From: Richard Henderson <richard.henderson@linaro.org>
To: qemu-devel@nongnu.org
Cc: zhiwei_liu@linux.alibaba.com, tangtiancheng.ttc@alibaba-inc.com,
liwei1518@gmail.com, bmeng.cn@gmail.com
Subject: [PATCH 03/12] tcg/i386: Split out tcg_out_vex_modrm_type
Date: Sat, 7 Sep 2024 19:26:23 -0700 [thread overview]
Message-ID: <20240908022632.459477-4-richard.henderson@linaro.org> (raw)
In-Reply-To: <20240908022632.459477-1-richard.henderson@linaro.org>
Helper function to handle setting of VEXL based
on the type of the operation.
Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
---
tcg/i386/tcg-target.c.inc | 38 +++++++++++++++-----------------------
1 file changed, 15 insertions(+), 23 deletions(-)
diff --git a/tcg/i386/tcg-target.c.inc b/tcg/i386/tcg-target.c.inc
index 9a54ef7f8d..af71a397b1 100644
--- a/tcg/i386/tcg-target.c.inc
+++ b/tcg/i386/tcg-target.c.inc
@@ -711,6 +711,15 @@ static void tcg_out_vex_modrm(TCGContext *s, int opc, int r, int v, int rm)
tcg_out8(s, 0xc0 | (LOWREGMASK(r) << 3) | LOWREGMASK(rm));
}
+static void tcg_out_vex_modrm_type(TCGContext *s, int opc,
+ int r, int v, int rm, TCGType type)
+{
+ if (type == TCG_TYPE_V256) {
+ opc |= P_VEXL;
+ }
+ tcg_out_vex_modrm(s, opc, r, v, rm);
+}
+
/* Output an opcode with a full "rm + (index<<shift) + offset" address mode.
We handle either RM and INDEX missing with a negative value. In 64-bit
mode for absolute addresses, ~RM is the size of the immediate operand
@@ -904,8 +913,7 @@ static bool tcg_out_dup_vec(TCGContext *s, TCGType type, unsigned vece,
TCGReg r, TCGReg a)
{
if (have_avx2) {
- int vex_l = (type == TCG_TYPE_V256 ? P_VEXL : 0);
- tcg_out_vex_modrm(s, avx2_dup_insn[vece] + vex_l, r, 0, a);
+ tcg_out_vex_modrm_type(s, avx2_dup_insn[vece], r, 0, a, type);
} else {
switch (vece) {
case MO_8:
@@ -3231,10 +3239,7 @@ static void tcg_out_vec_op(TCGContext *s, TCGOpcode opc,
goto gen_simd;
gen_simd:
tcg_debug_assert(insn != OPC_UD2);
- if (type == TCG_TYPE_V256) {
- insn |= P_VEXL;
- }
- tcg_out_vex_modrm(s, insn, a0, a1, a2);
+ tcg_out_vex_modrm_type(s, insn, a0, a1, a2, type);
break;
case INDEX_op_cmp_vec:
@@ -3250,10 +3255,7 @@ static void tcg_out_vec_op(TCGContext *s, TCGOpcode opc,
case INDEX_op_andc_vec:
insn = OPC_PANDN;
- if (type == TCG_TYPE_V256) {
- insn |= P_VEXL;
- }
- tcg_out_vex_modrm(s, insn, a0, a2, a1);
+ tcg_out_vex_modrm_type(s, insn, a0, a2, a1, type);
break;
case INDEX_op_shli_vec:
@@ -3281,10 +3283,7 @@ static void tcg_out_vec_op(TCGContext *s, TCGOpcode opc,
goto gen_shift;
gen_shift:
tcg_debug_assert(vece != MO_8);
- if (type == TCG_TYPE_V256) {
- insn |= P_VEXL;
- }
- tcg_out_vex_modrm(s, insn, sub, a0, a1);
+ tcg_out_vex_modrm_type(s, insn, sub, a0, a1, type);
tcg_out8(s, a2);
break;
@@ -3361,19 +3360,12 @@ static void tcg_out_vec_op(TCGContext *s, TCGOpcode opc,
gen_simd_imm8:
tcg_debug_assert(insn != OPC_UD2);
- if (type == TCG_TYPE_V256) {
- insn |= P_VEXL;
- }
- tcg_out_vex_modrm(s, insn, a0, a1, a2);
+ tcg_out_vex_modrm_type(s, insn, a0, a1, a2, type);
tcg_out8(s, sub);
break;
case INDEX_op_x86_vpblendvb_vec:
- insn = OPC_VPBLENDVB;
- if (type == TCG_TYPE_V256) {
- insn |= P_VEXL;
- }
- tcg_out_vex_modrm(s, insn, a0, a1, a2);
+ tcg_out_vex_modrm_type(s, OPC_VPBLENDVB, a0, a1, a2, type);
tcg_out8(s, args[3] << 4);
break;
--
2.43.0
next prev parent reply other threads:[~2024-09-08 2:27 UTC|newest]
Thread overview: 20+ messages / expand[flat|nested] mbox.gz Atom feed top
2024-09-08 2:26 [PATCH 00/12] tcg: Improve support for cmpsel_vec Richard Henderson
2024-09-08 2:26 ` [PATCH 01/12] tcg: Fix iteration step in 32-bit gvec operation Richard Henderson
2024-09-10 6:57 ` Philippe Mathieu-Daudé
2024-09-10 14:46 ` Richard Henderson
2024-09-08 2:26 ` [PATCH 02/12] tcg: Export vec_gen_6 Richard Henderson
2024-09-10 6:57 ` Philippe Mathieu-Daudé
2024-09-08 2:26 ` Richard Henderson [this message]
2024-09-10 6:59 ` [PATCH 03/12] tcg/i386: Split out tcg_out_vex_modrm_type Philippe Mathieu-Daudé
2024-09-08 2:26 ` [PATCH 04/12] tcg/i386: Do not expand cmp_vec early Richard Henderson
2024-09-08 2:26 ` [PATCH 05/12] tcg/i386: Do not expand cmpsel_vec early Richard Henderson
2024-09-08 2:26 ` [PATCH 06/12] tcg/optimize: Fold movcond with true and false values identical Richard Henderson
2024-09-10 7:00 ` Philippe Mathieu-Daudé
2024-09-08 2:26 ` [PATCH 07/12] tcg/optimize: Optimize cmp_vec and cmpsel_vec Richard Henderson
2024-09-08 2:26 ` [PATCH 08/12] tcg/optimize: Optimize bitsel_vec Richard Henderson
2024-09-08 2:26 ` [PATCH 09/12] tcg/i386: Optimize cmpsel with constant 0 arguments Richard Henderson
2024-09-08 2:26 ` [PATCH 10/12] tcg/i386: Implement cmp_vec with avx512 insns Richard Henderson
2024-09-08 2:26 ` [PATCH 11/12] tcg/i386: Add predicate parameters to tcg_out_evex_opc Richard Henderson
2024-09-10 7:04 ` Philippe Mathieu-Daudé
2024-09-08 2:26 ` [PATCH 12/12] tcg/i386: Implement cmpsel_vec with avx512 insns Richard Henderson
2024-09-09 23:37 ` [PATCH 00/12] tcg: Improve support for cmpsel_vec Richard Henderson
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20240908022632.459477-4-richard.henderson@linaro.org \
--to=richard.henderson@linaro.org \
--cc=bmeng.cn@gmail.com \
--cc=liwei1518@gmail.com \
--cc=qemu-devel@nongnu.org \
--cc=tangtiancheng.ttc@alibaba-inc.com \
--cc=zhiwei_liu@linux.alibaba.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
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 NNTP newsgroup(s).