netbsd-7 amd64 DRMKMS radeon crash after a while

[fukumoto%imasy.or.jp@localhost]

I tried netbsd-7 amd64 DRMKMS kernel with radeon driver a little.  So
far it tend to panic for lockdebug after several minutes of use with
X11.  Most of time it's in ttm_bo_release_list like this:

crash> bt
_KERNEL_OPT_NARCNET() at 0
_KERNEL_OPT_ACPI_SCANPCI() at _KERNEL_OPT_ACPI_SCANPCI+0x7
db_reboot_cmd() at db_reboot_cmd
db_command() at db_command+0xf0
db_command_loop() at db_command_loop+0xd0
db_trap() at db_trap+0xe9
kdb_trap() at kdb_trap+0xe5
trap() at trap+0xd7
--- trap (number 1) ---
breakpoint() at breakpoint+0x5
vpanic() at vpanic+0x13c
snprintf() at snprintf
lockdebug_wantlock() at lockdebug_wantlock
mutex_destroy() at mutex_destroy+0x30
ttm_bo_release_list() at ttm_bo_release_list+0xbb
ttm_bo_delayed_delete() at ttm_bo_delayed_delete+0x139
ttm_bo_delayed_workqueue() at ttm_bo_delayed_workqueue+0x16
linux_worker() at linux_worker+0xf0
workqueue_worker() at workqueue_worker+0x89
crash>


One time it crashed earlier, almost immediately after X server was
started.

crash> bt
_KERNEL_OPT_NARCNET() at 0
_KERNEL_OPT_ACPI_SCANPCI() at _KERNEL_OPT_ACPI_SCANPCI+0x7
db_reboot_cmd() at db_reboot_cmd
db_command() at db_command+0xf0
db_command_loop() at db_command_loop+0xd0
db_trap() at db_trap+0xe9
kdb_trap() at kdb_trap+0xe5
trap() at trap+0xd7
--- trap (number 1) ---
breakpoint() at breakpoint+0x5
vpanic() at vpanic+0x13c
snprintf() at snprintf
lockdebug_more() at lockdebug_more
_kernel_lock() at _kernel_lock+0x1a7
intr_biglock_wrapper() at intr_biglock_wrapper+0x12
Xintr_ioapic_level3() at Xintr_ioapic_level3+0xf2
--- interrupt ---
x86_stihlt() at x86_stihlt+0x6
acpicpu_cstate_idle_enter() at acpicpu_cstate_idle_enter+0xc2
acpicpu_cstate_idle() at acpicpu_cstate_idle+0xb6
idle_loop() at idle_loop+0x18c
crash> 



					FUKUMOTO Atsushi
					fukumoto%imasy.or.jp@localhost