Subject: Re: uvm_fault() in pmap_page_remove() panic
To: None <port-i386@netbsd.org>
From: David Brownlee <abs@NetBSD.org>
List: port-i386
Date: 04/04/2006 21:30:34
 	Bizarrely I've just had my home gateway drop into DDB with a
 	similar message:

 	uvm_fault(0xb4408540, 0x3f4b3000, 0, 1) -> 0xe
 	kernel: page fault trap, code=0
 	Stopped in pid 505.1 (ppp) at   netbsd:Xsyscall+0x61: cmpl  $0,%eax
 	db> bt
 	Xsyscall(0,1,9dbf6e3f,9db2f770,ab9b6cc) at netbsd:Xsycall+0x61
 	Bad user frame pointer: 0x9dbff09c
 	db> c

 	[works]

On Tue, 4 Apr 2006, David Brownlee wrote:

> 	Just had a production web/tomcat server panic with the following.
> 	It was over a light weekend so no obvious load implications.
> 	Machine has a GB of memory, runs 3.0_STABLE with an ACPI kernel,
> 	and normally has plenty of swap free.
>
> 	No crash dump (it had to be powercycled by remote hands). Anyone
> 	have any thoughts as to what I could do to get more information?
>
> uvm_fault(0x80757b40, 0x7fc20000, 0, 2) -> 0xe
> fatal page fault in supervisor mode
> trap type 6 code 2 eip 803e4cfc cs 8 eflags 10206 cr2 7fc209f8 ilevel 0
> panic: trap
> Begin traceback...
> trap() at netbsd:trap+0x171
> --- trap (number 6) ---
> pmap_page_remove(812822c0,95f8800c,95313e10,95313e30,95f88d54) at 
> netbsd:pmap_page_remove+0xac
> uvm_anfree(9527e100,9610f008,0,979e7f24,979e7f24) at netbsd:uvm_anfree+0x72
> amap_wipeout(95f88d54,0,13f,0,10000) at netbsd:amap_wipeout+0x76
> uvm_unmap_detach(97b2cab0,0,7fc00000,95313e94,0) at 
> netbsd:uvm_unmap_detach+0x8b
> uvmspace_free(963fb9a8,943750c0,0,0,8070a564) at netbsd:uvmspace_free+0x95
> exit1(96559224,0,96d17b28,96559224,0) at netbsd:exit1+0x226
> sys_exit(96559224,95313f64,95313f5c,96d17b28,96bba3c8) at 
> netbsd:sys_exit+0x29
> syscall_plain() at netbsd:syscall_plain+0x85
> --- syscall (number 1) ---
> 0x7d9dd54f:
> End traceback...
>
> 	thanks.
>
>

-- 
 		David/absolute       -- www.NetBSD.org: No hype required --