target/10383: sparc-sun-solaris2.7 G++ testsuite failures in g++.dg/eh/simd-1.C & g++.dg/eh/simd-2.C

[ghazi at caip dot rutgers dot edu]

>Number:         10383
>Category:       target
>Synopsis:       sparc-sun-solaris2.7 G++ testsuite failures in g++.dg/eh/simd-1.C & g++.dg/eh/simd-2.C
>Confidential:   no
>Severity:       serious
>Priority:       medium
>Responsible:    unassigned
>State:          open
>Class:          ice-on-legal-code
>Submitter-Id:   net
>Arrival-Date:   Sat Apr 12 04:16:00 UTC 2003
>Closed-Date:
>Last-Modified:
>Originator:     Kaveh Ghazi
>Release:        gcc version 3.4 20030411 (experimental)
>Organization:
>Environment:
sparc-sun-solaris2.7

>Description:
As noted here:
http://gcc.gnu.org/ml/gcc-testresults/2003-04/msg00685.html
I'm getting testsuite failures in g++.dg/eh/simd-1.C and g++.dg/eh/simd-2.C with mainline gcc-3.4.

Both involve compiler crashes that look like this:

simd-1.C:11: internal compiler error: in function_arg_slotno, at config/sparc/sparc.c:4526

Running simd-1.C under gdb, it gets into sparc.c:function_arg_slotno and enters a switch on a mode of V2SImode.  This mode isn't handled so it goes into the default abort clause and crashes.

The simd-2.C test does the same, except it enters with a mode of V4SImode.


>How-To-Repeat:
Target sparc-sun-solaris2.7 and compile simd-1.ii with:

cc1plus -fpreprocessed simd-1.ii -quiet -dumpbase simd-1.C -auxbase simd-1 -O -version -fmessage-length=0 -o simd-1.s

or compile simd-2.ii with:

cc1plus -fpreprocessed simd-2.ii -quiet -dumpbase simd-2.C -auxbase simd-2 -O -version -fmessage-length=0 -o simd-2.s


>Fix:

>Release-Note:
>Audit-Trail:
>Unformatted:
----gnatsweb-attachment----
Content-Type: application/octet-stream; name="simd-2.ii"
Content-Transfer-Encoding: base64
Content-Disposition: attachment; filename="simd-2.ii"