On Wed, 30 Nov 2005 13:53:13 -0500, Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>
wrote:
>The way the attached patch attacks this is for the shared-lock access
>case to simply set the page's LRU counter to zero, without bumping up
>the LRU counters of the other pages as the normal adjustment would do.

If you still plan to do this, you might also want to revert the
micro-optimisation intruduced by the original SLRU patch:

| Apart from refactoring I made a little change to SlruRecentlyUsed,
| formerly ClogRecentlyUsed:  It now skips incrementing lru_counts, if
| slotno is already the LRU slot, thus saving a few CPU cycles.

|+#define SlruRecentlyUsed(shared, slotno)	\
|+	do { \
|+		if ((shared)->page_lru_count[slotno] != 0) { \
|+			int		iilru; \
|+			for (iilru = 0; iilru < NUM_CLOG_BUFFERS; iilru++) \
|+				(shared)->page_lru_count[iilru]++; \
|+			(shared)->page_lru_count[slotno] = 0; \
|+		} \
|+	} while (0)

Otherwise you could end up with a stable state of several pages having
lru_count == 0.
Servus
 Manfred

In response to

• Shared locking in slru.c at 2005-11-30 18:53:13 from Tom Lane

Responses

• Re: Shared locking in slru.c at 2005-12-02 14:18:01 from Tom Lane

pgsql-hackers by date

Next:	From: Anuj Tripathi	Date: 2005-12-02 10:45:55
	Subject: Graphics in postgress using GTK
Previous:	From: Csaba Nagy	Date: 2005-12-02 10:07:06
	Subject: Re: generalizing the planner knobs