properties: increase shutdown timeout to 10s
By default, we have:
- 3s to thermal shutdown/reboot
- 6s to cleanly shutdown/reboot
The reboot procedure is quite CPU intensive, we have stop all processes, then unmount lots of partitions and APEXes.
The process stopping already takes over 3s:
[ 296.122460][ T1] init: Ready to unmount apexes. So far shutdown sequence took 3388ms
Unmounting the APEXes takes longer than 3s. Because of this, we always hit the time-out during reboot, which logs the following warning:
[ 299.213450][ T1] init: umount timeout, last resort, kill all and try
[ 299.528401][ T1] sysrq: Show backtrace of all active CPUs
[ 299.534068][ T1] sysrq: CPU0:
[ 299.537287][ T1] Call trace:
[ 299.540428][ T1] dump_backtrace+0xfc/0x1f8
[ 299.544910][ T1] sysrq_handle_showallcpus+0x94/0x13c
[ 299.550232][ T1] __handle_sysrq+0x108/0x208
[ 299.554765][ T1] write_sysrq_trigger+0xfc/0x160
[ 299.559642][ T1] proc_reg_write+0xf8/0x16c
[ 299.564095][ T1] vfs_write+0x138/0x34c
[ 299.568200][ T1] ksys_write+0x7c/0xe8
[ 299.572208][ T1] __arm64_sys_write+0x20/0x2c
[ 299.576822][ T1] invoke_syscall+0x5c/0x114
[ 299.581271][ T1] el0_svc_common+0xb8/0xf8
[ 299.585627][ T1] do_el0_svc+0x30/0xb8
[ 299.589637][ T1] el0_svc+0x28/0x9c
[ 299.593407][ T1] el0t_64_sync_handler+0x7c/0xe4
[ 299.598286][ T1] el0t_64_sync+0x1a4/0x1a8
[ 299.602669][ C1] sysrq: CPU1: backtrace skipped as idling
[ 299.602669][ C3] sysrq: CPU3: backtrace skipped as idling
[ 299.602669][ C2] sysrq: CPU2: backtrace skipped as idling
[ 299.602724][ T1] sysrq: Show Blocked State
[ 299.624165][ T1] sysrq: Kill All Tasks
Remove the warning by increasing the reboot timer to 10s.
Signed-off-by: Mattijs Korpershoek mkorpershoek@baylibre.com
Edited by Mattijs Korpershoek