Re: [Qemu-devel] ping: [PATCH v13] block/raw-posix.c: Make physical devices usable in QEMU under Mac OS X host

[Eric Blake <eblake@redhat.com>]

[-- Attachment #1: Type: text/plain, Size: 1584 bytes --]

On 02/02/2016 12:10 PM, Programmingkid wrote:

>> There was/is no leak because it qdict_get_str() returns 'const char *' and
>> so nothing needs freeing. So your change is still a backwards steps IMHO.
> 
> char filename[MAXPATHLEN];
> snprintf(filename, MAXPATHLEN, "%s", qdict_get_str(options, "filename"));
> 
> So you want the above to be changed so that it goes back to this:
> 
> const char *filename = qdict_get_str(options, "filename");

Correct.

> 
> then this code would have to be changed:
> 
> snprintf(filename, MAXPATHLEN, "%s", bsd_path);
> qdict_put(options, "filename", qstring_from_str(filename));
> 
> I could change it to this:
> 
> qdict_put(options, "filename", qstring_from_str(bsd_path));

Correct.

> 
> If I did that, then this part would not be accurate anymore:
> 
> if (strncmp(filename, "/dev/", 5) == 0) {
>             print_unmounting_directions(filename);
>             return -1;
>         }
> 
> filename would be just "/dev/cdrom" for when the user uses the optical drive. This would print incorrect unmounting directions. 

So fix it to call print_unmounting_directions(bsd_path), after hoisting
the declaration of bsd_path to be earlier to have a long enough scope.

> 
> I could add another variable that keeps track of the real device name, but that would consume more memory. It is so much easier and simpler to just use the fixed array.

And why isn't bsd_path usable for that purpose?

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


[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 604 bytes --]