[AArch64] Fix move_lo_quad_<mode> move
Sofiane Naci
sofiane.naci@arm.com
Tue Jun 11 13:45:00 GMT 2013
Hi,
This patch fixes a bug in the move_lo_quad_<mode> pattern.
The pattern, shown below, issues a scalar MOV instruction for vector modes:
(define_insn "move_lo_quad_<mode>"
[(set (match_operand:VQ 0 "register_operand" "=w,w,w")
(vec_concat:VQ
(match_operand:<VHALF> 1 "register_operand" "w,r,r")
(vec_duplicate:<VHALF> (const_int 0))))]
"TARGET_SIMD"
"@
mov\\t%d0, %d1
fmov\t%d0, %1
...
This is fixed by using DUP for the first alternative instead.
This passes the full regression test suite in aarch64-elf.
OK for trunk?
-----
Thanks
Sofiane
-------------- next part --------------
A non-text attachment was scrubbed...
Name: aarch64-mov-lo-quad-fix.patch
Type: application/octet-stream
Size: 938 bytes
Desc: not available
URL: <http://gcc.gnu.org/pipermail/gcc-patches/attachments/20130611/81b4547c/attachment.obj>
More information about the Gcc-patches
mailing list