on sparc-linux: FAIL: ld-elf/loadaddr2

[Christian Joensson]

hmm, with the patch
http://sourceware.org/ml/binutils-cvs/2006-08/msg00110.html, I get two
failures:

/usr/local/src/trunk/objdir-binutils32/ld/../gas/as-new    -o
tmpdir/dump0.o /usr/local/src/trunk/src/ld/testsuite/ld-elf/loadaddr.s
./ld-new  -L/usr/local/src/trunk/src/ld/testsuite/ld-elf  -T
loadaddr1.t -z max-page-size=0x200000 -o tmpdir/dump tmpdir/dump0.o
succeeded with: <lt-ld-new: warning: dot moved backwards before `.data'
lt-ld-new: warning: dot moved backwards before `.data'
lt-ld-new: warning: dot moved backwards before `.data'>, expected: <>
lt-ld-new: warning: dot moved backwards before `.data'
lt-ld-new: warning: dot moved backwards before `.data'
lt-ld-new: warning: dot moved backwards before `.data'
succeeded with: <lt-ld-new: warning: dot moved backwards before `.data'
lt-ld-new: warning: dot moved backwards before `.data'
lt-ld-new: warning: dot moved backwards before `.data'>, expected: <>
FAIL: ld-elf/loadaddr1
/usr/local/src/trunk/objdir-binutils32/ld/../gas/as-new    -o
tmpdir/dump0.o /usr/local/src/trunk/src/ld/testsuite/ld-elf/loadaddr.s
./ld-new  -L/usr/local/src/trunk/src/ld/testsuite/ld-elf  -T
loadaddr2.t -z max-page-size=0x200000 -o tmpdir/dump tmpdir/dump0.o
/usr/local/src/trunk/objdir-binutils32/ld/../binutils/readelf  -l
--wide tmpdir/dump > tmpdir/dump.out
regexp_diff match failure
regexp "^  LOAD +0x302000 0xf*80302000 0xf*80302000 0x0*1 0x0*1 RW  0x200000$"
line   "  LOAD           0x302000 0x80302000 0x802f3000 0x00001
0x00001 RW  0x200000"
FAIL: ld-elf/loadaddr2


without the patch, I only get the second failure

On 8/21/06, H. J. Lu <hjl@lucon.org> wrote:
> On Mon, Aug 21, 2006 at 10:13:55AM +0200, Christian Joensson wrote:
> > On 8/21/06, Christian Joensson <christian.joensson@gmail.com> wrote:
> > >I just tried binutils daily snapshot, 20060821, and found a new
> > >testsuite failure, not present on the 20060818 snapshot, siince it was
> > >added 20060818... http://gcc.gnu.org/ml/gcc-cvs/2006-06/msg00792.html
> >
> > bah, a mixup, it should have been
> > http://sourceware.org/ml/binutils-cvs/2006-08/msg00093.html
> >
> > >/usr/local/src/trunk/objdir-binutils32/ld/../gas/as-new    -o
> > >tmpdir/dump0.o /usr/local/src/trunk/src/ld/testsuite/ld-elf/loadaddr.s
> > >./ld-new  -L/usr/local/src/trunk/src/ld/testsuite/ld-elf  -T
> > >loadaddr2.t -z max-page-size=0x200000 -o tmpdir/dump tmpdir/dump0.o
> > >/usr/local/src/trunk/objdir-binutils32/ld/../binutils/readelf  -l
> > >--wide tmpdir/dump > tmpdir/dump.out
> > >regexp_diff match failure
> > >regexp "^  LOAD +0x302000 0xf*80302000 0xf*80302000 0x0*1 0x0*1 RW
> > >0x200000$"
> > >line   "  LOAD           0x302000 0x80302000 0x802f3000 0x00001
> > >0x00001 RW  0x200000"
> > >FAIL: ld-elf/loadaddr2
> > >
> > >Any ideas of what is wrong?
> > >
> > >Would you like me to provide more informaton?
> > >
>
> It is
>
> http://sources.redhat.com/bugzilla/show_bug.cgi?id=3052
>
> The documented behavior of setting output section's load memory address
> in
>
> SECTION :
> {
> }
>
> has been changed, depending on the lma/vma of the previous output
> section. The documented behavior is at least used by the x86-64
> Linix kernel. There are 2 different usages of the documented
> behavior. The one used by the x86-64 Linix kernel is fixed. But the
> other one isn't. I believe that it is a regression if the document
> isn't updated. That is the failure you saw.
>
>
> H.J.
>


-- 
Cheers,

/ChJ