Re: kern/39774: (not really) dirhash kernel panic

[Hauke Fath <hauke%Espresso.Rhein-Neckar.DE@localhost>]

The following reply was made to PR kern/39774; it has been noted by GNATS.

From: Hauke Fath <hauke%Espresso.Rhein-Neckar.DE@localhost>
To: David Holland <dholland-bugs%netbsd.org@localhost>
Cc: gnats-bugs%netbsd.org@localhost, kern-bug-people%netbsd.org@localhost, 
gnats-admin%netbsd.org@localhost
Subject: Re: kern/39774: (not really) dirhash kernel panic
Date: Fri, 24 Oct 2008 09:08:32 +0200

 At 2:10 Uhr +0000 23.10.2008, David Holland wrote:
 >Am I right in thinking this means it croaked in whatever function is
 >at 0xf0006254 (inside a module?), which was called from offset 0x204
 >into ufs_lookup?
 >
 >I don't see anything in ufs_lookup, or at least anything near enough
 >to the top to be at offset 0x204, that ought to be a call into a
 >module, so I'm not sure what to make of this.
 
 Another one, and a disassembly:
 
 db{1}> t
 cpu_Debugger(0xf0275310, 0xf0101870, 0xf913e7f4, 0xf0291000, 0x100,
 0xf02c5400) at netbsd:mem_access_fault4m+0x204
 mem_access_fault4m(0x9, 0x5336, 0xf913e7f4, 0xf6215b10, 0x0, 0x325c) at
 0xf0006254
 0xf0006254(0x1, 0x27f0, 0x1, 0x200, 0x0, 0x78000) at netbsd:ufs_lookup+0x490
 ufs_lookup(0x1e7e0, 0x3fff, 0x2, 0xf6215e94, 0x0, 0x200) at
 netbsd:VOP_LOOKUP+0x24
 VOP_LOOKUP(0xf7caa630, 0xf6215e80, 0xf6215e94, 0x0, 0xf7caa630, 0xf4bad806)
 at netbsd:lookup+0x32c
 lookup(0xf6215e70, 0x20002, 0xf6215cfc, 0x20, 0x0, 0xf4bad800) at
 netbsd:namei+0xf8
 namei(0x0, 0x0, 0x0, 0xf6215e94, 0xf6215e70, 0xf14b8080) at
 netbsd:sys___lstat30+0x2c
 sys___lstat30(0xf5229308, 0xf6215f20, 0xf6215f40, 0x218054, 0xffffffff,
 0xf02c3000) at netbsd:syscall_plain+0xe4
 syscall_plain(0x585, 0xf6215fb0, 0x401505f8, 0xf6215fb0, 0x0, 0x325c) at
 0xf0006574
 db{1}> ps
  PID           PPID     PGRP        UID S   FLAGS LWPS          COMMAND    WAIT
 >854          29836    20850          0 2  0x4000    1             find
  54           20850    20850         17 2  0x4100    1         sendmail  piperd
  738          20850    20850          0 2  0x4000    1              tee  piperd
  29836        20850    20850          0 2  0x4000    1               sh    wait
 
 [...]
 
 db{1}> x/i ufs_lookup+0x480
 netbsd:ufs_lookup+0x480:        ld              [%fp - 0x34], %g2
 db{1}>
 netbsd:ufs_lookup+0x484:        sub             %g1, %g2, %g1
 db{1}>
 netbsd:ufs_lookup+0x488:        ba              netbsd:ufs_lookup+0x37c
 db{1}>
 netbsd:ufs_lookup+0x48c:        st              %g1, [%fp - 0x30]
 db{1}>
 netbsd:ufs_lookup+0x490:        call            netbsd:memcmp
 db{1}>
 netbsd:ufs_lookup+0x494:        add             %l0, 0x8, %o1
 db{1}>
 netbsd:ufs_lookup+0x498:        subcc           %o0, 0x0, %g0
 db{1}> x/i 0xf0006254
 0xf0006254:     call            netbsd:mem_access_fault4m
 db{1}>
 
 Since I seem to get those panics nightly right now, is there anything else
 that I should explore from the debugger prompt?
 
        hauke
 
 --
 "It's never straight up and down"     (DEVO)