On Tue, Jun 09, 2015 at 10:27:09PM +0100, Russell King - ARM Linux wrote: > It's because there exist platforms where specific IRQs are permanently > asserted, and unexpectedly claiming such an interrupt (eg, through IRQ > probing) would lock the system - at least before we ended up with the > detection in genirq (which has its roots in the ARM code.)
Ah, yes - now I remember. I just had a vauge recollection of broken probing. > I'd suggest people think carefully about applying these patches. They > have only been around for a matter of hours, so to rush to apply them > when they haven't been revewied as a whole to assess whether the approach > is the right one is rather hasty. I'm fairly comfortable with the patch for regmap as that code is architecture neutral, should already have been consistent and the non-ARM path has definitely been tested (I've done it myself).
signature.asc
Description: Digital signature
