Re: [PATCH glibc 5/9] glibc: Perform rseq(2) registration at C startup and thread creation (v17)

[Szabolcs Nagy]

The 04/29/2020 10:18, Florian Weimer wrote:
> * Szabolcs Nagy:
> 
> > The 04/28/2020 10:58, Mathieu Desnoyers wrote:
> >> ----- On Apr 28, 2020, at 8:54 AM, Florian Weimer fw@xxxxxxxxxxxxx wrote:
> >> > That one definitely should work.
> >> > 
> >> > I expect you might see this if libgcc_s.so.1 is installed into a
> >> > multiarch subdirectory that upstream glibc does not search.  (The
> >> > Debian patches are unfortunately not upstream.)
> >> 
> >> My test environment is a Ubuntu 18.04.1 LTS.
> >> 
> >> > 
> >> > I think on my system, the built glibc can find the system libgcc_s via
> >> > /etc/ld.so.cache, so I haven't seen this issue yet.
> >> 
> >> On my system, libgcc_s is provided here:
> >> 
> >> /lib/x86_64-linux-gnu/libgcc_s.so.1
> >> 
> >> by this package:
> >> 
> >> Package: libgcc1
> >> Architecture: amd64
> >> Version: 1:8.4.0-1ubuntu1~18.04
> >
> > before running the tests
> >
> > cp `$CC --print-file-name libgcc_s.so.1` glibc/build/dir
> > cp `$CC --print-file-name libstdc++.so.6` glibc/build/dir
> >
> > so those toolchain libs are in the search path
> > of the newly built libc when running tests.
> 
> Do you actually see the need for these steps yourself?
> 
> I guess the correct fix would be to upstream the Debian multiarch
> changes and activate them automatically with a configure check on
> systems that use multiarch paths.

cancel tests work for me on an ubuntu system because
of /etc/ld.so.cache, but that may not be present
or the system may not be glibc based at all.

i always do the cp because i build gcc myself (usually
close to current master) and don't install it to the
system path which means at compile time and runtime
different libraries are used if i dont copy