Re: [PATCH] irq/irq_domain: Quit ignoring error returns from irq_alloc_desc_from().

[Grant Likely <grant.likely@secretlab.ca>]

On Mon, 09 Apr 2012 09:56:30 -0700, David Daney <ddaney.cavm@gmail.com> wrote:
> On 04/06/2012 06:26 PM, Grant Likely wrote:
> > On Thu,  5 Apr 2012 16:52:13 -0700, David Daney<ddaney.cavm@gmail.com>  wrote:
> >> From: David Daney<david.daney@cavium.com>
> >>
> >> In commit 4bbdd45a (irq_domain/powerpc: eliminate irq_map; use
> >> irq_alloc_desc() instead) code was added that ignores error returns
> >> from irq_alloc_desc_from() by (silently) casting the return value to
> >> unsigned.  The negitive value error return now suddenly looks like a
> >> valid irq number.
> >>
> >> Commits cc79ca69 (irq_domain: Move irq_domain code from powerpc to
> >> kernel/irq) and 1bc04f2c (irq_domain: Add support for base irq and
> >> hwirq in legacy mappings) move this code to its current location in
> >> irqdomain.c
> >>
> >> The result of all of this is a null pointer dereference OOPS if one of
> >> the error cases is hit.
> >>
> >> The fix: Don't cast away the negativeness of the return value and then
> >> check for errors.
> >>
> >> Signed-off-by: David Daney<david.daney@cavium.com>
> >> ---
> >>   kernel/irq/irqdomain.c |   11 ++++++-----
> >>   1 files changed, 6 insertions(+), 5 deletions(-)
> >>
> >> diff --git a/kernel/irq/irqdomain.c b/kernel/irq/irqdomain.c
> >> index af48e59..9d3e3ae 100644
> >> --- a/kernel/irq/irqdomain.c
> >> +++ b/kernel/irq/irqdomain.c
> >> @@ -351,6 +351,7 @@ unsigned int irq_create_mapping(struct irq_domain *domain,
> >>   				irq_hw_number_t hwirq)
> >>   {
> >>   	unsigned int virq, hint;
> >> +	int irq;
> >
> > Merged, but I've dropped the new variable in favour of making virq an
> > int.  Makes for a smaller diffstat.
> >
> 
> Thanks Grant,
> 
> I had thought about that too, but since virq throughout all the rest of 
> the code is unsigned, I didn't want to introduce an inconsistency.
> 
> After a little more thought, I think that the domain of virq and the irq 
> used by the rest of the kernel are the same, so it might make sense to 
> change virq to be int universally, and use the kernel convention that 
> negative numbers indicate error conditions.  But that would be a much 
> larger patch.

... touching pretty much *every* driver in the kernel!  Blech!

Yeah, that's not going to happen.  As a rule, irq numbers are always
unsigned, but there are a few apis that can return either '0' meaning
no irq, or a negative value indicating an error.  The irq_alloc_desc
apis unfortunately are one such case.

g.