Re: kern/56925: Amd64 server randomly panics

[Taylor R Campbell <riastradh%NetBSD.org@localhost>]

The following reply was made to PR kern/56925; it has been noted by GNATS.

From: Taylor R Campbell <riastradh%NetBSD.org@localhost>
To: joel.bertrand%systella.fr@localhost
Cc: gnats-bugs%NetBSD.org@localhost
Subject: Re: kern/56925: Amd64 server randomly panics
Date: Thu, 14 Jul 2022 11:24:36 +0000

 This is a null pointer dereference by the instruction at address
 0xffffffff8022800c.  I'm guessing it's in wake_ccb but I'm not sure.
 
 Can you try to obtain the following information from your kernel?
 
 1. The disassembly of the function where rip 0xffffffff8022800c lies,
    and, if possible, the line number of 0xffffffff8022800c.
 
 2. The disassembly of ccb_timeout, and, if possible, the line number
    of ccb_timeout+0xf0.
 
 If you have a netbsd.gdb file, you can get these with:
 
 $ gdb ./netbsd.gdb
 (gdb) info line *(0xffffffff8022800c)
 (gdb) disas 0xffffffff8022800c
 (gdb) info line *(ccb_timeout+0xf0)
 (gdb) disas ccb_timeout
 
 If not, you can at least use the `disas' commands in the netbsd
 kernel.
 
 It may help to set the sysctl variables ddb.onpanic=1 and
 ddb.commandonenter="bt;show registers", or
 
 ddb.commandonenter="bt;show registers;sync"
 
 if you want ddb to force a crash dump (with the caveat that the latter
 means manually entering ddb will immediately force a crash dump with
 no interaction).  It may also help to set `options ISCSI_DEBUG' and
 set hw.iscsi.debug=1 or higher (with the caveat that it might produce
 a lot of log spew).