On Fri, Mar 13, 2009 at 16:11, Karel Zak <kzak@xxxxxxxxxx> wrote: > On Fri, Mar 13, 2009 at 03:40:10PM +0100, Kay Sievers wrote: >> On Fri, Mar 13, 2009 at 15:16, Karel Zak <kzak@xxxxxxxxxx> wrote: >> > If anyone has any objections, please let us know. >> > >> > This change has been requested by many people. We already have >> > really bad experience with more valid FS signatures on one device. >> > >> > [CC: Ted and Kay; we had a discussion about this change few months ago] >> >> Looks good to me. >> >> > /dev/sdb1 originally initialized by cryptsetup: >> > >> > and OLD mkswap: >> > >> > # vol_id /dev/sdb1 | grep TYPE >> > ID_FS_TYPE=swap >> >> Just to check: That's an old version of vol_id, right? The current one > > Yeah, the old version. > >> should refuse to return anything for conflicting signatures. >> >> > # blkid -s TYPE /dev/sdb1 >> > /dev/sdb1: TYPE="crypt_LUKS" >> >> > So, we have two different *valid* signatures on the device now! >> >> Your version of blkid, which we will switch-over to next month, would >> also not return any probing result here, right? > > Right. There are two functions, blkid_do_probe() and > blkid_do_safeprobe() -- safeprobe checks for collisions. Oh, I see. What is the blkid tool using by default? I'm pretty sure, we must have the safe thing as the default. Something that returns a single value without checking for conflicts is nothing anybody should still use today, I think. Otherwise we are back to the weird probe-ordering issues, we had in the past and which are all gone now. Kay -- 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
