There's a bug in truncate if you are using softdeps.
I've been running with a patch from Chuq for about a week now
and it seems to have fixed the problem.  I guess he's in India, so
maybe I should commit the fix.

    -- Lennart


Michael Graff wrote:

> NetBSD/i386 (dns-01.gdib.nominum.com) (tty00)
>
> login: panic: vref used where vget required, vp 0xf7417834
> Stopped in thing_server at     cpu_Debugger+0x4:       leave
>
> cpu_Debugger(f7417834,0,f7417834,f73c5cfc,c0156651) at
> cpu_Debugger+0x4
> panic(c021d200,f7417834,f7417834,f73c5d18,c01c87aa) at panic+0x64
> vref(f7417834,9572c,8,9572c,0) at vref+0x19
> uvm_vnp_uncache(f7417834,f7417834,0,0) at uvm_vnp_uncache+0x32
> ffs_truncate(f73c5ea4,f7417834,f72e4ca4,f7417888,f73c5ebc) at
> ffs_truncate+0x27e
>
> ufs_inactive(f73c5ed8,f7417834,480fc000,0,c0248b40) at
> ufs_inactive+0xa0
> vrele(f7417834,f7417834,0,0,0) at vrele+0xc9
> uvn_detach(f7417834) at uvn_detach+0xa9
> uvm_unmap_detach(f7301f4c,0,f72e4ca4,f73c5f88,f73c5f80) at
> uvm_unmap_detach+0x5c
>
> sys_munmap(f72e4ca4,f73c5f88,f73c5f80,0,816c060) at sys_munmap+0x123
> syscall() at syscall+0x133
> --- syscall (number 73) ---