Re: isp(4) annoying problems on PWS500au

["Matthew Jacob" <lydianconcepts%gmail.com@localhost>]

Fascinating.

I wonder what Herr Drochner has to say about this.

On Thu, Nov 27, 2008 at 10:33 AM, Nicolas Joly <njoly%pasteur.fr@localhost> 
wrote:
> On Mon, Nov 24, 2008 at 03:13:50PM +0100, Nicolas Joly wrote:
>> On Mon, Nov 24, 2008 at 01:44:57PM +0100, Urban Boquist wrote:
>> > Hi all,
>> >
>> > >>>>> Nicolas Joly writes:
>> >
>> > Nicolas> waiting for pkc0.7.0.1004.0 to start...
>> > Nicolas> waiting for pkc0.7.0.1004.0 to start...
>> > Nicolas> waiting for pkc0.7.0.1004.0 to start...
>> > Nicolas> waiting for pkc0.7.0.1004.0 to start...
>> >
>> > I just tried a 5.0_BETA kernel on my PWS500au and I see this too. I
>> > have not seen any disk errors or anything else strange when the
>> > machine is running, but as soon as I reboot after having run on the
>> > 5.0 kernel it gets stuck in the SRM. Very annoying indeed! ;-)
>> >
>> > Nicolas - did you file a PR?
>>
>> Not yet. I was wondering if this was a more problem on my
>> machine... Thanks for confirming.
>>
>> > This machine is still running 4.0_STABLE so I could probably try to
>> > boot kernels between 4.0 and 5.0 if that could help nailing down the
>> > problem?
>>
>> I'll try to pinpoint it, if i time permit.
>
> Got it. This is not really an isp(4) problem, but a pmf(9) side effect
> ... It appeared with the pmf_system_shutdown introduction in the
> following commit:
>
> http://mail-index.netbsd.org/source-changes/2008/02/28/msg002955.html
>
> The main problem, if i understand correctly, is that with
> shutdown/halt command all pmf registered hooks are called ... And
> especially the one that disable all devices on a pci bus. This stops
> the SCSI controller which the PWS500au SRM dislike.
>
> Not sure how this can be fixed, but disabling pmf_system_shutdown make
> reboot/halt commands work as expected.
>
> --
> Nicolas Joly
>
> Biological Software and Databanks.
> Institut Pasteur, Paris.
>