udev gets tmpfs kernel support test wrong

[Debian Bug Importer]

Automatically imported from Debian bug report #286756 http://bugs.debian.org/286756

[Marco d'Itri (md)]

reassign 286756 kernel-source-2.6.9

On Dec 22, "John R. McPherson" <email address hidden> wrote:

> Debian sarge/x86, custom kernel 2.6.9.
>
> I built a kernel and didn't set CONFIG_TMPFS, and after installing the
> kernel and rebooting, udev mounts an unusable tmpfs over /dev, so
> no device files are accessible. The mount doesn't fail, but the
This is a known kernel bug (mount is supposed to fail), but I tought
that it had been fixed in 2.6.8 (2.6.9?).
I'm reassigning this bug to the kernel package.

(I also accept suggestions for a better test for tmpfs presence...)

--
ciao, |
Marco | [9974 spV5nrM59g86E]

[John R. McPherson (jrm+udevbug)]

On Wed, Dec 22, 2004 at 01:57:13AM +0100, Marco d'Itri wrote:
> reassign 286756 kernel-source-2.6.9
>
> On Dec 22, "John R. McPherson" <email address hidden> wrote:
>
> > Debian sarge/x86, custom kernel 2.6.9.
> >
> > I built a kernel and didn't set CONFIG_TMPFS, and after installing the
> > kernel and rebooting, udev mounts an unusable tmpfs over /dev, so
> > no device files are accessible. The mount doesn't fail, but the
> This is a known kernel bug (mount is supposed to fail), but I tought
> that it had been fixed in 2.6.8 (2.6.9?).
> I'm reassigning this bug to the kernel package.
>
> (I also accept suggestions for a better test for tmpfs presence...)

If /proc/config.gz exists, you can 'zgrep TMPFS /proc/config.gz' and
see if it is ="y"/"m" vs "n".

That won't help for kernels that don't have CONFIG_IKCONFIG_PROC set, so
don't have a /proc/config.gz. But if the file exists, you can tell
whether or not the kernel has tmpfs support.

John McPherson

[Debian Bug Importer (debzilla)]

Automatically imported from Debian bug report #286756 http://bugs.debian.org/286756

[Debian Bug Importer (debzilla)]

Message-ID: <email address hidden>
Date: Wed, 22 Dec 2004 13:39:16 +1300
From: "John R. McPherson" <email address hidden>
To: <email address hidden>
Subject: udev gets tmpfs kernel support test wrong

Package: udev
Version: 0.048-2
Severity: grave
Justification: renders computer unusable

Debian sarge/x86, custom kernel 2.6.9.

I built a kernel and didn't set CONFIG_TMPFS, and after installing the
kernel and rebooting, udev mounts an unusable tmpfs over /dev, so
no device files are accessible. The mount doesn't fail, but the
directory is unusable.
If I use 'strace', any accesses to files or dirs in "/dev" return ENODIR.
The init process complains that it can't mount filesystems and only
single-user login prompt is available.
Changing this kernel option and recompiling/installed fixed it.

The problem seems to be that the kernel lists "tmpfs" in
/proc/filesystems even if CONFIG_TMPFS is not set, and /etc/init.d/udev
tests for that to assume tmpfs can be used.

As seen in /usr/src/linux/Documentation/filesystems/tmpfs.txt:
   1) There is always a kernel internal mount which you will not see at
     all. This is used for shared anonymous mappings and SYSV shared
     memory.

     This mount does not depend on CONFIG_TMPFS. If CONFIG_TMPFS is not
     set, the user visible part of tmpfs is not build. But the internal
     mechanisms are always present.

So maybe the kernel will list tmpfs in /proc/filesystems if other
options are set, even if tmpfs ramdisks are not available?

I can provide a kernel config if more information is required.

John McPherson

[Debian Bug Importer (debzilla)]

Message-ID: <email address hidden>
Date: Wed, 22 Dec 2004 01:57:13 +0100
From: <email address hidden> (Marco d'Itri)
To: "John R. McPherson" <email address hidden>,
 <email address hidden>
Cc: <email address hidden>, <email address hidden>
Subject: Re: Bug#286756: udev gets tmpfs kernel support test wrong

--bg08WKrSYDhXBjb5
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
Content-Transfer-Encoding: quoted-printable

reassign 286756 kernel-source-2.6.9

On Dec 22, "John R. McPherson" <email address hidden> wrote:

> Debian sarge/x86, custom kernel 2.6.9.
>=20
> I built a kernel and didn't set CONFIG_TMPFS, and after installing the
> kernel and rebooting, udev mounts an unusable tmpfs over /dev, so
> no device files are accessible. The mount doesn't fail, but the
This is a known kernel bug (mount is supposed to fail), but I tought
that it had been fixed in 2.6.8 (2.6.9?).
I'm reassigning this bug to the kernel package.

(I also accept suggestions for a better test for tmpfs presence...)

--=20
ciao, |
Marco | [9974 spV5nrM59g86E]

--bg08WKrSYDhXBjb5
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)

iD8DBQFByMZpFGfw2OHuP7ERAto9AJ98tYfz7E1PtbEMsd+Hqhat0cvKJwCfduk7
VFREkD/LLXi4vc6hDMV3QAc=
=3/8Y
-----END PGP SIGNATURE-----

--bg08WKrSYDhXBjb5--

[Debian Bug Importer (debzilla)]

Message-ID: <email address hidden>
Date: Wed, 22 Dec 2004 14:14:34 +1300
From: "John R. McPherson" <email address hidden>
To: <email address hidden>, <email address hidden>
Subject: Re: Bug#286756: udev gets tmpfs kernel support test wrong

On Wed, Dec 22, 2004 at 01:57:13AM +0100, Marco d'Itri wrote:
> reassign 286756 kernel-source-2.6.9
>
> On Dec 22, "John R. McPherson" <email address hidden> wrote:
>
> > Debian sarge/x86, custom kernel 2.6.9.
> >
> > I built a kernel and didn't set CONFIG_TMPFS, and after installing the
> > kernel and rebooting, udev mounts an unusable tmpfs over /dev, so
> > no device files are accessible. The mount doesn't fail, but the
> This is a known kernel bug (mount is supposed to fail), but I tought
> that it had been fixed in 2.6.8 (2.6.9?).
> I'm reassigning this bug to the kernel package.
>
> (I also accept suggestions for a better test for tmpfs presence...)

If /proc/config.gz exists, you can 'zgrep TMPFS /proc/config.gz' and
see if it is ="y"/"m" vs "n".

That won't help for kernels that don't have CONFIG_IKCONFIG_PROC set, so
don't have a /proc/config.gz. But if the file exists, you can tell
whether or not the kernel has tmpfs support.

John McPherson

[Andres Salomon (dilinger-deactivatedaccount)]

Unfortunately, the kernel registers tmpfs whether CONFIG_TMPFS is
actually enabled or not; this is because mm/shmfs.c and mm/tiny-shmfs.c,
while offering functionality for shm even if CONFIG_TMPFS is unset,
decides to call itself tmpfs. I'm not sure the logic (if there is any)
behind this; I'm betting someone wrote shmfs, then decided it would be
nice to abstract it out into tmpfs.

I don't have any solutions for this right now; probably the "correct"
thing to do would be to split apart tmpfs and shmfs stuff, have
fs/tmpfs*, and have shm depend upon this (and provide its own shmfs
stubs if tmpfs is disabled?).

--
Andres Salomon <email address hidden>

[Debian Bug Importer (debzilla)]

Message-Id: <1103767770.11785.4.camel@localhost>
Date: Wed, 22 Dec 2004 21:09:30 -0500
From: Andres Salomon <email address hidden>
To: <email address hidden>
Cc: <email address hidden>
Subject: udev gets tmpfs kernel support test wrong

--=-QJeYn54Zln2k6XogrtNw
Content-Type: text/plain
Content-Transfer-Encoding: quoted-printable

Unfortunately, the kernel registers tmpfs whether CONFIG_TMPFS is
actually enabled or not; this is because mm/shmfs.c and mm/tiny-shmfs.c,
while offering functionality for shm even if CONFIG_TMPFS is unset,
decides to call itself tmpfs. I'm not sure the logic (if there is any)
behind this; I'm betting someone wrote shmfs, then decided it would be
nice to abstract it out into tmpfs.

I don't have any solutions for this right now; probably the "correct"
thing to do would be to split apart tmpfs and shmfs stuff, have
fs/tmpfs*, and have shm depend upon this (and provide its own shmfs
stubs if tmpfs is disabled?).

--=20
Andres Salomon <email address hidden>

--=-QJeYn54Zln2k6XogrtNw
Content-Type: application/pgp-signature; name=signature.asc
Content-Description: This is a digitally signed message part

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

iD8DBQBByija78o9R9NraMQRAqJdAJ9Q9Z6O9dGfQ3qsJvqnekWC5XaHmgCfaTYB
+k6U1j+w4+z7eh73EzSRYE8=
=YIFq
-----END PGP SIGNATURE-----

--=-QJeYn54Zln2k6XogrtNw--

[Justin Pryzby (justinpryzby-users)]

Couldn't we just use a dummy mount? Bind mount /bin/ as /tmp/`mktemp
-d`/, then test retval=diropen(/tmp/`mktemp`)? (Or
fopen("/tmp/`mktemp`/ls") or sth similar).

(I didn't say it was clean ... :) But this test is probably better
than nothing, as it prevents /dev/ from being unusable. Just abstract
the test to some bashscript or a few lines of C, stick it in
/usr/lib/udev/ until some point when the kernel tells the truth.

Justin

[Debian Bug Importer (debzilla)]

Message-ID: <20050121185514.GA31559@andromeda>
Date: Fri, 21 Jan 2005 13:55:14 -0500
From: Justin Pryzby <email address hidden>
To: <email address hidden>
Cc: <email address hidden>
Subject: udev tmpfs test

Couldn't we just use a dummy mount? Bind mount /bin/ as /tmp/`mktemp
-d`/, then test retval=diropen(/tmp/`mktemp`)? (Or
fopen("/tmp/`mktemp`/ls") or sth similar).

(I didn't say it was clean ... :) But this test is probably better
than nothing, as it prevents /dev/ from being unusable. Just abstract
the test to some bashscript or a few lines of C, stick it in
/usr/lib/udev/ until some point when the kernel tells the truth.

Justin

[Martin Michlmayr (tbm)]

reassign 279648 kernel-source-2.6.10
reassign 280091 kernel-source-2.6.10
reassign 281209 kernel-source-2.6.10
reassign 281439 kernel-source-2.6.10
reassign 281495 kernel-source-2.6.10
reassign 281517 kernel-source-2.6.10
reassign 281912 kernel-source-2.6.10
reassign 282023 kernel-source-2.6.10
reassign 282059 kernel-source-2.6.10
reassign 282083 kernel-source-2.6.10
reassign 282088 kernel-source-2.6.10
reassign 282119 kernel-source-2.6.10
reassign 282234 kernel-source-2.6.10
reassign 282318 kernel-source-2.6.10
reassign 282741 kernel-source-2.6.10
reassign 282964 kernel-source-2.6.10
reassign 283325 kernel-source-2.6.10
reassign 283852 kernel-source-2.6.10
reassign 284600 kernel-source-2.6.10
reassign 284680 kernel-source-2.6.10
reassign 286756 kernel-source-2.6.10
reassign 287690 kernel-source-2.6.10
reassign 289687 kernel-source-2.6.10
reassign 291107 kernel-source-2.6.10
reassign 295627 kernel-source-2.6.10
reassign 295948 kernel-source-2.6.10

reassign 285696 kernel-patch-debian-2.6.10

--
Martin Michlmayr
http://www.cyrius.com/

[Debian Bug Importer (debzilla)]

Message-ID: <email address hidden>
Date: Tue, 1 Mar 2005 11:40:57 +0800
From: Martin Michlmayr <email address hidden>
To: <email address hidden>
Subject: 2.6.9 -> 2.6.10

reassign 279648 kernel-source-2.6.10
reassign 280091 kernel-source-2.6.10
reassign 281209 kernel-source-2.6.10
reassign 281439 kernel-source-2.6.10
reassign 281495 kernel-source-2.6.10
reassign 281517 kernel-source-2.6.10
reassign 281912 kernel-source-2.6.10
reassign 282023 kernel-source-2.6.10
reassign 282059 kernel-source-2.6.10
reassign 282083 kernel-source-2.6.10
reassign 282088 kernel-source-2.6.10
reassign 282119 kernel-source-2.6.10
reassign 282234 kernel-source-2.6.10
reassign 282318 kernel-source-2.6.10
reassign 282741 kernel-source-2.6.10
reassign 282964 kernel-source-2.6.10
reassign 283325 kernel-source-2.6.10
reassign 283852 kernel-source-2.6.10
reassign 284600 kernel-source-2.6.10
reassign 284680 kernel-source-2.6.10
reassign 286756 kernel-source-2.6.10
reassign 287690 kernel-source-2.6.10
reassign 289687 kernel-source-2.6.10
reassign 291107 kernel-source-2.6.10
reassign 295627 kernel-source-2.6.10
reassign 295948 kernel-source-2.6.10

reassign 285696 kernel-patch-debian-2.6.10

--
Martin Michlmayr
http://www.cyrius.com/

[Martin Michlmayr (tbm)]

reassign 312901 linux-2.6
reassign 304617 linux-2.6
reassign 297832 linux-2.6
reassign 301425 linux-2.6
reassign 317798 linux-2.6
reassign 303491 linux-2.6
reassign 325004 linux-2.6
reassign 291107 linux-2.6
reassign 318044 linux-2.6
reassign 253802 linux-2.6
reassign 295627 linux-2.6
reassign 282964 linux-2.6
reassign 297086 linux-2.6
reassign 307909 linux-2.6
reassign 310244 linux-2.6
reassign 301063 linux-2.6
reassign 324368 linux-2.6
reassign 305557 linux-2.6
reassign 295245 linux-2.6
reassign 308649 linux-2.6
reassign 325006 linux-2.6
reassign 312320 linux-2.6
reassign 289687 linux-2.6
reassign 301514 linux-2.6
reassign 303550 linux-2.6
reassign 294303 linux-2.6
reassign 322164 linux-2.6
reassign 317309 linux-2.6
reassign 313552 linux-2.6
reassign 284680 linux-2.6
reassign 286756 linux-2.6
reassign 305212 linux-2.6
reassign 285696 linux-2.6
reassign 317007 linux-2.6
reassign 311185 linux-2.6
reassign 307459 linux-2.6
reassign 315796 linux-2.6
reassign 314707 linux-2.6
reassign 314637 linux-2.6
reassign 289948 linux-2.6
reassign 293195 linux-2.6
reassign 307549 linux-2.6
reassign 297627 linux-2.6
reassign 298643 linux-2.6
reassign 315865 linux-2.6
reassign 305234 linux-2.6
reassign 309407 linux-2.6
reassign 312677 linux-2.6
reassign 311758 linux-2.6
reassign 317756 linux-2.6
reassign 311815 linux-2.6
reassign 317258 linux-2.6
reassign 287867 linux-2.6
reassign 319982 linux-2.6
reassign 317312 linux-2.6
reassign 312845 linux-2.6
reassign 316591 linux-2.6
reassign 291191 linux-2.6
reassign 292516 linux-2.6
reassign 283852 linux-2.6
reassign 326446 linux-2.6
reassign 316681 linux-2.6
reassign 279648 linux-2.6
reassign 303490 linux-2.6
reassign 328135 linux-2.6
reassign 317581 linux-2.6
reassign 328122 linux-2.6
reassign 312457 linux-2.6
reassign 299014 linux-2.6
reassign 304507 linux-2.6
reassign 320953 linux-2.6
reassign 325053 linux-2.6
reassign 308946 linux-2.6
reassign 281517 linux-2.6
reassign 311412 linux-2.6
reassign 290925 linux-2.6
reassign 325014 linux-2.6
reassign 309551 linux-2.6
reassign 297834 linux-2.6
reassign 296955 linux-2.6
reassign 281209 linux-2.6
reassign 309961 linux-2.6
reassign 296808 linux-2.6
reassign 291029 linux-2.6
reassign 324823 linux-2.6
reassign 311820 linux-2.6
reassign 322878 linux-2.6
reassign 309206 linux-2.6
reassign 291464 linux-2.6
reassign 325011 linux-2.6
reassign 282318 linux-2.6
reassign 304302 linux-2.6
reassign 316560 linux-2.6
reassign 281912 linux-2.6

--
Martin Michlmayr
http://www.cyrius.com/

[Debian Bug Importer (debzilla)]

Message-ID: <email address hidden>
Date: Fri, 11 Nov 2005 20:14:50 +0000
From: Martin Michlmayr <email address hidden>
To: <email address hidden>
Subject: kernel 2.6

reassign 312901 linux-2.6
reassign 304617 linux-2.6
reassign 297832 linux-2.6
reassign 301425 linux-2.6
reassign 317798 linux-2.6
reassign 303491 linux-2.6
reassign 325004 linux-2.6
reassign 291107 linux-2.6
reassign 318044 linux-2.6
reassign 253802 linux-2.6
reassign 295627 linux-2.6
reassign 282964 linux-2.6
reassign 297086 linux-2.6
reassign 307909 linux-2.6
reassign 310244 linux-2.6
reassign 301063 linux-2.6
reassign 324368 linux-2.6
reassign 305557 linux-2.6
reassign 295245 linux-2.6
reassign 308649 linux-2.6
reassign 325006 linux-2.6
reassign 312320 linux-2.6
reassign 289687 linux-2.6
reassign 301514 linux-2.6
reassign 303550 linux-2.6
reassign 294303 linux-2.6
reassign 322164 linux-2.6
reassign 317309 linux-2.6
reassign 313552 linux-2.6
reassign 284680 linux-2.6
reassign 286756 linux-2.6
reassign 305212 linux-2.6
reassign 285696 linux-2.6
reassign 317007 linux-2.6
reassign 311185 linux-2.6
reassign 307459 linux-2.6
reassign 315796 linux-2.6
reassign 314707 linux-2.6
reassign 314637 linux-2.6
reassign 289948 linux-2.6
reassign 293195 linux-2.6
reassign 307549 linux-2.6
reassign 297627 linux-2.6
reassign 298643 linux-2.6
reassign 315865 linux-2.6
reassign 305234 linux-2.6
reassign 309407 linux-2.6
reassign 312677 linux-2.6
reassign 311758 linux-2.6
reassign 317756 linux-2.6
reassign 311815 linux-2.6
reassign 317258 linux-2.6
reassign 287867 linux-2.6
reassign 319982 linux-2.6
reassign 317312 linux-2.6
reassign 312845 linux-2.6
reassign 316591 linux-2.6
reassign 291191 linux-2.6
reassign 292516 linux-2.6
reassign 283852 linux-2.6
reassign 326446 linux-2.6
reassign 316681 linux-2.6
reassign 279648 linux-2.6
reassign 303490 linux-2.6
reassign 328135 linux-2.6
reassign 317581 linux-2.6
reassign 328122 linux-2.6
reassign 312457 linux-2.6
reassign 299014 linux-2.6
reassign 304507 linux-2.6
reassign 320953 linux-2.6
reassign 325053 linux-2.6
reassign 308946 linux-2.6
reassign 281517 linux-2.6
reassign 311412 linux-2.6
reassign 290925 linux-2.6
reassign 325014 linux-2.6
reassign 309551 linux-2.6
reassign 297834 linux-2.6
reassign 296955 linux-2.6
reassign 281209 linux-2.6
reassign 309961 linux-2.6
reassign 296808 linux-2.6
reassign 291029 linux-2.6
reassign 324823 linux-2.6
reassign 311820 linux-2.6
reassign 322878 linux-2.6
reassign 309206 linux-2.6
reassign 291464 linux-2.6
reassign 325011 linux-2.6
reassign 282318 linux-2.6
reassign 304302 linux-2.6
reassign 316560 linux-2.6
reassign 281912 linux-2.6

--
Martin Michlmayr
http://www.cyrius.com/

[Ben Collins (ben-collins)]

This bug has been flagged because it is old and possibly inactive. It may or may
not be fixed in the latest release (Breezy Badger 5.10). It is being marked as
"NEEDSINFO". In two weeks time, if the bug is not updated back to "NEW" and
validated against Breezy, it will be closed.

This is needed in order to help manage the current bug list for the kernel. We
would like to fix all bugs, but need users to test and help with debugging.

If this change was in error for this bug, please respond and make the
appropriate change (or email <email address hidden> if you cannot make the
change).

Thanks for your help.

[Sven Luther (sven-luther)]

severity 286756 important
thanks

Hello,

This is not really an RC bug report, since it affects a very small minority of
users (those who recompile their own kernels and don set TMPFS), and there
seems to be a tentative workaround possible with the info in this bug report,
so i downgrade it to important.

Friendly,

Sven Luther

[Debian Bug Importer (debzilla)]

Message-ID: <email address hidden>
Date: Mon, 28 Nov 2005 02:31:13 +0100
From: Sven Luther <email address hidden>
To: <email address hidden>
Subject: This is not really RC ...

severity 286756 important
thanks

Hello,

This is not really an RC bug report, since it affects a very small minority of
users (those who recompile their own kernels and don set TMPFS), and there
seems to be a tentative workaround possible with the info in this bug report,
so i downgrade it to important.

Friendly,

Sven Luther

[Dennis Kaarsemaker (dennis)]

Not for Ubuntu

[David Schmitt (david-schmitt)]

Hi John!

Could you please retest with a current kernel (testing:2.6.12,
unstable:2.6.14) and a current udev (>= 0.076)?

Thanks for your time and work, David
--
- hallo... wie gehts heute?
- *hust* gut *rotz* *keuch*
- gott sei dank kommunizieren wir über ein septisches medium ;)
 -- Matthias Leeb, Uni f. angewandte Kunst, 2005-02-15

[Debian Bug Importer (debzilla)]

Message-Id: <email address hidden>
Date: Mon, 2 Jan 2006 21:36:23 +0100
From: David Schmitt <email address hidden>
To: "John R. McPherson" <email address hidden>
Cc: <email address hidden>
Subject: Re: udev gets tmpfs kernel support test wrong

Hi John!

Could you please retest with a current kernel (testing:2.6.12,=20
unstable:2.6.14) and a current udev (>=3D 0.076)?

Thanks for your time and work, David
=2D-=20
=2D hallo... wie gehts heute?
=2D *hust* gut *rotz* *keuch*
=2D gott sei dank kommunizieren wir =FCber ein septisches medium ;)
 -- Matthias Leeb, Uni f. angewandte Kunst, 2005-02-15

[John R. McPherson (jrm+udevbug)]

On Mon, Jan 02, 2006 at 09:36:23PM +0100, David Schmitt wrote:
> Hi John!
>
> Could you please retest with a current kernel (testing:2.6.12,
> unstable:2.6.14) and a current udev (>= 0.076)?

Hi,
I thought this was fixed some time ago... the /etc/init.d/udev
script now checks if files can successfully be created in the tmpfs
/dev dir:

  # using ln to test if /dev works, because touch is in /usr/bin/
  if ln -s test /dev/test-file; then
    rm /dev/test-file
    log_end_msg 0
  else
    log_failure_msg "FATAL: udev requires tmpfs support, not started."
    umount /etc/udev
    umount /dev
    log_end_msg 1
    exit 1
  fi

(This is from udev version 0.060-1ubuntu1, I'm not sure which version
in debian fixed it).
The problem only occurred if someone was using a kernel that they compiled
themselves, since the debian provided kernels had the required support.

So it's fixed for udev, at least. I thought it had been re-assigned to
linux-2.6 because someone thought the bug was for listing tmpfs in
/proc/filesystems when tmpfs wasn't available.

John

[Debian Bug Importer (debzilla)]

Message-ID: <email address hidden>
Date: Tue, 3 Jan 2006 16:52:50 +1300
From: "John R. McPherson" <email address hidden>
To: <email address hidden>
Subject: Re: udev gets tmpfs kernel support test wrong

On Mon, Jan 02, 2006 at 09:36:23PM +0100, David Schmitt wrote:
> Hi John!
>
> Could you please retest with a current kernel (testing:2.6.12,
> unstable:2.6.14) and a current udev (>= 0.076)?

Hi,
I thought this was fixed some time ago... the /etc/init.d/udev
script now checks if files can successfully be created in the tmpfs
/dev dir:

  # using ln to test if /dev works, because touch is in /usr/bin/
  if ln -s test /dev/test-file; then
    rm /dev/test-file
    log_end_msg 0
  else
    log_failure_msg "FATAL: udev requires tmpfs support, not started."
    umount /etc/udev
    umount /dev
    log_end_msg 1
    exit 1
  fi

(This is from udev version 0.060-1ubuntu1, I'm not sure which version
in debian fixed it).
The problem only occurred if someone was using a kernel that they compiled
themselves, since the debian provided kernels had the required support.

So it's fixed for udev, at least. I thought it had been re-assigned to
linux-2.6 because someone thought the bug was for listing tmpfs in
/proc/filesystems when tmpfs wasn't available.

John

[maximilian attems (maks-debian)]

Version: linux-image-2.6.15

On Tue, 03 Jan 2006, John R. McPherson wrote:

> On Mon, Jan 02, 2006 at 09:36:23PM +0100, David Schmitt wrote:
> > Hi John!
> >
> > Could you please retest with a current kernel (testing:2.6.12,
> > unstable:2.6.14) and a current udev (>= 0.076)?
>
> Hi,
> I thought this was fixed some time ago... the /etc/init.d/udev
> script now checks if files can successfully be created in the tmpfs
> /dev dir:
>
> # using ln to test if /dev works, because touch is in /usr/bin/
> if ln -s test /dev/test-file; then
> rm /dev/test-file
> log_end_msg 0
> else
> log_failure_msg "FATAL: udev requires tmpfs support, not started."
> umount /etc/udev
> umount /dev
> log_end_msg 1
> exit 1
> fi
>
> (This is from udev version 0.060-1ubuntu1, I'm not sure which version
> in debian fixed it).

ok cool.

> The problem only occurred if someone was using a kernel that they compiled
> themselves, since the debian provided kernels had the required support.

well then he gains some experience about badly chosen .configs ;)

> So it's fixed for udev, at least. I thought it had been re-assigned to
> linux-2.6 because someone thought the bug was for listing tmpfs in
> /proc/filesystems when tmpfs wasn't available.

egrep tmpfs /proc/filesystems
nodev tmpfs

closing therefor

--
maks

[Debian Bug Importer (debzilla)]

Message-ID: <20060111104322.GI7691@nancy>
Date: Wed, 11 Jan 2006 11:43:22 +0100
From: maximilian attems <email address hidden>
To: <email address hidden>
Subject: Re: Bug#286756: udev gets tmpfs kernel support test wrong

Version: linux-image-2.6.15

On Tue, 03 Jan 2006, John R. McPherson wrote:

> On Mon, Jan 02, 2006 at 09:36:23PM +0100, David Schmitt wrote:
> > Hi John!
> >
> > Could you please retest with a current kernel (testing:2.6.12,
> > unstable:2.6.14) and a current udev (>= 0.076)?
>
> Hi,
> I thought this was fixed some time ago... the /etc/init.d/udev
> script now checks if files can successfully be created in the tmpfs
> /dev dir:
>
> # using ln to test if /dev works, because touch is in /usr/bin/
> if ln -s test /dev/test-file; then
> rm /dev/test-file
> log_end_msg 0
> else
> log_failure_msg "FATAL: udev requires tmpfs support, not started."
> umount /etc/udev
> umount /dev
> log_end_msg 1
> exit 1
> fi
>
> (This is from udev version 0.060-1ubuntu1, I'm not sure which version
> in debian fixed it).

ok cool.

> The problem only occurred if someone was using a kernel that they compiled
> themselves, since the debian provided kernels had the required support.

well then he gains some experience about badly chosen .configs ;)

> So it's fixed for udev, at least. I thought it had been re-assigned to
> linux-2.6 because someone thought the bug was for listing tmpfs in
> /proc/filesystems when tmpfs wasn't available.

egrep tmpfs /proc/filesystems
nodev tmpfs

closing therefor

--
maks