LinuxQuestions.org

LinuxQuestions.org (/questions/)
-   Slackware (http://www.linuxquestions.org/questions/slackware-14/)
-   -   Slackware64-current: Creation of LUKS encrypted partition segfaults (http://www.linuxquestions.org/questions/slackware-14/slackware64-current-creation-of-luks-encrypted-partition-segfaults-728159/)

titopoquito 05-24-2009 08:31 AM

Slackware64-current: Creation of LUKS encrypted partition segfaults
 
Hi,

has anyone successfully created a LUKS partition with cryptsetup on Slackware64-current?

Doing this when booting from the CD*, I just get a segfault. Before that, cryptsetup asks the same questions as usual, asks for uppercase YES to confirm the creation, but then stops. Opening an existing LUKS encrypted partitions works though.


* I created a miniiso disc with Eric's script from the Slackware64-current tree. I don't know if this can be the source for my cryptsetup problem.

Alien Bob 05-25-2009 05:52 PM

THis is a confirmed bug. It only happens with the cryptsetup's static binary (which is in the installer) and running it on x86_64 and trying to apply the luksFormat command.
Other distros suffered from the same, and all seem to have reverted to using the dynamic version of cryptsetup.
While that runs the risk of killing cryptsetup when we upgrade something like gpg libraries, the current situation sucks more because you can not create a LUKS volume in the slackware64 installer.

I guess we'll swap the static binary with the dynamically linked equivalent soon.

Eric

titopoquito 05-25-2009 06:00 PM

Thanks Eric for confirming this. I already wondered if I had created one more PEBCAC problem. Nice to hear it wasn't so ;)

gargamel 06-03-2009 01:12 AM

Just wanted to add, that it occurs not only in the installer, but also when I try to crypt a USB device from a running slackware64-current system.

gargamel

GazL 06-03-2009 04:12 AM

Any idea where the underlying problem is? Is it cryptsetup itself, one of the libraries it's linked against or a compiler/linker bug? It just seems a little strange that the bug is specific to the staticly compiled version.

arny 06-05-2009 01:20 PM

Quote:

Originally Posted by Alien Bob (Post 3552456)
THis is a confirmed bug. It only happens with the cryptsetup's static binary (which is in the installer) and running it on x86_64 and trying to apply the luksFormat command.
Eric

Are you sure about this Eric? :)

http://arny.ro/stuff/cryptsetup/

LE: That is a BW64-current with the latest updates (Mon Jun 1 17:46:46 EEST 2009)

Alien Bob 06-05-2009 01:40 PM

Don't be stupid.
This is a topic about what happens in slackware64-current as opposed to slackware-current.
If you have something constructive to say about the slackware64 issue, fine, go ahead. Otherwise, your comments are irrelevant to this discussion.

Eric

kenjiro 06-05-2009 01:52 PM

Quote:

Other distros suffered from the same, and all seem to have reverted to using the dynamic version of cryptsetup.
Very polite, Eric.

I believe arny was trying to tell that not all distros using static cryptsetup suffered from that bug. I used your howto (Luks+LVM) on a BW64-12.2 box and had no problems (fresh new install)

But then, I take the opportunity to ask this: So slackware64-current has more differences from slackware-current than just the x86_64 architecture? I admit my ignorance here, because I thought slack64-current was slack-current, just recompiled to 64bits.

By the way, perhaps instead of calling people stupid, you guys could ask those who have things working about how they fixed the problems. Perhaps it could save time and sweat ;)

Kenjiro

Alien Bob 06-05-2009 02:00 PM

Another post irrelevant to this topic. You should open a new thread if you want to have a discussion about slackware64 and the underlying philosophy for developing it the way it is now.

Eric

gargamel 07-25-2009 06:37 AM

With the latest updates this seems not yet to be fixed. After creating a Linux partition on my external USB hard disk I do:

Code:

# cryptsetup -s 256 -y luksFormat /dev/sdd1

WARNING!
========
This will overwrite data on /dev/sdd1 irrevocably.

Are you sure? (Type uppercase yes): YES
Segmentation fault

I just post this, because this functionality is very important for me, and I hope it's not dropped as a feature for Slackware64 13.0. I do not mean to urge anyone, just want to express that I hope it's gonna be working sooner or later.

gargamel

sahko 07-25-2009 02:07 PM

Quote:

Originally Posted by gargamel (Post 3619791)
I just post this, because this functionality is very important for me, and I hope it's not dropped as a feature for Slackware64 13.0. I do not mean to urge anyone, just want to express that I hope it's gonna be working sooner or later.

Couldn't agree more.

Ilgar 07-25-2009 03:31 PM

Hmm I did my slackware64-current installation based on a June 23rd snapshot and I created a LUKS partition without any problems. This is on an AMD Mobile Sempron 3500+ processor btw.

Edit: I used the twofish encryption algorithm. Maybe the segfault happens with the default aes encryption only.

gargamel 07-25-2009 04:25 PM

Hmm, I tried it several hours ago with all the latest patches applied on an external USB hard disk, and it failed. On Slackware 12.2 (32-bit) it worked just fine, as usual.

gargamel

chuckie 07-25-2009 08:49 PM

this worked for me

cryptsetup.dynamic -s 256 -y luksFormat /dev/sd_ _

Ilgar 07-26-2009 03:01 AM

Quote:

Originally Posted by gargamel (Post 3620205)
Hmm, I tried it several hours ago with all the latest patches applied on an external USB hard disk, and it failed. On Slackware 12.2 (32-bit) it worked just fine, as usual.

gargamel

Did you try twofish, or was it aes?


All times are GMT -5. The time now is 07:51 AM.