| Message ID | 20220809055138.101470-1-zhiguangni01@gmail.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | KVM:make cpuid_entry2_find more efficient | expand |
On Tue, Aug 09, 2022 at 01:51:38PM +0800, Liam Ni wrote: > Compared with the way of obtaining the pointer by > fetching the value of the array and then fetching the pointer, > the way of obtaining the pointer by the pointer offset is more efficient. How did you determine that? Hint: look at the generated assembler before and after your change and see if there are any differences. :-)
diff --git a/arch/x86/kvm/cpuid.c b/arch/x86/kvm/cpuid.c index de6d44e07e34..3bf82a891564 100644 --- a/arch/x86/kvm/cpuid.c +++ b/arch/x86/kvm/cpuid.c @@ -75,7 +75,7 @@ static inline struct kvm_cpuid_entry2 *cpuid_entry2_find( int i; for (i = 0; i < nent; i++) { - e = &entries[i]; + e = entries + i; if (e->function == function && (!(e->flags & KVM_CPUID_FLAG_SIGNIFCANT_INDEX) || e->index == index))
Compared with the way of obtaining the pointer by fetching the value of the array and then fetching the pointer, the way of obtaining the pointer by the pointer offset is more efficient. Signed-off-by: Liam Ni <zhiguangni01@gmail.com> --- arch/x86/kvm/cpuid.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-)