From: Jan Bobek <jan.bobek@gmail.com>
To: qemu-devel@nongnu.org
Cc: "Alex Bennée" <alex.bennee@linaro.org>,
"Richard Henderson" <richard.henderson@linaro.org>,
"Jan Bobek" <jan.bobek@gmail.com>
Subject: [Qemu-devel] [RFC PATCH v1 22/22] target/i386: reimplement (V)P(EQ, CMP)(B, W, D)
Date: Wed, 31 Jul 2019 13:57:02 -0400 [thread overview]
Message-ID: <20190731175702.4916-23-jan.bobek@gmail.com> (raw)
In-Reply-To: <20190731175702.4916-1-jan.bobek@gmail.com>
Use the gvec infrastructure to achieve the desired functionality.
Signed-off-by: Jan Bobek <jan.bobek@gmail.com>
---
target/i386/ops_sse.h | 13 -------
| 8 -----
target/i386/translate.c | 66 ++++++++++++++++++++++++++++++++----
3 files changed, 60 insertions(+), 27 deletions(-)
diff --git a/target/i386/ops_sse.h b/target/i386/ops_sse.h
index 75ff686bb6..b6ace9410f 100644
--- a/target/i386/ops_sse.h
+++ b/target/i386/ops_sse.h
@@ -337,11 +337,6 @@ static inline int satsw(int x)
}
}
-#define FCMPGTB(a, b) ((int8_t)(a) > (int8_t)(b) ? -1 : 0)
-#define FCMPGTW(a, b) ((int16_t)(a) > (int16_t)(b) ? -1 : 0)
-#define FCMPGTL(a, b) ((int32_t)(a) > (int32_t)(b) ? -1 : 0)
-#define FCMPEQ(a, b) ((a) == (b) ? -1 : 0)
-
#define FMULLW(a, b) ((a) * (b))
#define FMULHRW(a, b) (((int16_t)(a) * (int16_t)(b) + 0x8000) >> 16)
#define FMULHUW(a, b) ((a) * (b) >> 16)
@@ -350,14 +345,6 @@ static inline int satsw(int x)
#define FAVG(a, b) (((a) + (b) + 1) >> 1)
#endif
-SSE_HELPER_B(helper_pcmpgtb, FCMPGTB)
-SSE_HELPER_W(helper_pcmpgtw, FCMPGTW)
-SSE_HELPER_L(helper_pcmpgtl, FCMPGTL)
-
-SSE_HELPER_B(helper_pcmpeqb, FCMPEQ)
-SSE_HELPER_W(helper_pcmpeqw, FCMPEQ)
-SSE_HELPER_L(helper_pcmpeql, FCMPEQ)
-
SSE_HELPER_W(helper_pmullw, FMULLW)
#if SHIFT == 0
SSE_HELPER_W(helper_pmulhrw, FMULHRW)
--git a/target/i386/ops_sse_header.h b/target/i386/ops_sse_header.h
index 9c7451d28e..d8e33dff6b 100644
--- a/target/i386/ops_sse_header.h
+++ b/target/i386/ops_sse_header.h
@@ -60,14 +60,6 @@ DEF_HELPER_3(glue(pslldq, SUFFIX), void, env, Reg, Reg)
#define SSE_HELPER_Q(name, F)\
DEF_HELPER_3(glue(name, SUFFIX), void, env, Reg, Reg)
-SSE_HELPER_B(pcmpgtb, FCMPGTB)
-SSE_HELPER_W(pcmpgtw, FCMPGTW)
-SSE_HELPER_L(pcmpgtl, FCMPGTL)
-
-SSE_HELPER_B(pcmpeqb, FCMPEQ)
-SSE_HELPER_W(pcmpeqw, FCMPEQ)
-SSE_HELPER_L(pcmpeql, FCMPEQ)
-
SSE_HELPER_W(pmullw, FMULLW)
#if SHIFT == 0
SSE_HELPER_W(pmulhrw, FMULHRW)
diff --git a/target/i386/translate.c b/target/i386/translate.c
index d08d2cedce..729509e1ff 100644
--- a/target/i386/translate.c
+++ b/target/i386/translate.c
@@ -2783,9 +2783,9 @@ static const SSEFunc_0_epp sse_op_table1[256][4] = {
[0x61] = MMX_OP2(punpcklwd),
[0x62] = MMX_OP2(punpckldq),
[0x63] = MMX_OP2(packsswb),
- [0x64] = MMX_OP2(pcmpgtb),
- [0x65] = MMX_OP2(pcmpgtw),
- [0x66] = MMX_OP2(pcmpgtl),
+ [0x64] = { SSE_TOMBSTONE, SSE_TOMBSTONE },
+ [0x65] = { SSE_TOMBSTONE, SSE_TOMBSTONE },
+ [0x66] = { SSE_TOMBSTONE, SSE_TOMBSTONE },
[0x67] = MMX_OP2(packuswb),
[0x68] = MMX_OP2(punpckhbw),
[0x69] = MMX_OP2(punpckhwd),
@@ -2802,9 +2802,9 @@ static const SSEFunc_0_epp sse_op_table1[256][4] = {
[0x71] = { SSE_SPECIAL, SSE_SPECIAL }, /* shiftw */
[0x72] = { SSE_SPECIAL, SSE_SPECIAL }, /* shiftd */
[0x73] = { SSE_SPECIAL, SSE_SPECIAL }, /* shiftq */
- [0x74] = MMX_OP2(pcmpeqb),
- [0x75] = MMX_OP2(pcmpeqw),
- [0x76] = MMX_OP2(pcmpeql),
+ [0x74] = { SSE_TOMBSTONE, SSE_TOMBSTONE },
+ [0x75] = { SSE_TOMBSTONE, SSE_TOMBSTONE },
+ [0x76] = { SSE_TOMBSTONE, SSE_TOMBSTONE },
[0x77] = { SSE_DUMMY }, /* emms */
[0x78] = { NULL, SSE_SPECIAL, NULL, SSE_SPECIAL }, /* extrq_i, insertq_i */
[0x79] = { NULL, gen_helper_extrq_r, NULL, gen_helper_insertq_r },
@@ -3216,6 +3216,30 @@ static inline void gen_gvec_ld_modrm_3(CPUX86State *env, DisasContext *s,
#define gen_vpmaxu_xmm(env, s, modrm, vece) gen_gvec_ld_modrm_vxmm((env), (s), (modrm), (vece), tcg_gen_gvec_umax, 0123)
#define gen_vpmaxu_ymm(env, s, modrm, vece) gen_gvec_ld_modrm_vymm((env), (s), (modrm), (vece), tcg_gen_gvec_umax, 0123)
+static inline void gen_gvec_cmpeq(unsigned vece, uint32_t dofs,
+ uint32_t aofs, uint32_t bofs,
+ uint32_t oprsz, uint32_t maxsz)
+{
+ tcg_gen_gvec_cmp(TCG_COND_EQ, vece, dofs, aofs, bofs, oprsz, maxsz);
+}
+
+#define gen_pcmpeq_mm(env, s, modrm, vece) gen_gvec_ld_modrm_mm ((env), (s), (modrm), (vece), gen_gvec_cmpeq, 0112)
+#define gen_pcmpeq_xmm(env, s, modrm, vece) gen_gvec_ld_modrm_xmm ((env), (s), (modrm), (vece), gen_gvec_cmpeq, 0112)
+#define gen_vpcmpeq_xmm(env, s, modrm, vece) gen_gvec_ld_modrm_vxmm((env), (s), (modrm), (vece), gen_gvec_cmpeq, 0123)
+#define gen_vpcmpeq_ymm(env, s, modrm, vece) gen_gvec_ld_modrm_vymm((env), (s), (modrm), (vece), gen_gvec_cmpeq, 0123)
+
+static inline void gen_gvec_cmpgt(unsigned vece, uint32_t dofs,
+ uint32_t aofs, uint32_t bofs,
+ uint32_t oprsz, uint32_t maxsz)
+{
+ tcg_gen_gvec_cmp(TCG_COND_GT, vece, dofs, aofs, bofs, oprsz, maxsz);
+}
+
+#define gen_pcmpgt_mm(env, s, modrm, vece) gen_gvec_ld_modrm_mm ((env), (s), (modrm), (vece), gen_gvec_cmpgt, 0112)
+#define gen_pcmpgt_xmm(env, s, modrm, vece) gen_gvec_ld_modrm_xmm ((env), (s), (modrm), (vece), gen_gvec_cmpgt, 0112)
+#define gen_vpcmpgt_xmm(env, s, modrm, vece) gen_gvec_ld_modrm_vxmm((env), (s), (modrm), (vece), gen_gvec_cmpgt, 0123)
+#define gen_vpcmpgt_ymm(env, s, modrm, vece) gen_gvec_ld_modrm_vymm((env), (s), (modrm), (vece), gen_gvec_cmpgt, 0123)
+
#define gen_pand_mm(env, s, modrm) gen_gvec_ld_modrm_mm ((env), (s), (modrm), MO_64, tcg_gen_gvec_and, 0112)
#define gen_pand_xmm(env, s, modrm) gen_gvec_ld_modrm_xmm ((env), (s), (modrm), MO_64, tcg_gen_gvec_and, 0112)
#define gen_vpand_xmm(env, s, modrm) gen_gvec_ld_modrm_vxmm((env), (s), (modrm), MO_64, tcg_gen_gvec_and, 0123)
@@ -3451,6 +3475,36 @@ static void gen_sse(CPUX86State *env, DisasContext *s, int b)
case 0xee | M_0F | P_66 | VEX_128: gen_vpmaxs_xmm(env, s, modrm, MO_16); return;
case 0xee | M_0F | P_66 | VEX_256: gen_vpmaxs_ymm(env, s, modrm, MO_16); return;
+ case 0x64 | M_0F: gen_pcmpgt_mm(env, s, modrm, MO_8); return;
+ case 0x64 | M_0F | P_66: gen_pcmpgt_xmm(env, s, modrm, MO_8); return;
+ case 0x64 | M_0F | P_66 | VEX_128: gen_vpcmpgt_xmm(env, s, modrm, MO_8); return;
+ case 0x64 | M_0F | P_66 | VEX_256: gen_vpcmpgt_ymm(env, s, modrm, MO_8); return;
+
+ case 0x65 | M_0F: gen_pcmpgt_mm(env, s, modrm, MO_16); return;
+ case 0x65 | M_0F | P_66: gen_pcmpgt_xmm(env, s, modrm, MO_16); return;
+ case 0x65 | M_0F | P_66 | VEX_128: gen_vpcmpgt_xmm(env, s, modrm, MO_16); return;
+ case 0x65 | M_0F | P_66 | VEX_256: gen_vpcmpgt_ymm(env, s, modrm, MO_16); return;
+
+ case 0x66 | M_0F: gen_pcmpgt_mm(env, s, modrm, MO_32); return;
+ case 0x66 | M_0F | P_66: gen_pcmpgt_xmm(env, s, modrm, MO_32); return;
+ case 0x66 | M_0F | P_66 | VEX_128: gen_vpcmpgt_xmm(env, s, modrm, MO_32); return;
+ case 0x66 | M_0F | P_66 | VEX_256: gen_vpcmpgt_ymm(env, s, modrm, MO_32); return;
+
+ case 0x74 | M_0F: gen_pcmpeq_mm(env, s, modrm, MO_8); return;
+ case 0x74 | M_0F | P_66: gen_pcmpeq_xmm(env, s, modrm, MO_8); return;
+ case 0x74 | M_0F | P_66 | VEX_128: gen_vpcmpeq_xmm(env, s, modrm, MO_8); return;
+ case 0x74 | M_0F | P_66 | VEX_256: gen_vpcmpeq_ymm(env, s, modrm, MO_8); return;
+
+ case 0x75 | M_0F: gen_pcmpeq_mm(env, s, modrm, MO_16); return;
+ case 0x75 | M_0F | P_66: gen_pcmpeq_xmm(env, s, modrm, MO_16); return;
+ case 0x75 | M_0F | P_66 | VEX_128: gen_vpcmpeq_xmm(env, s, modrm, MO_16); return;
+ case 0x75 | M_0F | P_66 | VEX_256: gen_vpcmpeq_ymm(env, s, modrm, MO_16); return;
+
+ case 0x76 | M_0F: gen_pcmpeq_mm(env, s, modrm, MO_32); return;
+ case 0x76 | M_0F | P_66: gen_pcmpeq_xmm(env, s, modrm, MO_32); return;
+ case 0x76 | M_0F | P_66 | VEX_128: gen_vpcmpeq_xmm(env, s, modrm, MO_32); return;
+ case 0x76 | M_0F | P_66 | VEX_256: gen_vpcmpeq_ymm(env, s, modrm, MO_32); return;
+
case 0xdb | M_0F: gen_pand_mm(env, s, modrm); return;
case 0xdb | M_0F | P_66: gen_pand_xmm(env, s, modrm); return;
case 0xdb | M_0F | P_66 | VEX_128: gen_vpand_xmm(env, s, modrm); return;
--
2.20.1
next prev parent reply other threads:[~2019-07-31 18:07 UTC|newest]
Thread overview: 44+ messages / expand[flat|nested] mbox.gz Atom feed top
2019-07-31 17:56 [Qemu-devel] [RFC PATCH v1 00/22] reimplement (some) x86 vector instructions using tcg-gvec Jan Bobek
2019-07-31 17:56 ` [Qemu-devel] [RFC PATCH v1 01/22] target/i386: Push rex_r into DisasContext Jan Bobek
2019-07-31 17:56 ` [Qemu-devel] [RFC PATCH v1 02/22] target/i386: Push rex_w " Jan Bobek
2019-07-31 17:56 ` [Qemu-devel] [RFC PATCH v1 03/22] target/i386: Use prefix, aflag and dflag from DisasContext Jan Bobek
2019-07-31 19:41 ` Aleksandar Markovic
2019-07-31 20:04 ` Aleksandar Markovic
2019-08-02 13:20 ` Jan Bobek
2019-07-31 17:56 ` [Qemu-devel] [RFC PATCH v1 04/22] target/i386: Simplify gen_exception arguments Jan Bobek
2019-07-31 17:56 ` [Qemu-devel] [RFC PATCH v1 05/22] target/i386: introduce gen_ld_modrm_* helpers Jan Bobek
2019-07-31 19:08 ` Richard Henderson
2019-08-02 13:26 ` Jan Bobek
2019-07-31 17:56 ` [Qemu-devel] [RFC PATCH v1 06/22] target/i386: introduce gen_gvec_ld_modrm_* helpers Jan Bobek
2019-07-31 22:47 ` Richard Henderson
2019-08-02 13:34 ` Jan Bobek
2019-07-31 17:56 ` [Qemu-devel] [RFC PATCH v1 07/22] target/i386: add vector register file alignment constraints Jan Bobek
2019-07-31 19:14 ` Richard Henderson
2019-07-31 17:56 ` [Qemu-devel] [RFC PATCH v1 08/22] target/i386: reimplement (V)PAND, (V)ANDPS, (V)ANDPD Jan Bobek
2019-07-31 19:35 ` Richard Henderson
2019-07-31 20:27 ` Aleksandar Markovic
2019-07-31 21:21 ` Richard Henderson
2019-08-02 13:53 ` Jan Bobek
2019-07-31 17:56 ` [Qemu-devel] [RFC PATCH v1 09/22] target/i386: reimplement (V)POR, (V)ORPS, (V)ORPD Jan Bobek
2019-07-31 17:56 ` [Qemu-devel] [RFC PATCH v1 10/22] target/i386: reimplement (V)PXOR, (V)XORPS, (V)XORPD Jan Bobek
2019-07-31 17:56 ` [Qemu-devel] [RFC PATCH v1 11/22] target/i386: reimplement (V)PANDN, (V)ANDNPS, (V)ANDNPD Jan Bobek
2019-07-31 17:56 ` [Qemu-devel] [RFC PATCH v1 12/22] target/i386: reimplement (V)PADD(B, W, D, Q) Jan Bobek
2019-07-31 17:56 ` [Qemu-devel] [RFC PATCH v1 13/22] target/i386: reimplement (V)PSUB(B, " Jan Bobek
2019-07-31 17:56 ` [Qemu-devel] [RFC PATCH v1 14/22] target/i386: reimplement (V)PADDS(B, W) Jan Bobek
2019-07-31 17:56 ` [Qemu-devel] [RFC PATCH v1 15/22] target/i386: reimplement (V)PADDUS(B, W) Jan Bobek
2019-07-31 17:56 ` [Qemu-devel] [RFC PATCH v1 16/22] target/i386: reimplement (V)PSUBS(B, W) Jan Bobek
2019-07-31 17:56 ` [Qemu-devel] [RFC PATCH v1 17/22] target/i386: reimplement (V)PSUBUS(B, W) Jan Bobek
2019-07-31 17:56 ` [Qemu-devel] [RFC PATCH v1 18/22] target/i386: reimplement (V)PMINSW Jan Bobek
2019-07-31 17:56 ` [Qemu-devel] [RFC PATCH v1 19/22] target/i386: reimplement (V)PMINUB Jan Bobek
2019-07-31 17:57 ` [Qemu-devel] [RFC PATCH v1 20/22] target/i386: reimplement (V)PMAXSW Jan Bobek
2019-07-31 17:57 ` [Qemu-devel] [RFC PATCH v1 21/22] target/i386: reimplement (V)PMAXUB Jan Bobek
2019-07-31 17:57 ` Jan Bobek [this message]
2019-07-31 19:50 ` [Qemu-devel] [RFC PATCH v1 22/22] target/i386: reimplement (V)P(EQ, CMP)(B, W, D) Richard Henderson
2019-07-31 20:09 ` Aleksandar Markovic
2019-07-31 21:31 ` Richard Henderson
2019-08-02 14:07 ` Jan Bobek
2019-08-02 14:18 ` Aleksandar Markovic
2019-07-31 18:20 ` [Qemu-devel] [RFC PATCH v1 00/22] reimplement (some) x86 vector instructions using tcg-gvec no-reply
2019-07-31 19:21 ` no-reply
2019-07-31 19:21 ` no-reply
2019-08-01 15:46 ` no-reply
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=20190731175702.4916-23-jan.bobek@gmail.com \
--to=jan.bobek@gmail.com \
--cc=alex.bennee@linaro.org \
--cc=qemu-devel@nongnu.org \
--cc=richard.henderson@linaro.org \
/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).