Il 13/04/2017 18:41, Martin Gräßlin ha scritto: > Am 2017-04-13 17:19, schrieb Germano Massullo: >> Il 13/04/2017 16:49, Martin Gräßlin ha scritto: >>> Am 2017-04-13 16:08, schrieb Germano Massullo: >>>> Il 13/04/2017 13:12, David Edmundson ha scritto: >>>>> "echo 0 | sudo tee /proc/sys/kernel/yama/ptrace_scope" (or run GDB >>>>> as root) >>>>> >>>>> David >>>> As root in tty3 I attached GDB to kwin_wayland and as soon I got >>>> back to >>>> the graphical session, the system got stuck and I could not neither >>>> return back to tty3, so I had to use Magic SysRq key to restart the >>>> system. >>>> The wiki says to try to use SSH, but it would be useless too since >>>> I am >>>> not able to use the graphical session if kwin_wayland is attached >>>> to GDB >>> >>> Of course you need to continue once you attached gdb. I'm not sure >>> what you want to achieve but debugging the running display manager on >>> the same system - well it's even more impossible than debugging the >>> running window manager on X11. I'm working with nested kwin_wayland >>> running in gdb to do debugging. >>> >>> Cheers >>> Martin >>> >> >> A running from bash Plasma, when it crashes[1], sends a message like >>> The Wayland connection broke. Did the Wayland compositor die? >> so I was trying to figure out if kwin_wayland crashes too >> [1]: https://bugs.kde.org/show_bug.cgi?id=378707 > > I think you are looking at a false positive here. Given the backtrace > I think the message is a side effect not the cause. If the Wayland > server had problems everything would die not just Plasma. And it would > not crash - at least not directly. It could only crash in the cleanup > handler and that's not the case here. > > Cheers > Martin >
Mmh ok. Does the bugreport contain anything useful?
