[Bug build/20621] New: glibc 2.27 fails to compile on x86_64 with "operand type mismatch" in fpu/e_expl.S

["michael.tremer at ipfire dot org" <sourceware-bugzilla at sourceware dot org>]

https://sourceware.org/bugzilla/show_bug.cgi?id=20621

            Bug ID: 20621
           Summary: glibc 2.27 fails to compile on x86_64 with "operand
                    type mismatch" in fpu/e_expl.S
           Product: glibc
           Version: unspecified
            Status: UNCONFIRMED
          Severity: normal
          Priority: P2
         Component: build
          Assignee: unassigned at sourceware dot org
          Reporter: michael.tremer at ipfire dot org
                CC: carlos at redhat dot com
  Target Milestone: ---

I am compiling an unmodified version of glibc on my machine and receive the
following issue on x86_64:

> ../sysdeps/x86_64/fpu/e_expl.S: Assembler messages:
> ../sysdeps/x86_64/fpu/e_expl.S:84: Error: operand type mismatch for `jae'
> /builddir/source/glibc-build/sysd-rules:1651: recipe for target '/builddir/source/glibc-build/math/s_expm1l.o' failed

Relevant code:
https://sourceware.org/git/?p=glibc.git;a=blob;f=sysdeps/x86_64/fpu/e_expl.S;h=a4ef023b2bfd634fad2793f4cf10ec98ecb97201;hb=HEAD#l77

GCC is 5.3.0 with both binutils 2.26 and 2.27

I think it is unlikely that this is a compiler bug and the file hasn't been
changed in quite a while. Glibc 2.23 is compiling fine with the same compiler
and toolchain.

Would be happy to get any advice on how to find the problem here.

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