On Tue, May 17, 2016 at 11:20 AM, Ilya Enkovich <enkovich....@gmail.com> wrote:
> Hi,
>
> This patch fixes an insertion point for instructions generated by
> validize_mem in dimode_scalar_chain::convert_op. Patch also fixes
> scanned instruction pattern to fix test for Darwin.
>
> I bootstrapped and regtested it on x86_64-pc-linus gnu.
> Dominique d'Humieres tested it for Darwin. OK for trunk?
>
> Thanks,
> Ilya
> --
> gcc/
>
> 2016-05-17 Ilya Enkovich <ilya.enkov...@intel.com>
>
> PR target/71114
> * config/i386/i386.c (dimode_scalar_chain::convert_op): Fix
> insertion point for instructions generated by validize_mem.
>
> gcc/testsuite/
>
> 2016-05-17 Dominique d'Humieres <domi...@lps.ens.fr>
> Ilya Enkovich <ilya.enkov...@intel.com>
>
> PR target/71114
> * gcc.target/i386/pr70799-1.c: Fix scan for Darwin.
OK.
Thanks,
Uros.
>
> diff --git a/gcc/config/i386/i386.c b/gcc/config/i386/i386.c
> index 046717b..47d2e7d 100644
> --- a/gcc/config/i386/i386.c
> +++ b/gcc/config/i386/i386.c
> @@ -3688,7 +3688,13 @@ dimode_scalar_chain::convert_op (rtx *op, rtx_insn
> *insn)
> gen_rtvec (2, *op, const0_rtx));
>
> if (!standard_sse_constant_p (vec_cst, V2DImode))
> - vec_cst = validize_mem (force_const_mem (V2DImode, vec_cst));
> + {
> + start_sequence ();
> + vec_cst = validize_mem (force_const_mem (V2DImode, vec_cst));
> + rtx_insn *seq = get_insns ();
> + end_sequence ();
> + emit_insn_before (seq, insn);
> + }
>
> emit_insn_before (gen_move_insn (tmp, vec_cst), insn);
> *op = tmp;
> diff --git a/gcc/testsuite/gcc.target/i386/pr70799-1.c
> b/gcc/testsuite/gcc.target/i386/pr70799-1.c
> index 0abbfb9..f4e42fa 100644
> --- a/gcc/testsuite/gcc.target/i386/pr70799-1.c
> +++ b/gcc/testsuite/gcc.target/i386/pr70799-1.c
> @@ -3,7 +3,7 @@
> /* { dg-options "-O2 -march=slm" } */
> /* { dg-final { scan-assembler "pxor" } } */
> /* { dg-final { scan-assembler "pcmpeqd" } } */
> -/* { dg-final { scan-assembler "movdqa\[ \\t\]+.LC0" } } */
> +/* { dg-final { scan-assembler "movdqa\[ \\t\]+.?LC0" } } */
>
> long long a, b, c;
>
