Quoting Daniele Ceraolo Spurio (2019-07-13 17:44:32) > > > On 7/13/2019 3:00 AM, Chris Wilson wrote: > > We only employ a single task for log capture, and created a workqueue > > for the purpose of ensuring we had a high priority queue for low > > latency. We can simply use the system_highpri_wq and avoid the > > complication with creating our own admist the maze of mutexes. > > (Currently we create the wq early before we even know we need it in > > order to avoid trying to create it on demand while we hold the logging > > mutex.) > > Michal had suggested the same, but I wasn't sure about it since I have a > very vague recollection that on a very busy system the system_wq wasn't > low-latency enough at high verbosity levels with smaller buffer sizes. > However, the same could apply to a dedicated wq and we can now enable > bigger log buffers that should significantly reduce the amount of > flushes (and thus the latency requirements), so: We use the system_highpri_wq which should be no different from the wq we allocated for ourselves. If we need lower latency than that, we run into trouble :) -Chris _______________________________________________ Intel-gfx mailing list Intel-gfx@xxxxxxxxxxxxxxxxxxxxx https://lists.freedesktop.org/mailman/listinfo/intel-gfx
