Re: The 2.38 branch has been created.

["H.J. Lu" <hjl.tools@gmail.com>]

On Tue, Jan 25, 2022 at 4:39 AM Matthias Klose <doko@ubuntu.com> wrote:
>
> On 1/22/22 13:58, Nick Clifton via Binutils wrote:
> > Hi Everyone,
> >
> >   The 2.38 branch has now been created:
> >
> >      git clone git://sourceware.org/git/binutils-gdb.git -b binutils-2_38-branch
> >
> >   A snapshot of the sources is also available here:
> >
> >     https://sourceware.org/pub/binutils/snapshots/binutils-2.37.90.tar.xz
> >
> >   Please could all patches for the branch be run by me.
> >   The rules for the branch are:
> >
> >     * No new features.
> >     * Target specific bug fixes are OK.
> >     * Generic bug fixes are OK if they are important and widely tested.
> >     * Documentation updates/fixes are OK.
> >     * Translation updates are OK.
> >     * Fixes for testsuite failures are OK.
> >
> >   Ideally I would like to make the release happen in two weeks time,
> >   i.e. Saturday 5th Feb 2022.  Which I hope will be enough time for
> >   everyone to get their final fixes in.
>
> some issues:
>
> https://sourceware.org/bugzilla/show_bug.cgi?id=28819
> has a reproducer to show that a ld change between Jan 06 and Jan 19 builds
> binaries without RELRO support on at least some architectures.
>
> this is a build configured with -enable-pgo-build=lto.
>
> on any arch:
> make[5]: Entering directory '/<<PKGBUILDDIR>>/builddir-single/libbacktrace'
> FAIL: allocfail.sh
>
>
> libbacktrace seems not to be built during the build, but only when running the
> tests. Is this expected?
>
> test results for various architectures:
>
> amd64, ppc64, hppa, powerpc, x32: no test failures
>
>
> arm64, regression from 2.37:
> Running /<<PKGBUILDDIR>>/ld/testsuite/ld-aarch64/aarch64-elf.exp ...
> FAIL: ld-aarch64/tls-relax-gdesc-le-now
>
> armel, armhf:
> Running /<<PKGBUILDDIR>>/ld/testsuite/ld-elf/elf.exp ...
> FAIL: .noinit sections
> FAIL: .persistent sections
> Running /<<PKGBUILDDIR>>/ld/testsuite/ld-elf/tls.exp ...
> FAIL: Build pr22263-1
> Running /<<PKGBUILDDIR>>/ld/testsuite/ld-ifunc/ifunc.exp ...
> FAIL: Run pr18841 with libpr18841b.so
> FAIL: Run pr18841 with libpr18841c.so
> FAIL: Run pr18841 with libpr18841bn.so (-z now)
> FAIL: Run pr18841 with libpr18841cn.so (-z now)
> FAIL: Run pr23169a
> FAIL: Run pr23169d
>
>
> i686 (regressions):
> Running /<<PKGBUILDDIR>>/ld/testsuite/ld-i386/i386.exp ...
> FAIL: Run pr19031
> FAIL: Run got1
> FAIL: Undefined weak symbol (-fPIE -no-pie)
> FAIL: Undefined weak symbol (-fPIE -pie)
>

I can't reproduce this with --enable-pgo-build=lto using GCC 11.2.
Which GCC are you using? Does it fail without --enable-pgo-build=lto?

-- 
H.J.