Re: [Xen-devel] [PATCH RFC] xen: arm: context switch vtimer PPI state.

[Ian Campbell <ian.campbell@xxxxxxxxxx>]

On Tue, 2015-03-03 at 12:18 +0000, Ian Campbell wrote:
> > > > > > +    /* Xen s/w state */
> > > > > > +    unsigned long inprogress:1;
> > > > > > +};
> > > > 
> > > > Using a uint32_t bitmask would be more in line the rest of the code
> > > > style, but it is just a matter of taste.
> > > 
> > > You mean "uint32_t inprogress:1" for each? Or
> > > #define ENAVBLED 1
> > > #define PENDING 2
> > > etc
> > > and test_set_bit stuff?
> >  
> > Something like status in xen/include/asm-arm/vgic.h:struct pending_irq
> 
> OK.

Actually, what is the benefit? I think pending_irq.status is done this
way because it is used lock free from multiple contexts and this allows
for test_and_set_bit etc to be used.

The hwppi state isn't accessed like that, so I think using the normal
appropriate types like we do in stuct arch_{domain,vcpu} is the
appropriate style. struct hwppi_state is even defined in domain.h.

Ian.


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