<div dir="ltr"><div class="gmail_default" style="font-family:arial,helvetica,sans-serif"><br></div><div class="gmail_extra"><br><div class="gmail_quote">2015-06-23 8:28 GMT+02:00 Michel Dänzer <span dir="ltr"><<a href="mailto:michel@daenzer.net" target="_blank">michel@daenzer.net</a>></span>:<br><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left-width:1px;border-left-color:rgb(204,204,204);border-left-style:solid;padding-left:1ex"><span class="">On 23.06.2015 04:19, Piotr Redlewski wrote:<br> ><br> ><br> > 2015-06-22 12:01 GMT+02:00 Piotr Redlewski <<a href="mailto:predlewski@gmail.com">predlewski@gmail.com</a><br> </span>> <mailto:<a href="mailto:predlewski@gmail.com">predlewski@gmail.com</a>>>:<br> <span class="">><br> ><br> ><br> > 2015-06-22 9:49 GMT+02:00 Michel Dänzer <<a href="mailto:michel@daenzer.net">michel@daenzer.net</a><br> </span>> <mailto:<a href="mailto:michel@daenzer.net">michel@daenzer.net</a>>>:<br> <span class="">><br> > On <a href="tel:21.06.2015%2023" value="+12106201523">21.06.2015 23</a> <tel:21.06.2015%2023>:30, Piotr Redlewski wrote:<br> > > When turning the outputs off (DPMSModeOff), CRTCs stayed enabled. This led<br> > > to higher card temperatures than with fglrx driver.<br> > ><br> </span>> > Signed-off-by: Piotr Redlewski <<a href="mailto:predlewski@gmail.com">predlewski@gmail.com</a> <mailto:<a href="mailto:predlewski@gmail.com">predlewski@gmail.com</a>>><br> <div><div class="h5">> > ---<br> > > src/drmmode_display.c | 9 ++++++++-<br> > > 1 file changed, 8 insertions(+), 1 deletion(-)<br> > ><br> > > diff --git a/src/drmmode_display.c b/src/drmmode_display.c<br> > > index 3564853..04b4341 100644<br> > > --- a/src/drmmode_display.c<br> > > +++ b/src/drmmode_display.c<br> > > @@ -342,7 +342,14 @@ drmmode_do_crtc_dpms(xf86CrtcPtr crtc, int mode)<br> > > static void<br> > ><br> > <br> > drmmode_crtc_dpms(xf86CrtcPtr crtc, int mode)<br> > > {<br> > > - /* Nothing to do. drmmode_do_crtc_dpms() is called as appropriate */<br> > > + drmmode_crtc_private_ptr drmmode_crtc = crtc->driver_private;<br> > > + drmmode_ptr drmmode = drmmode_crtc->drmmode;<br> > > +<br> > > + /* Disable unused CRTCs */<br> > > + if (!crtc->enabled && mode != DPMSModeOn) {<br> ><br> > Maybe this should be<br> ><br> > if (!crtc->enabled || mode != DPMSModeOn) {<br> ><br> > or can it actually happen that<br> > <br> > !crtc->enabled and mode == DPMSModeOn?<br> ><br> > Just wanted to be safe here, but I think you are right. I'll send<br> > v2 later today.<br> ><br> > I wanted to test changes before sending v2 and I stumbled on one issue.<br> > Everything works as expected when xrandr is used, but problem starts<br> > when screensaver<br> > comes into play. X calls<br> > drmmode_crtc_dpms() to on/off screensaver, so when it is supposed to be<br> > enabled, function is called with mode=DPMSModeOff and we<br> > call drmModeSetCrtc() to set crtc's mode to NULL. Problem is when X<br> > wants to disable screensaver: drmmode_crtc_dpms() is called with<br> > mode=DPMSModeOn and... nothing happens as there is no logic for this<br> > scenario - we should call drmModeSetCrtc() to set valid mode, but we<br> > don't know what this mode should be.<br> <br> </div></div>Isn't crtc->mode valid at that point?<br></blockquote><div><div class="gmail_default" style="font-family:arial,helvetica,sans-serif;display:inline">I don't think so, but I'll double check it</div> </div><div><div class="gmail_default" style="font-family:arial,helvetica,sans-serif"></div><br></div><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left-width:1px;border-left-color:rgb(204,204,204);border-left-style:solid;padding-left:1ex"> <br> If it's not, maybe we can store the mode in <div class="gmail_default" style="font-family:arial,helvetica,sans-serif;display:inline"></div>drmmode_crtc and use that<br> instead.<br></blockquote><div><div class="gmail_default" style="font-family:arial,helvetica,sans-serif;display:inline">crtc mode is one thing, but we need other parameters as well: x/y crtc offsets,</div></div><div><div class="gmail_default" style="font-family:arial,helvetica,sans-serif;display:inline">buffer id, connectors list - do you think that these also can be stored in </div></div><div><div class="gmail_default" style="font-family:arial,helvetica,sans-serif;display:inline"></div>drmmode_crtc<div class="gmail_default" style="font-family:arial,helvetica,sans-serif;display:inline">?</div> </div><div><br></div><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left-width:1px;border-left-color:rgb(204,204,204);border-left-style:solid;padding-left:1ex"> <span class=""><br> <br> > Because of the situation with the screensaver I think that original<br> > version of the patch is ok (it solves the problem with CRTC staying<br> > enabled when using xrandr to disable the output).<br> <br> </span>It would be nice to get the same effect for DPMS off though, wouldn't it?<br></blockquote><div><div class="gmail_default" style="font-family:arial,helvetica,sans-serif;display:inline">Of course, if I find a way to implement it I'll do it.</div></div><div><div class="gmail_default" style="font-family:arial,helvetica,sans-serif;display:inline"></div> </div><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left-width:1px;border-left-color:rgb(204,204,204);border-left-style:solid;padding-left:1ex"> <div class=""><div class="h5"><br> <br> --<br> Earthling Michel Dänzer | <a href="http://www.amd.com" rel="noreferrer" target="_blank">http://www.amd.com</a><br> Libre software enthusiast | Mesa and X developer<br> <br> </div></div></blockquote></div><br></div></div>