[dm-crypt] Cryptsetup bug

Michael Kjörling michael at kjorling.se
Fri Dec 30 00:13:09 CET 2016


On 29 Dec 2016 12:53 -0500, from hd4pix at gmail.com (Hunter DeMeyer):
> I believe that under certain circumstances cryptsetup will encrypt more
> than one partion on a drive. I have attached the process that produced this
> bug.

I suspect that a full strace output from the cryptsetup run would be
useful in attempting to track this down.

However, the first thing that strikes me is that your log does not
show conclusively that the kernel is working with current data,
particularly partition offsets data.

I can see no reason why cryptsetup would be working with any device
other than the one named.

Assuming udev (otherwise adjust as appropriate), a udev device re-scan
between creating the partitions and luksFormat would be a good first
step. Also confirm that doing so refreshes the partition device nodes,
for example by first wiping all partitioning information and showing
conclusively that no partitions are found by the kernel (for example
by `hdparm -z` and showing us the relevant `dmesg` output and /dev
contents), then partitioning and showing conclusively that the newly
created partitions are found and reflected in the device nodes.

-- 
Michael Kjörling • https://michael.kjorling.semichael at kjorling.se
                 “People who think they know everything really annoy
                 those of us who know we don’t.” (Bjarne Stroustrup)


More information about the dm-crypt mailing list