Messages in this thread | | | From | Sedat Dilek <> | Date | Fri, 31 Jul 2020 16:58:55 +0200 | Subject | Re: [PATCH v5 0/7] x86/boot: Remove run-time relocations from compressed kernel |
| |
On Fri, Jul 31, 2020 at 4:53 PM Arvind Sankar <nivedita@alum.mit.edu> wrote: > > On Fri, Jul 24, 2020 at 04:25:20PM -0700, Kees Cook wrote: > > On Sat, Jul 18, 2020 at 08:44:50AM +0300, Ard Biesheuvel wrote: > > > On Fri, 17 Jul 2020 at 21:17, Nick Desaulniers <ndesaulniers@google.com> wrote: > > > > > > > > On Fri, Jul 17, 2020 at 6:46 AM Arvind Sankar <nivedita@alum.mit.edu> wrote: > > > > > > > > > > On Tue, Jul 14, 2020 at 08:41:26PM -0400, Arvind Sankar wrote: > > > > > > The compressed kernel currently contains bogus run-time relocations in > > > > > > the startup code in head_{32,64}.S, which are generated by the linker, > > > > > > but must not actually be processed at run-time. > > > > > > > > > > > > This generates warnings when linking with the BFD linker, and errors > > > > > > with LLD, which defaults to erroring on run-time relocations in read-only > > > > > > sections. It also requires the -z noreloc-overflow hack for the 64-bit > > > > > > kernel, which prevents us from linking it as -pie on an older BFD linker > > > > > > (<= 2.26) or on LLD, because the locations that are to be apparently > > > > > > relocated are only 32-bits in size and so cannot really have > > > > > > R_X86_64_RELATIVE relocations. > > > > > > > > > > > > This series aims to get rid of these relocations. I've build- and > > > > > > boot-tested with combinations of clang/gcc-10 with lld/bfd-2.34, and > > > > > > gcc-4.9.0 with bfd-2.24, skipping clang on 32-bit because it currently > > > > > > has other issues [0]. > > > > > > > > > > > > > > > > Hi Thomas, Ingo, Borislav, would you be able to take a look over this > > > > > series in time for 5.9? > > > > > > > > Hi Arvind, thanks for the series; I'm behind on testing. When I try > > > > to apply this series on top of linux-next, I get a collision in > > > > drivers/firmware/efi/libstub/Makefile:27 when applying "0002 > > > > x86/boot/compressed: Force hidden visibility for all symbol > > > > references". Would you mind refreshing the series to avoid that > > > > collision? > > > > > > That is not the right way to deal with conflicts against -next. > > > > > > This series targets the -tip tree, and applies fine against it. If you > > > want to apply it on some other tree and test it, that is fine, and > > > highly appreciated, but 'refreshing' the series against -next means it > > > no longer applies to -tip, and may be based on unidentified conflict > > > resolutions performed by Stephen that the maintainers will have to > > > deal with. > > > > > > Boris, Ingo, Thomas, > > > > > > Mind taking v5 of this series? (With Nick's Tested-by) I think these > > > patches have been simmering long enough. Do note there is a conflict > > > against the kbuild tree, but the resolution should be straightforward. > > > > I would love that; I need to rebase my orphan series on this too... > > > > -- > > Kees Cook > > Ping?
I just tested v5 against Linux v5.8-rc7 and the conflict you mention is with:
commit da05b143a308bd6a7a444401f9732678ae63fc70 x86/boot: Don't add the EFI stub to targets
If I forgot... Tested-by: Sedat Dilek <sedat.dilek@gmail.com>
- Sedat -
| |