Re: netbsd-7 amd64 DRMKMS radeon crash after a while

[Timo Buhrmester <timo%math.uni-bonn.de@localhost>]

I'm not sure if this is the same problem, but I'm also experiencing fairly frequent (every couple days) panics using radeon on netbsd-current/amd64.

DRM error in radeon_gem_object_create: Failed to allocate GEM object (16613376, 2, 4096, -12)
panic: kernel diagnostic assertion "mo->mo_refcnt > 0" failed: file "/usr/src.head/sys/kern/kern_mutex_obj.c", line 135 mutex_obj_free: lock 0xfffffe809ba31980: mo->mo_refcnt (0) ==
 0
fatal breakpoint trap in supervisor mode
trap type 1 code 0 rip ffffffff80192d5d cs 8 rflags 246 cr2 7f7ff7bd6020 ilevel 0 rsp fffffe804574abd0
curlwp 0xfffffe80aaea4880 pid 456.1 lowest kstack 0xfffffe80457472c0
Stopped in pid 456.1 (Xorg) at  netbsd:breakpoint+0x5:  leave
breakpoint() at netbsd:breakpoint+0x5
vpanic() at netbsd:vpanic+0x13c
kern_assert() at netbsd:kern_assert+0x4f
mutex_obj_free() at netbsd:mutex_obj_free+0x9d
ttm_bo_release() at netbsd:ttm_bo_release+0x3b
radeon_bo_unref() at netbsd:radeon_bo_unref+0x3f
radeon_gem_object_free() at netbsd:radeon_gem_object_free+0x21
drm_gem_object_handle_unreference_unlocked() at netbsd:drm_gem_object_handle_unr
eference_unlocked+0x92
drm_gem_handle_delete() at netbsd:drm_gem_handle_delete+0x88
drm_ioctl() at netbsd:drm_ioctl+0x122
sys_ioctl() at netbsd:sys_ioctl+0x17e
syscall() at netbsd:syscall+0x9a
--- syscall (number 54) ---
7f7ff40cf7fa:
db{1}>

It doesn't always throw me into ddb, most of the time the machine reboots or hangs, sometimes in the middle of printing the call stack.

I'm using netbsd-current as of a few weeks ago, rebuilding latest head now.


Cheers,
Timo