path: root/xcb-sync.vapi
diff options
authorPřemysl Eric Janouch <p@janouch.name>2020-10-02 08:23:31 +0200
committerPřemysl Eric Janouch <p@janouch.name>2021-07-05 21:58:43 +0200
commit4b0f2d1cdf480c9d516532361b13476083cbbec0 (patch)
tree83dc91e47fff554b62a37f830ee10bbdd1bf331f /xcb-sync.vapi
parent8f350388ee2f6115bf4e1117c4968b2d51b73121 (diff)
WIP: Track suspension/hibernationHEADmaster
WIP: should check for changes in the IDLETIME counter, as well as in the SERVERTIME counter. - xorg-server/Xext/sync.c - always uses XSyncCounterNeverDecreases even though os/utils.c works even without CLOCK_MONOTONIC (not present or retrieval fails) - what am I actually trying to use this knowledge for? - I expect the IDLETIME counter to keep running monotonically, otherwise I'd need to account for any change - it doesn't look like it is changed by anything other than input devices - actually Xext/saver.c does something but it doesn't look like it'd mess up anything (hw/xfree86/common/xf86Events.c upon VT switching bumps idle timers) - hw/xfree86/common/xf86PM.c also appears to bump the idle timer, when the computer is resumed from suspend :/ - I expect the SERVERTIME counter to either jump or not - exactly what will happen: really should not jump but might - so far it looks like the suspend check is useful in cases that: - an automated (or external) suspend comes before we declare user idleness: we can mark that period as idle - I've written misc/toys/x11-timers.c: read out the timers in a loop - everything behaves as expected, there's just an interesting delay between the monotonic timer starting to run and the system continuing to function WIP: somehow write the above part down in the source as a comment WIP: need to: a/ take a delay lock, b/ register for PrepareForSleep/PrepareForShutdown signals
Diffstat (limited to 'xcb-sync.vapi')
0 files changed, 0 insertions, 0 deletions