Re: [PATCH][RFC] Proposal for a new watchdog interface using sysfs

From: Patrick Mochel (
Date: Thu Feb 13 2003 - 14:19:54 EST

On 13 Feb 2003, Rusty Lynch wrote:

> On Thu, 2003-02-13 at 11:07, Dave Jones wrote:
> > On Thu, Feb 13, 2003 at 07:51:45AM -0800, Rusty Lynch wrote:
> >
> > > > You could regard them as 'system' devices, and have them show up in
> > > > devices/sys/, which would make more sense than 'legacy'.
> > > Ok, system device is the winner.
> >
> > Why? Stop for a second and look what we have in those dirs.
> > They both contain things that are essentially motherboard resources.
> >
> > These are add-on cards we're talking about. Surely a more sensible
> > place for them to live is somewhere under devices/pci0/ or whatever
> > bus-type said card is for.
> >
> > Whilst there are some watchdogs which _are_ part of the motherboard
> > chipset (which is arguably 'system'), these still show up in PCI
> > space as regular PCI devices.
> >
> > Lumping them all into the same category as things like rtc, pic,
> > fdd etc is just _wrong_.
> >
> > Dave
> >
> The thing I would like to see is an easy way for a user space
> application to see the available watchdog devices without searching
> every possible bus type. If we had that one location to find all
> watchdog devices, then each device could just be a symbolic link to the
> device in it's real bus.

Create a watchdog timer class. That will contain all watchdog timers, no
matter what bus they are on.

I apologize for leading you astray with suggesting you treat them as
system devices; I was under the assumption they were more important. :)
They should always be in the most accurate place in the tree. Don't worry
about what the user sees; consistency and accuracy are more important..


To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to
More majordomo info at
Please read the FAQ at

This archive was generated by hypermail 2b29 : Sat Feb 15 2003 - 22:00:48 EST