SUBREG_PROMOTED_VAR_P handling in expand_direct_optab_fn

This is needed by the later SVE LAST reductions, where an 8-bit
or 16-bit result is zero- rather than sign-extended to 32 bits.
I think it could occur in other situations too.

2017-09-19  Richard Sandiford  <richard.sandiford@linaro.org>
	    Alan Hayward  <alan.hayward@arm.com>
	    David Sherwood  <david.sherwood@arm.com>

gcc/
	* internal-fn.c (expand_direct_optab_fn): Don't assign directly
	to a SUBREG_PROMOTED_VAR.

Co-Authored-By: Alan Hayward <alan.hayward@arm.com>
Co-Authored-By: David Sherwood <david.sherwood@arm.com>

From-SVN: r253992
This commit is contained in:
Richard Sandiford 2017-10-22 21:41:57 +00:00 committed by Richard Sandiford
parent bb06a2d855
commit b7753f75ea
2 changed files with 16 additions and 1 deletions

View File

@ -1,3 +1,10 @@
2017-10-22 Richard Sandiford <richard.sandiford@linaro.org>
Alan Hayward <alan.hayward@arm.com>
David Sherwood <david.sherwood@arm.com>
* internal-fn.c (expand_direct_optab_fn): Don't assign directly
to a SUBREG_PROMOTED_VAR.
2017-10-22 Richard Sandiford <richard.sandiford@linaro.org>
Alan Hayward <alan.hayward@arm.com>
David Sherwood <david.sherwood@arm.com>

View File

@ -2606,7 +2606,15 @@ expand_direct_optab_fn (internal_fn fn, gcall *stmt, direct_optab optab,
tree lhs = gimple_call_lhs (stmt);
tree lhs_type = TREE_TYPE (lhs);
rtx lhs_rtx = expand_expr (lhs, NULL_RTX, VOIDmode, EXPAND_WRITE);
create_output_operand (&ops[0], lhs_rtx, insn_data[icode].operand[0].mode);
/* Do not assign directly to a promoted subreg, since there is no
guarantee that the instruction will leave the upper bits of the
register in the state required by SUBREG_PROMOTED_SIGN. */
rtx dest = lhs_rtx;
if (GET_CODE (dest) == SUBREG && SUBREG_PROMOTED_VAR_P (dest))
dest = NULL_RTX;
create_output_operand (&ops[0], dest, insn_data[icode].operand[0].mode);
for (unsigned int i = 0; i < nargs; ++i)
{