projects / GitHub / LineageOS / android_kernel_motorola_exynos9610.git / commit
? search:
summary | shortlog | log | commit | commitdiff | tree
(parent: 5cfc7e0) | patch
KVM: emulate: avoid per-byte copying in instruction fetches
authorPaolo Bonzini <pbonzini@redhat.com>
Tue, 6 May 2014 11:05:25 +0000 (13:05 +0200)
committerPaolo Bonzini <pbonzini@redhat.com>
Fri, 11 Jul 2014 07:14:03 +0000 (09:14 +0200)
commit9506d57de3bc8277a4e306e0d439976862f68c6d
treeaac16beac03ddc018f7ffec0d453d1a89d4a11f0tree | snapshot (tar.gz zip)
parent5cfc7e0f5e5e1adf998df94f8e36edaf5d30d38ecommit | diff
KVM: emulate: avoid per-byte copying in instruction fetches

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 diff | blob | blame | history