Closed x86-sec closed 2 years ago
Ignore the comment on WORKSPACE - edk2 setup has its own conflicting WORKSPACE variable and we should probably not rely on the kAFL variable to be intact.
Does it make sense to use the script root as working directory and clone to there? TARGET_ROOT already points there.
Ignore the comment on WORKSPACE - edk2 setup has its own conflicting WORKSPACE variable and we should probably not rely on the kAFL variable to be intact.
Does it make sense to use the script root as working directory and clone to there? TARGET_ROOT already points there.
Do you propose to clone KAFL_ROOT/edk2.git to KAFL_ROOT/targets/uefi_ovmf_64/edk2.git ?And TARGET_ROOT is an alias to KAFL_ROOT/targets/uefi_ovmf_64/ ?
Yes. TARGET_ROOT is already set for this and I suppose as user you anyway use this as your main/working directory, or not?
My goal is just that a user easily finds the edk2 folder. It should be obvious after ./compile.sh edk2.
Accidentally pushed to your branch. These fix the above items and some more. I can run your dxe_null sample and it finds the crash - but the TestBMP and TestDecompress do not build? :-/
I'm happy to merge (still an improvement) but would be great to get some more / non-trivial samples working again.
Also noted that build_ovmf() and build_platform() are kind of redundant. Can we merge them?
Fixes UEFI APP harnessing Adds DXE phase harnessing support