Re: Configurable interrupt sources, and DT bindings

[David Gibson <david-xT8FGy+AXnRB3Ne2BGzF6laj5H9X9Tb+@public.gmane.org>]

On Wed, Nov 30, 2011 at 09:33:06AM +0000, Mark Brown wrote:
> On Wed, Nov 30, 2011 at 04:13:49PM +1100, David Gibson wrote:
> > On Tue, Nov 29, 2011 at 10:55:38AM +0000, Mark Brown wrote:
> > > On Tue, Nov 29, 2011 at 12:30:55PM +1100, David Gibson wrote:
> 
> > > > Um.. why?  These new properties don't appear to give any information
> > > > that isn't already in the interrupts property (albeit in irq
> > > > controller dependent form).
> 
> > > If nothing else because doing a custom binding for every single
> > > interrupt controller out there is depressingly redundant; even if we've
> > > got a bunch of legacy devices with odd bindings it's going to be better
> > > all round if we have something new devices can just pick up.  Pretty
> > > much every single interrupt controller in the embedded space is going to
> > > need this information.
> 
> > Ok, so feel free to define a semi-standard way of encoding this
> > information in interrupt specifiers, for use by all new PIC bindings.
> > But these new properties still duplicate information that already has
> > to be in the interrupts property.  Worse, it provides no way of
> > specifying different information for each irq, if the device has
> > several.
> 
> I'm sorry, I'm not following you.  In what way is this duplicating
> information that's already there,

Each interrupt source device has an 'interrupts' property which is an
array of irq specifiers.  Each irq specifier contains enough
information for the irq controller driver to handle it - that is, both
the irq number and the type (if applicable for this irq controller).
The irq controller driver must already know how to interpret these irq
specifiers, or nothing would work.

> and for what reason is it not possible
> to specify different information per IRQ?

IIUC, the binding you proposed had individual boolean properties for
the trigger types / polarities.  Those properties would have to be on
or off for the whole node, but that node might have multiple irqs.

>  If the binding is custom to
> each device then presumably it's not possible make any general comment
> about the bindings...

They're only custom within certain constraints.

> Clearly if people are defining per-chip bindings that can't cope with
> configuring different interrupts differently that's insane

Not necessarily.  If the irq controller only handles one irq type in
hardware, there is no need for its irq specifiers in the DT to be able
to specify other types.

> (and all the
> more reason to come up with a standard binding so they don't have to
> think)

Well, we kind of have one already.  I believe the binding for several
recently defined PICs is source # in the first cell, trigger/polarity
in the second, encoded using the same values as the corresponding
Linux constants.

> but this all seems so far off base that I fear there must be some
> fundamental misunderstaniding here.

I tend to agree.

> > > Actually there's a large element of things being fixed by the board but
> > > the hardware being able to run with a wide range of board configurations
> > > depending on what amuses the electrical engineers and what restrictions
> > > the other devices in the system have.  For example, if there are pulls
> > > on the board we would want the idle configuration to be whatever the
> > > pull value is.
> 
> > Sure, but that's nothing new.  Board constraints are part of the fixed
> > hardware information that can already be specified by fixed interrupt
> > specifiers in the DT.
> 
> But according to what you say above each interrupt controller is on its
> own when it comes to coming up with a mechanism for specifying this?

Well, they  could just take the  irq specifier format  from a suitably
similar existing PIC binding.  e.g. UIC or IPIC (which use the same
format IIRC).

I really don't see what problem you're trying to solve.

-- 
David Gibson			| I'll have my music baroque, and my code
david AT gibson.dropbear.id.au	| minimalist, thank you.  NOT _the_ _other_
				| _way_ _around_!
http://www.ozlabs.org/~dgibson