https://bugs.kde.org/show_bug.cgi?id=522329
--- Comment #21 from Sanfod Rockowitz <[email protected]> --- Yes, further testing is in order. I would urge as many people as possible to test powerdevil behaviour with the current ddcutil 2.2.8-dev branch. Note that the meanings of several utility option flags in 2.2.8-dev have changed. The new algorithms, previously enabled by utility option flags, are all now the defaults. The utility option flags now select the old way of doing things, and their current meaning is as follows: --f25: use old recheck queue algorithm --f32: use polling instead of event-based blocking waits in watch thread --f33: disable single poll wait in interruptible sleep function dw_split_sleep() --f35: disable skipping /dev/i2c open when DRM connector status is disconnected --f36: Do not disable display change scan while EACCESS is seen The concern with option --f35 is that the DRM status value has been seen to not always be reliable with old nvidia drivers. I believe I have a sufficient test of when DRM status can be relied on, but it's possible that the test is insufficient. Finally, there's --i13: EACCESS diagnostics interval --i13 sets the number of seconds between executions of full EACCES diagnostics . The default is 10. My suspicion is that is overly large. If you decrease or increase the interval, at what point to you start to see a sequence of multiple EACESS diagnostics in the system log, and at what point do you perceive an impact on performance? -- You are receiving this mail because: You are watching all bug changes.
