On Fri, Jun 12, 2009 at 07:39:11AM -0700, Pan, Jacob jun wrote:
> >> +#if defined(CONFIG_X86_IO_APIC) && defined(CONFIG_SFI)
> >> + /* For platforms only have IOAPIC, the PCI irq line is 1:1 mapped to
> >> + * IOAPIC RTE entries, so we just enable RTE for the device.
> >> + */
> >> + if (platform_has(X86_PLATFORM_FEATURE_IOAPIC) &&
> >> + !platform_has(X86_PLATFORM_FEATURE_8259) &&
> >> + !platform_has(X86_PLATFORM_FEATURE_ACPI) &&
> >> + !platform_has(X86_PLATFORM_FEATURE_BIOS) &&
> >> + pin) {
> >> + ioapic = mp_sfi_find_ioapic(dev->irq);
> >> + io_apic_set_pci_routing(ioapic,
> >> + dev->irq, /* IOAPIC pin */
> >> + dev->irq, /* IRQ line */
> >> + 1, /* level trigger */
> >> + 1); /* polarity active low */
> >> + }
> >> +#endif
> >
> >Does there really need to be a #ifdef here? Wouldn't the first
> >platform_has() catch things? And what is CONFIG_SFI?
> >
> [[JPAN]] I agree, platform_has() check should be sufficient. Perhaps
> we should have a dummy io_apic_set_pci_routing() when
> CONFIG_X86_IO_APIC is not defined?
Yes, that would be a good idea to have.
> SFI is the Simple Firmware Interface, it provides static system tables
> similar to ACPI. System devices such as APIC and timers are included
> in the SFI tables.
But I do not see CONFIG_SFI or the sfi.h header file in Linus's kernel
tree. Has it been submitted to some subsystem tree already?
And can you build a single kernel image that supports SFI and ACPI? It
looks from this patch that this is not the case, which is not a good
thing for the distros at all.
thanks,
greg k-h
--
To unsubscribe from this list: send the line "unsubscribe linux-pci" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
