The 2.38 branch has been created.

[Alan Modra]

On Sun, Jan 30, 2022 at 02:32:11PM +0100, Matthias Klose wrote:
> On 1/29/22 06:29, Alan Modra wrote:
> > On Tue, Jan 25, 2022 at 03:37:08PM +0000, Nick Clifton via Binutils wrote:
> >>> ppc64el:
> >>> Running /<<PKGBUILDDIR>>/ld/testsuite/ld-bootstrap/bootstrap.exp ...
> >>> FAIL: bootstrap
> >>> FAIL: bootstrap with strip
> >>> FAIL: bootstrap with -Wl,--traditional-format
> >>> FAIL: bootstrap with -Wl,--no-keep-memory
> >>> FAIL: bootstrap with -Wl,--relax
> >>> FAIL: bootstrap with -Wl,--max-cache-size=-1
> >>
> >> I think that I will let Alan look into these...
> > 
> > I would if I could reproduce it.  My builds are OK.  Matthias,
> > ld/ld.log might give me a clue, if that is available somewhere.
> 
> Attached. This is triggered by configuring with --enable-pgo-build=lto.  The
> tests pass when binutils is configured without that option.

My --enable-pgo-build=lto was fine, but your compiler was different to
mine.  The log shows failures caused by gcc inform messages like:

../bfd/../../bfd/elfcode.h: In function 'bfd_elf64_swap_symbol_in':
../bfd/../../bfd/elfcode.h:189:16: note: the layout of aggregates containing vectors with 2-byte alignment has changed in GCC 5
  189 |   dst->st_info = H_GET_8 (abfd, src->st_info);

Don't ask me why you're getting that inform message.  There are no
vector fields let alone 2-byte aligned vector fields in either
elf_internal_sym or Elf64_External_Sym.

Anyway, dejagnu fails a test on any unexpected output that isn't
trimmed by prune_warnings.  So not actually a linker failure.

-- 
Alan Modra
Australia Development Lab, IBM