linux-kernel - Re: dm-crypt mount returns EINVAL

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Date:	Tue, 28 Aug 2007 10:45:53 -0400
From:	Paul Albrecht <albrecht@...1.com>
To:	Satyam Sharma <satyam@...radead.org>
Cc:	LKML <linux-kernel@...r.kernel.org>
Subject: Re: dm-crypt mount returns EINVAL

Hi Satyam,

On Tue, 2007-08-28 at 20:45 +0530, Satyam Sharma wrote: 
> Hi Paul,
> 
> 
> On Tue, 28 Aug 2007, Paul Albrecht wrote:
> 
> > Hi,
> > 
> > I have recently upgraded my kernel to 2.6.21 from 2.6.20 and that has
> > caused a problem mounting an encrypted partition using dm-crypt.
> > 
> > In my init script, I use cryptsetup to create device which I use to
> > mount an encrypted partition. Since I changed to 2.6.21 the mount fails
> > returning an EINVAL status which doesn't make much sense because the
> > init script is unchanged and works if reverted to 2.6.20.
> > 
> > Any ideas? How can I debug this problem? I'm not subscribed to lkml so
> > please cc in your response.
> 
> What was the error you got from cryptsetup()? Also, please post the
> kernel's dmesg log and the .config of the 2.6.21 kernel (new one).
> 

There's no problem with cryptsetup; it sets up the device for decryption
and returns good status.

Can't run dmesg because the problem occurs in my init script so I don't
have a usable system after the error occurs.

I can get some usable debug output via the serial console. Attached is a
copy of output from serial console for 2.6.20 and 2.6.21. I have added
straces for cryptsetup and mount commands to my init script for debug.

> Also, compare the .config's of the two kernels and ensure that
> CONFIG_DM_CRYPT (or possibly one of the _CRYPTO_ config options)
> didn't get lost or automatically unset during the upgrade/oldconfig.
> 

Compared both config's and don't see any difference that would account
for the mount failure. 

Paul Albrecht

View attachment "minicom.cap.lkml" of type "text/plain" (85219 bytes)

View attachment "oldconfig" of type "text/plain" (46642 bytes)

View attachment "newconfig" of type "text/plain" (46706 bytes)