Re: port-amd64/43236: Removing a file while it is accessed through NFS crashes the server

[Peter Kerwien <peter%kerwien.homeip.net@localhost>]

The following reply was made to PR kern/43236; it has been noted by GNATS.

From: Peter Kerwien <peter%kerwien.homeip.net@localhost>
To: <gnats-bugs%NetBSD.org@localhost>
Cc: 
Subject: Re: port-amd64/43236: Removing a file while it is accessed through NFS
 crashes the server
Date: Mon, 03 May 2010 21:35:57 +0200

 On Mon,  3 May 2010 05:40:03 +0000 (UTC), David Holland
 >  You can also turn on the built-in debugger by doing
 >  
 >     sysctl -w ddb.onpanic=1
 >  
 
 OK, here it comes. This time it was reproduced on a NetBSD-5.0.2
 installation without any modifications.
 
 uvm_fault(0xffff80004e9ab2e0, 0x0, 1) -> e
 fatal page fault in supervisor mode
 trap type 6 code 0 rip ffffffff803b4172 cs 8 rflags 10246 cr2 c0 cpl 0 rsp
 ffff8000ef7eb710
 kernel: page fault trap, code=0
 Stopped in pid 208.2 (nfsd) at netbsd:ufs_fhtovp+0x42: cmpw $0,0xc0(%rax)
 db{1}>bt
 ufs_hvtovp() at netbsd:ufs_hvtovp+0x42
 ufs_hvtovp() at netbsd:ufs_hvtovp+0x58
 VFS_HVTOVP() at netbsd:VFS_HVTOVP+0x33
 nfsrv_hvtovp() at netbsd:nfsrv_hvtovp+0xa8
 nfsrv_read() at netbsd:nfsrv_read+0x273
 nfssvc_nfsd() at netbsd:nfssvc_nfsd+0x1c5
 sys_nfssvc() at netbsd:sys_nfssvc+0x28f
 syscall() at netbsd:syscall+0xb6
 db{1}>
 
 There might be typos. I took some pictures of the screen when it crashed.
 Are there any easier way to dump the information from the kernel debugger?
 
 I hope you can extract some useful information from this. Let me know if
 you need more. I seems to easily reproduce this even on my lab-computer.
 
 /Peter