No more than 255 device files with Debian Sarge

Hello Kevin,

I did diff my Debian kernel configuration with the one from
Ubuntu. Lots of differences but nothing really in the filesystem
area 8-(

I found two differences worth checking: Ubuntu has no DEVFS
support compiled/activated. And there was a raw files parameter
set to 255. Neither throwing out DEVFS nor increasing the raw
file number to 8192 did help...

Problem is that the driver concept of one of our projects
depends on the ability to generate more than 255 minor
numbers. Sigh...

Thanks anyhow for your reply!

Greetings,
Stefan

--
Stefan Sperling
Network Administration / Software Development

====================================================
Gesellschaft für Netzwerk- und Automatisierungstechnologie mbH (N.A.T.)
Kamillenweg 22
53757 Sankt Augustin
Germany

Fon: ++49 - 2241 - 3989 - 0
Fax: ++49 - 2241 - 3989 - 10
http://www.nateurope.com

N.A.T. is an ISO 9001 certified company.

--

0

Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.

No more than 255 device files with Debian Sarge

On Fri, Nov 24, 2006 at 06:36:06PM +0100, Stefan Sperling wrote:
> Hello Kevin,
>
> I did diff my Debian kernel configuration with the one from
> Ubuntu. Lots of differences but nothing really in the filesystem
> area 8-(
>
> I found two differences worth checking: Ubuntu has no DEVFS
> support compiled/activated. And there was a raw files parameter
> set to 255. Neither throwing out DEVFS nor increasing the raw
> file number to 8192 did help...
>
> Problem is that the driver concept of one of our projects
> depends on the ability to generate more than 255 minor
> numbers. Sigh...
>
> Thanks anyhow for your reply!
>
> Greetings,
Hi Stefan,
I went to #debian-kernel on irc.debian.org and this is not an issue for
an testing or unstable system like mine. 2.6.8 was the debian kernel and
2.6.16 is not an official one (maybe backports.org or your own). In
either case, I have not tested a chroot with a sarge kenel, but this
issue is related to your system and not the current ones. It is either
specific to the sarge kernel or some other issue of your system. I may
try to try a chroot with a sarge kerel to see if I can reproduce it.
HTH,
Kev
--
| .''`. == Debian GNU/Linux == | my web site: |
| : :' : The Universal | debian.home.pipeline.com |
| `. `' Operating System | go to counter.li.org and |
| `- http://www.debian.org/ | be counted! #238656 |
| my keysever: pgp.mit.edu | my NPO: cfsg.org |

No more than 255 device files with Debian Sarge

Hello!

I just wanted to report that I finally found a way to
generate more than 256 device nodes for a major device:
Making a dist-upgrade to Debian Etch ("testing")!

To avoid any bad sideeffects that could postpone my project any
further all installations have been made inside a VMware Server
environment. These are the steps that have been done:

1. Installation of a "pure-as-possible" Debian Sarge. No X, no
compiler - only what is necessary to boot a small Linux. After
the installation all possible updates have been done.

Tested if the phenomenon happens:

mknod testnode1 c 200 255
mknod testnode2 c 200 256

This generates the following two files

testnode1 200 255
testnode2 201 0

Conclusion: the effect occurs.

2. Modifying /etc/apt/sources.conf: changed all appearences of
"stable" to "testing". Made an "apt-get update" to freshen
the package cache.

3. Finally did a "apt-get dist-upgrade". This found about 190 packages
that needed to be upgraded. Confirmed every question with its default
option.

After all was updated the machine was rebooted.

Kernel was still the old kernel 2.6.8-2.

Now repeated the test:

mknod testnode1 c 200 255
mknod testnode2 c 200 256

generates the following two files

testnode1 200 255
testnode2 201 256

Therefore it is assumed that upgrading a Debian Sarge to Etch
fixes the device file minor number limitation!

=> Problem is solved (but not yet understood)

My question now is this: Does anybody out there have a
glimpse of a clue about the real cause of this behaviour?

I bet that it has something to do with libc6 - but I am
not sure as I have not found a proof for this yet. (Even
the libc6 changelogs are unhelpful in this case.)

So far I think that this can solve the problem for our
customer. Thank you's go out to everyone who helped in
this case!!!

Best regards,

Stefan Sperling
N.A.T. GmbH
Software Developer / Net Admin

--

Syndicate content