Description of problem: I have used dm-crypt with LUKS on an entire hard disc connected via an external USB drive bay. blkid does not attempt to identify the hard disc. This worked fine in F7. Version-Release number of selected component (if applicable): 1.40.2-10.fc8 How reproducible: 100% Steps to Reproduce: 1. # cat /proc/partitions major minor #blocks name 8 0 78150744 sda 8 1 104391 sda1 8 2 78043770 sda2 8 16 732574584 sdb 253 0 10485760 dm-0 253 1 2097152 dm-1 (sdb is a 750 GB hard disc used for backups with dm-crypt (LUKS). 2. # cryptsetup luksUUID /dev/sdb a3f9a41b-e3ab-4620-9d29-4e6e7f44b0a1 3. # blkid -c /dev/null -t UUID=a3f9a41b-e3ab-4620-9d29-4e6e7f44b0a1 # blkid | grep sdb # blkid | grep -i a3f9a41b-e3ab-4620-9d29-4e6e7f44b0a1 Actual results: No joy :( strace indicates that blkid does not attempt to even open /dev/sdb Expected results: For it to find the device. Additional info: Attached is the strace output of: blkid -c /dev/null -t UUID=a3f9a41b-e3ab-4620-9d29-4e6e7f44b0a1
Created attachment 269391 [details] strace of blkid
I know about this bug. I found the bug when I read libblkid code few weeks ago. Sorry, forgot to fill BZ. The problem is in devname.c, probe_all(); where is heuristic that doesn't work correctly with unpartitioned disks when the disk is followed by another device<n>. For example: sda loop0
Created attachment 269931 [details] proposed change I think this should address the issue... hopefully without breaking anything else. Karel, care to give it a look-over, since you've recently read this code as well? With that much review I'll send it up to Ted. :) Thanks, -Eric
Seems good. Go ahead ;-)
Sent upstream: http://marc.info/?l=linux-ext4&m=119636478426226&w=2
patch is upstream now, I'll get it to e2fsprogs for fedora soon http://git.kernel.org/?p=fs/ext2/e2fsprogs.git;a=commitdiff;h=38361bbbb55735cf0f4ed6d56079c70fb3328036
e2fsprogs-1.40.4-1.fc8 has been pushed to the Fedora 8 stable repository. If problems still persist, please make note of it in this bug report.