Re: [PATCH 7/7] x86: print out irq nr for msi/ht -v2

[Ingo Molnar <mingo@elte.hu>]

* Bjorn Helgaas <bjorn.helgaas@hp.com> wrote:

> On Thursday 25 September 2008 11:07:17 am Yinghai Lu wrote:
> > On Thu, Sep 25, 2008 at 8:00 AM, Bjorn Helgaas <bjorn.helgaas@hp.com> wrote:
> > > On Thursday 25 September 2008 12:13:43 am Yinghai Lu wrote:
> > >> v2: fix hpet compiling
> > >>
> > >> Signed-off-by: Yinghai Lu <yhlu.kernel@gmail.com>
> > >> ---
> > >>  arch/x86/kernel/hpet.c    |    3 +++
> > >>  arch/x86/kernel/io_apic.c |    7 +++++++
> > >>  2 files changed, 10 insertions(+), 0 deletions(-)
> > >>
> > >> diff --git a/arch/x86/kernel/hpet.c b/arch/x86/kernel/hpet.c
> > >> index 422c577..686505a 100644
> > >> --- a/arch/x86/kernel/hpet.c
> > >> +++ b/arch/x86/kernel/hpet.c
> > >> @@ -467,6 +467,9 @@ static int hpet_setup_irq(struct hpet_dev *dev)
> > >>       irq_set_affinity(dev->irq, cpumask_of_cpu(dev->cpu));
> > >>       enable_irq(dev->irq);
> > >>
> > >> +     printk(KERN_DEBUG "hpet: %s is using irq %#x aka %d for MSI\n",
> > >> +                      dev->name, dev->irq, dev->irq);
> > >> +
> > >
> > > Why do we want to print the irq as hex?  Do we do that anywhere else?
> > > I think we should pick one format and stick with it, and I think
> > > decimal is the logical choice because it's the most common.
> > 
> > when sparseirq is enabled, irq for msi will be bus_nr/dev/func + 12bits.
> > so show it as hex is more straight. but we can not use hex in
> > /proc/interrupts --ABI want decimal.
> > 
> > just try to print out...in log, so user could get some idea that who
> > is using it...
> 
> I think it'd be better to just print the MSI IRQ number in decimal
> using a dev_printk().  Then the connection is obvious and we don't
> need to make everything ugly by having IRQs in both decimal and hex.
> 
> I know it's a nuisance to make hpet use dev_printk because it's
> pretty muddled as far as getting a struct device.  But I'd rather
> fix that than add confusion like this.

incidentally, in tip/timers/hpet-percpu [ new feature tree: MSI HPETs 
with per-CPU clockevents ] we have grown a struct device in that place, 
so it was easy :)

	Ingo