Version-Release number of selected component: systemd-238-7.fc28.1 Additional info: reporter: libreport-2.9.5 backtrace_rating: 4 cmdline: /usr/lib/systemd/systemd --user crash_function: log_assert_failed_unreachable_realm executable: /usr/lib/systemd/systemd journald_cursor: s=c914debce927456089778c6d919880d6;i=1e5b;b=6b4a5aed11934dcd85ff384cb355d82e;m=110894b48;t=56b66f158ac6b;x=7231780b4bdc469d kernel: 4.16.5-300.fc28.x86_64 rootdir: / runlevel: N 5 type: CCpp uid: 42 Truncated backtrace: Thread no. 1 (8 frames) #2 log_assert_failed_unreachable_realm at ../src/basic/log.c:838 #3 bus_process_internal at ../src/libsystemd/sd-bus/sd-bus.c:2978 #4 sd_bus_process at ../src/libsystemd/sd-bus/sd-bus.c:2994 #5 bus_ensure_running at ../src/libsystemd/sd-bus/sd-bus.c:2056 #6 sd_bus_flush at ../src/libsystemd/sd-bus/sd-bus.c:3102 #7 destroy_bus at ../src/core/dbus.c:1100 #8 bus_done at ../src/core/dbus.c:1138 #9 manager_free at ../src/core/manager.c:1203
Created attachment 1432237 [details] File: backtrace
Created attachment 1432238 [details] File: cgroup
Created attachment 1432239 [details] File: core_backtrace
Created attachment 1432240 [details] File: cpuinfo
Created attachment 1432241 [details] File: dso_list
Created attachment 1432242 [details] File: environ
Created attachment 1432243 [details] File: limits
Created attachment 1432244 [details] File: maps
Created attachment 1432245 [details] File: mountinfo
Created attachment 1432246 [details] File: open_fds
Created attachment 1432247 [details] File: proc_pid_status
Created attachment 1432248 [details] File: var_log_messages
*** Bug 1576084 has been marked as a duplicate of this bug. ***
*** Bug 1575350 has been marked as a duplicate of this bug. ***
There was a significant reworking of the bus code in https://github.com/systemd/systemd/commit/4502c40399. That patch is in systemd-238, and I was confused, because the backtraces in those three bugs show the old call sequence. So what must have happened here, is that an upgrade from older systemd was made to systemd-238, and then the older systemd crashed during reexec. That said, there seems to be a memory corruption here, and destroy_bus is called on a bus that has already been deallocated. I think/hope the patch above fixes this. There's another issue with the code, uncovered by this failure. See upstream PR https://github.com/systemd/systemd/pull/9640. Nevertheless, that patch wouldn't fix this problem anyway. I'll close this on the assumption that 4502c40399 is enough to fix this. If a similar backtrace happens again, this will mean I was wrong ;)