Re: [libvirt] [PATCH 0/3] Introduce RW locks to virDomainObjList

On Sun, Jul 23, 2017 at 02:33:49PM +0200, Michal Privoznik wrote: > On 07/23/2017 02:10 PM, John Ferlan wrote: >> >> >> On 07/19/2017 10:31 AM, Michal Privoznik wrote: >>> While this is not that critical (hash tables have O(1) time complexity for >>> lookups), it lays down path towards making virDomainObj using RW locks instead >>> of mutexes. Still, in my limited testing this showed slight improvement. >>> >>> Michal Privoznik (3): >>> virthread: Introduce virRWLockInitPreferWriter >>> virobject: Introduce virObjectRWLockable >>> virdomainobjlist: Use virObjectRWLockable >>> >>> src/conf/virdomainobjlist.c | 24 ++++---- >>> src/libvirt_private.syms | 4 ++ >>> src/util/virobject.c | 144 ++++++++++++++++++++++++++++++++++---------- >>> src/util/virobject.h | 16 +++++ >>> src/util/virthread.c | 35 +++++++++++ >>> src/util/virthread.h | 1 + >>> 6 files changed, 180 insertions(+), 44 deletions(-) >>> >> >> This could be a "next step" in the work I've been doing towards a common >> object: > > Sure. If we have just one common object the change can be done in one > place instead of many. I don't care in what order are the changes merged. I'm still not sure about the implementation that you are heading to,

I would actually prefer something similar to the current virDomainObjList implementation, create a new module in utils called virObjectList and make it somehow generic that it could be used by our objects. I personally don't like the fact that there will be two new classes, one that enables using the other one.