Re: [PATCH for 4.5 v7 1/1] Add mmio_hole_size

[Konrad Rzeszutek Wilk <konrad.wilk@oracle.com>]

On Tue, Oct 21, 2014 at 08:46:16AM +0100, Jan Beulich wrote:
> >>> On 21.10.14 at 00:01, <dslutz@verizon.com> wrote:
> > On 10/20/14 10:17, Ian Campbell wrote:
> >> On Mon, 2014-10-13 at 08:51 -0400, Don Slutz wrote:
> >>> --- a/tools/libxl/libxl_types.idl
> >>> +++ b/tools/libxl/libxl_types.idl
> >>> @@ -391,6 +391,7 @@ libxl_domain_build_info = Struct("domain_build_info",[
> >>>                                          ("timeoffset",       string),
> >>>                                          ("hpet",             libxl_defbool),
> >>>                                          ("vpt_align",        libxl_defbool),
> >>> +                                       ("mmio_hole_size",   uint64),
> >> Please make this a MemKB at the libxl interface level and convert
> >> internally to whatever hvmloader expects. It seems that a more
> >> conventional name would also have a _memkb suffix. Perhaps
> >> mmio_hole_memkb? (size seems to be implicit to me).
> >>
> > 
> > Ok, Jan Beulich had proposed the name with the _size (on 27 Jun 2014), 
> > but he
> > did also say "(whether the xl config option should also get renamed I'm not
> > sure - the xl maintainers will know )", (so I am taking this as ok from 
> > Jan) I will
> > rename to mmio_hole_memkb here.
> 
> My remark was really intended for you to actively inquire about the
> config option name.
> 
> >>> +            /*
> >>> +             * With HVM_BELOW_4G_RAM_END == 0xF0000000, mmio_hole_size
> >>> +             * must be >= 256 MiB and <= 3840 MiB.
> >> Isn't this just restating the if condition in a way which is liable to
> >> get out of sync if the values of HVM_BELOW_4G_* ever changes?
> > 
> > Well, Konrad asked for this:
> > [...]
> > So I do not know which way to go here.
> 
> I have to admit that I think Konrad is asking for too much commentary
> now and then. But in any event - Ian being one of the maintainers of
> the affected code, what he asks you to do generally overrules what
> any non-maintainer may have asked for.

<chuckles>

Jan and Ian are correct - you can ignore my comments when they have
an different idea - as they are ultimately on the hook.

> 
> Jan
>