Bug#808181: libc6: Upgrade can make the linker unusable
reassign 808181 libc6-dev
reassign 808205 libc6-dev
forcemerge 808181 808205
forcemerge 808181 808206
severity 808181 serious
thanks
On 2015-12-16 23:36, Aurelien Jarno wrote:
> On 2015-12-16 13:15, Dima Kogan wrote:
> > Package: libc6
> > Severity: normal
> >
> > Hi. I had
> >
> > libc6 = 2.19-22
> > binutils = 2.25-4
> >
> > and all was well. Then I upgraded to libc6 = 2.21-4 (currently latest in
> > sid). As a result, even the most basic build-time linking would fail.
> > For instance, with a trivial hello-world program:
> >
> > $ gcc-5 -o tst tst.c
> >
> > /usr/bin/ld: /usr/lib/gcc/x86_64-linux-gnu/5/../../../x86_64-linux-gnu/crti.o: unrecognized relocation (0x2a) in section `.init'
> > /usr/bin/ld: final link failed: Bad value
> > collect2: error: ld returned 1 exit status
> >
> > This would happen with gcc-5 and with gcc-4.9. Downgrading libc6 would
> > fix it. After some fiddling I realized that upgrading to binutils =
> > 2.25.90.20151209-1 (currently latest in sid) fixes it. I.e. with the
> > latest libc6 and the latest binutils packages things work.
>
> The problem is not introduced by the glibc, but just by the fact that
> it has been built with a recent binutils version which adds new
> relocation types on i386 and amd64. This means that ALL static libraries
> are affected by this problem.
>
> > Can the broken combination be prevented with some Conflicts: tags?
> > Currently this is a trap for the unwary.
>
> I therefore don't think we need to fix that at the glibc level. Either
> we just ignore the problem saying we don't support partial upgrades or
> we try to find a global way to fix the dependencies for all libraries.
We are working to migrate binutils version 2.25.90.20151209-1 into
testing asap. If everything goes well it should be the case after the
13:52 UTC dinstall run, so a few hours after that on the mirrors.
In the meantime fetching and installing this version from sid, should
solve the issue.
Aurelien
--
Aurelien Jarno GPG: 4096R/1DDD8C9B
aurelien@aurel32.net http://www.aurel32.net
Reply to: