[testsuite commit] Fix -m32 results [Re: [PATCH] Remove some KPASSes in gdb.base/callfuncs.exp]

[Mark Kettenis]

> Date: Wed, 24 Oct 2012 15:48:28 +0200
> From: Jan Kratochvil <jan.kratochvil@redhat.com>
> 
> On Tue, 23 Oct 2012 20:27:40 +0200, Mark Kettenis wrote:
> > Just fixed the bugs.  Committed as obvious.
> 
> It is a testsuite results regression in -m32 mode:

You have a funny defenition of regression.

The problem is just that the setup_kfail doesn't cater for non-default
compiler options.

Your "fix" has an annoying side-effect.  Took me a while to figure out
where the:

Executing on host: gcc  -c  -o ilp3216410.o ilp3216410.c    (timeout = 300)
spawn gcc -c -o ilp3216410.o ilp3216410.c
Executing on host: gcc  -c  -o reg6416410.o reg6416410.s    (timeout = 300)
spawn gcc -c -o reg6416410.o reg6416410.s
reg6416410.s: Assembler messages:
reg6416410.s:9: Error: bad register name `%r8'
reg6416410.s:10: Error: bad register name `%r9'
reg6416410.s:11: Error: bad register name `%r10'
reg6416410.s:12: Error: bad register name `%r11'
reg6416410.s:13: Error: bad register name `%r12'
reg6416410.s:14: Error: bad register name `%r13'
reg6416410.s:15: Error: bad register name `%r14'
reg6416410.s:16: Error: bad register name `%r15'
compiler exited with status 1
output is:
reg6416410.s: Assembler messages:
reg6416410.s:9: Error: bad register name `%r8'
reg6416410.s:10: Error: bad register name `%r9'
reg6416410.s:11: Error: bad register name `%r10'
reg6416410.s:12: Error: bad register name `%r11'
reg6416410.s:15: Error: bad register name `%r14'
reg6416410.s:16: Error: bad register name `%r15'

in the middle of my test output came from.