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

2017-07-24 Thread John Ferlan
On 07/24/2017 02:33 PM, Pavel Hrdina wrote: >> EXACTLY! This is why I started down this path... Of course I want far >> too generic for some people's preferences by going with primaryKey and >> secondaryKey type nomenclature, so I was forced by review to use >> UUID/Name which are far less gene

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

2017-07-24 Thread Pavel Hrdina
On Mon, Jul 24, 2017 at 11:12:01AM -0400, John Ferlan wrote: > > > On 07/24/2017 09:04 AM, Michal Privoznik wrote: > > On 07/23/2017 11:33 PM, John Ferlan wrote: > >> > >> > >> On 07/23/2017 04:46 PM, Michal Privoznik wrote: > >>> On 07/23/2017 07:21 PM, Pavel Hrdina wrote: > On Sun, Jul 23,

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

2017-07-24 Thread John Ferlan
On 07/24/2017 09:04 AM, Michal Privoznik wrote: > On 07/23/2017 11:33 PM, John Ferlan wrote: >> >> >> On 07/23/2017 04:46 PM, Michal Privoznik wrote: >>> On 07/23/2017 07:21 PM, Pavel Hrdina wrote: On Sun, Jul 23, 2017 at 02:33:49PM +0200, Michal Privoznik wrote: > On 07/23/2017 02:10 PM

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

2017-07-24 Thread Michal Privoznik
On 07/23/2017 11:33 PM, John Ferlan wrote: > > > On 07/23/2017 04:46 PM, Michal Privoznik wrote: >> On 07/23/2017 07:21 PM, Pavel Hrdina wrote: >>> 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:

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

2017-07-23 Thread Pavel Hrdina
On Sun, Jul 23, 2017 at 10:46:15PM +0200, Michal Privoznik wrote: > On 07/23/2017 07:21 PM, Pavel Hrdina wrote: > > 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: > >

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

2017-07-23 Thread John Ferlan
On 07/23/2017 04:46 PM, Michal Privoznik wrote: > On 07/23/2017 07:21 PM, Pavel Hrdina wrote: >> 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

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

2017-07-23 Thread Michal Privoznik
On 07/23/2017 07:21 PM, Pavel Hrdina wrote: > 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 >

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

2017-07-23 Thread Pavel Hrdina
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 makin

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

2017-07-23 Thread Michal Privoznik
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

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

2017-07-23 Thread John Ferlan
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 Priv

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

2017-07-19 Thread Michal Privoznik
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