[dm-crypt] problem cryptdisks_start

[dm-crypt] problem cryptdisks_start

[Benno Rieger]

Most time (but not everytime) when i call cryptdisks_start it reports:
 
#cryptdisks_start extern0

Starting crypto disk...extern0 (starting)...
semid 2293761: semop failed for cookie 0xd4d580e: incorrect semaphore
state
Failed to set a proper state for notification semaphore identified by
cookie value 223172622 (0xd4d580e) to initialize waiting for incoming
notifications.
The node /dev/mapper/extern0_unformatted should have been renamed
to /dev/mapper/extern0 by udev but new node is not present. Falling back
to direct node rename.
extern0 (started)...done.

I made checks every startup of the disk but there are no errors on it.
Is the "semop failed" a real problem or only a information?

System: debian-sid
Linux 2.6.35-trunk-amd64

hard dirve, on top lvm, on top dmcrypt/luks(using aes-xts-plain 64) with
ext4 file system

Version:
cryptsetup 1.1.3


Best Regards, 
Benno 

Re: [dm-crypt] problem cryptdisks_start

[Milan Broz]

On 12/23/2010 10:39 PM, Benno Rieger wrote:
> Most time (but not everytime) when i call cryptdisks_start it reports:
>  
> #cryptdisks_start extern0
> 
> Starting crypto disk...extern0 (starting)...
> semid 2293761: semop failed for cookie 0xd4d580e: incorrect semaphore
> state

This is problem in device-mapper library (which cryptsetup uses).

Which version you have installed? What architecture it is?
(I remember similar bug on ppc here which as fixed upstream)

Milan

Re: [dm-crypt] problem cryptdisks_start

[Milan Broz]

On 12/24/2010 10:28 AM, Milan Broz wrote:
> On 12/23/2010 10:39 PM, Benno Rieger wrote:
>> Most time (but not everytime) when i call cryptdisks_start it reports:
>>  
>> #cryptdisks_start extern0
>>
>> Starting crypto disk...extern0 (starting)...
>> semid 2293761: semop failed for cookie 0xd4d580e: incorrect semaphore
>> state
> 
> This is problem in device-mapper library (which cryptsetup uses).
> 
> Which version you have installed? What architecture it is?
> (I remember similar bug on ppc here which as fixed upstream)

FYI - check that this patch (included in device-mapper 1.02.49)

Fix semctl parameter (union) to avoid misaligned parameter on some arches.
http://sourceware.org/cgi-bin/cvsweb.cgi/LVM2/libdm/libdm-common.c.diff?cvsroot=lvm2&r1=1.69&r2=1.70

(and possible others) is included (I see on Debian unstable still use 1.02.48).

If not, please report bug to Debian. It is not easy to convince Debian lvm/dm
maintainer to do an update there, though:-)

Milan