On Mon, Apr 28, 2014 at 11:55 PM, Peter Zijlstra <peterz@xxxxxxxxxxxxx> wrote:
> What version,
the code snipt in question is extracted from v3.15-rc3.
for the (1) check in previous email, its assembled code looks like:
--- (1) snipt ---
mov (%rdi),%rax (a)
test $0x80,%ah (b)
jne 754 <put_compound_page+0x74> (c)
--- (1) snipt ---
(a) %rdi is the struct page pointer
(b) check if PG_tail(0x80) set(likely not set, we tell the compiler)
(c) if set, jump; not set, fall through (good, credit to our hint)
===================================================
for the (3) check in previous email, its assembled code looks like:
--- (3) snipt ---
mov (%rdi),%rax (A)
mov %rdi,%r12
test $0x80,%ah (B)
jne 8f8 <put_compound_page+0x218> (C)
--- (3) snipt ---
(A) %rdi is the struct page pointer
(B) check if PG_tail(0x80) set(likely set in this case, but we
tell compiler unlikely)
(C) if set, jump; not set, fall through (god! it would better
not jump if set, but we
tell compiler unlikely, so it happily did as we told it)
# all code are compiled by gcc (GCC) 4.8.2
> and why didn't your changelog include this useful information?
Sorry, I would have done so. I will resend the patch.
Thanks,
Jianyu Zhan
--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@xxxxxxxxx. For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@xxxxxxxxx";> email@xxxxxxxxx </a>