Re: [PATCH/RFC 1/2] simple SPI framework

[Mark Underwood <basicmark@yahoo.com>]

--- David Brownell <david-b@pacbell.net> wrote:

> > > >+/* Suspend/resume in "struct device_driver" don't really need that
> > > >+ * strange third parameter, so we just make it a constant and expect
> > > >+ * SPI drivers to ignore it just like most platform drivers do.
> > > >+ *
> > >
> > > So you just ignored my letter on that subject :(
> > > The fact that you don't need it doesn't mean that other people won't.
> > > The fact that there's no clean way to suspend USB doesn't mean that 
> > > there shouldn't be one for SPI.
> >
> > The third parameter is obsolete and should only be used to select _one_
> > of the tree suspend calls you will get.
> 
> Vitaly ... comments from Russell and Pavel both addresses your comments
> about that obsolete parameter.  What letter?  The one I remember was
> one responding to Mark Underwood (?) where you complained about issuing
> three calls for one suspend event.  You can't have it both ways!!
> Either that parameter should be used in the documented way (call the
> suspend method three times, one right after another) or it should be used
> more sanely (parameter is constant.

Yes, that was in reply to my SPI subsystem patch set (in which Vitaly didn't like the fact that I
call suspend/resume 3 times) and then in the same thread (in answer to David's response of
dropping this as he didn't think anyone would mind this) Vitaly said that you can't do this.

> 
> USB can suspend just fine, thank you, though starting with 2.6.12 some
> bugs seem to have crept in; fixes are in the 2.6.15 prepatchces.
> 
> 
> > Any additional suspend calls should _not_ create extra usage of this
> > parameter.  It's a left over from Pat's first driver model incarnation
> > which is specific to the platform device drivers.  (Mainly it exists
> > because no one can be bothered to clean it up.)
> 
> Most folk who've considered the question would like to see it go away.
> Except ... making sure every driver in a few dozen architectures still
> builds after removing that parameter is more than the usual amount of
> janitorial work!
> 
> Progress could start by updating Documentation/driver-model/driver.txt to
> say "don't test that parameter", reducing future confusion on this point.

Thank you! That would clear up this confusion :).

> 
> - Dave
> 
> -
> 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/
> 



		
___________________________________________________________ 
To help you stay safe and secure online, we've developed the all new Yahoo! Security Centre. http://uk.security.yahoo.com