[gcc r12-7190] [nvptx] Handle asm insn in prevent_branch_around_nothing

Tom de Vries vries@gcc.gnu.org
Fri Feb 11 07:46:19 GMT 2022


https://gcc.gnu.org/g:fd64b09217fbe8fa33b559e61564071e8aca71e5

commit r12-7190-gfd64b09217fbe8fa33b559e61564071e8aca71e5
Author: Tom de Vries <tdevries@suse.de>
Date:   Thu Feb 10 11:26:16 2022 +0100

    [nvptx] Handle asm insn in prevent_branch_around_nothing
    
    With GOMP_NVPTX_JIT=-00 and -mptx=3.1, I run into:
    ...
    FAIL: libgomp.oacc-c/../libgomp.oacc-c-c++-common/acc_prof-version-1.c \
      -DACC_DEVICE_TYPE_nvidia=1 -DACC_MEM_SHARED=0 -foffload=nvptx-none  -O2 \
      execution test
    ...
    
    The problem is that we're generating a diverging branch around nothing:
    ...
            {
                    .reg.u32        %x;
                    mov.u32 %x, %tid.x;
                    setp.ne.u32     %r23, %x, 0;
            }
            @%r23   bra     $L2;
    $L2:
    ...
    which the driver JIT has problems with at -O0, so consequently we run into the
    nvptx_uniform_warp_check.
    
    Fix this by handling asm ("") and alike in prevent_branch_around_nothing.
    
    Tested on x86_64 with nvptx accelerator.
    
    gcc/ChangeLog:
    
    2022-02-10  Tom de Vries  <tdevries@suse.de>
    
            PR target/104456
            * config/nvptx/nvptx.cc (prevent_branch_around_nothing): Handle asm
            insn.

Diff:
---
 gcc/config/nvptx/nvptx.cc | 8 ++++++++
 1 file changed, 8 insertions(+)

diff --git a/gcc/config/nvptx/nvptx.cc b/gcc/config/nvptx/nvptx.cc
index 5b26c0f4c7d..afbad5bdde6 100644
--- a/gcc/config/nvptx/nvptx.cc
+++ b/gcc/config/nvptx/nvptx.cc
@@ -5257,6 +5257,14 @@ prevent_branch_around_nothing (void)
 	    case CODE_FOR_nvptx_join:
 	    case CODE_FOR_nop:
 	      continue;
+	    case -1:
+	      /* Handle asm ("") and similar.  */
+	      if (GET_CODE (PATTERN (insn)) == ASM_INPUT
+		  || GET_CODE (PATTERN (insn)) == ASM_OPERANDS
+		  || (GET_CODE (PATTERN (insn)) == PARALLEL
+		      && asm_noperands (PATTERN (insn)) >= 0))
+		continue;
+	      /* FALLTHROUGH.  */
 	    default:
 	      seen_label = NULL;
 	      continue;


More information about the Gcc-cvs mailing list