Commit ed438a93 authored by Jan Beulich's avatar Jan Beulich Committed by Jan Beulich

x86: fold certain AVX512 rotate and shift templates

Their memory forms were bogusly using VexLWP instead of VexNDD. Adjust
VexNDD handling to cope with these, allowing their register and memory
forms to be folded.
parent 454172a9
2018-03-08 Jan Beulich <jbeulich@suse.com>
* config/tc-i386.c (build_modrm_byte): Make VexNDD handling cope
with 3rd (immediate) operand.
2018-03-08 Jan Beulich <jbeulich@suse.com>
* config/tc-i386.c (XMMWORD_MNEM_SUFFIX, YMMWORD_MNEM_SUFFIX,
......
......@@ -7157,9 +7157,10 @@ build_modrm_byte (void)
}
else
{
/* There are only 2 operands. */
gas_assert (op < 2 && i.operands == 2);
vex_reg = 1;
/* There are only 2 non-immediate operands. */
gas_assert (op < i.imm_operands + 2
&& i.operands == i.imm_operands + 2);
vex_reg = i.imm_operands + 1;
}
}
else
......
2018-03-08 Jan Beulich <jbeulich@suse.com>
* i386-opc.tbl (vexpandpd, vexpandps, vmovapd, vmovaps,
vmovdqa32, vmovdqa64, vmovdqu32, vmovdqu64, vmovupd, vmovups,
vpexpandd, vpexpandq): Fold AFX512VF templates.
* i386-tlb.h: Re-generate.
2018-03-08 Jan Beulich <jbeulich@suse.com>
* i386-opc.tbl (vgf2p8affineinvqb, vgf2p8affineqb, vgf2p8mulb):
......
This diff is collapsed.
This diff is collapsed.
Markdown is supported
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment