>Number:         3721
>Category:       kern
>Synopsis:       an accounting file can't be on nfs.
>Confidential:   no
>Severity:       non-critical
>Priority:       low
>Responsible:    kern-bug-people (Kernel Bug People)
>State:          open
>Class:          sw-bug
>Submitter-Id:   net
>Arrival-Date:   Sun Jun  8 01:35:02 1997
>Last-Modified:
>Originator:     enami tsugutomo
>Organization:
>Release:        NetBSD-current 1997 June 06
>Environment:
System: NetBSD pavlov.enami.ba2.so-net.or.jp 1.2F NetBSD 1.2F (PAVLOV) #282: Sun Jun 8 16:48:33 JST 1997 enami@pavlov.enami.ba2.so-net.or.jp:/b/netbsd/kernel/compile/PAVLOV i386


>Description:
	An accounting file /var/account/acct can't be on nfs.  If do so,
	kernel panics because of null pointer access (curproc is NULL
	when sosend is called).

>How-To-Repeat:
	For example, put /var on nfs and enable accounting.  System starts
	normally, but it will panic after 15 seconds.

	Here is panic message and ddb trace:

clearing /tmp
updating motd.
turning on accounting
standard daemons: update cron.
starting network daemons: routed timed printer sendmail rarpd inetd.
starting local daemons: sshd.
Sun Jun  8 16:51:26 JST 1997

NetBSD/i386 (pavlov.enami.ba2.so-net.or.jp) (tty00)

login: kernel: page fault trap, code=0
Stopped at      _sosend+0x8b:   movl    0x18(%ecx),%eax
db> trace
_sosend(f88b8700,f88b9180,0,f891b380,0) at _sosend+0x8b
_nfs_send(f88b8700,f88b9180,f891b380,f891a640,f891b700) at _nfs_send+0x91
_nfs_request(f88b9580,f8918c80,12,0,f891bb00) at _nfs_request+0x335
_nfs_statfs(f888c000,fefb8c6c,0,0,fefb8c84) at _nfs_statfs+0x387
_acctwatch(0) at _acctwatch+0xc9
_softclock(f886f7a0,f8908000,0,fefb8da8,f81bf837) at _softclock+0x68
_hardclock(fefb8db4,fefb8db0,f8100d3b,fefb8db4,0) at _hardclock+0x3e3
_clockintr(fefb8db4) at _clockintr+0xb
_Xresume0() at _Xresume0+0x2e
--- interrupt ---
_idle(0,0,4,100,fefb8f40) at _idle+0xd
bpendtsleep(f81d3944,118,f8127633,fa63,f81c9eb4) at bpendtsleep
_sys_select(f8908000,fefb8f88,fefb8f80,0,0) at _sys_select+0x31e
_syscall() at _syscall+0x154
--- syscall (number 93) ---
0x40043083:
db>

>Fix:
	Not known, Probably, disk capacity check need to be done at least
	out of interrupt time.
>Audit-Trail:
>Unformatted: