Re: [Libvir] PATCH: 11/16: directory/fs/netfs backend
On Fri, Feb 15, 2008 at 09:37:53PM +0100, Jim Meyering wrote:
It'd be good to pull the duplicate capacity update out of the
if- and else- blocks.
def->capacity *= ...
Also, just in case the file data is bogus enough to make the resulting
product overflow, you can detect that with e.g.,
if (def->capacity > ULLONG_MAX / fileTypeInfo[i].sizeMultiplier)
continue;
def->capacity *= fileTypeInfo[i].sizeMultiplier;
If the test fails, I'd be tempted to give a diagnostic
about the corrupt "size" data.
Originally I thought there would be no overflow risk, since the only file
types with a sizeMultiplier were those with a 4 byte size field. But I see
now that actually VMDK 4 format has an 8 byte size and a 512 multiplier.
So I'll add the sanity check...
Dan.
--
|=- Red Hat, Engineering, Emerging Technologies, Boston. +1 978 392 2496 -=|
|=- Perl modules: http://search.cpan.org/~danberr/ -=|
|=- Projects: http://freshmeat.net/~danielpb/ -=|
|=- GnuPG: 7D3B9505 F3C9 553F A1DA 4AC2 5648 23C1 B3DF F742 7D3B 9505 -=|