* [PATCH 0/3] gas/x86: operand handling (mainly when swapping)
@ 2022-07-05 6:51 Jan Beulich
2022-07-05 6:53 ` [PATCH 1/3] x86: fix 3-operand insn reverse-matching Jan Beulich
` (3 more replies)
0 siblings, 4 replies; 5+ messages in thread
From: Jan Beulich @ 2022-07-05 6:51 UTC (permalink / raw)
To: Binutils
While preparing the last patch I did notice the bug fixed here first.
The middle patch addresses a minor inefficiency I've been wondering
about for a long time, somewhat related to the other work here.
1: fix 3-operand insn reverse-matching
2: fold two switch() statements in match_template()
3: make D attribute usable for XOP and FMA4 insns
Jan
^ permalink raw reply [flat|nested] 5+ messages in thread
* [PATCH 1/3] x86: fix 3-operand insn reverse-matching
2022-07-05 6:51 [PATCH 0/3] gas/x86: operand handling (mainly when swapping) Jan Beulich
@ 2022-07-05 6:53 ` Jan Beulich
2022-07-05 6:53 ` [PATCH 2/3] x86: fold two switch() statements in match_template() Jan Beulich
` (2 subsequent siblings)
3 siblings, 0 replies; 5+ messages in thread
From: Jan Beulich @ 2022-07-05 6:53 UTC (permalink / raw)
To: Binutils
The middle operand would have gone entirely unchecked, allowing e.g.
vmovss %xmm0, %esp, %xmm2
to assemble successfully, or e.g.
vmovss %xmm0, $4, %xmm2
causing an internal error. Alongside dealing with this also drop a
related comment, which hasn't been applicable anymore since the
introduction of 3-operand patterns with D set (and which perhaps never
had been logical to be there, as reverse-matched insns don't make it
there in the first place).
--- a/gas/config/tc-i386.c
+++ b/gas/config/tc-i386.c
@@ -6724,8 +6724,12 @@ match_template (char mnem_suffix)
/* Try reversing direction of operands. */
overlap0 = operand_type_and (i.types[0], operand_types[i.operands - 1]);
overlap1 = operand_type_and (i.types[i.operands - 1], operand_types[0]);
+ overlap2 = operand_type_and (i.types[1], operand_types[1]);
+ gas_assert (t->operands != 3 || !check_register);
if (!operand_type_match (overlap0, i.types[0])
|| !operand_type_match (overlap1, i.types[i.operands - 1])
+ || (t->operands == 3
+ && !operand_type_match (overlap2, i.types[1]))
|| (check_register
&& !operand_type_register_match (i.types[0],
operand_types[i.operands - 1],
@@ -6797,8 +6801,6 @@ match_template (char mnem_suffix)
continue;
/* Fall through. */
case 3:
- /* Here we make use of the fact that there are no
- reverse match 3 operand instructions. */
if (!operand_type_match (overlap2, i.types[2])
|| ((check_register & 5) == 5
&& !operand_type_register_match (i.types[0],
--- a/gas/testsuite/gas/i386/inval-avx.l
+++ b/gas/testsuite/gas/i386/inval-avx.l
@@ -2,20 +2,13 @@
.*:4: Error: .*
.*:5: Error: .*
.*:6: Error: .*
-.*:9: Error:.* ambiguous .* `vcvtpd2dq'
-.*:10: Error:.* ambiguous .* `vcvtpd2ps'
-.*:11: Error:.* ambiguous .* `vcvttpd2dq'
+.*:8: Error: .*
+.*:9: Error: .*
+.*:10: Error: .*
+.*:11: Error: .*
+.*:12: Error: .*
+.*:15: Error:.* ambiguous .* `vcvtpd2dq'
+.*:16: Error:.* ambiguous .* `vcvtpd2ps'
+.*:17: Error:.* ambiguous .* `vcvttpd2dq'
GAS LISTING .*
-
-
-[ ]*1[ ]+\# Check illegal AVX instructions
-[ ]*2[ ]+\.text
-[ ]*3[ ]+_start:
-[ ]*4[ ]+vcvtpd2dq \(%ecx\),%xmm2
-[ ]*5[ ]+vcvtpd2ps \(%ecx\),%xmm2
-[ ]*6[ ]+vcvttpd2dq \(%ecx\),%xmm2
-[ ]*7[ ]+
-[ ]*8[ ]+\.intel_syntax noprefix
-[ ]*9[ ]+vcvtpd2dq xmm2,\[ecx\]
-[ ]*10[ ]+vcvtpd2ps xmm2,\[ecx\]
-[ ]*11[ ]+vcvttpd2dq xmm2,\[ecx\]
+#pass
--- a/gas/testsuite/gas/i386/inval-avx.s
+++ b/gas/testsuite/gas/i386/inval-avx.s
@@ -5,6 +5,12 @@ _start:
vcvtpd2ps (%ecx),%xmm2
vcvttpd2dq (%ecx),%xmm2
+ vmovss %xmm0, (%esp), %xmm2
+ vmovss %xmm0, $4, %xmm2
+ vmovss %xmm0, %cr0, %xmm2
+ vmovss %xmm0, %ymm4, %xmm2
+ vmovss %xmm0, %mm4, %xmm2
+
.intel_syntax noprefix
vcvtpd2dq xmm2,[ecx]
vcvtpd2ps xmm2,[ecx]
^ permalink raw reply [flat|nested] 5+ messages in thread
* [PATCH 2/3] x86: fold two switch() statements in match_template()
2022-07-05 6:51 [PATCH 0/3] gas/x86: operand handling (mainly when swapping) Jan Beulich
2022-07-05 6:53 ` [PATCH 1/3] x86: fix 3-operand insn reverse-matching Jan Beulich
@ 2022-07-05 6:53 ` Jan Beulich
2022-07-05 6:54 ` [PATCH 3/3] x86: make D attribute usable for XOP and FMA4 insns Jan Beulich
2022-07-05 15:40 ` [PATCH 0/3] gas/x86: operand handling (mainly when swapping) H.J. Lu
3 siblings, 0 replies; 5+ messages in thread
From: Jan Beulich @ 2022-07-05 6:53 UTC (permalink / raw)
To: Binutils
I don't see why two of them were introduced (very long ago) using
similar fall-through logic.
--- a/gas/config/tc-i386.c
+++ b/gas/config/tc-i386.c
@@ -6763,22 +6763,7 @@ match_template (char mnem_suffix)
switch (t->operands)
{
case 5:
- overlap4 = operand_type_and (i.types[4],
- operand_types[4]);
- /* Fall through. */
- case 4:
- overlap3 = operand_type_and (i.types[3],
- operand_types[3]);
- /* Fall through. */
- case 3:
- overlap2 = operand_type_and (i.types[2],
- operand_types[2]);
- break;
- }
-
- switch (t->operands)
- {
- case 5:
+ overlap4 = operand_type_and (i.types[4], operand_types[4]);
if (!operand_type_match (overlap4, i.types[4])
|| !operand_type_register_match (i.types[3],
operand_types[3],
@@ -6787,6 +6772,7 @@ match_template (char mnem_suffix)
continue;
/* Fall through. */
case 4:
+ overlap3 = operand_type_and (i.types[3], operand_types[3]);
if (!operand_type_match (overlap3, i.types[3])
|| ((check_register & 0xa) == 0xa
&& !operand_type_register_match (i.types[1],
@@ -6801,6 +6787,7 @@ match_template (char mnem_suffix)
continue;
/* Fall through. */
case 3:
+ overlap2 = operand_type_and (i.types[2], operand_types[2]);
if (!operand_type_match (overlap2, i.types[2])
|| ((check_register & 5) == 5
&& !operand_type_register_match (i.types[0],
^ permalink raw reply [flat|nested] 5+ messages in thread
* [PATCH 3/3] x86: make D attribute usable for XOP and FMA4 insns
2022-07-05 6:51 [PATCH 0/3] gas/x86: operand handling (mainly when swapping) Jan Beulich
2022-07-05 6:53 ` [PATCH 1/3] x86: fix 3-operand insn reverse-matching Jan Beulich
2022-07-05 6:53 ` [PATCH 2/3] x86: fold two switch() statements in match_template() Jan Beulich
@ 2022-07-05 6:54 ` Jan Beulich
2022-07-05 15:40 ` [PATCH 0/3] gas/x86: operand handling (mainly when swapping) H.J. Lu
3 siblings, 0 replies; 5+ messages in thread
From: Jan Beulich @ 2022-07-05 6:54 UTC (permalink / raw)
To: Binutils
This once again allows to reduce redundancy in (and size of) the opcode
table.
Don't go as far as also making D work on the two 5-operand XOP insns:
This would significantly complicate the code, as there the first
(immediate) operand would need special treatment in several places.
Note that the .s suffix isn't being enabled to have any effect, for
being deprecated. Whereas neither {load} nor {store} pseudo prefixes
make sense here, as the respective operands are inputs (loads) only
anyway, regardless of order. Hence there is (as before) no way for the
programmer to request the alternative encoding to be used for register-
only insns.
Note further that it is always the first original template which is
retained (and altered), to make sure the same encoding as before is
used for register-only insns. This has the slightly odd (but pre-
existing) effect of XOP register-only insns having XOP.W clear, but FMA4
ones having VEX.W set.
--- a/gas/config/tc-i386.c
+++ b/gas/config/tc-i386.c
@@ -2087,12 +2087,18 @@ operand_size_match (const insn_template
}
/* Check reverse. */
- gas_assert (i.operands >= 2 && i.operands <= 3);
+ gas_assert ((i.operands >= 2 && i.operands <= 3)
+ || t->opcode_modifier.vexsources);
for (j = 0; j < i.operands; j++)
{
unsigned int given = i.operands - j - 1;
+ /* For 4- and 5-operand insns VEX.W controls just the first two
+ register operands. */
+ if (t->opcode_modifier.vexsources)
+ given = j < 2 ? 1 - j : j;
+
if (t->operand_types[j].bitfield.class == Reg
&& !match_operand_size (t, j, given))
goto mismatch;
@@ -6722,18 +6728,19 @@ match_template (char mnem_suffix)
if (!(size_match & MATCH_REVERSE))
continue;
/* Try reversing direction of operands. */
- overlap0 = operand_type_and (i.types[0], operand_types[i.operands - 1]);
- overlap1 = operand_type_and (i.types[i.operands - 1], operand_types[0]);
+ j = t->opcode_modifier.vexsources ? 1 : i.operands - 1;
+ overlap0 = operand_type_and (i.types[0], operand_types[j]);
+ overlap1 = operand_type_and (i.types[j], operand_types[0]);
overlap2 = operand_type_and (i.types[1], operand_types[1]);
gas_assert (t->operands != 3 || !check_register);
if (!operand_type_match (overlap0, i.types[0])
- || !operand_type_match (overlap1, i.types[i.operands - 1])
+ || !operand_type_match (overlap1, i.types[j])
|| (t->operands == 3
&& !operand_type_match (overlap2, i.types[1]))
|| (check_register
&& !operand_type_register_match (i.types[0],
- operand_types[i.operands - 1],
- i.types[i.operands - 1],
+ operand_types[j],
+ i.types[j],
operand_types[0])))
{
/* Does not match either direction. */
@@ -6745,6 +6752,11 @@ match_template (char mnem_suffix)
found_reverse_match = 0;
else if (operand_types[0].bitfield.tbyte)
found_reverse_match = Opcode_FloatD;
+ else if (t->opcode_modifier.vexsources)
+ {
+ found_reverse_match = Opcode_VexW;
+ goto check_operands_345;
+ }
else if (operand_types[0].bitfield.xmmword
|| operand_types[i.operands - 1].bitfield.xmmword
|| operand_types[0].bitfield.class == RegMMX
@@ -6760,6 +6772,7 @@ match_template (char mnem_suffix)
else
{
/* Found a forward 2 operand match here. */
+ check_operands_345:
switch (t->operands)
{
case 5:
@@ -6928,8 +6941,12 @@ match_template (char mnem_suffix)
i.tm.operand_types[addr_prefix_disp]
= operand_types[addr_prefix_disp];
- if (found_reverse_match)
+ switch (found_reverse_match)
{
+ case 0:
+ break;
+
+ default:
/* If we found a reverse match we must alter the opcode direction
bit and clear/flip the regmem modifier one. found_reverse_match
holds bits to change (different for int & float insns). */
@@ -6946,6 +6963,17 @@ match_template (char mnem_suffix)
= i.tm.opcode_modifier.modrm && i.tm.opcode_modifier.d
&& i.tm.operands > 2U - i.tm.opcode_modifier.sse2avx
&& !i.tm.opcode_modifier.regmem;
+ break;
+
+ case Opcode_VexW:
+ /* Only the first two register operands need reversing, alongside
+ flipping VEX.W. */
+ i.tm.opcode_modifier.vexw ^= VEXW0 ^ VEXW1;
+
+ j = i.tm.operand_types[0].bitfield.imm8;
+ i.tm.operand_types[j] = operand_types[j + 1];
+ i.tm.operand_types[j + 1] = operand_types[j];
+ break;
}
return t;
--- a/opcodes/i386-opc.h
+++ b/opcodes/i386-opc.h
@@ -925,6 +925,9 @@ typedef struct insn_template
#define Opcode_FloatD 0x400 /* Direction bit for float insns. */
#define Opcode_SIMD_FloatD 0x1 /* Direction bit for SIMD fp insns. */
#define Opcode_SIMD_IntD 0x10 /* Direction bit for SIMD int insns. */
+/* The next value is arbitrary, as long as it's non-zero and distinct
+ from all other values above. */
+#define Opcode_VexW 0xf /* Operand order controlled by VEX.W. */
/* (Fake) base opcode value for pseudo prefixes. */
#define PSEUDO_PREFIX 0
--- a/opcodes/i386-opc.tbl
+++ b/opcodes/i386-opc.tbl
@@ -1928,46 +1928,26 @@ shrx, 0xf2f7, None, CpuBMI2, Modrm|Check
// FMA4 instructions
-vfmaddpd, 0x6669, None, CpuFMA4, Modrm|Vex|Space0F3A|VexVVVV=1|VexW=2|VexSources=2|CheckRegSize|No_bSuf|No_wSuf|No_lSuf|No_sSuf|No_qSuf|No_ldSuf, {Unspecified|BaseIndex|RegXMM|RegYMM, RegXMM|RegYMM, RegXMM|RegYMM, RegXMM|RegYMM }
-vfmaddpd, 0x6669, None, CpuFMA4, Modrm|Vex|Space0F3A|VexVVVV=1|VexW=1|VexSources=2|CheckRegSize|No_bSuf|No_wSuf|No_lSuf|No_sSuf|No_qSuf|No_ldSuf, {RegXMM|RegYMM, Unspecified|BaseIndex|RegXMM|RegYMM, RegXMM|RegYMM, RegXMM|RegYMM }
-vfmaddps, 0x6668, None, CpuFMA4, Modrm|Vex|Space0F3A|VexVVVV=1|VexW=2|VexSources=2|CheckRegSize|No_bSuf|No_wSuf|No_lSuf|No_sSuf|No_qSuf|No_ldSuf, {Unspecified|BaseIndex|RegXMM|RegYMM, RegXMM|RegYMM, RegXMM|RegYMM, RegXMM|RegYMM }
-vfmaddps, 0x6668, None, CpuFMA4, Modrm|Vex|Space0F3A|VexVVVV=1|VexW=1|VexSources=2|CheckRegSize|No_bSuf|No_wSuf|No_lSuf|No_sSuf|No_qSuf|No_ldSuf, {RegXMM|RegYMM, Unspecified|BaseIndex|RegXMM|RegYMM, RegXMM|RegYMM, RegXMM|RegYMM }
-vfmaddsd, 0x666b, None, CpuFMA4, Modrm|VexLIG|Space0F3A|VexVVVV|VexW1|VexSources=2|No_bSuf|No_wSuf|No_lSuf|No_sSuf|No_qSuf|No_ldSuf, { Qword|Unspecified|BaseIndex|RegXMM, RegXMM, RegXMM, RegXMM }
-vfmaddsd, 0x666b, None, CpuFMA4, Modrm|VexLIG|Space0F3A|VexVVVV|VexW0|VexSources=2|No_bSuf|No_wSuf|No_lSuf|No_sSuf|No_qSuf|No_ldSuf, { RegXMM, Qword|Unspecified|BaseIndex|RegXMM, RegXMM, RegXMM }
-vfmaddss, 0x666a, None, CpuFMA4, Modrm|VexLIG|Space0F3A|VexVVVV|VexW1|VexSources=2|No_bSuf|No_wSuf|No_lSuf|No_sSuf|No_qSuf|No_ldSuf, { Dword|Unspecified|BaseIndex|RegXMM, RegXMM, RegXMM, RegXMM }
-vfmaddss, 0x666a, None, CpuFMA4, Modrm|VexLIG|Space0F3A|VexVVVV|VexW0|VexSources=2|No_bSuf|No_wSuf|No_lSuf|No_sSuf|No_qSuf|No_ldSuf, { RegXMM, Dword|Unspecified|BaseIndex|RegXMM, RegXMM, RegXMM }
-vfmaddsubpd, 0x665d, None, CpuFMA4, Modrm|Vex|Space0F3A|VexVVVV=1|VexW=2|VexSources=2|CheckRegSize|No_bSuf|No_wSuf|No_lSuf|No_sSuf|No_qSuf|No_ldSuf, {Unspecified|BaseIndex|RegXMM|RegYMM, RegXMM|RegYMM, RegXMM|RegYMM, RegXMM|RegYMM }
-vfmaddsubpd, 0x665d, None, CpuFMA4, Modrm|Vex|Space0F3A|VexVVVV=1|VexW=1|VexSources=2|CheckRegSize|No_bSuf|No_wSuf|No_lSuf|No_sSuf|No_qSuf|No_ldSuf, {RegXMM|RegYMM, Unspecified|BaseIndex|RegXMM|RegYMM, RegXMM|RegYMM, RegXMM|RegYMM }
-vfmaddsubps, 0x665c, None, CpuFMA4, Modrm|Vex|Space0F3A|VexVVVV=1|VexW=2|VexSources=2|CheckRegSize|No_bSuf|No_wSuf|No_lSuf|No_sSuf|No_qSuf|No_ldSuf, {Unspecified|BaseIndex|RegXMM|RegYMM, RegXMM|RegYMM, RegXMM|RegYMM, RegXMM|RegYMM }
-vfmaddsubps, 0x665c, None, CpuFMA4, Modrm|Vex|Space0F3A|VexVVVV=1|VexW=1|VexSources=2|CheckRegSize|No_bSuf|No_wSuf|No_lSuf|No_sSuf|No_qSuf|No_ldSuf, {RegXMM|RegYMM, Unspecified|BaseIndex|RegXMM|RegYMM, RegXMM|RegYMM, RegXMM|RegYMM }
-vfmsubaddpd, 0x665f, None, CpuFMA4, Modrm|Vex|Space0F3A|VexVVVV=1|VexW=2|VexSources=2|CheckRegSize|No_bSuf|No_wSuf|No_lSuf|No_sSuf|No_qSuf|No_ldSuf, {Unspecified|BaseIndex|RegXMM|RegYMM, RegXMM|RegYMM, RegXMM|RegYMM, RegXMM|RegYMM }
-vfmsubaddpd, 0x665f, None, CpuFMA4, Modrm|Vex|Space0F3A|VexVVVV=1|VexW=1|VexSources=2|CheckRegSize|No_bSuf|No_wSuf|No_lSuf|No_sSuf|No_qSuf|No_ldSuf, {RegXMM|RegYMM, Unspecified|BaseIndex|RegXMM|RegYMM, RegXMM|RegYMM, RegXMM|RegYMM }
-vfmsubaddps, 0x665e, None, CpuFMA4, Modrm|Vex|Space0F3A|VexVVVV=1|VexW=2|VexSources=2|CheckRegSize|No_bSuf|No_wSuf|No_lSuf|No_sSuf|No_qSuf|No_ldSuf, {Unspecified|BaseIndex|RegXMM|RegYMM, RegXMM|RegYMM, RegXMM|RegYMM, RegXMM|RegYMM }
-vfmsubaddps, 0x665e, None, CpuFMA4, Modrm|Vex|Space0F3A|VexVVVV=1|VexW=1|VexSources=2|CheckRegSize|No_bSuf|No_wSuf|No_lSuf|No_sSuf|No_qSuf|No_ldSuf, {RegXMM|RegYMM, Unspecified|BaseIndex|RegXMM|RegYMM, RegXMM|RegYMM, RegXMM|RegYMM }
-vfmsubpd, 0x666d, None, CpuFMA4, Modrm|Vex|Space0F3A|VexVVVV=1|VexW=2|VexSources=2|CheckRegSize|No_bSuf|No_wSuf|No_lSuf|No_sSuf|No_qSuf|No_ldSuf, {Unspecified|BaseIndex|RegXMM|RegYMM, RegXMM|RegYMM, RegXMM|RegYMM, RegXMM|RegYMM }
-vfmsubpd, 0x666d, None, CpuFMA4, Modrm|Vex|Space0F3A|VexVVVV=1|VexW=1|VexSources=2|CheckRegSize|No_bSuf|No_wSuf|No_lSuf|No_sSuf|No_qSuf|No_ldSuf, {RegXMM|RegYMM, Unspecified|BaseIndex|RegXMM|RegYMM, RegXMM|RegYMM, RegXMM|RegYMM }
-vfmsubps, 0x666c, None, CpuFMA4, Modrm|Vex|Space0F3A|VexVVVV=1|VexW=2|VexSources=2|CheckRegSize|No_bSuf|No_wSuf|No_lSuf|No_sSuf|No_qSuf|No_ldSuf, {Unspecified|BaseIndex|RegXMM|RegYMM, RegXMM|RegYMM, RegXMM|RegYMM, RegXMM|RegYMM }
-vfmsubps, 0x666c, None, CpuFMA4, Modrm|Vex|Space0F3A|VexVVVV=1|VexW=1|VexSources=2|CheckRegSize|No_bSuf|No_wSuf|No_lSuf|No_sSuf|No_qSuf|No_ldSuf, {RegXMM|RegYMM, Unspecified|BaseIndex|RegXMM|RegYMM, RegXMM|RegYMM, RegXMM|RegYMM }
-vfmsubsd, 0x666f, None, CpuFMA4, Modrm|VexLIG|Space0F3A|VexVVVV|VexW1|VexSources=2|No_bSuf|No_wSuf|No_lSuf|No_sSuf|No_qSuf|No_ldSuf, { Qword|Unspecified|BaseIndex|RegXMM, RegXMM, RegXMM, RegXMM }
-vfmsubsd, 0x666f, None, CpuFMA4, Modrm|VexLIG|Space0F3A|VexVVVV|VexW0|VexSources=2|No_bSuf|No_wSuf|No_lSuf|No_sSuf|No_qSuf|No_ldSuf, { RegXMM, Qword|Unspecified|BaseIndex|RegXMM, RegXMM, RegXMM }
-vfmsubss, 0x666e, None, CpuFMA4, Modrm|VexLIG|Space0F3A|VexVVVV|VexW1|VexSources=2|No_bSuf|No_wSuf|No_lSuf|No_sSuf|No_qSuf|No_ldSuf, { Dword|Unspecified|BaseIndex|RegXMM, RegXMM, RegXMM, RegXMM }
-vfmsubss, 0x666e, None, CpuFMA4, Modrm|VexLIG|Space0F3A|VexVVVV|VexW0|VexSources=2|No_bSuf|No_wSuf|No_lSuf|No_sSuf|No_qSuf|No_ldSuf, { RegXMM, Dword|Unspecified|BaseIndex|RegXMM, RegXMM, RegXMM }
-vfnmaddpd, 0x6679, None, CpuFMA4, Modrm|Vex|Space0F3A|VexVVVV=1|VexW=2|VexSources=2|CheckRegSize|No_bSuf|No_wSuf|No_lSuf|No_sSuf|No_qSuf|No_ldSuf, {Unspecified|BaseIndex|RegXMM|RegYMM, RegXMM|RegYMM, RegXMM|RegYMM, RegXMM|RegYMM }
-vfnmaddpd, 0x6679, None, CpuFMA4, Modrm|Vex|Space0F3A|VexVVVV=1|VexW=1|VexSources=2|CheckRegSize|No_bSuf|No_wSuf|No_lSuf|No_sSuf|No_qSuf|No_ldSuf, {RegXMM|RegYMM, Unspecified|BaseIndex|RegXMM|RegYMM, RegXMM|RegYMM, RegXMM|RegYMM }
-vfnmaddps, 0x6678, None, CpuFMA4, Modrm|Vex|Space0F3A|VexVVVV=1|VexW=2|VexSources=2|CheckRegSize|No_bSuf|No_wSuf|No_lSuf|No_sSuf|No_qSuf|No_ldSuf, {Unspecified|BaseIndex|RegXMM|RegYMM, RegXMM|RegYMM, RegXMM|RegYMM, RegXMM|RegYMM }
-vfnmaddps, 0x6678, None, CpuFMA4, Modrm|Vex|Space0F3A|VexVVVV=1|VexW=1|VexSources=2|CheckRegSize|No_bSuf|No_wSuf|No_lSuf|No_sSuf|No_qSuf|No_ldSuf, {RegXMM|RegYMM, Unspecified|BaseIndex|RegXMM|RegYMM, RegXMM|RegYMM, RegXMM|RegYMM }
-vfnmaddsd, 0x667b, None, CpuFMA4, Modrm|VexLIG|Space0F3A|VexVVVV|VexW1|VexSources=2|No_bSuf|No_wSuf|No_lSuf|No_sSuf|No_qSuf|No_ldSuf, { Qword|Unspecified|BaseIndex|RegXMM, RegXMM, RegXMM, RegXMM }
-vfnmaddsd, 0x667b, None, CpuFMA4, Modrm|VexLIG|Space0F3A|VexVVVV|VexW0|VexSources=2|No_bSuf|No_wSuf|No_lSuf|No_sSuf|No_qSuf|No_ldSuf, { RegXMM, Qword|Unspecified|BaseIndex|RegXMM, RegXMM, RegXMM }
-vfnmaddss, 0x667a, None, CpuFMA4, Modrm|VexLIG|Space0F3A|VexVVVV|VexW1|VexSources=2|No_bSuf|No_wSuf|No_lSuf|No_sSuf|No_qSuf|No_ldSuf, { Dword|Unspecified|BaseIndex|RegXMM, RegXMM, RegXMM, RegXMM }
-vfnmaddss, 0x667a, None, CpuFMA4, Modrm|VexLIG|Space0F3A|VexVVVV|VexW0|VexSources=2|No_bSuf|No_wSuf|No_lSuf|No_sSuf|No_qSuf|No_ldSuf, { RegXMM, Dword|Unspecified|BaseIndex|RegXMM, RegXMM, RegXMM }
-vfnmsubpd, 0x667d, None, CpuFMA4, Modrm|Vex|Space0F3A|VexVVVV=1|VexW=2|VexSources=2|CheckRegSize|No_bSuf|No_wSuf|No_lSuf|No_sSuf|No_qSuf|No_ldSuf, {Unspecified|BaseIndex|RegXMM|RegYMM, RegXMM|RegYMM, RegXMM|RegYMM, RegXMM|RegYMM }
-vfnmsubpd, 0x667d, None, CpuFMA4, Modrm|Vex|Space0F3A|VexVVVV=1|VexW=1|VexSources=2|CheckRegSize|No_bSuf|No_wSuf|No_lSuf|No_sSuf|No_qSuf|No_ldSuf, {RegXMM|RegYMM, Unspecified|BaseIndex|RegXMM|RegYMM, RegXMM|RegYMM, RegXMM|RegYMM }
-vfnmsubps, 0x667c, None, CpuFMA4, Modrm|Vex|Space0F3A|VexVVVV=1|VexW=2|VexSources=2|CheckRegSize|No_bSuf|No_wSuf|No_lSuf|No_sSuf|No_qSuf|No_ldSuf, {Unspecified|BaseIndex|RegXMM|RegYMM, RegXMM|RegYMM, RegXMM|RegYMM, RegXMM|RegYMM }
-vfnmsubps, 0x667c, None, CpuFMA4, Modrm|Vex|Space0F3A|VexVVVV=1|VexW=1|VexSources=2|CheckRegSize|No_bSuf|No_wSuf|No_lSuf|No_sSuf|No_qSuf|No_ldSuf, {RegXMM|RegYMM, Unspecified|BaseIndex|RegXMM|RegYMM, RegXMM|RegYMM, RegXMM|RegYMM }
-vfnmsubsd, 0x667f, None, CpuFMA4, Modrm|VexLIG|Space0F3A|VexVVVV|VexW1|VexSources=2|No_bSuf|No_wSuf|No_lSuf|No_sSuf|No_qSuf|No_ldSuf, { Qword|Unspecified|BaseIndex|RegXMM, RegXMM, RegXMM, RegXMM }
-vfnmsubsd, 0x667f, None, CpuFMA4, Modrm|VexLIG|Space0F3A|VexVVVV|VexW0|VexSources=2|No_bSuf|No_wSuf|No_lSuf|No_sSuf|No_qSuf|No_ldSuf, { RegXMM, Qword|Unspecified|BaseIndex|RegXMM, RegXMM, RegXMM }
-vfnmsubss, 0x667e, None, CpuFMA4, Modrm|VexLIG|Space0F3A|VexVVVV|VexW1|VexSources=2|No_bSuf|No_wSuf|No_lSuf|No_sSuf|No_qSuf|No_ldSuf, { Dword|Unspecified|BaseIndex|RegXMM, RegXMM, RegXMM, RegXMM }
-vfnmsubss, 0x667e, None, CpuFMA4, Modrm|VexLIG|Space0F3A|VexVVVV|VexW0|VexSources=2|No_bSuf|No_wSuf|No_lSuf|No_sSuf|No_qSuf|No_ldSuf, { RegXMM, Dword|Unspecified|BaseIndex|RegXMM, RegXMM, RegXMM }
+vfmaddpd, 0x6669, None, CpuFMA4, D|Modrm|Vex|Space0F3A|VexVVVV|VexW1|VexSources=2|CheckRegSize|No_bSuf|No_wSuf|No_lSuf|No_sSuf|No_qSuf|No_ldSuf, {Unspecified|BaseIndex|RegXMM|RegYMM, RegXMM|RegYMM, RegXMM|RegYMM, RegXMM|RegYMM }
+vfmaddps, 0x6668, None, CpuFMA4, D|Modrm|Vex|Space0F3A|VexVVVV|VexW1|VexSources=2|CheckRegSize|No_bSuf|No_wSuf|No_lSuf|No_sSuf|No_qSuf|No_ldSuf, {Unspecified|BaseIndex|RegXMM|RegYMM, RegXMM|RegYMM, RegXMM|RegYMM, RegXMM|RegYMM }
+vfmaddsd, 0x666b, None, CpuFMA4, D|Modrm|VexLIG|Space0F3A|VexVVVV|VexW1|VexSources=2|No_bSuf|No_wSuf|No_lSuf|No_sSuf|No_qSuf|No_ldSuf, { Qword|Unspecified|BaseIndex|RegXMM, RegXMM, RegXMM, RegXMM }
+vfmaddss, 0x666a, None, CpuFMA4, D|Modrm|VexLIG|Space0F3A|VexVVVV|VexW1|VexSources=2|No_bSuf|No_wSuf|No_lSuf|No_sSuf|No_qSuf|No_ldSuf, { Dword|Unspecified|BaseIndex|RegXMM, RegXMM, RegXMM, RegXMM }
+vfmaddsubpd, 0x665d, None, CpuFMA4, D|Modrm|Vex|Space0F3A|VexVVVV|VexW1|VexSources=2|CheckRegSize|No_bSuf|No_wSuf|No_lSuf|No_sSuf|No_qSuf|No_ldSuf, {Unspecified|BaseIndex|RegXMM|RegYMM, RegXMM|RegYMM, RegXMM|RegYMM, RegXMM|RegYMM }
+vfmaddsubps, 0x665c, None, CpuFMA4, D|Modrm|Vex|Space0F3A|VexVVVV|VexW1|VexSources=2|CheckRegSize|No_bSuf|No_wSuf|No_lSuf|No_sSuf|No_qSuf|No_ldSuf, {Unspecified|BaseIndex|RegXMM|RegYMM, RegXMM|RegYMM, RegXMM|RegYMM, RegXMM|RegYMM }
+vfmsubaddpd, 0x665f, None, CpuFMA4, D|Modrm|Vex|Space0F3A|VexVVVV|VexW1|VexSources=2|CheckRegSize|No_bSuf|No_wSuf|No_lSuf|No_sSuf|No_qSuf|No_ldSuf, {Unspecified|BaseIndex|RegXMM|RegYMM, RegXMM|RegYMM, RegXMM|RegYMM, RegXMM|RegYMM }
+vfmsubaddps, 0x665e, None, CpuFMA4, D|Modrm|Vex|Space0F3A|VexVVVV|VexW1|VexSources=2|CheckRegSize|No_bSuf|No_wSuf|No_lSuf|No_sSuf|No_qSuf|No_ldSuf, {Unspecified|BaseIndex|RegXMM|RegYMM, RegXMM|RegYMM, RegXMM|RegYMM, RegXMM|RegYMM }
+vfmsubpd, 0x666d, None, CpuFMA4, D|Modrm|Vex|Space0F3A|VexVVVV|VexW1|VexSources=2|CheckRegSize|No_bSuf|No_wSuf|No_lSuf|No_sSuf|No_qSuf|No_ldSuf, {Unspecified|BaseIndex|RegXMM|RegYMM, RegXMM|RegYMM, RegXMM|RegYMM, RegXMM|RegYMM }
+vfmsubps, 0x666c, None, CpuFMA4, D|Modrm|Vex|Space0F3A|VexVVVV|VexW1|VexSources=2|CheckRegSize|No_bSuf|No_wSuf|No_lSuf|No_sSuf|No_qSuf|No_ldSuf, {Unspecified|BaseIndex|RegXMM|RegYMM, RegXMM|RegYMM, RegXMM|RegYMM, RegXMM|RegYMM }
+vfmsubsd, 0x666f, None, CpuFMA4, D|Modrm|VexLIG|Space0F3A|VexVVVV|VexW1|VexSources=2|No_bSuf|No_wSuf|No_lSuf|No_sSuf|No_qSuf|No_ldSuf, { Qword|Unspecified|BaseIndex|RegXMM, RegXMM, RegXMM, RegXMM }
+vfmsubss, 0x666e, None, CpuFMA4, D|Modrm|VexLIG|Space0F3A|VexVVVV|VexW1|VexSources=2|No_bSuf|No_wSuf|No_lSuf|No_sSuf|No_qSuf|No_ldSuf, { Dword|Unspecified|BaseIndex|RegXMM, RegXMM, RegXMM, RegXMM }
+vfnmaddpd, 0x6679, None, CpuFMA4, D|Modrm|Vex|Space0F3A|VexVVVV|VexW1|VexSources=2|CheckRegSize|No_bSuf|No_wSuf|No_lSuf|No_sSuf|No_qSuf|No_ldSuf, {Unspecified|BaseIndex|RegXMM|RegYMM, RegXMM|RegYMM, RegXMM|RegYMM, RegXMM|RegYMM }
+vfnmaddps, 0x6678, None, CpuFMA4, D|Modrm|Vex|Space0F3A|VexVVVV|VexW1|VexSources=2|CheckRegSize|No_bSuf|No_wSuf|No_lSuf|No_sSuf|No_qSuf|No_ldSuf, {Unspecified|BaseIndex|RegXMM|RegYMM, RegXMM|RegYMM, RegXMM|RegYMM, RegXMM|RegYMM }
+vfnmaddsd, 0x667b, None, CpuFMA4, D|Modrm|VexLIG|Space0F3A|VexVVVV|VexW1|VexSources=2|No_bSuf|No_wSuf|No_lSuf|No_sSuf|No_qSuf|No_ldSuf, { Qword|Unspecified|BaseIndex|RegXMM, RegXMM, RegXMM, RegXMM }
+vfnmaddss, 0x667a, None, CpuFMA4, D|Modrm|VexLIG|Space0F3A|VexVVVV|VexW1|VexSources=2|No_bSuf|No_wSuf|No_lSuf|No_sSuf|No_qSuf|No_ldSuf, { Dword|Unspecified|BaseIndex|RegXMM, RegXMM, RegXMM, RegXMM }
+vfnmsubpd, 0x667d, None, CpuFMA4, D|Modrm|Vex|Space0F3A|VexVVVV|VexW1|VexSources=2|CheckRegSize|No_bSuf|No_wSuf|No_lSuf|No_sSuf|No_qSuf|No_ldSuf, {Unspecified|BaseIndex|RegXMM|RegYMM, RegXMM|RegYMM, RegXMM|RegYMM, RegXMM|RegYMM }
+vfnmsubps, 0x667c, None, CpuFMA4, D|Modrm|Vex|Space0F3A|VexVVVV|VexW1|VexSources=2|CheckRegSize|No_bSuf|No_wSuf|No_lSuf|No_sSuf|No_qSuf|No_ldSuf, {Unspecified|BaseIndex|RegXMM|RegYMM, RegXMM|RegYMM, RegXMM|RegYMM, RegXMM|RegYMM }
+vfnmsubsd, 0x667f, None, CpuFMA4, D|Modrm|VexLIG|Space0F3A|VexVVVV|VexW1|VexSources=2|No_bSuf|No_wSuf|No_lSuf|No_sSuf|No_qSuf|No_ldSuf, { Qword|Unspecified|BaseIndex|RegXMM, RegXMM, RegXMM, RegXMM }
+vfnmsubss, 0x667e, None, CpuFMA4, D|Modrm|VexLIG|Space0F3A|VexVVVV|VexW1|VexSources=2|No_bSuf|No_wSuf|No_lSuf|No_sSuf|No_qSuf|No_ldSuf, { Dword|Unspecified|BaseIndex|RegXMM, RegXMM, RegXMM, RegXMM }
// XOP instructions
@@ -1979,8 +1959,7 @@ vfrczpd, 0x81, None, CpuXOP, Modrm|Space
vfrczps, 0x80, None, CpuXOP, Modrm|SpaceXOP09|VexW=1|CheckRegSize|No_bSuf|No_wSuf|No_lSuf|No_sSuf|No_qSuf|No_ldSuf|Vex, { Unspecified|BaseIndex|RegXMM|RegYMM, RegXMM|RegYMM }
vfrczsd, 0x83, None, CpuXOP, Modrm|SpaceXOP09|VexW0|No_bSuf|No_wSuf|No_lSuf|No_sSuf|No_qSuf|No_ldSuf|Vex, { Qword|RegXMM|Unspecified|BaseIndex, RegXMM }
vfrczss, 0x82, None, CpuXOP, Modrm|SpaceXOP09|VexW0|No_bSuf|No_wSuf|No_lSuf|No_sSuf|No_qSuf|No_ldSuf|Vex, { Dword|RegXMM|Unspecified|BaseIndex, RegXMM }
-vpcmov, 0xa2, None, CpuXOP, Modrm|SpaceXOP08|VexSources=2|VexVVVV=1|VexW=1|CheckRegSize|No_bSuf|No_wSuf|No_lSuf|No_sSuf|No_qSuf|No_ldSuf|Vex, { RegXMM|RegYMM, Unspecified|BaseIndex|RegXMM|RegYMM, RegXMM|RegYMM, RegXMM|RegYMM }
-vpcmov, 0xa2, None, CpuXOP, Modrm|SpaceXOP08|VexSources=2|VexVVVV=1|VexW=2|CheckRegSize|No_bSuf|No_wSuf|No_lSuf|No_sSuf|No_qSuf|No_ldSuf|Vex, { Unspecified|BaseIndex|RegXMM|RegYMM, RegXMM|RegYMM, RegXMM|RegYMM, RegXMM|RegYMM }
+vpcmov, 0xa2, None, CpuXOP, D|Modrm|SpaceXOP08|VexSources=2|VexVVVV|VexW0|CheckRegSize|No_bSuf|No_wSuf|No_lSuf|No_sSuf|No_qSuf|No_ldSuf|Vex, { RegXMM|RegYMM, Unspecified|BaseIndex|RegXMM|RegYMM, RegXMM|RegYMM, RegXMM|RegYMM }
vpcom<xop_sign><xop_elem>, 0xcc | 0x<xop_sign:opc> | <xop_elem:opc>, None, CpuXOP, Modrm|Vex128|SpaceXOP08|VexVVVV|VexW0|No_bSuf|No_wSuf|No_lSuf|No_sSuf|No_qSuf|No_ldSuf, { Imm8, RegXMM|Unspecified|BaseIndex, RegXMM, RegXMM }
vpcom<xop_irel><xop_sign><xop_elem>, 0xcc | 0x<xop_sign:opc> | <xop_elem:opc>, <xop_irel:imm>, CpuXOP, Modrm|Vex128|SpaceXOP08|VexVVVV|VexW0|No_bSuf|No_wSuf|No_lSuf|No_sSuf|No_qSuf|No_ldSuf|ImmExt, { RegXMM|Unspecified|BaseIndex, RegXMM, RegXMM }
vpermil2pd, 0x6649, None, CpuXOP, Modrm|Space0F3A|VexVVVV=1|VexW=1|Vex|VexSources=2|CheckRegSize|No_bSuf|No_wSuf|No_lSuf|No_sSuf|No_qSuf|No_ldSuf, { Imm8, RegXMM|RegYMM, Unspecified|BaseIndex|RegXMM|RegYMM, RegXMM|RegYMM, RegXMM|RegYMM }
@@ -2014,15 +1993,11 @@ vpmacswd, 0x96, None, CpuXOP, Modrm|Spac
vpmacsww, 0x95, None, CpuXOP, Modrm|SpaceXOP08|VexSources=2|VexVVVV=1|VexW=1|No_bSuf|No_wSuf|No_lSuf|No_sSuf|No_qSuf|No_ldSuf|Vex, { RegXMM, RegXMM|Unspecified|BaseIndex, RegXMM, RegXMM }
vpmadcsswd, 0xa6, None, CpuXOP, Modrm|SpaceXOP08|VexSources=2|VexVVVV=1|VexW=1|No_bSuf|No_wSuf|No_lSuf|No_sSuf|No_qSuf|No_ldSuf|Vex, { RegXMM, RegXMM|Unspecified|BaseIndex, RegXMM, RegXMM }
vpmadcswd, 0xb6, None, CpuXOP, Modrm|SpaceXOP08|VexSources=2|VexVVVV=1|VexW=1|No_bSuf|No_wSuf|No_lSuf|No_sSuf|No_qSuf|No_ldSuf|Vex, { RegXMM, RegXMM|Unspecified|BaseIndex, RegXMM, RegXMM }
-vpperm, 0xa3, None, CpuXOP, Modrm|SpaceXOP08|VexSources=2|VexVVVV=1|VexW=1|No_bSuf|No_wSuf|No_lSuf|No_sSuf|No_qSuf|No_ldSuf|Vex, { RegXMM, RegXMM|Unspecified|BaseIndex, RegXMM, RegXMM }
-vpperm, 0xa3, None, CpuXOP, Modrm|SpaceXOP08|VexSources=2|VexVVVV=1|VexW=2|No_bSuf|No_wSuf|No_lSuf|No_sSuf|No_qSuf|No_ldSuf|Vex, { RegXMM|Unspecified|BaseIndex, RegXMM, RegXMM, RegXMM }
-vprot<xop_elem>, 0x90 | <xop_elem:opc>, None, CpuXOP, Modrm|Vex128|SpaceXOP09|VexW0|VexSources=1|No_wSuf|No_lSuf|No_sSuf|No_ldSuf|No_qSuf, { RegXMM, RegXMM|Unspecified|BaseIndex, RegXMM }
-vprot<xop_elem>, 0x90 | <xop_elem:opc>, None, CpuXOP, Modrm|Vex128|SpaceXOP09|VexW1|VexSources=1|No_wSuf|No_lSuf|No_sSuf|No_ldSuf|No_qSuf, { RegXMM|Unspecified|BaseIndex, RegXMM, RegXMM }
+vpperm, 0xa3, None, CpuXOP, D|Modrm|SpaceXOP08|VexSources=2|VexVVVV|VexW0|No_bSuf|No_wSuf|No_lSuf|No_sSuf|No_qSuf|No_ldSuf|Vex, { RegXMM, RegXMM|Unspecified|BaseIndex, RegXMM, RegXMM }
+vprot<xop_elem>, 0x90 | <xop_elem:opc>, None, CpuXOP, D|Modrm|Vex128|SpaceXOP09|VexW0|VexSources=1|No_wSuf|No_lSuf|No_sSuf|No_ldSuf|No_qSuf, { RegXMM, RegXMM|Unspecified|BaseIndex, RegXMM }
vprot<xop_elem>, 0xc0 | <xop_elem:opc>, None, CpuXOP, Modrm|Vex128|SpaceXOP08|VexW0|VexSources=1|No_wSuf|No_lSuf|No_sSuf|No_ldSuf|No_qSuf, { Imm8, RegXMM|Unspecified|BaseIndex, RegXMM }
-vpsha<xop_elem>, 0x98 | <xop_elem:opc>, None, CpuXOP, Modrm|Vex128|SpaceXOP09|VexW0|VexSources=1|No_wSuf|No_lSuf|No_sSuf|No_ldSuf|No_qSuf, { RegXMM, RegXMM|Unspecified|BaseIndex, RegXMM }
-vpsha<xop_elem>, 0x98 | <xop_elem:opc>, None, CpuXOP, Modrm|Vex128|SpaceXOP09|VexW1|VexSources=1|No_wSuf|No_lSuf|No_sSuf|No_ldSuf|No_qSuf, { RegXMM|Unspecified|BaseIndex, RegXMM, RegXMM }
-vpshl<xop_elem>, 0x94 | <xop_elem:opc>, None, CpuXOP, Modrm|Vex128|SpaceXOP09|VexW0|VexSources=1|No_wSuf|No_lSuf|No_sSuf|No_ldSuf|No_qSuf, { RegXMM, RegXMM|Unspecified|BaseIndex, RegXMM }
-vpshl<xop_elem>, 0x94 | <xop_elem:opc>, None, CpuXOP, Modrm|Vex128|SpaceXOP09|VexW1|VexSources=1|No_wSuf|No_lSuf|No_sSuf|No_ldSuf|No_qSuf, { RegXMM|Unspecified|BaseIndex, RegXMM, RegXMM }
+vpsha<xop_elem>, 0x98 | <xop_elem:opc>, None, CpuXOP, D|Modrm|Vex128|SpaceXOP09|VexW0|VexSources=1|No_wSuf|No_lSuf|No_sSuf|No_ldSuf|No_qSuf, { RegXMM, RegXMM|Unspecified|BaseIndex, RegXMM }
+vpshl<xop_elem>, 0x94 | <xop_elem:opc>, None, CpuXOP, D|Modrm|Vex128|SpaceXOP09|VexW0|VexSources=1|No_wSuf|No_lSuf|No_sSuf|No_ldSuf|No_qSuf, { RegXMM, RegXMM|Unspecified|BaseIndex, RegXMM }
// LWP instructions
^ permalink raw reply [flat|nested] 5+ messages in thread
* Re: [PATCH 0/3] gas/x86: operand handling (mainly when swapping)
2022-07-05 6:51 [PATCH 0/3] gas/x86: operand handling (mainly when swapping) Jan Beulich
` (2 preceding siblings ...)
2022-07-05 6:54 ` [PATCH 3/3] x86: make D attribute usable for XOP and FMA4 insns Jan Beulich
@ 2022-07-05 15:40 ` H.J. Lu
3 siblings, 0 replies; 5+ messages in thread
From: H.J. Lu @ 2022-07-05 15:40 UTC (permalink / raw)
To: Jan Beulich; +Cc: Binutils
On Mon, Jul 4, 2022 at 11:51 PM Jan Beulich <jbeulich@suse.com> wrote:
>
> While preparing the last patch I did notice the bug fixed here first.
> The middle patch addresses a minor inefficiency I've been wondering
> about for a long time, somewhat related to the other work here.
>
> 1: fix 3-operand insn reverse-matching
> 2: fold two switch() statements in match_template()
> 3: make D attribute usable for XOP and FMA4 insns
>
> Jan
OK to all.
Thanks.
--
H.J.
^ permalink raw reply [flat|nested] 5+ messages in thread
end of thread, other threads:[~2022-07-05 15:40 UTC | newest]
Thread overview: 5+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2022-07-05 6:51 [PATCH 0/3] gas/x86: operand handling (mainly when swapping) Jan Beulich
2022-07-05 6:53 ` [PATCH 1/3] x86: fix 3-operand insn reverse-matching Jan Beulich
2022-07-05 6:53 ` [PATCH 2/3] x86: fold two switch() statements in match_template() Jan Beulich
2022-07-05 6:54 ` [PATCH 3/3] x86: make D attribute usable for XOP and FMA4 insns Jan Beulich
2022-07-05 15:40 ` [PATCH 0/3] gas/x86: operand handling (mainly when swapping) H.J. Lu
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).