Re: [PATCH 4/5] drm/scheduler: Job timeout handler returns status (v2)

2020-12-07 Thread Luben Tuikov
On 2020-12-04 3:13 a.m., Christian König wrote: > Thinking more about that I came to the conclusion that the whole > approach here isn't correct. > > See even when the job has been completed or canceled we still want to > restart the timer. > > The reason for this is that the timer is then not

Re: [PATCH 4/5] drm/scheduler: Job timeout handler returns status (v2)

2020-12-07 Thread Andrey Grodzovsky
On 12/7/20 2:19 PM, Christian König wrote: Am 07.12.20 um 20:09 schrieb Andrey Grodzovsky: On 12/7/20 1:04 PM, Christian König wrote: Am 07.12.20 um 17:00 schrieb Andrey Grodzovsky: On 12/7/20 6:13 AM, Christian König wrote: Am 04.12.20 um 16:10 schrieb Andrey Grodzovsky: On 12/4/20

Re: [PATCH 4/5] drm/scheduler: Job timeout handler returns status (v2)

2020-12-07 Thread Christian König
Am 07.12.20 um 20:09 schrieb Andrey Grodzovsky: On 12/7/20 1:04 PM, Christian König wrote: Am 07.12.20 um 17:00 schrieb Andrey Grodzovsky: On 12/7/20 6:13 AM, Christian König wrote: Am 04.12.20 um 16:10 schrieb Andrey Grodzovsky: On 12/4/20 3:13 AM, Christian König wrote: Thinking more

Re: [PATCH 4/5] drm/scheduler: Job timeout handler returns status (v2)

2020-12-07 Thread Andrey Grodzovsky
On 12/7/20 1:04 PM, Christian König wrote: Am 07.12.20 um 17:00 schrieb Andrey Grodzovsky: On 12/7/20 6:13 AM, Christian König wrote: Am 04.12.20 um 16:10 schrieb Andrey Grodzovsky: On 12/4/20 3:13 AM, Christian König wrote: Thinking more about that I came to the conclusion that the whole

Re: [PATCH 4/5] drm/scheduler: Job timeout handler returns status (v2)

2020-12-07 Thread Christian König
Am 07.12.20 um 17:00 schrieb Andrey Grodzovsky: On 12/7/20 6:13 AM, Christian König wrote: Am 04.12.20 um 16:10 schrieb Andrey Grodzovsky: On 12/4/20 3:13 AM, Christian König wrote: Thinking more about that I came to the conclusion that the whole approach here isn't correct. See even when

Re: [PATCH 4/5] drm/scheduler: Job timeout handler returns status (v2)

2020-12-07 Thread Andrey Grodzovsky
On 12/7/20 6:13 AM, Christian König wrote: Am 04.12.20 um 16:10 schrieb Andrey Grodzovsky: On 12/4/20 3:13 AM, Christian König wrote: Thinking more about that I came to the conclusion that the whole approach here isn't correct. See even when the job has been completed or canceled we still

Re: [PATCH 4/5] drm/scheduler: Job timeout handler returns status (v2)

2020-12-07 Thread Christian König
Am 04.12.20 um 16:10 schrieb Andrey Grodzovsky: On 12/4/20 3:13 AM, Christian König wrote: Thinking more about that I came to the conclusion that the whole approach here isn't correct. See even when the job has been completed or canceled we still want to restart the timer. The reason for

Re: [PATCH 4/5] drm/scheduler: Job timeout handler returns status (v2)

2020-12-04 Thread Andrey Grodzovsky
On 12/4/20 3:13 AM, Christian König wrote: Thinking more about that I came to the conclusion that the whole approach here isn't correct. See even when the job has been completed or canceled we still want to restart the timer. The reason for this is that the timer is then not restarted for

Re: [PATCH 4/5] drm/scheduler: Job timeout handler returns status (v2)

2020-12-04 Thread Christian König
Thinking more about that I came to the conclusion that the whole approach here isn't correct. See even when the job has been completed or canceled we still want to restart the timer. The reason for this is that the timer is then not restarted for the current job, but for the next job in the

[PATCH 4/5] drm/scheduler: Job timeout handler returns status (v2)

2020-12-03 Thread Luben Tuikov
The driver's job timeout handler now returns status indicating back to the DRM layer whether the task (job) was successfully aborted or whether more time should be given to the task to complete. Default behaviour as of this patch, is preserved, except in obvious-by-comment case in the Panfrost