[PATCH v3 3/5] PCI: st: Provide support for the sti PCIe controller

[Gabriel Fernandez]

Hi

Thanks for reviewing.

On 11 April 2015 at 16:55, Arnd Bergmann <arnd at arndb.de> wrote:
> On Saturday 11 April 2015 12:17:57 Paul Bolle wrote:
>> Something I didn't spot in my first look at this patch.
>>
>> On Fri, 2015-04-10 at 11:12 +0200, Gabriel FERNANDEZ wrote:
>> > --- a/drivers/pci/host/Kconfig
>> > +++ b/drivers/pci/host/Kconfig
>> >
>> > +config PCI_ST
>> > +     bool "ST PCIe controller"
>> > +     depends on ARCH_STI || (ARM && COMPILE_TEST)
>> > +     select PCIE_DW
>> > +     help
>> > +       Enable PCIe controller support on ST Socs. This controller is based
>> > +       on Designware hardware and therefore the driver re-uses the
>> > +       Designware core functions to implement the driver.
>>
>> You can't have ARCH_STI without ARM, so ARM will always be set if this
>> driver is enabled. Correct?
>
> Right, though the ARM dependency could soon be dropped, once the PCIE_DW
> driver can use generic infrastructure in the few places it relies on
> ARM specific code today.
>
>> > --- /dev/null
>> > +++ b/drivers/pci/host/pci-st.c
>>
>> > +     if (IS_ENABLED(CONFIG_ARM)) {
>> > +             /*
>> > +              * We have to hook the abort handler so that we can intercept
>> > +              * bus errors when doing config read/write that return UR,
>> > +              * which is flagged up as a bus error
>> > +              */
>> > +             hook_fault_code(16+6, st_pcie_abort_handler, SIGBUS, 0,
>> > +                             "imprecise external abort");
>> > +     }
>>
>> So, unless I'm missing something obvious here, IS_ENABLED(CONFIG_ARM)
>> will always evaluate to 1. Can't that test be dropped?
>
> I would leave it in, as it's quite likely to get reused with ARM64 at some
> point in the future (no, I don't know anything about ST's product plans,
> but everybody seems to be doing this).
>
>         Arnd

Yes i agree with that.

Gabriel