| Message ID | 0fec13b74998216009d338fad3e143a3488e0b4a.1471151622.git.bhaktipriya96@gmail.com (mailing list archive) |
|---|---|
| State | Accepted |
| Headers | show |
On Sun, Aug 14, 2016 at 10:55:30AM +0530, Bhaktipriya Shridhar wrote: > The workqueue "workqueue" has a single workitem(&priv->work) and hence > doesn't require ordering. Also, it is not being used on a memory reclaim > path. Hence, the singlethreaded workqueue has been replaced with the use > of system_wq. > > System workqueues have been able to handle high level of concurrency > for a long time now and hence it's not required to have a singlethreaded > workqueue just to gain concurrency. Unlike a dedicated per-cpu workqueue > created with create_singlethread_workqueue(), system_wq allows multiple > work items to overlap executions even on the same CPU; however, a > per-cpu workqueue doesn't have any CPU locality or global ordering > guarantee unless the target CPU is explicitly specified and thus the > increase of local concurrency shouldn't make any difference. > > Workitem is sync cancelled in mc13783_ts_remove() to ensure that there > are no workitems pending when the driver is disconnected. > > Signed-off-by: Bhaktipriya Shridhar <bhaktipriya96@gmail.com> Acked-by: Tejun Heo <tj@kernel.org> Thanks.
On Sun, Aug 14, 2016 at 10:55:30AM +0530, Bhaktipriya Shridhar wrote: > The workqueue "workqueue" has a single workitem(&priv->work) and hence > doesn't require ordering. Also, it is not being used on a memory reclaim > path. Hence, the singlethreaded workqueue has been replaced with the use > of system_wq. > > System workqueues have been able to handle high level of concurrency > for a long time now and hence it's not required to have a singlethreaded > workqueue just to gain concurrency. Unlike a dedicated per-cpu workqueue > created with create_singlethread_workqueue(), system_wq allows multiple > work items to overlap executions even on the same CPU; however, a > per-cpu workqueue doesn't have any CPU locality or global ordering > guarantee unless the target CPU is explicitly specified and thus the > increase of local concurrency shouldn't make any difference. > > Workitem is sync cancelled in mc13783_ts_remove() to ensure that there > are no workitems pending when the driver is disconnected. > > Signed-off-by: Bhaktipriya Shridhar <bhaktipriya96@gmail.com> Applied, thank you. But: > @@ -233,7 +222,7 @@ static int mc13783_ts_remove(struct platform_device *pdev) > { > struct mc13783_ts_priv *priv = platform_get_drvdata(pdev); > > - destroy_workqueue(priv->workq); > + cancel_delayed_work_sync(&priv->work); This is not needed, we cancel work in mc13783_ts_close(). I dropped this bit. > input_unregister_device(priv->idev); > kfree(priv); > > -- > 2.1.4 >
diff --git a/drivers/input/touchscreen/mc13783_ts.c b/drivers/input/touchscreen/mc13783_ts.c index 913e25a..3b19035 100644 --- a/drivers/input/touchscreen/mc13783_ts.c +++ b/drivers/input/touchscreen/mc13783_ts.c @@ -37,7 +37,6 @@ struct mc13783_ts_priv { struct input_dev *idev; struct mc13xxx *mc13xxx; struct delayed_work work; - struct workqueue_struct *workq; unsigned int sample[4]; struct mc13xxx_ts_platform_data *touch; }; @@ -54,7 +53,7 @@ static irqreturn_t mc13783_ts_handler(int irq, void *data) * be rescheduled for immediate execution here. However the rearm * delay is HZ / 50 which is acceptable. */ - queue_delayed_work(priv->workq, &priv->work, 0); + schedule_delayed_work(&priv->work, 0); return IRQ_HANDLED; } @@ -106,7 +105,7 @@ static void mc13783_ts_report_sample(struct mc13783_ts_priv *priv) dev_dbg(&idev->dev, "report (%d, %d, %d)\n", x1, y1, 0x1000 - cr0); - queue_delayed_work(priv->workq, &priv->work, HZ / 50); + schedule_delayed_work(&priv->work, HZ / 50); } else dev_dbg(&idev->dev, "report release\n"); @@ -189,14 +188,6 @@ static int __init mc13783_ts_probe(struct platform_device *pdev) goto err_free_mem; } - /* - * We need separate workqueue because mc13783_adc_do_conversion - * uses keventd and thus would deadlock. - */ - priv->workq = create_singlethread_workqueue("mc13783_ts"); - if (!priv->workq) - goto err_free_mem; - idev->name = MC13783_TS_NAME; idev->dev.parent = &pdev->dev; @@ -215,14 +206,12 @@ static int __init mc13783_ts_probe(struct platform_device *pdev) if (ret) { dev_err(&pdev->dev, "register input device failed with %d\n", ret); - goto err_destroy_wq; + goto err_free_mem; } platform_set_drvdata(pdev, priv); return 0; -err_destroy_wq: - destroy_workqueue(priv->workq); err_free_mem: input_free_device(idev); kfree(priv); @@ -233,7 +222,7 @@ static int mc13783_ts_remove(struct platform_device *pdev) { struct mc13783_ts_priv *priv = platform_get_drvdata(pdev); - destroy_workqueue(priv->workq); + cancel_delayed_work_sync(&priv->work); input_unregister_device(priv->idev); kfree(priv);
The workqueue "workqueue" has a single workitem(&priv->work) and hence doesn't require ordering. Also, it is not being used on a memory reclaim path. Hence, the singlethreaded workqueue has been replaced with the use of system_wq. System workqueues have been able to handle high level of concurrency for a long time now and hence it's not required to have a singlethreaded workqueue just to gain concurrency. Unlike a dedicated per-cpu workqueue created with create_singlethread_workqueue(), system_wq allows multiple work items to overlap executions even on the same CPU; however, a per-cpu workqueue doesn't have any CPU locality or global ordering guarantee unless the target CPU is explicitly specified and thus the increase of local concurrency shouldn't make any difference. Workitem is sync cancelled in mc13783_ts_remove() to ensure that there are no workitems pending when the driver is disconnected. Signed-off-by: Bhaktipriya Shridhar <bhaktipriya96@gmail.com> --- drivers/input/touchscreen/mc13783_ts.c | 19 ++++--------------- 1 file changed, 4 insertions(+), 15 deletions(-) -- 2.1.4 -- To unsubscribe from this list: send the line "unsubscribe linux-input" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html