Skip site navigation (1)Skip section navigation (2)
Date:      Mon, 11 Jan 2010 03:28:24 +0900 (JST)
From:      Kazumaro Aoki <ka@flu.if0.org>
To:        freebsd-firewire@freebsd.org
Subject:   Re: reset with fwcontrol -R
Message-ID:  <20100111.032824.208936614.ka@flu.if0.org>
In-Reply-To: <1262639782.2209.305.camel@home-lappy>
References:  <20091228.001440.71123742.ka@flu.if0.org> <20100104.190107.74684647.ka@flu.if0.org> <1262639782.2209.305.camel@home-lappy>

next in thread | previous in thread | raw e-mail | index | archive | help
From: Sean Bruno <sean.bruno@dsl-only.net>
Subject: Re: reset with fwcontrol -R
Date: Mon, 04 Jan 2010 13:16:22 -0800

>> > I have several experiences that sudden reboot is happened just after
>> > "fwcontrol -R foo.dv" was invoked.  The operating system is FreeBSD
>> > 7.2R/i386.  Before I use 7.2R, I used 6.3R and I never experienced
>> > such phenomenon.
>> 
>> At December 27, I met the similar problem.  The system caused a panic.
(snip)
> I'm not clear what is going on from the attached information.  I have
> been looking around to see what could have changed to affect these
> problems.

Thank you for considering my problem.

> Can you try setting the firewire debug to 3 and testing?  
> sysctl debug.firewire_debug=3

Yes.  I changed the value described above and I observed the system
behavior.  The OS hardly panics, but I finally got the panic.
I copied the log message from messages without date and system name:
  start IR DMA 0x40000011
  
  
  Fatal trap 12: page fault while
  in kern
  el mode
  cpuid 
  = 0; apic id =
  00
  fault vi
  rtual addres
  s	= 0x22
  
  fault code		= 
  supervisor
  read, page 
  not presen
  t
  instruction 
  pointer	= 
  0x20:0xc05f6d8
  a
  stack poin
  ter	        
  =
  0x28:
  0xe5c3
  8c2c
  fra
  me poi
  nter	 
  
  = 0x28
  :0xe5c
  38c9c
  
  code seg
  ment		
  = base
  0x0, 
  limit 0x
  fffff,
  type 
  0x1b
  = D
  PL 0, 
  pres 1, def32 
  1, gran 1
  processor eflags	= interrup
  t enab
  led, r
  esume,
  IOPL 
  = 0
  curr
  ent pr
  ocess	
  = 42 
  (fw0_t
  askq)
  
  trap numbe
  kernel boot file is /boot/kernel/kernel

I feel that no helpful information is logged.

The panic was occurred January 7.  After this, I did not meet a panic.
At that time, I did not make a crash dump.  When the panic happens
next time, I can get a crash dump.  I used fwcontrol -R about 17 times
per day.

Ah, I forgot to mention the following fact.  When I used without -M dv
option, fwcontrol -R outputs:
  fwcontrol: detect_recv_fn: ioctl FW_SSTBUF: Bad file descriptor
So, I used fwcontrol implementation in FreeBSD 6.3R.  Moreover, when
I am using fwcontrol -R, I cannot use "fwcontrol" with no option.
When I used FreeBSD6.3R, I can see the result from fwcontrol, because
fwcontrol outputs:
  fwcontrol: main: Error opening firewire controller #0 /dev/fw0.0: Device busy
That is the another reason why I use fwcontrol implemented in FreeBSD
6.3R, though the operating system is FreeBSD7.2R.

Best regards,

Kazumaro



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?20100111.032824.208936614.ka>