public inbox for kvm@vger.kernel.org
 help / color / mirror / Atom feed
From: Avi Kivity <avi@redhat.com>
To: Marcelo Tosatti <mtosatti@redhat.com>, kvm@vger.kernel.org
Cc: Wei Xu <wexu2@cisco.com>
Subject: [PATCH v2 6/8] KVM: x86 emulator: Specialize decoding for insns with 66/f2/f3 prefixes
Date: Tue, 29 Mar 2011 14:54:01 +0200	[thread overview]
Message-ID: <1301403243-23524-7-git-send-email-avi@redhat.com> (raw)
In-Reply-To: <1301403243-23524-1-git-send-email-avi@redhat.com>

Most SIMD instructions use the 66/f2/f3 prefixes to distinguish between
different variants of the same instruction.  Usually the encoding is quite
regular, but in some cases (including non-SIMD instructions) the prefixes
generate very different instructions.  Examples include XCHG/PAUSE,
MOVQ/MOVDQA/MOVDQU, and MOVBE/CRC32.

Allow the emulator to handle these special cases by splitting such opcodes
into groups, with different decode flags and execution functions for different
prefixes.

Signed-off-by: Avi Kivity <avi@redhat.com>
---
 arch/x86/kvm/emulate.c |   26 +++++++++++++++++++++++++-
 1 files changed, 25 insertions(+), 1 deletions(-)

diff --git a/arch/x86/kvm/emulate.c b/arch/x86/kvm/emulate.c
index 7066cf8..458faea 100644
--- a/arch/x86/kvm/emulate.c
+++ b/arch/x86/kvm/emulate.c
@@ -75,6 +75,7 @@
 #define Stack       (1<<13)     /* Stack instruction (push/pop) */
 #define Group       (1<<14)     /* Bits 3:5 of modrm byte extend opcode */
 #define GroupDual   (1<<15)     /* Alternate decoding of mod == 3 */
+#define Prefix      (1<<16)     /* Instruction varies with 66/f2/f3 prefix */
 /* Misc flags */
 #define VendorSpecific (1<<22) /* Vendor specific instruction */
 #define NoAccess    (1<<23) /* Don't access memory (lea/invlpg/verr etc) */
@@ -106,6 +107,7 @@ struct opcode {
 		int (*execute)(struct x86_emulate_ctxt *ctxt);
 		struct opcode *group;
 		struct group_dual *gdual;
+		struct gprefix *gprefix;
 	} u;
 };
 
@@ -114,6 +116,13 @@ struct group_dual {
 	struct opcode mod3[8];
 };
 
+struct gprefix {
+	struct opcode pfx_no;
+	struct opcode pfx_66;
+	struct opcode pfx_f2;
+	struct opcode pfx_f3;
+};
+
 /* EFLAGS bit definitions. */
 #define EFLG_ID (1<<21)
 #define EFLG_VIP (1<<20)
@@ -2625,7 +2634,8 @@ x86_decode_insn(struct x86_emulate_ctxt *ctxt, void *insn, int insn_len)
 	struct decode_cache *c = &ctxt->decode;
 	int rc = X86EMUL_CONTINUE;
 	int mode = ctxt->mode;
-	int def_op_bytes, def_ad_bytes, dual, goffset;
+	int def_op_bytes, def_ad_bytes, dual, goffset, simd_prefix;
+	bool op_prefix = false;
 	struct opcode opcode, *g_mod012, *g_mod3;
 	struct operand memop = { .type = OP_NONE };
 
@@ -2662,6 +2672,7 @@ x86_decode_insn(struct x86_emulate_ctxt *ctxt, void *insn, int insn_len)
 	for (;;) {
 		switch (c->b = insn_fetch(u8, 1, c->eip)) {
 		case 0x66:	/* operand-size override */
+			op_prefix = true;
 			/* switch between 2/4 bytes */
 			c->op_bytes = def_op_bytes ^ 6;
 			break;
@@ -2742,6 +2753,19 @@ done_prefixes:
 		c->d |= opcode.flags;
 	}
 
+	if (c->d & Prefix) {
+		if (c->rep_prefix && op_prefix)
+			return X86EMUL_UNHANDLEABLE;
+		simd_prefix = op_prefix ? 0x66 : c->rep_prefix;
+		switch (simd_prefix) {
+		case 0x00: opcode = opcode.u.gprefix->pfx_no; break;
+		case 0x66: opcode = opcode.u.gprefix->pfx_66; break;
+		case 0xf2: opcode = opcode.u.gprefix->pfx_f2; break;
+		case 0xf3: opcode = opcode.u.gprefix->pfx_f3; break;
+		}
+		c->d |= opcode.flags;
+	}
+
 	c->execute = opcode.u.execute;
 
 	/* Unrecognised? */
-- 
1.7.1


  parent reply	other threads:[~2011-03-29 12:54 UTC|newest]

Thread overview: 13+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2011-03-29 12:53 [PATCH v2 0/8] SSE MMIO Avi Kivity
2011-03-29 12:53 ` [PATCH v2 1/8] KVM: extend in-kernel mmio to handle >8 byte transactions Avi Kivity
2011-03-29 14:51   ` Wei Xu
2011-03-29 14:55     ` Avi Kivity
2011-03-29 12:53 ` [PATCH v2 2/8] KVM: Split mmio completion into a function Avi Kivity
2011-03-29 12:53 ` [PATCH v2 3/8] KVM: 16-byte mmio support Avi Kivity
2011-03-29 12:53 ` [PATCH v2 4/8] KVM: x86 emulator: do not munge rep prefix Avi Kivity
2011-03-29 12:54 ` [PATCH v2 5/8] KVM: x86 emulator: define callbacks for using the guest fpu within the emulator Avi Kivity
2011-03-29 12:54 ` Avi Kivity [this message]
2011-03-29 12:54 ` [PATCH v2 7/8] KVM: x86 emulator: SSE support Avi Kivity
2011-06-28  6:49   ` Cyclonus J
2011-06-28  8:11     ` Avi Kivity
2011-03-29 12:54 ` [PATCH v2 8/8] KVM: x86 emulator: implement movdqu instruction (f3 0f 6f, f3 0f 7f) Avi Kivity

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=1301403243-23524-7-git-send-email-avi@redhat.com \
    --to=avi@redhat.com \
    --cc=kvm@vger.kernel.org \
    --cc=mtosatti@redhat.com \
    --cc=wexu2@cisco.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