[3/3] IRQ: Print "unexpected IRQ" messages consistently across architectures

Michael Ellerman mpe at ellerman.id.au
Thu Jul 16 14:12:29 AEST 2015


On Mon, 2015-07-13 at 13:35 -0500, Bjorn Helgaas wrote:
> On Sun, Jul 12, 2015 at 10:23 PM, Michael Ellerman <mpe at ellerman.id.au> wrote:
> > On Sun, 2015-12-07 at 22:02:11 UTC, Bjorn Helgaas wrote:
> >> Many architectures use a variant of "unexpected IRQ trap at vector %x" to
> >> log unexpected IRQs.  This is confusing because (a) it prints the Linux IRQ
> >> number, but "vector" more often refers to a CPU vector number, and (b) it
> >> prints the IRQ number in hex with no base indication, while Linux IRQ
> >> numbers are usually printed in decimal.
> >>
> >> Print the same text ("unexpected IRQ %d") across all architectures.
> >>
> >> No functional change other than the output text.
> >
> > There's already a fallback version in asm-generic, so shouldn't you instead
> > just delete all the versions that are identical to that?
> >
> > eg. on powerpc we have:
> >
> >>  static inline void ack_bad_irq(unsigned int irq)
> >>  {
> >> -     printk(KERN_CRIT "unexpected IRQ trap at vector %02x\n", irq);
> >> +     printk(KERN_CRIT "unexpected IRQ %d\n", irq);
> >>  }
> >
> > And the generic version is:
> >
> >>  #ifndef ack_bad_irq
> >>  static inline void ack_bad_irq(unsigned int irq)
> >>  {
> >> -     printk(KERN_CRIT "unexpected IRQ trap at vector %02x\n", irq);
> >> +     printk(KERN_CRIT "unexpected IRQ %d\n", irq);
> >>  }
> >>  #endif
> >
> > So we can just delete the powerpc version?
> 
> Wow, I really didn't do my homework here.  Not only is there a generic
> version already, but there's also print_irq_desc(), which prints way
> more information than any of the ack_bad_irq() implementations.

Even better :)

> I'll try again :)

Thanks.

cheers




More information about the Linuxppc-dev mailing list