FW: ITS#2183 IDL stack slab

["Howard Chu" <hyc@highlandsun.com>]

This patch has quite a dramatic effect. In a test with only read operations,
test008 run time drops from ~5.6 seconds to ~5.1 seconds. In a test with
read/write mix, the test008 runtime drops from ~22 seconds to as fast as ~11
seconds. At the moment the buffer size is hardcoded at 16 but again, we may
want to make it configurable by a config keyword.

  -- Howard Chu
  Chief Architect, Symas Corp.       Director, Highland Sun
  http://www.symas.com               http://highlandsun.com/hyc
  Symas: Premier OpenSource Development and Support

-----Original Message-----
From: owner-openldap-bugs@OpenLDAP.org
[mailto:owner-openldap-bugs@OpenLDAP.org]On Behalf Of
hyc@highlandsun.com
Sent: Tuesday, December 10, 2002 8:12 AM
To: openldap-its@OpenLDAP.org
Subject: ITS#2183 IDL stack slab


Thinking about this some more; it can be greatly simplified. Instead of
having a sequence of buckets of increasing length, just use a single
fixed-size buffer per thread. This would also avoid the need for any locking
on the structures. (See the use of ldap_pvt_thread_pool_getkey/setkey in
cache.c.) I can code this up pretty quickly for a test.

  -- Howard Chu
  Chief Architect, Symas Corp.       Director, Highland Sun
  http://www.symas.com               http://highlandsun.com/hyc
  Symas: Premier OpenSource Development and Support