Hi,
On Wed, May 16, 2007 at 01:04:48PM -0500, Shapor Naghibzadeh wrote:
>
> I noticed your patch reads the blkid cache when guessing the filesystem type,
> which is the one difference from my patch. I left that out because the cost
> of using libblkid is almost certainly more expensive than simply using the
> filesystem detection logic, especially with a large number of devices. Also,
> that way libblkid is used for what it is intended to optimize: caching uuids
> and labels to prevent having to scan n**2 devices in order to mount by uuid.
Good point. Unfortunately
blkid_get_tag_value(NULL, "TYPE", devname);
is bad solution, because the libblkid internally initializes the cache:
char *blkid_get_tag_value(blkid_cache cache, const char *tagname,
const char *devname)
{
......
if (!cache) {
if (blkid_get_cache(&c, NULL) < 0)
return NULL;
}
.......
if (!cache)
blkid_put_cache(c);
it means you save nothing if you don't call the blkid_get_cache() in
the mount(8).
I think possible workaround is:
blkid_get_cache(&cache, "/dev/null");
blkid_get_tag_value(&cache, "TYPE", devname);
(I'll try to use this solution if the cache is not initialized yet.)
The best solution would be to add interface for a direct FS probing to
libblkid API, something like blkid_verify(), but without the cache.
blkid_fsprobe(devname, &uuid, &label, &fstype);
Ted?
> See my test case documented at: http://www.shapor.com/libblkid/, with a link
> to the thread on linux-ext4 which I cc'ed the maintainer of util-linux on with
> no response (yet).
Nice test.
> Its refreshing to see work being done on util-linux with an active mailing
> list and some issues with core Linux infrastructure being attended to. Keep
> up the good work!
Thanks.
Karel
--
Karel Zak <kzak@xxxxxxxxxx>
-
To unsubscribe from this list: send the line "unsubscribe util-linux-ng" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
