On Thu, 17 Oct 2024 21:39:03 +0200 Quelrond <quelr...@gmail.com> said:
> Is there any way to debug this problem? > > Is there any way to stop E or X or both before sleep and then restart > them correctly without loosing the apps after resume? you cant stop x without losing apps. all your state for your dsplay is held in the xserver. you end that process - you lose all of that. as i already said - restarting e ends the previous e process and starts a new e process. so that already does that and leaves you with blank windows which makes me think your problem is lower down and not any state e is holding as e will have lost all state during it's restart as it's a whole new process (enlightenment_start handles all of this - its a small baby-sitter process to do these things and has nothing to do with display at all - just sit and watch it's child). so it's a whole new e wm+compositor process with a whole new xserver connection that starts on an existing xorg server - and then it can't get working pixmaps/textures for windows. that is why i suspect some issue in the xserver. > What another WM can I test to find the source of the problem? other wm's and compositors will likely be different in how they access window pixmaps - but you'd probably want a tiling wm with no borders at all and a egl+gles based compositor. > Peter > > On 17/10/2024 21:11, Carsten Haitzler wrote: > > On Thu, 17 Oct 2024 21:02:00 +0200 Quelrond <quelr...@gmail.com> said: > > > >> Hello, > >> > >> I've just upgraded my laptop from FreeBSD 13.3 to FreeBSD 14.1. The most > >> important progress is that suspend resume (almost) works with this > >> version. I have a strange problem with Enlightenment though. > >> > >> After the resume the performance of E seems to be lower than normally, > >> repainting of windows content is a bit slow. And if I try to restart E > >> (with Ctrl-Alt-End) - the windows content is completely lost: > >> > >> https://www.flytrace.com/static/download/shot-2024-10-14_22-27-20.jpg > >> > >> Any ideas? > > the fact that a restart of e loses content... implies the pixmaps of the > > window content somehow don't work or cannot be accessed. a e restart has e > > exit and start again - thus as x for pixmaps and so on... so it might be > > that the kernel and/or gpu driver somehow during suspend do something bad? > > they might migrate pixmaps from gpu vram to system memory before suspend so > > they can shut down gpu vram, but then on resume not copy them back to vram > > and switch back correctly? > > > -- ------------- Codito, ergo sum - "I code, therefore I am" -------------- Carsten Haitzler - ras...@rasterman.com _______________________________________________ enlightenment-users mailing list enlightenment-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/enlightenment-users
