From: | Alvaro Herrera <alvherre(at)commandprompt(dot)com> |
---|---|
To: | Hackers <pgsql-hackers(at)postgresql(dot)org> |
Subject: | Re: fixing dllist? |
Date: | 2007-03-22 14:09:08 |
Message-ID: | 20070322140908.GF4102@alvh.no-ip.org |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-hackers |
Another change that could be done to Dllist is removing the Dllist
pointer from the Dlelem struct:
Index: src/include/lib/dllist.h
===================================================================
RCS file: /home/alvherre/Code/cvs/pgsql/src/include/lib/dllist.h,v
retrieving revision 1.27
diff -c -p -r1.27 dllist.h
*** src/include/lib/dllist.h 5 Jan 2007 22:19:55 -0000 1.27
--- src/include/lib/dllist.h 18 Mar 2007 05:53:12 -0000
*************** typedef struct Dlelem
*** 50,56 ****
struct Dlelem *dle_next; /* next element */
struct Dlelem *dle_prev; /* previous element */
void *dle_val; /* value of the element */
- struct Dllist *dle_list; /* what list this element is in */
} Dlelem;
typedef struct Dllist
--- 49,54 ----
This means that to remove a element from a list or move it to the front of the
list, you need not only know the element pointer itself, but also the list
pointer. This capability is not used much however; the only patch of any
significance needed is below. The rest of the callers know the list pointer
already.
I tried to measure a performance difference with pgbench (using a test
small enough to fit in memory, fsync off and initialized with -s 10,
test runs with -c 5) but the differences seem to be way down in the
noise.
Are there objections to this change?
Index: src/backend/utils/cache/catcache.c
===================================================================
RCS file: /home/alvherre/Code/cvs/pgsql/src/backend/utils/cache/catcache.c,v
retrieving revision 1.136
diff -c -p -r1.136 catcache.c
*** src/backend/utils/cache/catcache.c 5 Jan 2007 22:19:42 -0000 1.136
--- src/backend/utils/cache/catcache.c 18 Mar 2007 06:38:23 -0000
*************** CatCachePrintStats(int code, Datum arg)
*** 327,332 ****
--- 327,335 ----
static void
CatCacheRemoveCTup(CatCache *cache, CatCTup *ct)
{
+ uint32 hashValue;
+ Index hashIndex;
+
Assert(ct->refcount == 0);
Assert(ct->my_cache == cache);
*************** CatCacheRemoveCTup(CatCache *cache, CatC
*** 343,349 ****
}
/* delink from linked list */
! DLRemove(&ct->cache_elem);
/* free associated tuple data */
if (ct->tuple.t_data != NULL)
--- 346,354 ----
}
/* delink from linked list */
! hashValue = CatalogCacheComputeTupleHashValue(cache, &ct->tuple);
! hashIndex = HASH_INDEX(hashValue, cache->cc_nbuckets);
! DLRemove(&cache->cc_bucket[hashIndex], &ct->cache_elem);
/* free associated tuple data */
if (ct->tuple.t_data != NULL)
--
Alvaro Herrera http://www.CommandPrompt.com/
PostgreSQL Replication, Consulting, Custom Development, 24x7 support
From | Date | Subject | |
---|---|---|---|
Next Message | Merlin Moncure | 2007-03-22 14:10:50 | Re: Function cache regeneration |
Previous Message | Merlin Moncure | 2007-03-22 13:25:04 | Re: CREATE INDEX and HOT - revised design |