https://bugs.kde.org/show_bug.cgi?id=516066
Sanfod Rockowitz <[email protected]> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |[email protected] --- Comment #2 from Sanfod Rockowitz <[email protected]> --- It's hard to know what to make of this report. The special Nvidia driver config parms are sometimes needed for I2C to work, but it's odd that the latest ddcutil development branch would trigger that need; perhaps it's triggered by increased use of /sys. I've requested additional information from YanEx. ddcutil issue https://github.com/rockowitz/ddcutil/issues/583 appears to be a permissions configuration problem on the user's nixOS system. Issues https://github.com/rockowitz/ddcutil/issues/581 and https://github.com/rockowitz/ddcutil/issues/582 are more critical. In particular, issue #581 hangs powerdevil at startup, because /dev/i2c permissions are lost. It appears that under certain circumstances, permission granted by udev token uaccess is not respected by linux api access(). Perhaps the problem is arch/cachos specific. I've recoded libddcutil to not use access(). Also reported in that issue thread was a display change detection hang caused when a laptop display was not enabled at startup and then enabled. Laptop displays don't support DDC/CI, as there's no point in libddcutil tracking connection changes for them, so I've simply disabled change detection for laptop displays. (A display reference for the laptop display is still reported by libddcutil, so I don't think this change affects powerdevil.) I've requested more information on issue #582 and haven't received a response. The problem appears related to VM usage. I'd very much appreciate people testing release 2.2.5 and branch 2.2.6-dev with powerdevil on as many distributions and hardware configurations as possible. I've not been able to replicate the failures. Their resolution is currently my top ddcutil priority. -- You are receiving this mail because: You are watching all bug changes.
