On Mon, Aug 1, 2022 at 3:31 AM Petr Mladek <pmladek@xxxxxxxx> wrote: > > On Sat 2022-07-30 15:32:58, Song Liu wrote: > > On Tue, Jul 26, 2022 at 8:54 PM Song Liu <song@xxxxxxxxxx> wrote: > > > > > > On Tue, Jul 26, 2022 at 4:33 PM Josh Poimboeuf <jpoimboe@xxxxxxxxxx> wrote: > > > > > > > > On Thu, Jul 21, 2022 at 10:51:47AM -0700, Song Liu wrote: > > > > > From: Miroslav Benes <mbenes@xxxxxxx> > > > > > > > > > > Josh reported a bug: > > > > > > > > > > When the object to be patched is a module, and that module is > > > > > rmmod'ed and reloaded, it fails to load with: > > > > > > > > > > module: x86/modules: Skipping invalid relocation target, existing value is nonzero for type 2, loc 00000000ba0302e9, val ffffffffa03e293c > > > > > livepatch: failed to initialize patch 'livepatch_nfsd' for module 'nfsd' (-8) > > > > > livepatch: patch 'livepatch_nfsd' failed for module 'nfsd', refusing to load module 'nfsd' > > > > > > > > > > The livepatch module has a relocation which references a symbol > > > > > in the _previous_ loading of nfsd. When apply_relocate_add() > > > > > tries to replace the old relocation with a new one, it sees that > > > > > the previous one is nonzero and it errors out. > > > > > > > > > > On ppc64le, we have a similar issue: > > > > > > > > > > module_64: livepatch_nfsd: Expected nop after call, got e8410018 at e_show+0x60/0x548 [livepatch_nfsd] > > > > > livepatch: failed to initialize patch 'livepatch_nfsd' for module 'nfsd' (-8) > > > > > livepatch: patch 'livepatch_nfsd' failed for module 'nfsd', refusing to load module 'nfsd' > > > > > > > > > 2) We can't only fix x86, powerpc also needs a fix. > > > > > > I have very little experience with powerpc. Would someone be willing to > > > help with powerpc part of this? > > > > I guess folks are all busy. Any suggestions on how to test powerpc changes? > > You might also send the patch and just mention that you were not able > to test it on powerpc. There are people with access to powerpc that > might check it. Let me give it a try. > > Another question is how to actually test it. I wonder how you do > it for x86. I tested it with a patch on an in-tree module. I was using kpatch-build and testing it on 5.18 kernels. This is tricky at the moment for 5.19 because of this issue: https://github.com/dynup/kpatch/issues/1277 > > Finally, also s390 supports livepatching. I guess that the problem is > there as well. Is is straightforward to write the revert code, please? v2 of the work doesn't have any real logic for s390. I wonder whether it is needed. Thanks, Song v2: https://lore.kernel.org/all/20190905124514.8944-1-mbenes@xxxxxxx/T/#u
