From mboxrd@z Thu Jan 1 00:00:00 1970 From: Zdenek Kabelac Date: Sun, 30 Jan 2011 21:54:11 +0100 Subject: [PATCH 12/24] Perf: Increase hash sizes In-Reply-To: <4D45CB78.8080208@redhat.com> References: <2483686e0452fbdc01bccd983c0d85672fc400ee.1296391340.git.zkabelac@redhat.com> <4D45CB78.8080208@redhat.com> Message-ID: <4D45CFF3.6000108@redhat.com> List-Id: To: lvm-devel@redhat.com MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit Dne 30.1.2011 21:35, Milan Broz napsal(a): > On 01/30/2011 01:57 PM, Zdenek Kabelac wrote: >> Modify size of created hash tables in a way, it could be uniquely >> identified in debug code - as hash will be rouned up to nearest power > > code obfuscation :) can we better add new has function or macro > which takes some identifier? (similar to pool create) > > Why do you need such hacks? Well I've been thinking about few solutions - for this - so take this just as an initial 'startup' - whether we want to extend API with some new function with string parameter for hash description - or as some 'debug - like' version, which is compiled only with some debug option enabled. Meanwhile - this was rather very easy 'quick and see' - as we don't have named hashes, like we have for pools. This hash part needs some polishing. Zdenek