Re: lockdep and threaded IRQs (was: ...)

[David Brownell <david-b@pacbell.net>]

On Monday 02 March 2009, Ingo Molnar wrote:
> 
> > > > What's unfortunate is that you prefer not to fix that
> > > > IRQF_DISABLED bug in lockdep, which you co-"maintain".
> > > > When running with lockdep, that bug (a) introduces bugs
> > > > in some drivers and (b) hides bugs in others.  You've
> > > > rejected even a minimal warning fix, to help minimize
> > > > the amount of time developers waste on (a) and (b).
> > > 
> > > I've come to the conclusion that the only technically sound solution is
> > > to do as I proposed today, utterly eliminate !IRQF_DISABLED handlers.
> > 
> > As you announced today.  If you truly believe that, then
> > you should at least submit a warning patch for 2.6.29-rc
> > ("driver X isn't setting IRQF_DISABLED, reimplement!")
> 
> i have changed the BUG_ON() to a WARN_ONCE() message so the 
> warning is in place now.

The patch Peter sent doesn't relate in the least to removing
the IRQF_DISABLED flag though.  Patches addressing that
would be in setup_irq() code paths not IRQ dispatch.


> > with a Documentation/feature-removal-schedule.txt plan for 
> > removing that mechanism. [...]
> 
> you are misunderstanding the workings and purpose of 
> feature-removal-schedule.txt. It is mainly used for 
> functionality that is user-visible.

If by "user" you include "kernel developers", yes;
otherwise, I'd dispute "mainly".  The first several
entries right now relate to kernel interfaces, as
do quite a lot of the others.


> It is sometimes used for  
> functionality that a subsystem has exported to a lot of drivers 
> consciously and which is being removed.

The IRQ framework has very consciously exported IRQF_DISABLED.
That functionality has been around for a very long time; I'm
thinking at least ten years now.

So removing IRQF_DISABLED -- if it's even agreed to be
a good idea, which seems to be a minority opinion so far
on this thread -- is very much the sort of thing one would
expect to appear in that schedule.

--
To unsubscribe from this list: send the line "unsubscribe linux-input" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html