[Freedreno] [PATCH v3] drm/msm: validate display and event threads

[Jeykumar Sankaran]

On 2018-10-09 07:24, Sean Paul wrote:
> On Mon, Oct 08, 2018 at 04:55:45PM -0700, Jeykumar Sankaran wrote:
>> While creating display and event threads per crtc, validate
>> them before setting their priorities.
>> 
>> changes in v2:
>> 	- use dev_warn (Abhinav Kumar)
>> changes in v3:
>> 	- fix compilation error
>> 
>> Change-Id: I1dda805286df981c0f0e2b26507d089d3a21ff6c
> 
> No Change-Id's please!
> 
> FWIW, checkpatch.pl would have caught this and the nits I found below. 
> It
> might
> be worth running patches through it before sending.
> 
>> Signed-off-by: Jeykumar Sankaran <jsanka at codeaurora.org>
>> ---
>>  drivers/gpu/drm/msm/msm_drv.c | 49
> ++++++++++++++-----------------------------
>>  1 file changed, 16 insertions(+), 33 deletions(-)
>> 
>> diff --git a/drivers/gpu/drm/msm/msm_drv.c
> b/drivers/gpu/drm/msm/msm_drv.c
>> index 4904d0d..ab1b0a9 100644
>> --- a/drivers/gpu/drm/msm/msm_drv.c
>> +++ b/drivers/gpu/drm/msm/msm_drv.c
>> @@ -553,17 +553,18 @@ static int msm_drm_init(struct device *dev, 
>> struct
> drm_driver *drv)
>>  			kthread_run(kthread_worker_fn,
>>  				&priv->disp_thread[i].worker,
>>  				"crtc_commit:%d",
> priv->disp_thread[i].crtc_id);
>> -		ret = sched_setscheduler(priv->disp_thread[i].thread,
>> -							SCHED_FIFO,
> &param);
>> -		if (ret)
>> -			pr_warn("display thread priority update failed:
> %d\n",
>> -
> ret);
>> -
>>  		if (IS_ERR(priv->disp_thread[i].thread)) {
>>  			dev_err(dev, "failed to create crtc_commit
> kthread\n");
>>  			priv->disp_thread[i].thread = NULL;
>> +			goto err_msm_uninit;
>>  		}
>> 
>> +		ret = sched_setscheduler(priv->disp_thread[i].thread,
>> +					 SCHED_FIFO, &param);
>> +		if (ret)
>> +			dev_warn(dev, "display thread priority update
> failed: %d\n",
> 
> Although this is wrapped, the line still exceeds 80 chars. Perhaps:
> 
>                         dev_warn(dev, "disp_thread set priority failed
> %d\n",
>                                  ret);
> 
I did run the checkpath.pl on this patch. Check patch usually will not 
complain
if the >80 exceeding line is the logging string.

https://01.org/linuxgraphics/gfx-docs/drm/process/coding-style.html#breaking-long-lines-and-strings

https://gitlab.freedesktop.org/seanpaul/dpu-staging/blob/for-next/scripts/checkpatch.pl#L3058

Anyway, rewording will be a good idea to avoid nits!

>> +				ret);
>> +
>>  		/* initialize event thread */
>>  		priv->event_thread[i].crtc_id = priv->crtcs[i]->base.id;
>>  		kthread_init_worker(&priv->event_thread[i].worker);
>> @@ -572,6 +573,12 @@ static int msm_drm_init(struct device *dev, 
>> struct
> drm_driver *drv)
>>  			kthread_run(kthread_worker_fn,
>>  				&priv->event_thread[i].worker,
>>  				"crtc_event:%d",
> priv->event_thread[i].crtc_id);
>> +		if (IS_ERR(priv->event_thread[i].thread)) {
>> +			dev_err(dev, "failed to create crtc_event
> kthread\n");
>> +			priv->event_thread[i].thread = NULL;
>> +			goto err_msm_uninit;
>> +		}
>> +
>>  		/**
>>  		 * event thread should also run at same priority as
> disp_thread
>>  		 * because it is handling frame_done events. A lower
> priority
>> @@ -580,34 +587,10 @@ static int msm_drm_init(struct device *dev, 
>> struct
> drm_driver *drv)
>>  		 * failure at crtc commit level.
>>  		 */
>>  		ret = sched_setscheduler(priv->event_thread[i].thread,
>> -							SCHED_FIFO,
> &param);
>> +					 SCHED_FIFO, &param);
>>  		if (ret)
>> -			pr_warn("display event thread priority update
> failed: %d\n",
>> -
> ret);
>> -
>> -		if (IS_ERR(priv->event_thread[i].thread)) {
>> -			dev_err(dev, "failed to create crtc_event
> kthread\n");
>> -			priv->event_thread[i].thread = NULL;
>> -		}
>> -
>> -		if ((!priv->disp_thread[i].thread) ||
>> -				!priv->event_thread[i].thread) {
>> -			/* clean up previously created threads if any */
>> -			for ( ; i >= 0; i--) {
>> -				if (priv->disp_thread[i].thread) {
>> -					kthread_stop(
>> -
> priv->disp_thread[i].thread);
>> -					priv->disp_thread[i].thread =
> NULL;
>> -				}
>> -
>> -				if (priv->event_thread[i].thread) {
>> -					kthread_stop(
>> -
> priv->event_thread[i].thread);
>> -					priv->event_thread[i].thread =
> NULL;
>> -				}
>> -			}
>> -			goto err_msm_uninit;
>> -		}
>> +			dev_warn(dev, "display event thread priority
> update failed:%d\n",
> 
> Same problem here,
>                         dev_warn(dev, "event thread set priority failed
> %d\n",
> 
> 
>> +				ret);
>>  	}
>> 
>>  	ret = drm_vblank_init(ddev, priv->num_crtcs);
>> --
>> The Qualcomm Innovation Center, Inc. is a member of the Code Aurora
> Forum,
>> a Linux Foundation Collaborative Project
>> 

-- 
Jeykumar S