Re: [Xen-devel] [PATCH v4 1/9] xen/vpci: introduce basic handlers to trap accesses to the PCI config space

["Jan Beulich" <jbeulich@xxxxxxxx>]

>>> Roger Pau Monné <roger.pau@xxxxxxxxxx> 07/14/17 6:42 PM >>>
>On Fri, Jul 14, 2017 at 10:01:54AM -0600, Jan Beulich wrote:
>> >>> On 14.07.17 at 17:33, <roger.pau@xxxxxxxxxx> wrote:
>> > On Thu, Jul 13, 2017 at 08:36:18AM -0600, Jan Beulich wrote:
>> >> >>> Roger Pau Monne <roger.pau@xxxxxxxxxx> 06/30/17 5:01 PM >>>
>> >> > --- a/xen/arch/arm/xen.lds.S
>> >> > +++ b/xen/arch/arm/xen.lds.S
>> >> > @@ -41,6 +41,9 @@ SECTIONS
>> >> >  
>> >> >    . = ALIGN(PAGE_SIZE);
>> >> >    .rodata : {
>> >> > +       __start_vpci_array = .;
>> >> > +       *(.rodata.vpci)
>> >> > +       __end_vpci_array = .;
>> >> 
>> >> Do you really need this (unconditionally)?
>> > 
>> > Right, this should have a ifdef CONFIG_PCI.
>> 
>> CONFIG_HAS_PCI for one, and then ARM doesn't select this at
>> all. Hence the question.
>
>I think it would be better to just add it now? The code is not really
>x86 specific (although it's only used by x86 ATM). IMHO adding a
>CONFIG_HAS_PCI to both linker scripts is the best solution.

Yeah, adding it with the conditional would be fine to me. Eventually
we'll want to break out common pieces like this anyway.

Jan


_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxx
https://lists.xen.org/xen-devel