This is a piece of information that may or may not relate to what you're seeing. Over 10 years ago, with Kmail running, one of the dbus-daemon processes would start consuming 95+% of CPU. This happened several times a day. I finally fixed it by creating my own daemon, which would use ps to find any dbus-daemon instances over 90%, and if it was still over 90% five seconds later, my daemon would kill it.
So, I recommend when this hangs, you take a look at the ps command. By the way, I gave up Kmail in 2012 and never had this problem again. SteveT Steve Litt March 2022 featured book: Making Mental Models: Advanced Edition http://www.troubleshooters.com/mmm Parke said on Wed, 9 Feb 2022 22:42:32 -0800 >On my system, the command `qemu-system-x86_64` prints no output and >simply hangs. No display window appears. Apparently QEMU hangs while >polling dbus. > >If I run `strace qemu-system-x86_64`, the end of the output is: > >socket(AF_UNIX, SOCK_STREAM|SOCK_CLOEXEC, 0) = 8 >connect(8, {sa_family=AF_UNIX, sun_path="/tmp/dbus-4fpXSf1PZL"}, 22) = >0 fcntl(8, F_GETFL) = 0x2 (flags O_RDWR) >fcntl(8, F_SETFL, O_RDWR|O_NONBLOCK) = 0 >geteuid() = 1000 >getsockname(8, {sa_family=AF_UNIX}, [128 => 2]) = 0 >poll([{fd=8, events=POLLOUT}], 1, 0) = 1 ([{fd=8, revents=POLLOUT}]) >sendto(8, "\0", 1, MSG_NOSIGNAL, NULL, 0) = 1 >sendto(8, "AUTH EXTERNAL 31303030\r\n", 24, MSG_NOSIGNAL, NULL, 0) = 24 >poll([{fd=8, events=POLLIN}], 1, -1 > >My system is running Ubuntu, but I installed Ubuntu manually (using >custom scripts I wrote), so it is entirely possible that dbus is not >behaving as expected. (I have deliberately disabled or removed many >Ubuntu components that are present on a "standard" installation of >Ubuntu.) > >Is it possible to prevent QEMU from attempting to use dbus? > >Does anyone have any suggestions for diagnosing the problem? QEMU is >not printing any error or diagnostic messages. If dbus writes any >error or log messages, I don't know where they would be. > >If I move the socket /tmp/dbus-4fpXSf1PZL to some other path, then it >appears that QEMU runs successfully. (Presumably, in this case QEMU >skips trying to connect to dbus.) > >Thank you, > >Parke >
