Re: [PATCH RFC 4/4] hvmloader:ovmf: setup E820 map

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

On Wed, Nov 20, 2013 at 10:12:11AM +0000, Ian Campbell wrote:
> On Tue, 2013-11-19 at 20:05 +0000, Wei Liu wrote:
> > On Tue, Nov 19, 2013 at 02:56:31PM -0500, Konrad Rzeszutek Wilk wrote:
> > > On Fri, Nov 15, 2013 at 03:59:25PM +0000, Wei Liu wrote:
> > > > E820 map will be used by OVMF to create memory map.
> > > > 
> > > > Signed-off-by: Wei Liu <wei.liu2@citrix.com>
> > > > ---
> > > >  tools/firmware/hvmloader/ovmf.c |   13 ++++++++++++-
> > > >  1 file changed, 12 insertions(+), 1 deletion(-)
> > > > 
> > > > diff --git a/tools/firmware/hvmloader/ovmf.c b/tools/firmware/hvmloader/ovmf.c
> > > > index c253083..52ccd0d 100644
> > > > --- a/tools/firmware/hvmloader/ovmf.c
> > > > +++ b/tools/firmware/hvmloader/ovmf.c
> > > > @@ -128,6 +128,17 @@ static void ovmf_create_smbios_tables(void)
> > > >          SMBIOS_PHYSICAL_END);
> > > >  }
> > > >  
> > > > +static void ovmf_setup_e820(void)
> > > > +{
> > > > +    struct ovmf_info *info = (void *)OVMF_INFO_PHYSICAL_ADDRESS;
> > > > +    struct e820entry *e820 = scratch_alloc(sizeof(struct e820entry)*16, 0);
> > > 
> > > 16? Why not 128 (the default in most kernels)?
> > 
> > Aye, sir. :-)
> 
> This is an internal datastructure which hvmloader generates an e820
> into, it only needs to be large enough to hold whatever hvmloader
> currently generates and not the theoretical maximum.
> 
> I think 16 is more than sufficient for now. In fact by looking at
> build_e820_table it seems that it currently generates at most 9 entries.
> 
> If you really wanted to change something here then either expose a
> suitable table size from e820.[ch] (e.g. as a #define) to
> {seabios,ovmf}.c or pass in the allocated size and validate it against
> nr in build_e820_table.

OK, I was thinking to make e820_hole parameter work with HVM. Which would
mean that the E820 can be up to 128.

I can do the right change to "allocated size and validate it against
nr in build_e820_table" when I get to this point.
> 
> Ian.
>