Commit 098fb254 authored by Kent Overstreet's avatar Kent Overstreet

bcache: Delete some slower inline asm

Never saw a profile of bset_search_tree() where it wasn't bottlenecked
on memory until I got my new Haswell machine, but when I tried it there
it was suddenly burning 20% of the cpu in the inner loop on shrd...

Turns out, the version of shrd that takes 64 bit operands has a 9 cycle
latency. hah.
Signed-off-by: default avatarKent Overstreet <kmo@daterainc.com>
parent 28935ab5
...@@ -481,16 +481,8 @@ static struct bkey *table_to_bkey(struct bset_tree *t, unsigned cacheline) ...@@ -481,16 +481,8 @@ static struct bkey *table_to_bkey(struct bset_tree *t, unsigned cacheline)
static inline uint64_t shrd128(uint64_t high, uint64_t low, uint8_t shift) static inline uint64_t shrd128(uint64_t high, uint64_t low, uint8_t shift)
{ {
#ifdef CONFIG_X86_64
asm("shrd %[shift],%[high],%[low]"
: [low] "+Rm" (low)
: [high] "R" (high),
[shift] "ci" (shift)
: "cc");
#else
low >>= shift; low >>= shift;
low |= (high << 1) << (63U - shift); low |= (high << 1) << (63U - shift);
#endif
return low; return low;
} }
......
Markdown is supported
0%
or
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment