Re: aic7xxx driver update

Marcelo Tosatti (marcelo@conectiva.com.br)
Tue, 24 Jun 2003 18:08:12 -0300 (BRT)


On Tue, 24 Jun 2003, Justin T. Gibbs wrote:

> > Justin,
> >
> > Would you mind sending me a inlined patch to update aic7xxx to the current
> > sources, with detailed explanation?
>
> Can you be more specific about what you need? I take it the BK send output
> on my website is not sufficient? I can probably figure out the proper
> BK magic to generate a patch for it, but I would hope that you would,
> after review, just accept the BK data so that revision history is not
> lost.

Oh that is fine. I thought you only had tarballs with your newer code.
Thats why I asked for a patch. But we're fine, then.

> The BK data has lots of changelog information since there are 108 changesets
> in my local BK tree that are not in your tree. I can provide more details
> about any particular changeset that peaks your interest.
>
> > Are there any known issues with the new code, and what are them.
>
> For both drivers, I'm currently working on an error recovery handler that
> will override the standard one. The standard handler is broken in
> so many ways that it would be hard to list them briefly. The short
> of it is that if you get a timeout, it may take 20 minutes for you
> to recover and recovery is rare. Timeouts occur in almost any system
> should you run it long enough, so this is unacceptable. Unfortunately,
> having the mid-layer do all of the recovery will likely never work.
> The mid-layer can't determine the most likely transaction outstanding
> that is causing the timeout whereas the end driver can. I hope to have
> the new error handler tested sufficiently for release by the end of this
> week.
>
> For the aic79xx driver, we are going through qualification of new drives
> and new drive firmware. There are a few issues that we are investigating
> now that should also be resolved this week.
>
> My suggestion would be to shoot for say next Monday as a good time to
> sync up. I can provide whatever material you require for released
> changesets today so that you can review all but the latest changes
> before Monday.

I plan to release -pre2 today or tomorrow and I think that having your
changes in it, even without the new error handler, is worth it for obvious
reasons.

WHats your website again? :)
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/