Re: Xen/pv_ops features/patches for 2.6.29? (updating wiki)

[Jeremy Fitzhardinge <jeremy@goop.org>]

Nakajima, Jun wrote:
> I noticed that the changes like
>         if (xen_initial_domain())
>                 ...
>
> For example,
> static inline unsigned int io_apic_read(unsigned int apic, unsigned int reg)
> {
>         struct io_apic __iomem *io_apic;
>
>         if (xen_initial_domain())
>                 return xen_io_apic_read(apic, reg);
>
>         io_apic = io_apic_base(apic);
>         writel(reg, &io_apic->index);
>         return readl(&io_apic->data);
> }
>
>
> Do you think those need to be wrapped by pv_ops? Probably it's better to replace such functions (e.g. io_apic_read/write) with Xen-specific ones, rather than adding hooks.
>   

That's been a point of contention.  Originally I had some io_apic_ops to 
do that, but I got pushback.  The current patch is more to make a point 
than something I'd expect to be upstreamed (but who knows?).  It isn't 
actually too bad this way, because xen_initial_domain() compiles down to 
a constant 0 if CONFIG_XEN_DOM0 is disabled, and its just a test of a 
variable if its enabled.

    J