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

[fukumoto%imasy.or.jp@localhost]

From: fukumoto%imasy.or.jp@localhost
Subject: Re: netbsd-7 amd64 DRMKMS radeon crash after a while
Date: Thu, 19 Mar 2015 20:29:53 +0900 (JST)

> 
> Updated to today's source and rebuilt, and it feels stable than
> before.  At least it hasn't crashed after a few hours use.  Thank you!


I guess it was too early to judge...  I got another crash.

uvm_fault(0xffffffff81046240, 0x0, 1) -> e
fatal page fault in supervisor mode
trap type 6 code 0 rip ffffffff805cab2d cs 8 rflags 10286 cr2 8 ilevel 0 rsp fff
ffe8040065cd8
curlwp 0xfffffe81071399a0 pid 0.35 lowest kstack 0xfffffe80400632c0
kernel: page fault trap, code=0
Stopped in pid 0.35 (system) at netbsd:mutex_oncpu.part.0+0x8:  movq    18(%rdi)
,%rax
db{0}> bt
mutex_oncpu.part.0() at netbsd:mutex_oncpu.part.0+0x8
mutex_vector_enter() at netbsd:mutex_vector_enter+0x93
uvm_obj_unwirepages() at netbsd:uvm_obj_unwirepages+0x1a
ttm_tt_swapout() at netbsd:ttm_tt_swapout+0x1b
ttm_bus_dma_unpopulate() at netbsd:ttm_bus_dma_unpopulate+0x35
ttm_tt_destroy() at netbsd:ttm_tt_destroy+0x2f
ttm_bo_cleanup_memtype_use() at netbsd:ttm_bo_cleanup_memtype_use+0x3a
ttm_bo_cleanup_refs_and_unlock() at netbsd:ttm_bo_cleanup_refs_and_unlock+0xfa
ttm_bo_delayed_delete() at netbsd:ttm_bo_delayed_delete+0x1c4
ttm_bo_delayed_workqueue() at netbsd:ttm_bo_delayed_workqueue+0x16
linux_worker() at netbsd:linux_worker+0xd0
workqueue_worker() at netbsd:workqueue_worker+0x89
db{0}> 



					FUKUMOTO Atsushi
					fukumoto%imasy.or.jp@localhost