On 10 July 2015 at 15:53, Matt Fleming <m...@codeblueprint.co.uk> wrote: > On Thu, 09 Jul, at 05:58:36PM, Ard Biesheuvel wrote: >> >> Sure. In the meantime, I have cooked up a slightly more elaborate >> series that also fixes the >> EFI_IMAGE_REL_BASED_LOW/EFI_IMAGE_REL_BASED_HIGH issues (even though >> they don't seem to be used widely) and does some cleanups. This is >> unlikely to be adopted as-is, since there is still an unresolved issue >> with inter-region relative relocations. >> >> Tree is here >> https://git.linaro.org/people/ard.biesheuvel/uefi-next.git/shortlog/refs/heads/pecoff-runtime-relocation > > Thanks Ard! This is really, super helpful. >
Sure, no problem. >> When running this, you should notice regions in the memory map with >> either the RO (0x20000) or XP (0x4000) bit set in the kernel log. >> For instance, >> >> """ >> [ 0.000000] efi: mem38: type=6, attr=0x800000000000400f, >> range=[0x0000000007e9a000-0x0000000007e9f000) (0MB) >> [ 0.000000] efi: mem39: type=5, attr=0x800000000002000f, >> range=[0x0000000007e9f000-0x0000000007ea4000) (0MB) >> [ 0.000000] efi: mem40: type=6, attr=0x800000000000400f, >> range=[0x0000000007ea4000-0x0000000007eaa000) (0MB) >> [ 0.000000] efi: mem41: type=5, attr=0x800000000002000f, >> range=[0x0000000007eaa000-0x0000000007eae000) (0MB) >> """ > > Yep, I see this and it looks good. > >> And no crash .... (hopefully) > > With a standard kernel, I don't see a crash. However, with the minimal > top 2 patches in, > > https://git.kernel.org/cgit/linux/kernel/git/mfleming/efi.git/log/?h=memmap > > I see a page fault while executing SetVirtualAddressMap(). It appears to > be caused by the firmware performing relocation fixups on an address > that's mapped as EfiRuntimeServicesCode, which is in the EFI memory map > as EFI_MEMORY_RO. > > I suspect you don't see this issue on aarch64 because you call > SetVirtualAddressMap() so early in boot. > Indeed. > I'll try and find some time to dig into this issue this coming week, but > if anyone wants to beat me to it, feel free ;-) > > The first question is: Why is the image's reloc section in a > EfiRuntimeServiceCode region? > Are you sure it is the .reloc section itself? It could well be the target of a relocation fixup that is inside a code region, which cannot be applied due to the page permissions. -- Ard. ------------------------------------------------------------------------------ Don't Limit Your Business. Reach for the Cloud. GigeNET's Cloud Solutions provide you with the tools and support that you need to offload your IT needs and focus on growing your business. Configured For All Businesses. Start Your Cloud Today. https://www.gigenetcloud.com/ _______________________________________________ edk2-devel mailing list edk2-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/edk2-devel
