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

On Tuesday 03 March 2009, Ingo Molnar wrote:

i think you severely over-estimate the importance and ratio of 
drivers that enable irqs within irq handlers. (Nor does anyone 
want to break them really - we want to have a sane default and 
we want to flag the broken cases as broken.)

For the record, I've been running for some time now
with a patch that issues a warning for each IRQ that
lockdep forces to use IRQF_DISABLED.

On my x86 systems, pretty much every driver triggers
that warning.  Which makes me think maybe that shoe
is being placed on the wrong foot:  use of IRQF_DISABLED
is the *EXCEPTION* not the rule.  At least on one major
Linux platform...