mount - losetup uses not available /proc/crypto/cipher

[Debian Bug Importer]

Automatically imported from Debian bug report #285353 http://bugs.debian.org/285353

[Debian Bug Importer (debzilla)]

Automatically imported from Debian bug report #285353 http://bugs.debian.org/285353

[Debian Bug Importer (debzilla)]

Message-ID: <email address hidden>
Date: Sun, 12 Dec 2004 19:46:19 +0100
From: Bastian Blank <email address hidden>
To: <email address hidden>
Subject: mount - losetup uses not available /proc/crypto/cipher

--HcAYCG3uE/tztfnV
Content-Type: text/plain; charset=utf-8
Content-Disposition: inline
Content-Transfer-Encoding: quoted-printable

Package: mount
Version: 2.13j-3
Severity: grave

This version of losetup begins to use /proc/crypto/cipher which does not
exist on any official (upstream and debian) kernel release. This breaks
any cryptoloop partition.

Bastian

--=20
A princess should not be afraid -- not with a brave knight to protect her.
  -- McCoy, "Shore Leave", stardate 3025.3

--HcAYCG3uE/tztfnV
Content-Type: application/pgp-signature; name="signature.asc"
Content-Description: Digital signature
Content-Disposition: inline

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.5 (GNU/Linux)

iEYEARECAAYFAkG8kfsACgkQnw66O/MvCNERagCeLfUmSg05eYqv3IlHCDDASt1Y
HXMAoIZYKdrk8nJwKEj95cZ5V40qhjfZ
=tebo
-----END PGP SIGNATURE-----

--HcAYCG3uE/tztfnV--

[LaMont Jones (lamont)]

severity 285353 serious
--

On Sun, Dec 12, 2004 at 07:46:19PM +0100, Bastian Blank wrote:
> This version of losetup begins to use /proc/crypto/cipher which does not
> exist on any official (upstream and debian) kernel release. This breaks
> any cryptoloop partition.

Actually, that code dates back to the 2.11n-2 release, to detect
kerneli implementations.

Before any progress can be made on this bug, I'll need some actual
details of what kind of kernel, what cryptoloop system your using,
what losetup/mount command you're using, etc.

FWIW, my cryptoloop device works just fine with the 2.6 cryptoloop
backport in debian 2.4 kernels, as well as 2.6 kernels.

lamont

[Debian Bug Importer (debzilla)]

Message-ID: <email address hidden>
Date: Sun, 12 Dec 2004 13:47:30 -0700
From: LaMont Jones <email address hidden>
To: Bastian Blank <email address hidden>, <email address hidden>
Cc: <email address hidden>
Subject: Re: Bug#285353: mount - losetup uses not available /proc/crypto/cipher

severity 285353 serious
--

On Sun, Dec 12, 2004 at 07:46:19PM +0100, Bastian Blank wrote:
> This version of losetup begins to use /proc/crypto/cipher which does not
> exist on any official (upstream and debian) kernel release. This breaks
> any cryptoloop partition.

Actually, that code dates back to the 2.11n-2 release, to detect
kerneli implementations.

Before any progress can be made on this bug, I'll need some actual
details of what kind of kernel, what cryptoloop system your using,
what losetup/mount command you're using, etc.

FWIW, my cryptoloop device works just fine with the 2.6 cryptoloop
backport in debian 2.4 kernels, as well as 2.6 kernels.

lamont

[Bastian Blank (waldi)]

On Sun, Dec 12, 2004 at 01:47:30PM -0700, LaMont Jones wrote:
> Before any progress can be made on this bug, I'll need some actual
> details of what kind of kernel, what cryptoloop system your using,
> what losetup/mount command you're using, etc.

Kernel: 2.6.9, vanilla, included cryptoloop on ppc.
losetup call: losetup -p 0 -e aes -k 256 /dev/loop0 file

It works after I hacked that piece of code to see /proc/crypto or loose
the -e aes.

Bastian

--
There are always alternatives.
  -- Spock, "The Galileo Seven", stardate 2822.3

[Debian Bug Importer (debzilla)]

Message-ID: <email address hidden>
Date: Mon, 13 Dec 2004 12:17:03 +0100
From: Bastian Blank <email address hidden>
To: LaMont Jones <email address hidden>
Cc: <email address hidden>
Subject: Re: Bug#285353: mount - losetup uses not available /proc/crypto/cipher

--fUYQa+Pmc3FrFX/N
Content-Type: text/plain; charset=utf-8
Content-Disposition: inline
Content-Transfer-Encoding: quoted-printable

On Sun, Dec 12, 2004 at 01:47:30PM -0700, LaMont Jones wrote:
> Before any progress can be made on this bug, I'll need some actual
> details of what kind of kernel, what cryptoloop system your using,
> what losetup/mount command you're using, etc.

Kernel: 2.6.9, vanilla, included cryptoloop on ppc.
losetup call: losetup -p 0 -e aes -k 256 /dev/loop0 file

It works after I hacked that piece of code to see /proc/crypto or loose
the -e aes.

Bastian

--=20
There are always alternatives.
  -- Spock, "The Galileo Seven", stardate 2822.3

--fUYQa+Pmc3FrFX/N
Content-Type: application/pgp-signature; name="signature.asc"
Content-Description: Digital signature
Content-Disposition: inline

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.5 (GNU/Linux)

iEYEARECAAYFAkG9ei8ACgkQnw66O/MvCNHZcwCcD85m1H+Qpz6EHPYUyNMi7kqi
MgIAnjzyIDf9XSxnIdpcEOCb/2EUA0xc
=rPPo
-----END PGP SIGNATURE-----

--fUYQa+Pmc3FrFX/N--

[LaMont Jones (lamont)]

On Mon, Dec 13, 2004 at 12:17:03PM +0100, Bastian Blank wrote:
> On Sun, Dec 12, 2004 at 01:47:30PM -0700, LaMont Jones wrote:
> Kernel: 2.6.9, vanilla, included cryptoloop on ppc.
> losetup call: losetup -p 0 -e aes -k 256 /dev/loop0 file
>
> It works after I hacked that piece of code to see /proc/crypto or loose
> the -e aes.

Which version of mount? There was a bug in 2.12j-1 that was fixed in
2.12j-2, specifically related to crypto devices.

I'm using a 2.6.9 kernel, 2.12j-3 mount, and a command line of:
    losetup -e aes /dev/loop0 file
without any issues. /dev/crypto access is accomplished by passing the
name to the kernel.

lamont

[Debian Bug Importer (debzilla)]

Message-ID: <email address hidden>
Date: Mon, 13 Dec 2004 07:15:42 -0700
From: LaMont Jones <email address hidden>
To: Bastian Blank <email address hidden>
Cc: LaMont Jones <email address hidden>, <email address hidden>
Subject: Re: Bug#285353: mount - losetup uses not available /proc/crypto/cipher

On Mon, Dec 13, 2004 at 12:17:03PM +0100, Bastian Blank wrote:
> On Sun, Dec 12, 2004 at 01:47:30PM -0700, LaMont Jones wrote:
> Kernel: 2.6.9, vanilla, included cryptoloop on ppc.
> losetup call: losetup -p 0 -e aes -k 256 /dev/loop0 file
>
> It works after I hacked that piece of code to see /proc/crypto or loose
> the -e aes.

Which version of mount? There was a bug in 2.12j-1 that was fixed in
2.12j-2, specifically related to crypto devices.

I'm using a 2.6.9 kernel, 2.12j-3 mount, and a command line of:
    losetup -e aes /dev/loop0 file
without any issues. /dev/crypto access is accomplished by passing the
name to the kernel.

lamont

[Bastian Kleineidam (calvin)]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Hi,

I suggest that the code to detect kerneli uses stat(2) instead of
open(2). Currently it uses:
int kerneli = open("/proc/crypto/cipher");
if (kerneli >= 0) {
    close(kerneli);
    kerneli=1;
}
else kerneli=0;

It should use:
struct stat stbuf;
int kerneli = (stat("/proc/crypto/cipher", &stbuf) == 0);

Regards,
  Bastian
- --
  ,''`. Bastian Kleineidam
 : :' : GnuPG Schlüssel
 `. `' gpg --keyserver wwwkeys.pgp.net --recv-keys 32EC6F3E
   `-

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.5 (GNU/Linux)

iD8DBQFBwGsJeBwlBDLsbz4RAlQfAJ9TYFVIc7Z86MM9c4jj6Hp7q4FPxACfeITM
kZbgaEOxhRHCqfynms1mBFg=
=i7S6
-----END PGP SIGNATURE-----

[Debian Bug Importer (debzilla)]

Message-ID: <20041215164913.GA10491@treasure>
Date: Wed, 15 Dec 2004 17:49:13 +0100
From: Bastian Kleineidam <email address hidden>
To: <email address hidden>, <email address hidden>
Subject: Re: mount - losetup uses not available /proc/crypto/cipher

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Hi,

I suggest that the code to detect kerneli uses stat(2) instead of
open(2). Currently it uses:
int kerneli = open("/proc/crypto/cipher");
if (kerneli >= 0) {
    close(kerneli);
    kerneli=1;
}
else kerneli=0;

It should use:
struct stat stbuf;
int kerneli = (stat("/proc/crypto/cipher", &stbuf) == 0);

Regards,
  Bastian
- --
  ,''`. Bastian Kleineidam
 : :' : GnuPG Schlüssel
 `. `' gpg --keyserver wwwkeys.pgp.net --recv-keys 32EC6F3E
   `-

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.5 (GNU/Linux)

iD8DBQFBwGsJeBwlBDLsbz4RAlQfAJ9TYFVIc7Z86MM9c4jj6Hp7q4FPxACfeITM
kZbgaEOxhRHCqfynms1mBFg=
=i7S6
-----END PGP SIGNATURE-----

[Thomas Hood (jdthood-aglu)]

forwarded 285353 Andries Brouwer <email address hidden>
thanks

Is there any bug here?

(Note: The user reported using "2.13j-3" but very probably meant
2.12j-3.)

--- Forwarded bug report ---

Date: Sun, 12 Dec 2004 19:46:19 +0100
From: Bastian Blank <email address hidden>
Subject: mount - losetup uses not available /proc/crypto/cipher

Package: mount
Version: 2.13j-3
Severity: grave

This version of losetup begins to use /proc/crypto/cipher which does not
exist on any official (upstream and debian) kernel release. This breaks
any cryptoloop partition.

[later ...]

Kernel: 2.6.9, vanilla, included cryptoloop on ppc.
losetup call: losetup -p 0 -e aes -k 256 /dev/loop0 file

It works after I hacked that piece of code to see /proc/crypto or loose
the -e aes.

--- Lamont Jones's reply ---

I'm using a 2.6.9 kernel, 2.12j-3 mount, and a command line of:
    losetup -e aes /dev/loop0 file
without any issues. /dev/crypto access is accomplished by passing the
name to the kernel.

[Debian Bug Importer (debzilla)]

Message-Id: <1103456623.22112.838.camel@thanatos>
Date: Sun, 19 Dec 2004 12:43:43 +0100
From: Thomas Hood <email address hidden>
To: Andries Brouwer <email address hidden>
Subject: Forwarded Debian bug report #285353: "mount - losetup uses not
 available /proc/crypto/cipher"

forwarded 285353 Andries Brouwer <email address hidden>
thanks

Is there any bug here?

(Note: The user reported using "2.13j-3" but very probably meant
2.12j-3.)

--- Forwarded bug report ---

Date: Sun, 12 Dec 2004 19:46:19 +0100
From: Bastian Blank <email address hidden>
Subject: mount - losetup uses not available /proc/crypto/cipher

Package: mount
Version: 2.13j-3
Severity: grave

This version of losetup begins to use /proc/crypto/cipher which does not
exist on any official (upstream and debian) kernel release. This breaks
any cryptoloop partition.

[later ...]

Kernel: 2.6.9, vanilla, included cryptoloop on ppc.
losetup call: losetup -p 0 -e aes -k 256 /dev/loop0 file

It works after I hacked that piece of code to see /proc/crypto or loose
the -e aes.

--- Lamont Jones's reply ---

I'm using a 2.6.9 kernel, 2.12j-3 mount, and a command line of:
    losetup -e aes /dev/loop0 file
without any issues. /dev/crypto access is accomplished by passing the
name to the kernel.

[Thomas Hood (jdthood-aglu)]

notforwarded 285353
thanks

On Sun, 2004-12-19 at 13:47, Andries Brouwer wrote:
> On Sun, Dec 19, 2004 at 12:43:43PM +0100, Thomas Hood wrote:
>
> > forwarded 285353 Andries Brouwer <email address hidden>
> > thanks
> >
> > Is there any bug here?
> >
> > (Note: The user reported using "2.13j-3" but very probably meant
> > 2.12j-3.)
> >
> > --- Forwarded bug report ---
> >
> > Date: Sun, 12 Dec 2004 19:46:19 +0100
> > From: Bastian Blank <email address hidden>
> > Subject: mount - losetup uses not available /proc/crypto/cipher
> >
> > Package: mount
> > Version: 2.13j-3
> > Severity: grave
> >
> > This version of losetup begins to use /proc/crypto/cipher which does not
> > exist on any official (upstream and debian) kernel release. This breaks
> > any cryptoloop partition.
>
> The util-linux source does not contain the string crypto or cipher.
> Thus, it sounds like this is a debian change.
>
> Andries

[Debian Bug Importer (debzilla)]

Message-Id: <1103460261.2065.53.camel@thanatos>
Date: Sun, 19 Dec 2004 13:44:22 +0100
From: Thomas Hood <email address hidden>
To: <email address hidden>
Subject: Re: Forwarded Debian bug report #285353: "mount - losetup uses not
 available /proc/crypto/cipher"

notforwarded 285353
thanks

On Sun, 2004-12-19 at 13:47, Andries Brouwer wrote:
> On Sun, Dec 19, 2004 at 12:43:43PM +0100, Thomas Hood wrote:
>
> > forwarded 285353 Andries Brouwer <email address hidden>
> > thanks
> >
> > Is there any bug here?
> >
> > (Note: The user reported using "2.13j-3" but very probably meant
> > 2.12j-3.)
> >
> > --- Forwarded bug report ---
> >
> > Date: Sun, 12 Dec 2004 19:46:19 +0100
> > From: Bastian Blank <email address hidden>
> > Subject: mount - losetup uses not available /proc/crypto/cipher
> >
> > Package: mount
> > Version: 2.13j-3
> > Severity: grave
> >
> > This version of losetup begins to use /proc/crypto/cipher which does not
> > exist on any official (upstream and debian) kernel release. This breaks
> > any cryptoloop partition.
>
> The util-linux source does not contain the string crypto or cipher.
> Thus, it sounds like this is a debian change.
>
> Andries

[Thomas Hood (jdthood-aglu)]

This bug is in the Debian diff.

[Debian Bug Importer (debzilla)]

Message-Id: <1103473934.28796.46.camel@thanatos>
Date: Sun, 19 Dec 2004 17:32:14 +0100
From: Thomas Hood <email address hidden>
To: <email address hidden>
Subject: in the Debian part

This bug is in the Debian diff.

[Thomas Hood (jdthood-aglu)]

FYI:
> util-linux (2.12k-2) unstable; urgency=lo
> [...]
> * Use stat instead of open in losetup. (From #285353)

What's your opinion about this bug?

--
Thomas Hood

[Debian Bug Importer (debzilla)]

Message-Id: <1103619732.28796.489.camel@thanatos>
Date: Tue, 21 Dec 2004 10:02:12 +0100
From: Thomas Hood <email address hidden>
To: Bastian Kleineidam <email address hidden>
Cc: <email address hidden>
Subject: Re: #285353: Use stat instead of open in losetup

FYI:
> util-linux (2.12k-2) unstable; urgency=lo
> [...]
> * Use stat instead of open in losetup. (From #285353)

What's your opinion about this bug?

--
Thomas Hood

[Bastian Kleineidam (calvin)]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Hi Thomas and Bastian,

On Tue, Dec 21, 2004 at 10:02:12AM +0100, Thomas Hood wrote:
> FYI:
> > util-linux (2.12k-2) unstable; urgency=lo
> > [...]
> > * Use stat instead of open in losetup. (From #285353)
>
>
> What's your opinion about this bug?

I think it can be closed. I did not have however any problems,
so it is up to Bastian Blank (the other Bastian ;) to say it is
ok to close.

Regards,
  Bastian

- --
  ,''`. Bastian Kleineidam
 : :' : GnuPG Schlüssel
 `. `' gpg --keyserver wwwkeys.pgp.net --recv-keys 32EC6F3E
   `-

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.5 (GNU/Linux)

iD8DBQFByAUIeBwlBDLsbz4RAjiGAJ0cFN9MIly0jp7tfgFUJ7mblW1n3QCfSNql
wMQY9IID9DTeBD+DZ0j9Fa4=
=4Z7m
-----END PGP SIGNATURE-----

[Debian Bug Importer (debzilla)]

Message-ID: <20041221111209.GA2137@treasure>
Date: Tue, 21 Dec 2004 12:12:09 +0100
From: Bastian Kleineidam <email address hidden>
To: Thomas Hood <email address hidden>, <email address hidden>
Cc: <email address hidden>
Subject: Re: #285353: Use stat instead of open in losetup

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Hi Thomas and Bastian,

On Tue, Dec 21, 2004 at 10:02:12AM +0100, Thomas Hood wrote:
> FYI:
> > util-linux (2.12k-2) unstable; urgency=lo
> > [...]
> > * Use stat instead of open in losetup. (From #285353)
>
>
> What's your opinion about this bug?

I think it can be closed. I did not have however any problems,
so it is up to Bastian Blank (the other Bastian ;) to say it is
ok to close.

Regards,
  Bastian

- --
  ,''`. Bastian Kleineidam
 : :' : GnuPG Schlüssel
 `. `' gpg --keyserver wwwkeys.pgp.net --recv-keys 32EC6F3E
   `-

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.5 (GNU/Linux)

iD8DBQFByAUIeBwlBDLsbz4RAjiGAJ0cFN9MIly0jp7tfgFUJ7mblW1n3QCfSNql
wMQY9IID9DTeBD+DZ0j9Fa4=
=4Z7m
-----END PGP SIGNATURE-----

[Bastian Blank (waldi)]

On Tue, Dec 21, 2004 at 12:12:09PM +0100, Bastian Kleineidam wrote:
> I think it can be closed. I did not have however any problems,
> so it is up to Bastian Blank (the other Bastian ;) to say it is
> ok to close.

It seems to work, so I'm closing the report.

Bastian

--
If I can have honesty, it's easier to overlook mistakes.
  -- Kirk, "Space Seed", stardate 3141.9

[Debian Bug Importer (debzilla)]

Message-ID: <email address hidden>
Date: Tue, 21 Dec 2004 15:07:02 +0100
From: Bastian Blank <email address hidden>
To: Bastian Kleineidam <email address hidden>
Cc: Thomas Hood <email address hidden>, <email address hidden>
Subject: Re: #285353: Use stat instead of open in losetup

On Tue, Dec 21, 2004 at 12:12:09PM +0100, Bastian Kleineidam wrote:
> I think it can be closed. I did not have however any problems,
> so it is up to Bastian Blank (the other Bastian ;) to say it is
> ok to close.

It seems to work, so I'm closing the report.

Bastian

--=20
If I can have honesty, it's easier to overlook mistakes.
  -- Kirk, "Space Seed", stardate 3141.9

[LaMont Jones (lamont)]

Fixed in 2.12m or so.