Re: [PATCH 1/3] Enable non page boundary BAR device assignment

["Michael S. Tsirkin" <mst@redhat.com>]

On Tue, Dec 15, 2009 at 11:50:23AM -0800, Chris Wright wrote:
> * Alexander Graf (agraf@suse.de) wrote:
> > >>>> @@ -405,7 +511,7 @@ static int assigned_dev_register_regions(PCIRegion *io_regions,
> > >>>>              } else {
> > >>>>                  pci_dev->v_addrs[i].u.r_virtbase =
> > >>>>                      mmap(NULL,
> > >>>> -                         (cur_region->size + 0xFFF) & 0xFFFFF000,
> > >>>> +                         cur_region->size,
> > >>>>     
> > >>>>         
> > >>> Hmm, one assumes this code did work at some point ...
> > >>> do you know why was it done this way?
> > >>>       
> > >> Nope :-).
> > >
> > > So maybe keep it that way or make the change in a separate patch.
> > 
> > Well we're sure the normal path is size aligned, no? In fact, you
> > recommended I should change the code to what this looks like ;-).
> 
> Sort of moot.  The typical case is page size aligned, and the kernel is
> going to round up anyway.

Why was this code here originally then? Any idea?

-- 
MST