Linux 2.4.21-rc1 : aic7xxx deadlock on boot on my machine

Linux 2.4.21-rc1 : aic7xxx deadlock on boot on my machine

[Eric Valette]

I just wants to point out on this list that since pre7 (or using 
pre4-acxx), I boot once out of two on my dual adaptec scsi machine.
The good news is that upgrading to more recent aic7xxx drivers fixes the 
problemas indeed the locking strategy has been changed... BTW, 2.5.68 
includes the new code also.

So I hope, it will be updated in rc2...

-- 
    __
   /  `                   	Eric Valette
  /--   __  o _.          	6 rue Paul Le Flem
(___, / (_(_(__         	35740 Pace

Tel: +33 (0)2 99 85 26 76	Fax: +33 (0)2 99 85 26 76
E-mail: eric.valette@free.fr

Re: Linux 2.4.21-rc1 : aic7xxx deadlock on boot on my machine

[Marc-Christian Petersen]

On Tuesday 22 April 2003 10:37, Eric Valette wrote:

Hi Eric,

> I just wants to point out on this list that since pre7 (or using
> pre4-acxx), I boot once out of two on my dual adaptec scsi machine.
> The good news is that upgrading to more recent aic7xxx drivers fixes the
> problemas indeed the locking strategy has been changed... BTW, 2.5.68
> includes the new code also.
> So I hope, it will be updated in rc2...
I'll _bet_ that the new well good code from Justin won't make it into 2.4 
earlier than 2.4.22-pre1.

ciao, Marc

Re: Linux 2.4.21-rc1 : aic7xxx deadlock on boot on my machine

[Eric Valette]

Marc-Christian Petersen wrote:

> I'll _bet_ that the new well good code from Justin won't make it into 2.4 
> earlier than 2.4.22-pre1.

I do not really see why because 2.4.20 -> 2.4.21 contains already such a 
driver change. Note that I mailed alan and justin on pre4-acxx. Justin 
suggested various fixes that did not resolve the problem and then 
decided to do a code inspection. Reading the Changelog, indeed the 
locking startegy has changed compared to the rc1 version. I mailed 
arcello again on pre7 even providing kdbg backtrace.

Never mind : the report is more informative to others as I already 
patched my kernel to the last version...

Thanks for responding, and providing integrated patch sets.

Have a good day.

-- 
    __
   /  `                   	Eric Valette
  /--   __  o _.          	6 rue Paul Le Flem
(___, / (_(_(__         	35740 Pace

Tel: +33 (0)2 99 85 26 76	Fax: +33 (0)2 99 85 26 76
E-mail: eric.valette@free.fr

Re: Linux 2.4.21-rc1 : aic7xxx deadlock on boot on my machine

[Alan Cox]

On Maw, 2003-04-22 at 09:36, Marc-Christian Petersen wrote:
> > So I hope, it will be updated in rc2...
> I'll _bet_ that the new well good code from Justin won't make it into 2.4 
> earlier than 2.4.22-pre1.

Its not the good code that worries me - its what bits of it turn out to
be buggy

Re: Linux 2.4.21-rc1 : aic7xxx deadlock on boot on my machine

[Eric Valette]

Alan Cox wrote:
> On Maw, 2003-04-22 at 09:36, Marc-Christian Petersen wrote:
> 
>>>So I hope, it will be updated in rc2...
>>
>>I'll _bet_ that the new well good code from Justin won't make it into 2.4 
>>earlier than 2.4.22-pre1.
> 
> 
> Its not the good code that worries me - its what bits of it turn out to
> be buggy

On the other hand, probably mainy/every server with  dual scsi cards 
(one for 160 Mbps and other for slow devices as tape, cdrom, ...) will 
probably not boot with actual good blessed code :-)

Never mind :
	1) I have warned and done my debugging duty by sending a kdbg backtrace,
	2) I suggested a possible fix,
	3) I have a solution for myself,



-- 
    __
   /  `                   	Eric Valette
  /--   __  o _.          	6 rue Paul Le Flem
(___, / (_(_(__         	35740 Pace

Tel: +33 (0)2 99 85 26 76	Fax: +33 (0)2 99 85 26 76
E-mail: eric.valette@free.fr

Re: Linux 2.4.21-rc1 : aic7xxx deadlock on boot on my machine

[Alan Cox]

On Maw, 2003-04-22 at 20:34, Eric Valette wrote:
> On the other hand, probably mainy/every server with  dual scsi cards 
> (one for 160 Mbps and other for slow devices as tape, cdrom, ...) will 
> probably not boot with actual good blessed code :-)
> 
> Never mind :
> 	1) I have warned and done my debugging duty by sending a kdbg backtrace,
> 	2) I suggested a possible fix,
> 	3) I have a solution for myself,

Indeed, and its on the list to try and work through. The size of the changes from
.0 to .6 make it non trivial though