<div dir="ltr"><div class="gmail_extra"><div class="gmail_quote">2017-09-27 9:18 GMT+02:00 Jani Nikula <span dir="ltr"><<a href="mailto:jani.nikula@linux.intel.com" target="_blank">jani.nikula@linux.intel.com</a>></span>:<br><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex"><span class="gmail-">On Wed, 27 Sep 2017, Karsten Wiese <<a href="mailto:fzuuzf@googlemail.com">fzuuzf@googlemail.com</a>> wrote:<br> > 2017-09-25 15:48 GMT+02:00 Jani Nikula <<a href="mailto:jani.nikula@linux.intel.com">jani.nikula@linux.intel.com</a>>:<br> ><br> >> On Tue, 19 Sep 2017, Karsten Wiese <<a href="mailto:fzuuzf@googlemail.com">fzuuzf@googlemail.com</a>> wrote:<br> >> > This makes poll work for the<br> >> > /sys/class/drm/cardX/<wbr>connectorY/dpms attributes.<br> >><br> >> I guess the question is, what are you trying to achieve? What is the<br> >> problem that this solves?<br> >><br> > <br> > The patch enables cpu cycle less waiting for dpms flag changes.<br> ><br> > Here it lets a screen brightness setting daemon know which monitor to<br> > handle. One of the attached screens gets confused if it is set just<br> > after being switched on, hence the daemon leaves it untouched until it<br> > has been active for some seconds.<br> <br> </span>Screen brightness settings daemon? <br></blockquote><div><div style="font-family:courier new,monospace;display:inline" class="gmail_default">Running on a laptop lacking an ambient light sensor employing it's webcam instead to measure ambient light and adjust monitors' brightnesses accordingly.</div><div style="font-family:courier new,monospace;display:inline" class="gmail_default"><br></div><div style="font-family:courier new,monospace;display:inline" class="gmail_default"><br></div></div><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex"> What exactly do you mean by "if it is<br> set"? What interface are you using to change brightness?</blockquote><div><div style="font-family:courier new,monospace;display:inline" class="gmail_default">The external monitor's brightness is adjusted <span class="gmail-st">by DDC/CI via the /dev/i2c-x the monitor is attached to.</span></div><div style="font-family:courier new,monospace;display:inline" class="gmail_default"><span class="gmail-st"> I there a saner interface to use?<br></span></div><div style="font-family:courier new,monospace;display:inline" class="gmail_default"><span class="gmail-st"><br></span></div></div><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex"> What happens<br> when the display "gets confused"?<br></blockquote><div><div style="font-family:courier new,monospace;display:inline" class="gmail_default">The monitor wrongly toggles sharpness if it receives the brightness adjusting</div><div style="font-family:courier new,monospace;display:inline" class="gmail_default"> <span class="gmail-st">DDC/CI soon after resuming from power saving state.<br></span></div></div><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex"> <br> My first instinct is that you're proposing a new kernel ABI to solve an<br> issue you shouldn't be solving in userspace to begin with.</blockquote><div><div style="font-family:courier new,monospace;display:inline" class="gmail_default">Calculating ambient light from pictures acquired via </div><div style="font-family:courier new,monospace;display:inline" class="gmail_default">v4l2 in kernel seamed wrong to me.</div><div style="font-family:courier new,monospace;display:inline" class="gmail_default"><br></div></div><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex"> Or that<br> perhaps the userspace should be doing this in cooperation with the drm<br> master, not standalone.<br></blockquote><div><div style="font-family:courier new,monospace;display:inline" class="gmail_default">Is there a way to call into the drm-master (xscreensaver/xserver here</div><div style="font-family:courier new,monospace;display:inline" class="gmail_default">) with the call only returning if a monitor's power state changed?</div><div style="font-family:courier new,monospace;display:inline" class="gmail_default"><br></div><div style="font-family:courier new,monospace;display:inline" class="gmail_default">There is DPMSInfo, but it returns immediately rendering the daemon less efficient.</div><div style="font-family:courier new,monospace;display:inline" class="gmail_default"><br></div><div style="font-family:courier new,monospace;display:inline" class="gmail_default"><br></div><div style="font-family:courier new,monospace;display:inline" class="gmail_default">BR,</div><div style="font-family:courier new,monospace;display:inline" class="gmail_default"><br></div><div style="font-family:courier new,monospace;display:inline" class="gmail_default">Karsten<br></div></div><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex"> <br> BR,<br> Jani.<br> <div class="gmail-HOEnZb"><div class="gmail-h5"><br> ><br> > Without the patch the daemon would have to actively read the dpms flag<br> > every once in a while.<br> ><br> ><br> >><br> >> We have zero sysfs_notify in all of drm AFAICT.<br> ><br> ><br> > Yes I noticed too and looked for some dbus signal to listen to but didn't<br> > find any.<br> ><br> > BR,<br> > Karsten<br> ><br> >><br> >> BR,<br> >> Jani.<br> >><br> >><br> >> ><br> >> > Tested with i915 suspended by XScreenServer and<br> >> > suspend to RAM.<br> >> ><br> >> > Signed-off-by: Karsten Wiese <<a href="mailto:fzuuzf@googlemail.com">fzuuzf@googlemail.com</a>><br> >> > ---<br> >> > drivers/gpu/drm/drm_atomic.c | 4 ++++<br> >> > drivers/gpu/drm/drm_atomic_<wbr>helper.c | 6 +++++-<br> >> > 2 files changed, 9 insertions(+), 1 deletion(-)<br> >> ><br> >> > diff --git a/drivers/gpu/drm/drm_atomic.c b/drivers/gpu/drm/drm_atomic.c<br> >> > index 2fd383d..b6fa87b 100644<br> >> > --- a/drivers/gpu/drm/drm_atomic.c<br> >> > +++ b/drivers/gpu/drm/drm_atomic.c<br> >> > @@ -1880,6 +1880,10 @@ int drm_atomic_connector_commit_<wbr>dpms(struct<br> >> drm_atomic_state *state,<br> >> > out:<br> >> > if (ret != 0)<br> >> > connector->dpms = old_mode;<br> >> > + else<br> >> > + if (connector->dpms != old_mode)<br> >> > + sysfs_notify(&connector->kdev-<wbr>>kobj, NULL,<br> >> "dpms");<br> >> > +<br> >> > return ret;<br> >> > }<br> >> ><br> >> > diff --git a/drivers/gpu/drm/drm_atomic_<wbr>helper.c<br> >> b/drivers/gpu/drm/drm_atomic_<wbr>helper.c<br> >> > index 4e53aae..6198772 100644<br> >> > --- a/drivers/gpu/drm/drm_atomic_<wbr>helper.c<br> >> > +++ b/drivers/gpu/drm/drm_atomic_<wbr>helper.c<br> >> > @@ -921,12 +921,16 @@ drm_atomic_helper_update_<wbr>legacy_modeset_state(struct<br> >> drm_device *dev,<br> >> > crtc = new_conn_state->crtc;<br> >> > if ((!crtc && old_conn_state->crtc) ||<br> >> > (crtc && drm_atomic_crtc_needs_modeset(<wbr>crtc->state)))<br> >> {<br> >> > - int mode = DRM_MODE_DPMS_OFF;<br> >> > + int old_mode, mode = DRM_MODE_DPMS_OFF;<br> >> ><br> >> > if (crtc && crtc->state->active)<br> >> > mode = DRM_MODE_DPMS_ON;<br> >> ><br> >> > + old_mode = connector->dpms;<br> >> > connector->dpms = mode;<br> >> > + if (old_mode != mode)<br> >> > + sysfs_notify(&connector->kdev-<wbr>>kobj,<br> >> > + NULL, "dpms");<br> >> > }<br> >> > }<br> >><br> >> --<br> >> Jani Nikula, Intel Open Source Technology Center<br> >><br> <br> --<br> Jani Nikula, Intel Open Source Technology Center<br> </div></div></blockquote></div><br></div></div>