Re: [PATCH RFC] watchdog: add a new driver for VIA chipsets

[Dmitry Artamonow]

On 15:12 Wed 23 Nov     , Marc Vertes wrote:
> Dmitry Artamonow <mad_soft@xxxxxxxx> wrote:
> 
> > "via_wdt" seems to be a too generic name. What if someone will want to
> > write driver for some incompatible watchdog also made by VIA?
> > Maybe it's worth to use something more specific - say, "via_vx8xx_wdt"?
> >
> According to Harald, it is the same watchdog on all via chipsets and
> back to Athlon chipsets. But why not.

In that case I think it's ok to leave via_wdt as is.

> > Why is to rely on BIOS setup? Is there really no information on what
> > registers to write to start watchdog? Or you was just lazy to find it?
> > It took me less than a minute to google up a datasheet[1] on VX800/VX820
> > and it seems to have all needed information to implement proper start()
> > and stop() functions. I suppose info on watchdog registers should be
> > applicable to VX855 as well - but it's better to check on your hardware
> > anyway.
> >
> > [1] http://linux.via.com.tw/support/beginDownload.action?eleid=161&fid=241
> >
> I rely on BIOS setup because it is the only working method on the
> hardware I could test on. This watchdog stuff exists on VIA cpus for
> years (back to Athlon chipsets) and still no working open
> source driver. Wonder why?
> 
> Of course I read the datasheet of the VX855 southbridge prior to
> write this driver. And the VX800 one, the CX700 one. Unfortunately,
> the instructions to start, stop, get and set counters simply
> do not work. The doc is false and/or incomplete. Harald
> already implemented a full version of this driver back in 2009
> (http://git.gnumonks.org/cgi-bin/gitweb.cgi?p=linux-2.6-via.git;a=shortlog;     h=refs/heads/via-wdt),
> but not working. Questions to VIA remain unanswered.
> 
> So I decided to keep in the source code only what really works, and not
> the well documented but non functional stuff. If you can test or find
> hardware where those instructions work as advertised, I will be happy
> to expose them again in an updated driver.
> 
> The current result, in addition of actually working, is safe,
> simple, and compatible with potential "correct" configurations.

Ok. Thanks to you and Harald for explaining! That's sad HW doesn't
behave according to datasheets. Maybe it's worth to add a note about this
in the driver to warn future hackers? :)

-- 
Best regards,
Dmitry "MAD" Artamonow

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