On Mon, Apr 22, 2013 at 09:34:36AM +1000, Michael Ellerman wrote:
>On Fri, Apr 19, 2013 at 05:32:45PM +0800, Gavin Shan wrote:
>> The EOI handler of MSI/MSI-X interrupts for P8 (PHB3) need additional
>> steps to handle the P/Q bits in IVE before EOIing the corresponding
>> interrupt. The patch changes the EOI handler to cover that.
Thanks for your time to review it, Michael. By the way, I think I need
rebase the patch since the patch fb1b55d654a7038ca6337fbf55839a308c9bc1a7
("Using bitmap to manage MSI") has been merged to linux-next.
>> diff --git a/arch/powerpc/sysdev/xics/icp-native.c
>> b/arch/powerpc/sysdev/xics/icp-native.c
>> index 48861d3..289355e 100644
>> --- a/arch/powerpc/sysdev/xics/icp-native.c
>> +++ b/arch/powerpc/sysdev/xics/icp-native.c
>> @@ -27,6 +27,10 @@
>> #include <asm/xics.h>
>> #include <asm/kvm_ppc.h>
>>
>> +#if defined(CONFIG_PPC_POWERNV) && defined(CONFIG_PCI_MSI)
>> +extern int pnv_pci_msi_eoi(unsigned int hw_irq);
>> +#endif
>
>You don't need to #ifdef the extern. But it should be in a header, not
>here.
>
Ok. I'll put it into asm/xics.h, but I want to confirm we needn't
#ifdef when moving it to asm/xics.h?
>> @@ -89,6 +93,24 @@ static void icp_native_eoi(struct irq_data *d)
>> icp_native_set_xirr((xics_pop_cppr() << 24) | hw_irq);
>> }
>>
>> +static void icp_p8_native_eoi(struct irq_data *d)
>> +{
>> + unsigned int hw_irq = (unsigned int)irqd_to_hwirq(d);
>> + int ret;
>> +
>> + /* Let firmware handle P/Q bits */
>> +#if defined(CONFIG_PPC_POWERNV) && defined(CONFIG_PCI_MSI)
>> + if (hw_irq != XICS_IPI) {
>> + ret = pnv_pci_msi_eoi(hw_irq);
>> + WARN_ON_ONCE(ret);
>> + }
>> +#endif
>
>Why the ifdef in here? You only ever hook this function up if those are
>true, so why do you need to check them again?
>
Right. I will remove #ifdef here in next version.
>> @@ -296,6 +318,15 @@ int __init icp_native_init(void)
>> if (found == 0)
>> return -ENODEV;
>>
>> + /* Change the EOI handler for P8 */
>> +#if defined(CONFIG_PPC_POWERNV) && defined(CONFIG_PCI_MSI)
>
>This would be neater if you created CONFIG_POWERNV_MSI, like we have
>CONFIG_PSERIES_MSI.
>
Sure. I'll introduce CONFIG_PSERIES_MSI in next version.
Thanks,
Gavin
_______________________________________________
Linuxppc-dev mailing list
Linuxppc-dev@lists.ozlabs.org
https://lists.ozlabs.org/listinfo/linuxppc-dev
