| Message ID | 20250226-uvc-granpower-ng-v4-4-3ec9be906048@chromium.org |
|---|---|
| State | Superseded |
| Headers | show |
| Series | media: uvcvideo: Implement Granular Power Saving | expand |
Hi, On 26-Feb-25 15:23, Ricardo Ribalda wrote: > Now that every ioctl takes care of their power management we can remove > the "global" power management. > > Despite its size, this is a relatively big change. We hope that there > are no size effects of it. If there are some specific devices that > miss-behave, we can add a small quirk for them. > > This patch introduces a behavioral change for the uvc "trigger" button. > It will not work unless the camera is streaming. We consider that this > the most common (if not the only) usecase and therefore we do not consider > it a regression. You may want to expand this part of the commit message a bit and make it clear that before the "trigger" button would work as long as an open has /dev/video# open and now it only works when the camera is actually streaming. I agree that in practice this is not an issue since apps only open the /dev/video# node for a longer period at a time (vs short open for query/probe) when they actually want to stream. Otherwise the patch looks good to me: Reviewed-by: Hans de Goede <hdegoede@redhat.com> Regards, Hans > > Signed-off-by: Ricardo Ribalda <ribalda@chromium.org> > --- > drivers/media/usb/uvc/uvc_v4l2.c | 8 -------- > 1 file changed, 8 deletions(-) > > diff --git a/drivers/media/usb/uvc/uvc_v4l2.c b/drivers/media/usb/uvc/uvc_v4l2.c > index 1c9ac72be58a..6af93e00b304 100644 > --- a/drivers/media/usb/uvc/uvc_v4l2.c > +++ b/drivers/media/usb/uvc/uvc_v4l2.c > @@ -652,7 +652,6 @@ static int uvc_v4l2_open(struct file *file) > { > struct uvc_streaming *stream; > struct uvc_fh *handle; > - int ret = 0; > > stream = video_drvdata(file); > uvc_dbg(stream->dev, CALLS, "%s\n", __func__); > @@ -662,12 +661,6 @@ static int uvc_v4l2_open(struct file *file) > if (!handle) > return -ENOMEM; > > - ret = uvc_pm_get(stream->dev); > - if (ret) { > - kfree(handle); > - return ret; > - } > - > v4l2_fh_init(&handle->vfh, &stream->vdev); > v4l2_fh_add(&handle->vfh); > handle->chain = stream->chain; > @@ -701,7 +694,6 @@ static int uvc_v4l2_release(struct file *file) > kfree(handle); > file->private_data = NULL; > > - uvc_pm_put(stream->dev); > return 0; > } > >
diff --git a/drivers/media/usb/uvc/uvc_v4l2.c b/drivers/media/usb/uvc/uvc_v4l2.c index 1c9ac72be58a..6af93e00b304 100644 --- a/drivers/media/usb/uvc/uvc_v4l2.c +++ b/drivers/media/usb/uvc/uvc_v4l2.c @@ -652,7 +652,6 @@ static int uvc_v4l2_open(struct file *file) { struct uvc_streaming *stream; struct uvc_fh *handle; - int ret = 0; stream = video_drvdata(file); uvc_dbg(stream->dev, CALLS, "%s\n", __func__); @@ -662,12 +661,6 @@ static int uvc_v4l2_open(struct file *file) if (!handle) return -ENOMEM; - ret = uvc_pm_get(stream->dev); - if (ret) { - kfree(handle); - return ret; - } - v4l2_fh_init(&handle->vfh, &stream->vdev); v4l2_fh_add(&handle->vfh); handle->chain = stream->chain; @@ -701,7 +694,6 @@ static int uvc_v4l2_release(struct file *file) kfree(handle); file->private_data = NULL; - uvc_pm_put(stream->dev); return 0; }
Now that every ioctl takes care of their power management we can remove the "global" power management. Despite its size, this is a relatively big change. We hope that there are no size effects of it. If there are some specific devices that miss-behave, we can add a small quirk for them. This patch introduces a behavioral change for the uvc "trigger" button. It will not work unless the camera is streaming. We consider that this the most common (if not the only) usecase and therefore we do not consider it a regression. Signed-off-by: Ricardo Ribalda <ribalda@chromium.org> --- drivers/media/usb/uvc/uvc_v4l2.c | 8 -------- 1 file changed, 8 deletions(-)