From: Paolo Bonzini <pbonzini@redhat.com>
To: linux-kernel@vger.kernel.org
Cc: bdas@redhat.com, gleb@kernel.org
Subject: [PATCH 21/25] KVM: emulate: avoid per-byte copying in instruction fetches
Date: Mon, 9 Jun 2014 14:59:09 +0200 [thread overview]
Message-ID: <1402318753-23362-22-git-send-email-pbonzini@redhat.com> (raw)
In-Reply-To: <1402318753-23362-1-git-send-email-pbonzini@redhat.com>
We do not need a memory copying loop anymore in insn_fetch; we
can use a byte-aligned pointer to access instruction fields directly
from the fetch_cache. This eliminates 50-150 cycles (corresponding to
a 5-10% improvement in performance) from each instruction.
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
---
arch/x86/kvm/emulate.c | 46 ++++++++++++++++++++++------------------------
1 file changed, 22 insertions(+), 24 deletions(-)
diff --git a/arch/x86/kvm/emulate.c b/arch/x86/kvm/emulate.c
index f81d7a5bc19b..6ec0be2d2461 100644
--- a/arch/x86/kvm/emulate.c
+++ b/arch/x86/kvm/emulate.c
@@ -707,7 +707,7 @@ static int segmented_read_std(struct x86_emulate_ctxt *ctxt,
* Prefetch the remaining bytes of the instruction without crossing page
* boundary if they are not in fetch_cache yet.
*/
-static int do_insn_fetch_bytes(struct x86_emulate_ctxt *ctxt, int op_size)
+static int __do_insn_fetch_bytes(struct x86_emulate_ctxt *ctxt, int op_size)
{
struct fetch_cache *fc = &ctxt->fetch;
int rc;
@@ -739,41 +739,39 @@ static int do_insn_fetch_bytes(struct x86_emulate_ctxt *ctxt, int op_size)
return X86EMUL_CONTINUE;
}
-static int do_insn_fetch(struct x86_emulate_ctxt *ctxt,
- void *__dest, unsigned size)
+static __always_inline int do_insn_fetch_bytes(struct x86_emulate_ctxt *ctxt,
+ unsigned size)
{
- struct fetch_cache *fc = &ctxt->fetch;
- u8 *dest = __dest;
- u8 *src = &fc->data[ctxt->_eip - fc->start];
-
/* We have to be careful about overflow! */
- if (unlikely(ctxt->_eip > fc->end - size)) {
- int rc = do_insn_fetch_bytes(ctxt, size);
- if (rc != X86EMUL_CONTINUE)
- return rc;
- }
-
- while (size--) {
- *dest++ = *src++;
- ctxt->_eip++;
- continue;
- }
- return X86EMUL_CONTINUE;
+ if (unlikely(ctxt->_eip > ctxt->fetch.end - size))
+ return __do_insn_fetch_bytes(ctxt, size);
+ else
+ return X86EMUL_CONTINUE;
}
/* Fetch next part of the instruction being emulated. */
#define insn_fetch(_type, _ctxt) \
-({ unsigned long _x; \
- rc = do_insn_fetch(_ctxt, &_x, sizeof(_type)); \
+({ _type _x; \
+ struct fetch_cache *_fc; \
+ \
+ rc = do_insn_fetch_bytes(_ctxt, sizeof(_type)); \
if (rc != X86EMUL_CONTINUE) \
goto done; \
- (_type)_x; \
+ _fc = &ctxt->fetch; \
+ _x = *(_type __aligned(1) *) &_fc->data[ctxt->_eip - _fc->start]; \
+ ctxt->_eip += sizeof(_type); \
+ _x; \
})
#define insn_fetch_arr(_arr, _size, _ctxt) \
-({ rc = do_insn_fetch(_ctxt, _arr, (_size)); \
+({ \
+ struct fetch_cache *_fc; \
+ rc = do_insn_fetch_bytes(_ctxt, _size); \
if (rc != X86EMUL_CONTINUE) \
goto done; \
+ _fc = &ctxt->fetch; \
+ memcpy(_arr, &_fc->data[ctxt->_eip - _fc->start], _size); \
+ ctxt->_eip += (_size); \
})
/*
@@ -4310,7 +4308,7 @@ int x86_decode_insn(struct x86_emulate_ctxt *ctxt, void *insn, int insn_len)
if (insn_len > 0)
memcpy(ctxt->fetch.data, insn, insn_len);
else {
- rc = do_insn_fetch_bytes(ctxt, 1);
+ rc = __do_insn_fetch_bytes(ctxt, 1);
if (rc != X86EMUL_CONTINUE)
return rc;
}
--
1.8.3.1
next prev parent reply other threads:[~2014-06-09 13:00 UTC|newest]
Thread overview: 28+ messages / expand[flat|nested] mbox.gz Atom feed top
2014-06-09 12:58 [PATCH 00/25] KVM: x86: Speed up emulation of invalid state Paolo Bonzini
2014-06-09 12:58 ` [PATCH 01/25] KVM: vmx: speed up emulation of invalid guest state Paolo Bonzini
2014-06-09 12:58 ` [PATCH 02/25] KVM: x86: return all bits from get_interrupt_shadow Paolo Bonzini
2014-06-09 12:58 ` [PATCH 03/25] KVM: x86: avoid useless set of KVM_REQ_EVENT after emulation Paolo Bonzini
2014-06-09 12:58 ` [PATCH 04/25] KVM: emulate: move around some checks Paolo Bonzini
2014-06-09 12:58 ` [PATCH 05/25] KVM: emulate: protect checks on ctxt->d by a common "if (unlikely())" Paolo Bonzini
2014-06-09 12:58 ` [PATCH 06/25] KVM: emulate: speed up emulated moves Paolo Bonzini
2014-06-09 12:58 ` [PATCH 07/25] KVM: emulate: simplify writeback Paolo Bonzini
2014-06-09 12:58 ` [PATCH 08/25] KVM: emulate: abstract handling of memory operands Paolo Bonzini
2014-06-09 12:58 ` [PATCH 09/25] KVM: export mark_page_dirty_in_slot Paolo Bonzini
2014-06-09 12:58 ` [PATCH 10/25] KVM: emulate: introduce memory_prepare callback to speed up memory access Paolo Bonzini
2014-06-09 12:58 ` [PATCH 11/25] KVM: emulate: activate memory access optimization Paolo Bonzini
2014-06-09 12:59 ` [PATCH 12/25] KVM: emulate: extend memory access optimization to stores Paolo Bonzini
2014-06-09 18:40 ` Bandan Das
2014-06-19 11:37 ` Paolo Bonzini
2014-06-09 12:59 ` [PATCH 13/25] KVM: emulate: move init_decode_cache to emulate.c Paolo Bonzini
2014-06-09 12:59 ` [PATCH 14/25] KVM: emulate: Remove ctxt->intercept and ctxt->check_perm checks Paolo Bonzini
2014-06-09 12:59 ` [PATCH 15/25] KVM: emulate: cleanup decode_modrm Paolo Bonzini
2014-06-09 12:59 ` [PATCH 16/25] KVM: emulate: clean up initializations in init_decode_cache Paolo Bonzini
2014-06-09 12:59 ` [PATCH 17/25] KVM: emulate: rework seg_override Paolo Bonzini
2014-06-09 12:59 ` [PATCH 18/25] KVM: emulate: do not initialize memopp Paolo Bonzini
2014-06-09 12:59 ` [PATCH 19/25] KVM: emulate: speed up do_insn_fetch Paolo Bonzini
2014-06-09 12:59 ` [PATCH 20/25] KVM: emulate: avoid repeated calls to do_insn_fetch_bytes Paolo Bonzini
2014-06-09 12:59 ` Paolo Bonzini [this message]
2014-06-09 12:59 ` [PATCH 22/25] KVM: emulate: put pointers in the fetch_cache Paolo Bonzini
2014-06-09 12:59 ` [PATCH 23/25] KVM: x86: use kvm_read_guest_page for emulator accesses Paolo Bonzini
2014-06-09 12:59 ` [PATCH 24/25] KVM: emulate: simplify BitOp handling Paolo Bonzini
2014-06-09 12:59 ` [PATCH 25/25] KVM: emulate: fix harmless typo in MMX decoding Paolo Bonzini
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=1402318753-23362-22-git-send-email-pbonzini@redhat.com \
--to=pbonzini@redhat.com \
--cc=bdas@redhat.com \
--cc=gleb@kernel.org \
--cc=linux-kernel@vger.kernel.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
all inboxes | Powered by JetHome®