github/ocf
4
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Optimize set_hot calls in eviction path

Browse Source 
Split traversation into two distinct phases: lookup()
and lru set_hot(). prepare_cachelines() now only calls
set_hot() once after lookup and insert are finished.
lookup() is called explicitly only once in
prepare_cachelines() at the very beginning of the
procedure. If request is a miss then then map()
performs operations equivalent to lookup() supplemented
by an attempt to map cachelines. Both lookup() and
set_hot() are called via traverse() from the engines
which do not attempt mapping, and thus do not call
prepare_clines().

Signed-off-by: Adam Rutkowski <adam.j.rutkowski@intel.com>
This commit is contained in:
Adam Rutkowski
2021-03-05 11:20:47 +01:00
parent 1c6168d82b
commit 7927b0b74f
2 changed files with 37 additions and 10 deletions
Download Patch File Download Diff File Expand all files Collapse all files

8
src/concurrency/ocf_metadata_concurrency.c
View File

@@ -423,7 +423,12 @@ void ocf_hb_id_prot_unlock_wr(struct ocf_metadata_lock *metadata_lock,
for (hash = _MIN_HASH(req); hash <= _MAX_HASH(req); \
hash = _HASH_NEXT(req, hash))
/* returns true if given core hash collides with request LBA range */
/* Returns true if the the given LBA (determined by core_id
* and core_line) resolves to a hash value that is within the
* set of hashes for the given request (i.e. after the request
* hash bucket are locked, the given core line is hash bucket
* locked as well).
*/
bool ocf_req_hash_in_range(struct ocf_request *req,
ocf_core_id_t core_id, uint64_t core_line)
{
@@ -438,7 +443,6 @@ bool ocf_req_hash_in_range(struct ocf_request *req,
return (hash >= _MIN_HASH(req) && hash <= _GAP_START(req)) ||
(hash > _GAP_START(req) + _GAP_VAL(req) &&
hash <= _MAX_HASH(req));
}
void ocf_hb_req_prot_lock_rd(struct ocf_request *req)