* [PATCH] v4l2-core/v4l2-ctrls: Filter NOOP CH_RANGE events @ 2015-11-11 11:58 Ricardo Ribalda Delgado 2015-11-11 12:11 ` Hans Verkuil 0 siblings, 1 reply; 4+ messages in thread From: Ricardo Ribalda Delgado @ 2015-11-11 11:58 UTC (permalink / raw) To: Hans Verkuil, Mauro Carvalho Chehab, linux-media, linux-kernel, Dimitrios Katsaros Cc: Ricardo Ribalda Delgado If modify_range is called but no range is changed, do not send the CH_RANGE event. Reported-by: Dimitrios Katsaros <patcherwork@gmail.com> Signed-off-by: Ricardo Ribalda Delgado <ricardo.ribalda@gmail.com> --- drivers/media/v4l2-core/v4l2-ctrls.c | 23 ++++++++++++++--------- 1 file changed, 14 insertions(+), 9 deletions(-) diff --git a/drivers/media/v4l2-core/v4l2-ctrls.c b/drivers/media/v4l2-core/v4l2-ctrls.c index 4a1d9fdd14bb..f9c0e8150bd1 100644 --- a/drivers/media/v4l2-core/v4l2-ctrls.c +++ b/drivers/media/v4l2-core/v4l2-ctrls.c @@ -3300,7 +3300,8 @@ EXPORT_SYMBOL(v4l2_ctrl_notify); int __v4l2_ctrl_modify_range(struct v4l2_ctrl *ctrl, s64 min, s64 max, u64 step, s64 def) { - bool changed; + bool value_changed; + bool range_changed = false; int ret; lockdep_assert_held(ctrl->handler->lock); @@ -3324,10 +3325,14 @@ int __v4l2_ctrl_modify_range(struct v4l2_ctrl *ctrl, default: return -EINVAL; } - ctrl->minimum = min; - ctrl->maximum = max; - ctrl->step = step; - ctrl->default_value = def; + if ((ctrl->minimum != min) || (ctrl->maximum != max) || + (ctrl->step != step) || ctrl->default_value != def) { + range_changed = true; + ctrl->minimum = min; + ctrl->maximum = max; + ctrl->step = step; + ctrl->default_value = def; + } cur_to_new(ctrl); if (validate_new(ctrl, ctrl->p_new)) { if (ctrl->type == V4L2_CTRL_TYPE_INTEGER64) @@ -3337,12 +3342,12 @@ int __v4l2_ctrl_modify_range(struct v4l2_ctrl *ctrl, } if (ctrl->type == V4L2_CTRL_TYPE_INTEGER64) - changed = *ctrl->p_new.p_s64 != *ctrl->p_cur.p_s64; + value_changed = *ctrl->p_new.p_s64 != *ctrl->p_cur.p_s64; else - changed = *ctrl->p_new.p_s32 != *ctrl->p_cur.p_s32; - if (changed) + value_changed = *ctrl->p_new.p_s32 != *ctrl->p_cur.p_s32; + if (value_changed) ret = set_ctrl(NULL, ctrl, V4L2_EVENT_CTRL_CH_RANGE); - else + else if (range_changed) send_event(NULL, ctrl, V4L2_EVENT_CTRL_CH_RANGE); return ret; } -- 2.6.2 ^ permalink raw reply related [flat|nested] 4+ messages in thread
* Re: [PATCH] v4l2-core/v4l2-ctrls: Filter NOOP CH_RANGE events 2015-11-11 11:58 [PATCH] v4l2-core/v4l2-ctrls: Filter NOOP CH_RANGE events Ricardo Ribalda Delgado @ 2015-11-11 12:11 ` Hans Verkuil 2015-11-11 12:20 ` Ricardo Ribalda Delgado 0 siblings, 1 reply; 4+ messages in thread From: Hans Verkuil @ 2015-11-11 12:11 UTC (permalink / raw) To: Ricardo Ribalda Delgado, Hans Verkuil, Mauro Carvalho Chehab, linux-media, linux-kernel, Dimitrios Katsaros On 11/11/15 12:58, Ricardo Ribalda Delgado wrote: > If modify_range is called but no range is changed, do not send the > CH_RANGE event. While not opposed to this patch, I do wonder what triggered this patch? Is it just a matter of efficiency? And since it is a driver that calls this, shouldn't the driver only call this function when something actually changes? In other words, can you give some background information? Regards, Hans PS: still haven't processed your V4L2_CTRL_WHICH_DEF_VAL patch series. Hope to do this next week at the latest. > > Reported-by: Dimitrios Katsaros <patcherwork@gmail.com> > Signed-off-by: Ricardo Ribalda Delgado <ricardo.ribalda@gmail.com> > --- > drivers/media/v4l2-core/v4l2-ctrls.c | 23 ++++++++++++++--------- > 1 file changed, 14 insertions(+), 9 deletions(-) > > diff --git a/drivers/media/v4l2-core/v4l2-ctrls.c b/drivers/media/v4l2-core/v4l2-ctrls.c > index 4a1d9fdd14bb..f9c0e8150bd1 100644 > --- a/drivers/media/v4l2-core/v4l2-ctrls.c > +++ b/drivers/media/v4l2-core/v4l2-ctrls.c > @@ -3300,7 +3300,8 @@ EXPORT_SYMBOL(v4l2_ctrl_notify); > int __v4l2_ctrl_modify_range(struct v4l2_ctrl *ctrl, > s64 min, s64 max, u64 step, s64 def) > { > - bool changed; > + bool value_changed; > + bool range_changed = false; > int ret; > > lockdep_assert_held(ctrl->handler->lock); > @@ -3324,10 +3325,14 @@ int __v4l2_ctrl_modify_range(struct v4l2_ctrl *ctrl, > default: > return -EINVAL; > } > - ctrl->minimum = min; > - ctrl->maximum = max; > - ctrl->step = step; > - ctrl->default_value = def; > + if ((ctrl->minimum != min) || (ctrl->maximum != max) || > + (ctrl->step != step) || ctrl->default_value != def) { > + range_changed = true; > + ctrl->minimum = min; > + ctrl->maximum = max; > + ctrl->step = step; > + ctrl->default_value = def; > + } > cur_to_new(ctrl); > if (validate_new(ctrl, ctrl->p_new)) { > if (ctrl->type == V4L2_CTRL_TYPE_INTEGER64) > @@ -3337,12 +3342,12 @@ int __v4l2_ctrl_modify_range(struct v4l2_ctrl *ctrl, > } > > if (ctrl->type == V4L2_CTRL_TYPE_INTEGER64) > - changed = *ctrl->p_new.p_s64 != *ctrl->p_cur.p_s64; > + value_changed = *ctrl->p_new.p_s64 != *ctrl->p_cur.p_s64; > else > - changed = *ctrl->p_new.p_s32 != *ctrl->p_cur.p_s32; > - if (changed) > + value_changed = *ctrl->p_new.p_s32 != *ctrl->p_cur.p_s32; > + if (value_changed) > ret = set_ctrl(NULL, ctrl, V4L2_EVENT_CTRL_CH_RANGE); > - else > + else if (range_changed) > send_event(NULL, ctrl, V4L2_EVENT_CTRL_CH_RANGE); > return ret; > } > ^ permalink raw reply [flat|nested] 4+ messages in thread
* Re: [PATCH] v4l2-core/v4l2-ctrls: Filter NOOP CH_RANGE events 2015-11-11 12:11 ` Hans Verkuil @ 2015-11-11 12:20 ` Ricardo Ribalda Delgado 2015-11-11 12:33 ` Hans Verkuil 0 siblings, 1 reply; 4+ messages in thread From: Ricardo Ribalda Delgado @ 2015-11-11 12:20 UTC (permalink / raw) To: Hans Verkuil Cc: Hans Verkuil, Mauro Carvalho Chehab, linux-media, LKML, Dimitrios Katsaros Hello Hans On my cameras, the framerate can be controlled in very fine steps (so you can be in sync with the conveyor belt). The exposure time is dependent on the framerate. I was calling modify_range every time there was a change on the framerate on all the different heads. Leading to invalid/noop events. I have already fixed it on my sensor code, but there is a lot of copy paste.... I thought that check belong to the framework. It also will give some consistency, because If my memory is good, set_ctrl does not send and event if the value is unchanged. Thanks :) And take as much time as you need for V4L2_CTRL_WHICH_DEF_VAL :-) On Wed, Nov 11, 2015 at 1:11 PM, Hans Verkuil <hverkuil@xs4all.nl> wrote: > On 11/11/15 12:58, Ricardo Ribalda Delgado wrote: >> If modify_range is called but no range is changed, do not send the >> CH_RANGE event. > > While not opposed to this patch, I do wonder what triggered this patch? > Is it just a matter of efficiency? And since it is a driver that calls > this, shouldn't the driver only call this function when something > actually changes? > > In other words, can you give some background information? > > Regards, > > Hans > > PS: still haven't processed your V4L2_CTRL_WHICH_DEF_VAL patch series. Hope > to do this next week at the latest. > >> >> Reported-by: Dimitrios Katsaros <patcherwork@gmail.com> >> Signed-off-by: Ricardo Ribalda Delgado <ricardo.ribalda@gmail.com> >> --- >> drivers/media/v4l2-core/v4l2-ctrls.c | 23 ++++++++++++++--------- >> 1 file changed, 14 insertions(+), 9 deletions(-) >> >> diff --git a/drivers/media/v4l2-core/v4l2-ctrls.c b/drivers/media/v4l2-core/v4l2-ctrls.c >> index 4a1d9fdd14bb..f9c0e8150bd1 100644 >> --- a/drivers/media/v4l2-core/v4l2-ctrls.c >> +++ b/drivers/media/v4l2-core/v4l2-ctrls.c >> @@ -3300,7 +3300,8 @@ EXPORT_SYMBOL(v4l2_ctrl_notify); >> int __v4l2_ctrl_modify_range(struct v4l2_ctrl *ctrl, >> s64 min, s64 max, u64 step, s64 def) >> { >> - bool changed; >> + bool value_changed; >> + bool range_changed = false; >> int ret; >> >> lockdep_assert_held(ctrl->handler->lock); >> @@ -3324,10 +3325,14 @@ int __v4l2_ctrl_modify_range(struct v4l2_ctrl *ctrl, >> default: >> return -EINVAL; >> } >> - ctrl->minimum = min; >> - ctrl->maximum = max; >> - ctrl->step = step; >> - ctrl->default_value = def; >> + if ((ctrl->minimum != min) || (ctrl->maximum != max) || >> + (ctrl->step != step) || ctrl->default_value != def) { >> + range_changed = true; >> + ctrl->minimum = min; >> + ctrl->maximum = max; >> + ctrl->step = step; >> + ctrl->default_value = def; >> + } >> cur_to_new(ctrl); >> if (validate_new(ctrl, ctrl->p_new)) { >> if (ctrl->type == V4L2_CTRL_TYPE_INTEGER64) >> @@ -3337,12 +3342,12 @@ int __v4l2_ctrl_modify_range(struct v4l2_ctrl *ctrl, >> } >> >> if (ctrl->type == V4L2_CTRL_TYPE_INTEGER64) >> - changed = *ctrl->p_new.p_s64 != *ctrl->p_cur.p_s64; >> + value_changed = *ctrl->p_new.p_s64 != *ctrl->p_cur.p_s64; >> else >> - changed = *ctrl->p_new.p_s32 != *ctrl->p_cur.p_s32; >> - if (changed) >> + value_changed = *ctrl->p_new.p_s32 != *ctrl->p_cur.p_s32; >> + if (value_changed) >> ret = set_ctrl(NULL, ctrl, V4L2_EVENT_CTRL_CH_RANGE); >> - else >> + else if (range_changed) >> send_event(NULL, ctrl, V4L2_EVENT_CTRL_CH_RANGE); >> return ret; >> } >> -- Ricardo Ribalda ^ permalink raw reply [flat|nested] 4+ messages in thread
* Re: [PATCH] v4l2-core/v4l2-ctrls: Filter NOOP CH_RANGE events 2015-11-11 12:20 ` Ricardo Ribalda Delgado @ 2015-11-11 12:33 ` Hans Verkuil 0 siblings, 0 replies; 4+ messages in thread From: Hans Verkuil @ 2015-11-11 12:33 UTC (permalink / raw) To: Ricardo Ribalda Delgado, Hans Verkuil Cc: Hans Verkuil, Mauro Carvalho Chehab, linux-media, LKML, Dimitrios Katsaros On 11/11/15 13:20, Ricardo Ribalda Delgado wrote: > Hello Hans > > On my cameras, the framerate can be controlled in very fine steps (so > you can be in sync with the conveyor belt). The exposure time is > dependent on the framerate. I was calling modify_range every time > there was a change on the framerate on all the different heads. > Leading to invalid/noop events. > > I have already fixed it on my sensor code, but there is a lot of copy > paste.... I thought that check belong to the framework. It also will > give some consistency, because If my memory is good, set_ctrl does not > send and event if the value is unchanged. Ah, OK. That makes sense. Thanks for the explanation. Hans > > > Thanks :) > > And take as much time as you need for V4L2_CTRL_WHICH_DEF_VAL :-) > > > > On Wed, Nov 11, 2015 at 1:11 PM, Hans Verkuil <hverkuil@xs4all.nl> wrote: >> On 11/11/15 12:58, Ricardo Ribalda Delgado wrote: >>> If modify_range is called but no range is changed, do not send the >>> CH_RANGE event. >> >> While not opposed to this patch, I do wonder what triggered this patch? >> Is it just a matter of efficiency? And since it is a driver that calls >> this, shouldn't the driver only call this function when something >> actually changes? >> >> In other words, can you give some background information? >> >> Regards, >> >> Hans >> >> PS: still haven't processed your V4L2_CTRL_WHICH_DEF_VAL patch series. Hope >> to do this next week at the latest. >> >>> >>> Reported-by: Dimitrios Katsaros <patcherwork@gmail.com> >>> Signed-off-by: Ricardo Ribalda Delgado <ricardo.ribalda@gmail.com> >>> --- >>> drivers/media/v4l2-core/v4l2-ctrls.c | 23 ++++++++++++++--------- >>> 1 file changed, 14 insertions(+), 9 deletions(-) >>> >>> diff --git a/drivers/media/v4l2-core/v4l2-ctrls.c b/drivers/media/v4l2-core/v4l2-ctrls.c >>> index 4a1d9fdd14bb..f9c0e8150bd1 100644 >>> --- a/drivers/media/v4l2-core/v4l2-ctrls.c >>> +++ b/drivers/media/v4l2-core/v4l2-ctrls.c >>> @@ -3300,7 +3300,8 @@ EXPORT_SYMBOL(v4l2_ctrl_notify); >>> int __v4l2_ctrl_modify_range(struct v4l2_ctrl *ctrl, >>> s64 min, s64 max, u64 step, s64 def) >>> { >>> - bool changed; >>> + bool value_changed; >>> + bool range_changed = false; >>> int ret; >>> >>> lockdep_assert_held(ctrl->handler->lock); >>> @@ -3324,10 +3325,14 @@ int __v4l2_ctrl_modify_range(struct v4l2_ctrl *ctrl, >>> default: >>> return -EINVAL; >>> } >>> - ctrl->minimum = min; >>> - ctrl->maximum = max; >>> - ctrl->step = step; >>> - ctrl->default_value = def; >>> + if ((ctrl->minimum != min) || (ctrl->maximum != max) || >>> + (ctrl->step != step) || ctrl->default_value != def) { >>> + range_changed = true; >>> + ctrl->minimum = min; >>> + ctrl->maximum = max; >>> + ctrl->step = step; >>> + ctrl->default_value = def; >>> + } >>> cur_to_new(ctrl); >>> if (validate_new(ctrl, ctrl->p_new)) { >>> if (ctrl->type == V4L2_CTRL_TYPE_INTEGER64) >>> @@ -3337,12 +3342,12 @@ int __v4l2_ctrl_modify_range(struct v4l2_ctrl *ctrl, >>> } >>> >>> if (ctrl->type == V4L2_CTRL_TYPE_INTEGER64) >>> - changed = *ctrl->p_new.p_s64 != *ctrl->p_cur.p_s64; >>> + value_changed = *ctrl->p_new.p_s64 != *ctrl->p_cur.p_s64; >>> else >>> - changed = *ctrl->p_new.p_s32 != *ctrl->p_cur.p_s32; >>> - if (changed) >>> + value_changed = *ctrl->p_new.p_s32 != *ctrl->p_cur.p_s32; >>> + if (value_changed) >>> ret = set_ctrl(NULL, ctrl, V4L2_EVENT_CTRL_CH_RANGE); >>> - else >>> + else if (range_changed) >>> send_event(NULL, ctrl, V4L2_EVENT_CTRL_CH_RANGE); >>> return ret; >>> } >>> > > > ^ permalink raw reply [flat|nested] 4+ messages in thread
end of thread, other threads:[~2015-11-11 12:43 UTC | newest] Thread overview: 4+ messages (download: mbox.gz follow: Atom feed -- links below jump to the message on this page -- 2015-11-11 11:58 [PATCH] v4l2-core/v4l2-ctrls: Filter NOOP CH_RANGE events Ricardo Ribalda Delgado 2015-11-11 12:11 ` Hans Verkuil 2015-11-11 12:20 ` Ricardo Ribalda Delgado 2015-11-11 12:33 ` Hans Verkuil
This is a public inbox, see mirroring instructions for how to clone and mirror all data and code used for this inbox; as well as URLs for NNTP newsgroup(s).