This is the mail archive of the binutils@sourceware.org mailing list for the binutils project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: Xfail some of PR ld/3809 MIPS failures (part 1)


> From: Hans-Peter Nilsson <hp@axis.com>
> Date: Mon, 5 Mar 2012 05:19:34 +0100

Ping?

> Namely, that for 32-bit embedded MIPS targets (well, all, see
> below), there are these test failures:
> 
> Running /mnt/test1/hp/axis/binutils/src/ld/testsuite/ld-mips-elf/mips-elf.exp ...
> FAIL: MIPS eh-frame 3
> FAIL: MIPS eh-frame 4
> 
> ("part 2" will deal with "FAIL: Global calls from mips16")
> 
> As seen in mips-elf.exp, the tests are for some reason gated on
> mips*-*-elf:
> 
> if {$embedded_elf} {
>     run_dump_test "eh-frame3"
>     run_dump_test "eh-frame4"
> }
> 
> But the *.d files speak of "the official LP64 version of EABI64"
> and "the semi-official ILP32 variation of EABI64" which AFAICT
> can only apply to 64-bit MIPSen.  At first I thought "better
> gate the tests on '#target: mips*64*-*-*' then", but the tests
> fail for mipsisa64-elf and mipsisa64-sde-elf too, so they don't
> really pass anywhere.  A guess would be that the given gas and
> ld parameters don't force 64-bittyness to happen, if at all
> applicable.  With a PR already in place, better xfail the tests
> until a MIPS maintainer can take a closer look.  (FWIW, this
> hasn't happened for the last five years that the PR has been
> open.)
> 
> Tested with 'RUNTESTFLAGS=mips-elf.exp' cross to mipsel-linux,
> mipsisa32r2el-linux, mipsisa32r2el-elf, mipsisa32r2-elf,
> mipsel-elf, mips-elf, mips-sgi-irix6, mips-sgi-irix5,
> mipsel-sde-elf, mips64-kfreebsd-gnu, mips-dde-sysv4.2MP (a
> "mips-*-sysv*" yields ECOFF; a gas/ld inconsistency wart?),
> mipsisa64-sde-elf, mipsisa64el-elf, mips-freebsd8.2,
> mips64el-freebsd7.3, mips64-freebsd8.1, mipsel-freebsd7.2.
> together with the other fixes to-be-posted: lots of failures for
> mips-sgi-irix6 but none remaining for this test.  Results for
> *-freebsd* are blank and uninteresting as they returned too
> early due to mips-elf.exp gating on is_elf_target; see
> <http://sourceware.org/ml/binutils/2012-03/msg00039.html>.
> 
> Ok?
> 
> ld/testsuite:
> 
> 	PR ld/3809
> 	* ld-mips-elf/eh-frame3.d, ld-mips-elf/eh-frame4.d: Xfail.
> 
> diff --git a/ld/testsuite/ld-mips-elf/eh-frame3.d b/ld/testsuite/ld-mips-elf/eh-frame3.d
> index 0328eb8..0150236 100644
> --- a/ld/testsuite/ld-mips-elf/eh-frame3.d
> +++ b/ld/testsuite/ld-mips-elf/eh-frame3.d
> @@ -1,4 +1,5 @@
>  #name: MIPS eh-frame 3
> +#xfail: *-*-*
>  #source: eh-frame1.s
>  #source: eh-frame1.s
>  #as: -EB -mips3 -mabi=eabi --defsym alignment=3 --defsym fill=0 --defsym foo=0x1020304050607080
> @@ -8,6 +9,7 @@
>  # This test is for the official LP64 version of EABI64, which uses a
>  # combination of 32-bit objects and 64-bit FDE addresses.
>  #
> +# Currently xfailed as per PR ld/3809.
>  
>  Contents of the \.eh_frame section:
>  
> diff --git a/ld/testsuite/ld-mips-elf/eh-frame4.d b/ld/testsuite/ld-mips-elf/eh-frame4.d
> index effb457..6aea872 100644
> --- a/ld/testsuite/ld-mips-elf/eh-frame4.d
> +++ b/ld/testsuite/ld-mips-elf/eh-frame4.d
> @@ -1,4 +1,5 @@
>  #name: MIPS eh-frame 4
> +#xfail: *-*-*
>  #source: eh-frame1.s
>  #source: eh-frame1.s
>  #as: -EB -mips3 -mabi=eabi --defsym alignment=2 --defsym fill=0 --defsym foo=0x50607080
> @@ -7,6 +8,7 @@
>  #
>  # This test is for the semi-official ILP32 variation of EABI64.
>  #
> +# Currently xfailed as per PR ld/3809.
>  
>  Contents of the \.eh_frame section:
>  
> brgds, H-P
> 


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]