Re: [Qemu-devel] [PATCH v2 3/4] raw-posix: Fix try_seek_hole()'s handling of SEEK_DATA failure

[Eric Blake]

On 11/13/2014 06:03 AM, Kevin Wolf wrote:
> Am 13.11.2014 um 11:17 hat Markus Armbruster geschrieben:
>> When SEEK_HOLE tells us we're in a hole, we try SEEK_DATA to find its
>> end.  When that fails, we pretend the hole extends to the end of file.
>> Wrong.
> 
> Wrong only in some cases, see below.
> 
>> Except when SEEK_END fails, we screw up and claim it extends
>> to offset -1.  More wrong.


>> +++ b/block/raw-posix.c
>> @@ -1494,8 +1494,9 @@ static int try_seek_hole(BlockDriverState *bs, off_t 
>> start, off_t *data,
>>      } else {
>>          /* On a hole.  We need another syscall to find its end.  */
>>          *data = lseek(s->fd, start, SEEK_DATA);
>> -        if (*data == -1) {
>> -            *data = lseek(s->fd, 0, SEEK_END);
>> +        if (*data < 0) {
>> +            /* no idea where the hole ends, give up (unlikely to happen) */
> 
> Not quite unlikely. If the file ends with a sparse area, we'll get
> -1/ENXIO here.
> 
> lseek() with SEEK_DATA starting in a hole when there is no data until
> EOF is actually the part that isn't documented in the man page, but
> ENXIO is what I'm seeing here on RHEL 7.

Here's the (proposed) POSIX wording:

http://austingroupbugs.net/view.php?id=415

And ENXIO is indeed the expected error for SEEK_DATA on a trailing hole,
so maybe we should special case it.

-- 
Eric Blake   eblake redhat com    +1-919-301-3266
Libvirt virtualization library http://libvirt.org

signature.asc
Description: OpenPGP digital signature