Re: pciehp borkage.

[Josh Boyer]

On 6/26/06, Greg KH <gregkh@xxxxxxx> wrote:
> On Mon, Jun 26, 2006 at 11:37:49PM -0400, Dave Jones wrote:
> > My head hurts..
> >
> > drivers/pci/pcie/Kconfig ..
> >
> > config HOTPLUG_PCI_PCIE
> >     tristate "PCI Express Hotplug driver"
> >     depends on HOTPLUG_PCI && PCIEPORTBUS && (BROKEN || ACPI)
> >
> >
> >
> > but drivers/pci/hotplug/Makefile has..
> >
> > pciehp-objs     :=  pciehp_core.o   \
> >                 pciehp_ctrl.o   \
> >                 pciehp_pci.o    \
> >                 pciehp_hpc.o
> >
> > So it gets built regardless of the option, which leaves ppc (among others)
> > totally busted..
>
> Yes, this driver does have issues on ppc, see the archives for Anton
> trying to fix it up to get it to build.  But as ppc currently doesn't
> _have_ pci express hotplug hardware it really doesn't matter much :)
>
> > In file included from include/acpi/platform/acenv.h:140,
> >                  from include/acpi/acpi.h:54,
> >                  from drivers/pci/hotplug/pciehp_hpc.c:41:
> > include/acpi/platform/aclinux.h:59:22: error: asm/acpi.h: No such file or directory
> > In file included from include/acpi/acpi.h:55,
> >                  from drivers/pci/hotplug/pciehp_hpc.c:41:
> > include/acpi/actypes.h:129: error: expected '=', ',', ';', 'asm' or '__attribute__' before 'UINT64'
> > include/acpi/actypes.h:130: error: expected '=', ',', ';', 'asm' or '__attribute__' before 'INT64'
> > make[3]: *** [drivers/pci/hotplug/pciehp_hpc.o] Error 1
> > make[2]: *** [drivers/pci/hotplug] Error 2
> > make[1]: *** [drivers/pci] Error 2
> >
> >
> > Should that Makefile be more along the lines of..
> >
> > pciehp-$(CONFIG_PCI_PCIE)     :=  pciehp_core.o   \
> >                 pciehp_ctrl.o   \
> >                 pciehp_pci.o    \
> >                 pciehp_hpc.o
> >
> > perhaps ?
>
> No, look up a bit higher:
>         obj-$(CONFIG_HOTPLUG_PCI_PCIE)          += pciehp.o
>
> which will build pciehp or not.  Just don't enable the option for now
> on ppc please.  Until people sanitize the ACPI headers for non-acpi
> arches (which is currently underway...)

Would it be sane to make the Kconfig refuse to enable the option for
archs that this is known to be broken on?

josh
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/