This is the mail archive of the
binutils@sources.redhat.com
mailing list for the binutils project.
Re: Miscellaneous fixes for testsuite failures on hppa64-hp-hpux11.00
- From: law at redhat dot com
- To: "John David Anglin" <dave at hiauly1 dot hia dot nrc dot ca>
- Cc: binutils at sources dot redhat dot com, amodra at bigpond dot net dot au
- Date: Thu, 06 Jun 2002 00:53:36 -0600
- Subject: Re: Miscellaneous fixes for testsuite failures on hppa64-hp-hpux11.00
- Reply-to: law at redhat dot com
In message <200206060412.g564CPRE021228@hiauly1.hia.nrc.ca>, "John David Anglin
" writes:
> Here is a bunch of small fixes for testsuite failures on hppa64-hp-hpux11.00
> .
> After this patch is applied, there are no unexpected failures in either
> the binutils or gas testsuites. There are 7 remaining fails in the ld
> testsuite when ld is built with the HP ansi compiler. These are:
>
> FAIL: bootstrap with --static
> FAIL: cdtest
> FAIL: cdtest with -Ur
> FAIL: selective1
> FAIL: selective2
> FAIL: selective4
> FAIL: selective5
>
> There is one additional failure with gcc which I am currently investigating.
>
> The "bootstrap with --static" is a result of the following error:
>
> /xxx/gnu/binutils-2.12.90/objdir64/ld/ld-new -m elf64hppa -o tmpdir/ld1 /us
> r/cc
> s/lib/pa20_64/crt0.o --static tmpdir/ld-partial.o ../bfd/.libs/libbfd.a ../l
> ibib
> erty/libiberty.a /opt/gnu64/lib/gcc-lib/hppa64-hp-hpux11.00/3.2/libgcc.a -
> lc /
> usr/lib/pa20_64/milli.a
> /xxx/gnu/binutils-2.12.90/objdir64/ld/ld-new: tmpdir/ld1: Not enough room fo
> r pr
> ogram headers (allocated 5, need 6)
> /xxx/gnu/binutils-2.12.90/objdir64/ld/ld-new: final link failed: Bad value
> FAIL: bootstrap with --static
>
> I'm also seeing this error recently in a number of libstdc++ testsuite
> failures under hppa-linux.
>
> The only non-obvious fix is the one to tc-hppa.c. In this case, the
> expected assembler warning wasn't being produced under hppa64-hp-hpux11.00
> because a forced relocation was being generated for the branch even
> though there was no possibility that the branch could reach a long branch
> stub. The check added only guarantees that the branch could reach
> a stub at the beginning of the current call_info block. This underestimates
> the actual distance of the branch but it's difficult to be more precise.
>
> Please apply if OK.
>
> Dave
> --
> J. David Anglin dave.anglin@nrc.ca
> National Research Council of Canada (613) 990-0752 (FAX: 952-66
> 05)
>
> 2002-06-06 John David Anglin <dave@hiauly1.hia.nrc.ca>
>
> * binutils/testsuite/binutils-all/objcopy.exp: clear xfail for
> "hppa*64*-*-hpux*".
> * binutils/testsuite/binutils-all/hppa/objdump.exp: Return if
> "*64*-*-*".
> * gas/config/tc-hppa.c (pa_ip): Replace "L$0\001" with FAKE_LABEL_NAME.
> (hppa_force_relocation): Check if a stub just before the start symbol
> of the last call_info is reachable before forcing relocation. Fix
> typo.
> * gas/testsuite/gas/hppa/basic/add.s: Set code level to 1.1.
> * gas/testsuite/gas/hppa/basic/branch.s: Likewise.
> * gas/testsuite/gas/hppa/basic/dcor.s: Likewise.
> * gas/testsuite/gas/hppa/basic/logical.s: Likewise.
> * gas/testsuite/gas/hppa/basic/shladd.s: Likewise.
> * gas/testsuite/gas/hppa/basic/sub.s: Likewise.
> * gas/testsuite/gas/hppa/basic/unit.s: Likewise.
> * gas/testsuite/gas/hppa/unsorted/unsorted.exp: Return if target
> "*-*-linux*".
> * ld/configure.host (hppa*64*-*-hpux11*): Define NATIVE_LIB_DIRS,
> HOSTING_CRT0 and HOSTING_LIBS.
Thanks. Installed.
jeff