Re: [Xen-devel] [PATCH v5 02/11] vpci: introduce basic handlers to trap accesses to the PCI config space

["Jan Beulich" <JBeulich@xxxxxxxx>]

>>> On 08.09.17 at 16:41, <roger.pau@xxxxxxxxxx> wrote:
> On Mon, Sep 04, 2017 at 09:38:11AM -0600, Jan Beulich wrote:
>> >>> On 14.08.17 at 16:28, <roger.pau@xxxxxxxxxx> wrote:
>> > +    /*
>> > +     * At this point we have the following layout:
>> > +     *
>> > +     * Note that this refers to the position of the variables,
>> > +     * but the value has already changed from the one given at
>> > +     * initialization time because write tests have been performed.
>> > +     *
>> > +     * 32    24    16     8     0
>> > +     *  +-----+-----+-----+-----+
>> > +     *  |          r0           | 0
>> > +     *  +-----+-----+-----+-----+
>> > +     *  | r7  |  r6 |  r5 |/////| 32
>> > +     *  +-----+-----+-----+-----|
>> > +     *  |///////////////////////| 64
>> > +     *  +-----------+-----------+
>> > +     *  |///////////|    r12    | 96
>> > +     *  +-----------+-----------+
>> > +     *             ...
>> > +     *  / = empty.
>> 
>> Maybe better "unwritten"?
> 
> I've been thinking about this, and I'm not sure unwritten is better,
> in fact the test will write to this registers, it's just that there's
> no backing handlers so writes will be discarded and reads will return
> ~0.
> 
> So I think "empty" or maybe "unhandled" is more descriptive.

"unhandled" then please - registers can't possibly be empty imo.

Jan


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