On Thu, Jun 13, 2013 at 10:06:43AM +0800, Fam Zheng wrote:
> On Mon, 06/10 16:32, Stefan Hajnoczi wrote:
> > On Mon, Jun 10, 2013 at 04:04:55PM +0200, Kevin Wolf wrote:
> > > Am 10.06.2013 um 11:07 hat Stefan Hajnoczi geschrieben:
> > > > Remember to byteswap VMDK4Header.desc_offset on big-endian machines.
> > > >
> > > > Cc: qemu-sta...@nongnu.org
> > > > Signed-off-by: Stefan Hajnoczi <stefa...@redhat.com>
> > >
> > > Thanks, applied to the block layer.
> > >
> > > > @@ -507,8 +507,11 @@ static int vmdk_open_vmdk4(BlockDriverState *bs,
> > > > if (ret < 0) {
> > > > return ret;
> > > > }
> > > > - if (header.capacity == 0 && header.desc_offset) {
> > > > - return vmdk_open_desc_file(bs, flags, header.desc_offset << 9);
> > > > + if (header.capacity == 0) {
> > > > + int64_t desc_offset = le64_to_cpu(header.desc_offset);
> > > > + if (desc_offset) {
> > > > + return vmdk_open_desc_file(bs, flags, desc_offset << 9);
> > > > + }
> > > > }
> > >
> > > Splitting up the if condition wouldn't have been necessary, strictly
> > > speaking. But I don't mind too much here.
> >
> > True. The reason I did it is because accessing header.desc_offset
> > directly is a bad habit. Someone modifying the code might conclude it's
> > safe to access directly when it actually only works for the limited
> > cases of zero and non-zero.
> >
>
> Not byteswapping header.capacity here, any reason?
Byteswapping header.capacity too is fine by me. This patch is focussed
on just header.desc_offset.
Stefan
