As the first step towards automated parsing, this change set replaces outdated BootGuard-related parsers with shiny new KaitaiStruct-based ones.
It also does the following:
- improves Intel FIT definitions by using the relevant specification
- adds sha1, sha384, sha512 and sm3 digest implementations
- updates LZMA SDK to v22.01
- moves GUIDs out of include files to prevent multiple instantiations
- enforces C++11
- adds Kaitai-based parsers for Intel FIT, BootGuard v1 and BootGuard v2 structures
- makes many small refactorings here, there and everywhere
+ FFSv3 support with large files and large sections
+ proper names for Flash Descriptor v2 regions (#89)
+ better alignment calculations (#96)
+ improved NVRAM parser
+ post IBB hash support for Boot Guard
+ bugfixes
+ companion tool updated
- human readable JEDEC ID
- NVRAM parser separated from FFS parser
- added support for LZMAF86 sections
- solved a bug with parsing of VSS variables with invalid sizes
- added wrappers over Qt classes for seamless replacement if Qt is not available
- added bstrlib as submodule
- only UEFIExtract works with this changes for now, others will followa bit later
- thanks to NVAR extended headers, item tails are back
- fixed a bug with "Extract uncompressed" producing incorrect results
for multiple sections
- fixed detection and parsing of NVRAM external defaults file
- fixed "Extract as is" on NVAR entries not including extended header
- small bugfixes and refactorings
- added support for all Phoenix SCT stuff from NVRAM volume: Intel
microcodes, CMDB block, SLIC pubkey/marker.
- added support for EVSA extended variables (bigger than 0xFFFF)
- better EVSA parsing (invalid data vars are shown now)
- FlashMap entries now have human-readable types
- various small fixes and refactorings
- added UI for VSS storage operations
- changed VSS storage/vars parsing order
- solved problems with GUIDs as filenames
- Fsys, _FDC and other NVRAM formats TBD
- added FfsBuilder code and UI, but reconstruction routines for volumes,
files and sections are still not ready
- FfsOps moved to common
- QVector and QPair aren't used anymore, replaces with std::vector and
std::pair
- common classes are now independent from QObject
- next step is to replace QString with CBString from bstrlib
Thanks to lordkag for #41:
- improved parsing of Intel flash descriptor
- improved detection of Tiano/EFI 1.1 compression type
- added 2 UEFI capsule GUIDs used by Lenovo
- solved potential crash on very low memory available
- UEFIExtract and UEFIFind update to include the latest parser changes
- added detection of VTF inside volume's non-UEFI data
- corrected ImageSize information for EFI capsule
- *.scap is added to the list of known Image file extensions
- all FIT candidates in a tree item are checked for being referenced
from the last VTF now
- added sanity checks during volume header parsing
- corrected various small FIT issues
- corrected bogus "unaligned volume" messages during capsule files
parsing
- solved a bug in offset calculation for Intel images inside a capsule
- fitParser code cleanup
- hopefully the last commit to this tree, time to start another branch
and rewrite the whole thing once again
- second pass of parsing to add physical memory addresses to all
uncompressed items in the tree
- TE image revisions can be detected now
- more builder routines, but still not ready for enabling
- PE and TE header info is back
