| romangg | |
| sebas |
| KWin | |
| Plasma |
This change hooks up the bits for changing the mode when a request to
change mode from kscreen is received. This includes:
Room for improvement:
Started both a QPainter/OpenGL session. Used kcmshell5 kscreen
to modify the resolution. Saw that it got applied, maximized a window to
verify it's correct. Changed back to original resolution. Verified it looked
good by maximizing a window.
| Automatic diff as part of commit; lint not applicable. |
| Automatic diff as part of commit; unit tests not applicable. |
Nice!
You are using AMS on your testing machine I assume? The mode change should currently not work in legacy mode, because m_modesetRequested gets only used by AMS. Without AMS you need to call directly setModeLegacy with a black buffer. Or maybe just test on m_modesetRequested in presentLegacy on the next present. This said is there the possibility that the next present won't come soon (also relevant for AMS, where the mode gets changed with the next present)?
| plugins/platforms/drm/drm_output.cpp | ||
|---|---|---|
| 725 | Also make sure modeIndex>=0. | |
| plugins/platforms/drm/egl_gbm_backend.cpp | ||
|---|---|---|
| 153 | Rename this function resetOutput and put the destroy surface logic from the lambda function below in here? Would look cleaner I think. | |
Yes. I'm using defaults :-)
The mode change should currently not work in legacy mode, because m_modesetRequested gets only used by AMS. Without AMS you need to call directly setModeLegacy with a black buffer. Or maybe just test on m_modesetRequested in presentLegacy on the next present. This said is there the possibility that the next present won't come soon (also relevant for AMS, where the mode gets changed with the next present)?
Yeah, I already got from the code that my change won't work with legacy. And I'm tempted to just not implement it there. At least we could consider to not implement it and wait for a bug report. If we don't get one we could drop the complete legacy mode settings in ~1 year. That's the xorg style of deprecation: known breakage, waiting for someone notices.
Rename method to resetOutput and move the complete recreate logic there.
Hope I didn't break anything with it.
Looks good to me. I couldn't test it though, because KScreen KCM only showed me one mode to select from.
| plugins/platforms/drm/drm_output.h | ||
|---|---|---|
| 152 | Nowhere (yet?) used. | |
My notebook also only reports one more. I tested this on "real" hardware.
| plugins/platforms/drm/drm_output.h | ||
|---|---|---|
| 152 | Yes, it is for the case that modesettings fails, so that we can go back. | |