| Message ID | 20160702082540.GA26570@Karyakshetra (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
On Sat, Jul 02, 2016 at 01:55:40PM +0530, Bhaktipriya Shridhar wrote: > The workqueue "workqueue" has only a single workitem(&ddata->ulps_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. > > dsicm_remove() calls dsicm_cancel_ulps_work which uses > cancel_delayed_work() to ensure that there are no pending tasks while > disconnecting the driver. > > Signed-off-by: Bhaktipriya Shridhar <bhaktipriya96@gmail.com> Looks good. Thanks.
On Sat, Jul 02, 2016 at 06:32:31AM -0500, Tejun Heo wrote: > On Sat, Jul 02, 2016 at 01:55:40PM +0530, Bhaktipriya Shridhar wrote: > > The workqueue "workqueue" has only a single workitem(&ddata->ulps_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. > > > > dsicm_remove() calls dsicm_cancel_ulps_work which uses > > cancel_delayed_work() to ensure that there are no pending tasks while > > disconnecting the driver. > > > > Signed-off-by: Bhaktipriya Shridhar <bhaktipriya96@gmail.com> > > Looks good. Thanks. Heh, should have been Acked-by: Tejun Heo <tj@kernel.org> Sorry about that. Thanks.
On 02/07/16 11:25, Bhaktipriya Shridhar wrote: > The workqueue "workqueue" has only a single workitem(&ddata->ulps_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. > > dsicm_remove() calls dsicm_cancel_ulps_work which uses > cancel_delayed_work() to ensure that there are no pending tasks while > disconnecting the driver. > > Signed-off-by: Bhaktipriya Shridhar <bhaktipriya96@gmail.com> > --- > drivers/video/fbdev/omap2/omapfb/displays/panel-dsi-cm.c | 14 ++------------ > 1 file changed, 2 insertions(+), 12 deletions(-) Thanks, queued for v4.9. Tomi
diff --git a/drivers/video/fbdev/omap2/omapfb/displays/panel-dsi-cm.c b/drivers/video/fbdev/omap2/omapfb/displays/panel-dsi-cm.c index 3414c26..281ac7e 100644 --- a/drivers/video/fbdev/omap2/omapfb/displays/panel-dsi-cm.c +++ b/drivers/video/fbdev/omap2/omapfb/displays/panel-dsi-cm.c @@ -76,8 +76,6 @@ struct panel_drv_data { bool intro_printed; - struct workqueue_struct *workqueue; - bool ulps_enabled; unsigned ulps_timeout; struct delayed_work ulps_work; @@ -233,7 +231,7 @@ static int dsicm_set_update_window(struct panel_drv_data *ddata, static void dsicm_queue_ulps_work(struct panel_drv_data *ddata) { if (ddata->ulps_timeout > 0) - queue_delayed_work(ddata->workqueue, &ddata->ulps_work, + schedule_delayed_work(&ddata->ulps_work, msecs_to_jiffies(ddata->ulps_timeout)); } @@ -1284,11 +1282,6 @@ static int dsicm_probe(struct platform_device *pdev) dev_dbg(dev, "Using GPIO TE\n"); } - ddata->workqueue = create_singlethread_workqueue("dsicm_wq"); - if (ddata->workqueue == NULL) { - dev_err(dev, "can't create workqueue\n"); - return -ENOMEM; - } INIT_DELAYED_WORK(&ddata->ulps_work, dsicm_ulps_work); dsicm_hw_reset(ddata); @@ -1302,7 +1295,7 @@ static int dsicm_probe(struct platform_device *pdev) dev, ddata, &dsicm_bl_ops, &props); if (IS_ERR(bldev)) { r = PTR_ERR(bldev); - goto err_bl; + goto err_reg; } ddata->bldev = bldev; @@ -1325,8 +1318,6 @@ static int dsicm_probe(struct platform_device *pdev) err_sysfs_create: if (bldev != NULL) backlight_device_unregister(bldev); -err_bl: - destroy_workqueue(ddata->workqueue); err_reg: return r; } @@ -1356,7 +1347,6 @@ static int __exit dsicm_remove(struct platform_device *pdev) omap_dss_put_device(ddata->in); dsicm_cancel_ulps_work(ddata); - destroy_workqueue(ddata->workqueue); /* reset, to be sure that the panel is in a valid state */ dsicm_hw_reset(ddata);
The workqueue "workqueue" has only a single workitem(&ddata->ulps_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. dsicm_remove() calls dsicm_cancel_ulps_work which uses cancel_delayed_work() to ensure that there are no pending tasks while disconnecting the driver. Signed-off-by: Bhaktipriya Shridhar <bhaktipriya96@gmail.com> --- drivers/video/fbdev/omap2/omapfb/displays/panel-dsi-cm.c | 14 ++------------ 1 file changed, 2 insertions(+), 12 deletions(-) -- 2.1.4 -- To unsubscribe from this list: send the line "unsubscribe linux-fbdev" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html