[Bug testsuite/48614] New: gcc.target/i386/pr46084.c should require avx_runtime

[Bug testsuite/48614] New: gcc.target/i386/pr46084.c should require avx_runtime

[zsojka at seznam dot cz]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=48614

           Summary: gcc.target/i386/pr46084.c should require avx_runtime
           Product: gcc
           Version: 4.7.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: testsuite
        AssignedTo: unassigned@gcc.gnu.org
        ReportedBy: zsojka@seznam.cz
              Host: x86_64-pc-linux-gnu
            Target: x86_64-pc-linux-gnu


Currently, that test fails with -ftree-vectorize in RUNTESTFLAGS:
$ gcc gcc.target/i386/pr46084.c -fsplit-stack -O2 -mavx -ftree-vectorize
$ gdb ./a.out
Program received signal SIGILL, Illegal instruction.
0x0000000000400d0e in use_buffer (buf=Unhandled dwarf expression opcode 0xf3
) at gcc.target/i386/pr46084.c:21
21        for (i = 0; i < c; ++i)
(gdb) disassemble
Dump of assembler code for function use_buffer:
...
   0x0000000000400d02 <+98>:    lea    0x1(%rax),%rdx
   0x0000000000400d06 <+102>:   mov    %rax,-0x10(%rsp)
   0x0000000000400d0b <+107>:   add    %rdi,%rcx
=> 0x0000000000400d0e <+110>:   vmovq  -0x10(%rsp),%xmm0
   0x0000000000400d14 <+116>:   vpinsrq $0x1,%rdx,%xmm0,%xmm1
   0x0000000000400d1a <+122>:   vmovdqa 0x10ae(%rip),%xmm13        # 0x401dd0
   0x0000000000400d22 <+130>:   xor    %edx,%edx
   0x0000000000400d24 <+132>:   vmovdqa 0x10b4(%rip),%xmm0        # 0x401de0
   0x0000000000400d2c <+140>:   vmovdqa 0x10bc(%rip),%xmm9        # 0x401df0
   0x0000000000400d34 <+148>:   vmovdqa 0x10c4(%rip),%xmm8        # 0x401e00
   0x0000000000400d3c <+156>:   vmovdqa 0x10cc(%rip),%xmm12        # 0x401e10
   0x0000000000400d44 <+164>:   vmovdqa 0x10d4(%rip),%xmm11        # 0x401e20
   0x0000000000400d4c <+172>:   jmp    0x400d54 <use_buffer+180>

I know I am using custom RUNTESTFLAGS, but if the fix is as easy as adding "{
dg-require-effective-target avx_runtime }", it would be nice to have it fixed.


Index: gcc/testsuite/gcc.target/i386/pr46084.c
===================================================================
--- gcc/testsuite/gcc.target/i386/pr46084.c     (revision 172439)
+++ gcc/testsuite/gcc.target/i386/pr46084.c     (working copy)
@@ -2,6 +2,7 @@
    only run on Unix.  */
 /* { dg-do run { target *-*-linux* *-*-solaris* *-*-darwin* } } */
 /* { dg-require-effective-target avx } */
+/* { dg-require-effective-target avx_runtime } */
 /* { dg-require-effective-target split_stack } */
 /* { dg-options "-fsplit-stack -O2 -mavx" } */

Prevents the failure, but I am not sure if the line above the change isn't now
redundant.

[Bug testsuite/48614] gcc.target/i386/pr46084.c should require avx_runtime

[jakub at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=48614

--- Comment #1 from Jakub Jelinek <jakub at gcc dot gnu.org> 2011-04-15 10:29:13 UTC ---
Author: jakub
Date: Fri Apr 15 10:29:09 2011
New Revision: 172484

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=172484
Log:
    PR target/48614
    * gcc.target/i386/pr46084.c: Require avx_runtime instead of
    just avx.

Modified:
    trunk/gcc/testsuite/ChangeLog
    trunk/gcc/testsuite/gcc.target/i386/pr46084.c

[Bug testsuite/48614] gcc.target/i386/pr46084.c should require avx_runtime

[jakub at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=48614

--- Comment #2 from Jakub Jelinek <jakub at gcc dot gnu.org> 2011-04-15 10:30:01 UTC ---
Author: jakub
Date: Fri Apr 15 10:29:58 2011
New Revision: 172485

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=172485
Log:
    PR target/48614
    * gcc.target/i386/pr46084.c: Require avx_runtime instead of
    just avx.

Modified:
    branches/gcc-4_6-branch/gcc/testsuite/ChangeLog
    branches/gcc-4_6-branch/gcc/testsuite/gcc.target/i386/pr46084.c

[Bug testsuite/48614] gcc.target/i386/pr46084.c should require avx_runtime

[jakub at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=48614

Jakub Jelinek <jakub at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |RESOLVED
                 CC|                            |jakub at gcc dot gnu.org
         Resolution|                            |FIXED

--- Comment #3 from Jakub Jelinek <jakub at gcc dot gnu.org> 2011-04-15 10:35:15 UTC ---
Fixed.