[Bug binutils/32340] New: strip has difficulty handling the relocations in .rel.llvm.call-graph-profile

[ptr1337 at cachyos dot org]

https://sourceware.org/bugzilla/show_bug.cgi?id=32340

            Bug ID: 32340
           Summary: strip has difficulty handling the relocations in
                    .rel.llvm.call-graph-profile
           Product: binutils
           Version: 2.43
            Status: UNCONFIRMED
          Severity: normal
          Priority: P2
         Component: binutils
          Assignee: unassigned at sourceware dot org
          Reporter: ptr1337 at cachyos dot org
  Target Milestone: ---

Hi together,

Recently there has been the Clang AutoFDO series posted to lkml. After testing
this series, I faced a bug in the packaging, as soon an AutoFDO profile was
passed.

Log:

```
make[2]: *** [scripts/Makefile.modinst:125: 
/tmp/makepkg/linux-cachyos-rc-autofdo/pkg/linux-cachyos-rc-autofdo/usr/lib/modules/6.12.0-rc5-5-cachyos-rc-autofdo/kernel/arch/x86/kvm/kvm.ko]
 
Error 1
make[2]: *** Deleting file 
'/tmp/makepkg/linux-cachyos-rc-autofdo/pkg/linux-cachyos-rc-autofdo/usr/lib/modules/6.12.0-rc5-5-cachyos-rc-autofdo/kernel/arch/x86/kvm/kvm.ko'
   INSTALL 
/tmp/makepkg/linux-cachyos-rc-autofdo/pkg/linux-cachyos-rc-autofdo/usr/lib/modules/6.12.0-rc5-5-cachyos-rc-autofdo/kernel/crypto/cryptd.ko
make[2]: *** Waiting for unfinished jobs....
```

Rong Xu found the issue in binutils after debugging:

Quote:
" 
In this case,
$ readelf -r kvm.ko |grep llvm.call-graph-profile
Relocation section '.rel.llvm.call-graph-profile' at offset 0xf62a00
contains 4 entries:

strip (v2.43.0) has difficulty handling the relocations in
.rel.llvm.call-graph-profile -- it silently failed with --strip-debug.
But strip (v.2.42) has no issue with kvm.ko. The strip in llvm (i.e.
llvm-strip) also passes with kvm.ko

I compared binutil/strip source code for version v2.43.0 and v2.42.
The different is around here:
In v2.42 of bfd/elfcode.h
   1618       if ((entsize == sizeof (Elf_External_Rela)
   1619            && ebd->elf_info_to_howto != NULL)
   1620           || ebd->elf_info_to_howto_rel == NULL)
   1621         res = ebd->elf_info_to_howto (abfd, relent, &rela);
   1622       else
   1623         res = ebd->elf_info_to_howto_rel (abfd, relent, &rela);

In v2.43.0 of bfd/elfcode.h
   1618       if (entsize == sizeof (Elf_External_Rela)
   1619           && ebd->elf_info_to_howto != NULL)
   1620         res = ebd->elf_info_to_howto (abfd, relent, &rela);
   1621       else if (ebd->elf_info_to_howto_rel != NULL)
   1622         res = ebd->elf_info_to_howto_rel (abfd, relent, &rela);

In the 2.43 strip, line 1618 is false and line 1621 is also false.
"res" is returned as false and the program exits with -1.

While in 2.42, line 1620 is true and we get "res" from line 1621 and
program functions correctly."

The commit, which introduced the issue is following:
https://github.com/bminor/binutils-gdb/commit/b20ab53f81db7eefa0db00d14f06c04527ac324c

Reverting this commit or using binutils 2.42 does not provide any issue and the
packaging is successful.

-- 
You are receiving this mail because:
You are on the CC list for the bug.