last byte of the region, matching the subsequent > rom_end comparison.
- Refactored signature-check log flow: collapsed the dual-return branches into a single if (signature != PCI_ROM_IMAGE_SIGNATURE) block, emitting the appropriate pci_info() based on expect_valid, then returning false; success path returns true at the end.
- Reorder pci_rom_is_data_struct_valid() to check the "PCIR" signature before reading data_len, so bad signatures are still logged.
- Collapse the signature branch to early-return on failure, matching the style of pci_rom_is_header_valid().
- Add PCI_ROM_DATA_STRUCT_MIN_LEN (0x18), the PCI 2.x baseline PCI Data Structure length.
- Reject data_len < PCI_ROM_DATA_STRUCT_MIN_LEN to keep the fixed-offset reads (PCI_ROM_IMAGE_LEN @0x10, PCI_ROM_LAST_IMAGE_INDICATOR @0x15) in pci_get_rom_size() inside the mapped ROM window.
- Cite PCI Firmware Spec r3.3 sec 5.1.3 Table 5-2 in the new macro's comment.

v10 -> v11:
- Change 'pci rom' to 'PCI ROM' of the tittle of the first patch.
- Add Andy Shevchenko's rb tag in the first patch, thanks.

v9 -> v10:
- Reorder the header files, and not touch kernel.h
- Change PCI_ROM_IMAGE_LEN_UNIT_BYTES to PCI_ROM_IMAGE_SECTOR_SIZE.
- Add a comment for PCI_ROM_DATA_STRUCT_SIGNATURE.

v8 -> v9:
- Supplemental explanation for the commit body of the first patch.
- Change PCI_ROM_IMAGE_LEN_UNIT_SZ_512 to PCI_ROM_IMAGE_LEN_UNIT_BYTES, and change it's definition to SZ_512.
- Use u16 and u32 for signature val instead of unsigned short/int.

v7 -> v8:
- Ordered header files alphabetically.
- Convert the literals too in the firt patch.
- Use local val to save signature instead of reading twice.

v6 -> v7:
- Put all named defines to a separate patch.
- Change PCI_ROM_IMAGE_LEN_UNIT_BYTES to PCI_ROM_IMAGE_LEN_UNIT_SZ_512.
- Named BIT(7) to PCI_ROM_LAST_IMAGE_INDICATOR_BIT.
- Fix all other comments from Ilpo, such as including header files, and alignment fault, Thanks.

v5 -> v6:
- Convert some magic number to named defines, suggested by Ilpo, thanks.

v4 -> v5:
- Add Andy Shevchenko's rb tag, thanks.
- Change u64 to unsigned long.
- Change pci_rom_header_valid() to pci_rom_is_header_valid() and change pci_rom_data_struct_valid() to pci_rom_is_data_struct_valid().
- Change rom_end from rom+size to rom+size-1 for more readble, and also change header_end >= rom_end to header_end > rom_end, same as data structure end.
- Change if(!last_image) to if (last_image)..
- Use U16_MAX instead of 0xffff.
- Split check_add_overflow() from data_len checking.
- Remove !!() when reading last_image, and Use BIT(7) instead of 0x80.

v3 -> v4:
- Use "u64" instead of "uintptr_t".
- Invert the if statement to avoid excessive indentation.
- Add comment for alignment checking.
- Change last_image's type from int to bool.

v2 -> v3:
- Add pci_rom_header_valid() helper for checking image addr and signature.
- Add pci_rom_data_struct_valid() helper for checking data struct add and signature.
- Handle overflow issue when adding addr with size.
- Handle alignment fault when running on arm64.

v1 -> v2:
- Fix commit body problems, such as blank line in "Call Trace" both sides, thanks, (Andy Shevchenko).
- Remove every step checking, just check the addr is in header or data struct.
- Add Suggested-by: Guanghui Feng <[email protected]> tag.

Guixin Liu (2):
PCI: Introduce named defines for PCI ROM
PCI: Check ROM header and data structure addr before accessing

 drivers/pci/rom.c | 154 +++++++++++++++++++++++++++++++++++++++-------
 1 file changed, 131 insertions(+), 23 deletions(-)
Signature · 1 line
2.43.7