fast_hash: avoid indirect function calls

By default the arch_fast_hash hashing function pointers are initialized
to jhash(2). If during boot-up a CPU with SSE4.2 is detected they get
updated to the CRC32 ones. This dispatching scheme incurs a function
pointer lookup and indirect call for every hashing operation.

rhashtable as a user of arch_fast_hash e.g. stores pointers to hashing
functions in its structure, too, causing two indirect branches per
hashing operation.

Using alternative_call we can get away with one of those indirect branches.

Acked-by: Daniel Borkmann <dborkman@redhat.com>
Cc: Thomas Graf <tgraf@suug.ch>
Signed-off-by: Hannes Frederic Sowa <hannes@stressinduktion.org>
Signed-off-by: David S. Miller <davem@davemloft.net>
This commit is contained in:
Hannes Frederic Sowa 2014-11-05 00:23:04 +01:00 committed by David S. Miller
parent 2c99cd914d
commit e5a2c89995
6 changed files with 98 additions and 93 deletions

View file

@ -84,38 +84,4 @@ static inline u32 hash32_ptr(const void *ptr)
return (u32)val;
}
struct fast_hash_ops {
u32 (*hash)(const void *data, u32 len, u32 seed);
u32 (*hash2)(const u32 *data, u32 len, u32 seed);
};
/**
* arch_fast_hash - Caclulates a hash over a given buffer that can have
* arbitrary size. This function will eventually use an
* architecture-optimized hashing implementation if
* available, and trades off distribution for speed.
*
* @data: buffer to hash
* @len: length of buffer in bytes
* @seed: start seed
*
* Returns 32bit hash.
*/
extern u32 arch_fast_hash(const void *data, u32 len, u32 seed);
/**
* arch_fast_hash2 - Caclulates a hash over a given buffer that has a
* size that is of a multiple of 32bit words. This
* function will eventually use an architecture-
* optimized hashing implementation if available,
* and trades off distribution for speed.
*
* @data: buffer to hash (must be 32bit padded)
* @len: number of 32bit words
* @seed: start seed
*
* Returns 32bit hash.
*/
extern u32 arch_fast_hash2(const u32 *data, u32 len, u32 seed);
#endif /* _LINUX_HASH_H */