On Wed, 2009-11-11 at 21:51 +0100, Rafał Miłecki wrote: > I attach my very hackish patch to downclock GPU. > > Could someone tell me please, why first method in this patch works > while second locks up my GPU? > > I've run > cat radeon_pm_info > 100x times with first method. No problem. Second method fails on every > try, checked it with 5 reboots. > > Jerome told sth about engine idle, but does it really matter? Why then > first method works fine every time? I don't really believe that is the > real issue.
Does calling radeon_pm_reclock 2 times in a row leads to lockup too ? radeon_pm_reclock((unsigned long)rdev); radeon_pm_reclock((unsigned long)rdev); If so then it's most likely because with the timer engine didn't had enough time to settle after reclock. Sadly atombios seems to be too much conservative about that so we might need to find how to figure if engine is settle at the new frequency. The need for idle comes from AMD documentation and it makes sense, you don't want to change clock while the GPU is doing somethings. Cheers, Jerome Glisse ------------------------------------------------------------------------------ Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day trial. Simplify your report design, integration and deployment - and focus on what you do best, core application coding. Discover what's new with Crystal Reports now. http://p.sf.net/sfu/bobj-july -- _______________________________________________ Dri-devel mailing list Dri-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/dri-devel
