[gentoo-user] make dev / udev question (1394 related)

[gentoo-user] make dev / udev question (1394 related)

[Mark Knecht]

Hi there,
   When I emerged libraw1394 it tells me there is no /dev/raw1394 device:

********************************************
Required /dev/raw1394 device file not found.
Run 'make dev' to create it.
********************************************

which is correct:

mark@godzilla ~ $ ls /dev/raw1394
ls: /dev/raw1394: No such file or directory
mark@godzilla ~ $

Question is how should that be created on a udev only system? The
Linux1394 web site gives this set of commands:

# Compile libraw1394:
cd /where/you/downloaded/libraw1394
tar xvfz libraw1394-1.1.0.tar.gz
cd libraw1394-1.1.0
./configure
make
make install
# Create the raw device (/dev/raw1394):
make dev
# Reboot:
shutdown -r now 

which would have made the device. Is the fact that the device isn't
created a problem with the ebuild? Even if it was, how should this
really be created for a udev system? A rule somewhere in some
udev.conf file?

Thanks,
Mark

-- 
gentoo-user@gentoo.org mailing list

Re: [gentoo-user] make dev / udev question (1394 related)

[Mike Williams]

On Sunday 17 July 2005 22:35, Mark Knecht wrote:
> which would have made the device. Is the fact that the device isn't
> created a problem with the ebuild? Even if it was, how should this
> really be created for a udev system? A rule somewhere in some
> udev.conf file?

Dynamic dev filesystems (devfs and udev) do what they say on the tin, creating 
device nodes dynamically.
The warning will be being generated by the package, not the ebuild.

Load the module and see what happens.

-- 
Mike Williams
-- 
gentoo-user@gentoo.org mailing list

Re: [gentoo-user] make dev / udev question (1394 related)

[Zac Medico]

Mike Williams wrote:
> On Sunday 17 July 2005 22:35, Mark Knecht wrote:
> 
>>which would have made the device. Is the fact that the device isn't
>>created a problem with the ebuild? Even if it was, how should this
>>really be created for a udev system? A rule somewhere in some
>>udev.conf file?
> 
> 
> Dynamic dev filesystems (devfs and udev) do what they say on the tin, creating 
> device nodes dynamically.
> The warning will be being generated by the package, not the ebuild.
> 
> Load the module and see what happens.
> 

On mine, the device is created automatically when I modprobe raw1394.

Zac
-- 
gentoo-user@gentoo.org mailing list

Re: [gentoo-user] make dev / udev question (1394 related)

[Daniel Drake]

Hi,

Mark Knecht wrote:
> which would have made the device. Is the fact that the device isn't
> created a problem with the ebuild? Even if it was, how should this
> really be created for a udev system? A rule somewhere in some
> udev.conf file?

Its a kernel "bug" which is fixed in 2.6.12. If you upgrade, it will start 
working automatically.

Daniel
-- 
gentoo-user@gentoo.org mailing list

Re: [gentoo-user] make dev / udev question (1394 related)

[Mark Knecht]

Zac & Mike,
   Thanks for the responses. I am getting a device /dev/raw/raw1394
created when I load the module and removed when I rmmod the module:

godzilla ~ # ls -la /dev/raw
ls: /dev/raw: No such file or directory
godzilla ~ # modprobe raw1394
godzilla ~ # ls -la /dev/raw
total 0
drwxr-xr-x   2 root root     60 Jul 17 15:05 .
drwxr-xr-x  18 root root  13660 Jul 17 15:05 ..
crw-rw----   1 root disk 171, 0 Jul 17 15:05 raw1394
godzilla ~ #

What's throwing me is that the libraw1394 emerge makes this limited complaint:

********************************************
Required /dev/raw1394 device file not found.
Run 'make dev' to create it.
********************************************

It seems to me that if the device was moved by either udev or the
Gentoo developers then shouldn't this package have been modified to
understand that and not throw a message like this? Or is there
possibly supposed to be a link of some type from /dev/raw1394 to
/dev/raw/raw1394?

Or is my device not the same as yours?

Thanks,
Mark


On 7/17/05, Zac Medico <zmedico@gmail.com> wrote:
> Mike Williams wrote:
> > On Sunday 17 July 2005 22:35, Mark Knecht wrote:
> >
> >>which would have made the device. Is the fact that the device isn't
> >>created a problem with the ebuild? Even if it was, how should this
> >>really be created for a udev system? A rule somewhere in some
> >>udev.conf file?
> >
> >
> > Dynamic dev filesystems (devfs and udev) do what they say on the tin, creating
> > device nodes dynamically.
> > The warning will be being generated by the package, not the ebuild.
> >
> > Load the module and see what happens.
> >
> 
> On mine, the device is created automatically when I modprobe raw1394.
> 
> Zac
> --
> gentoo-user@gentoo.org mailing list
> 
>

-- 
gentoo-user@gentoo.org mailing list

Re: [gentoo-user] make dev / udev question (1394 related)

[Mark Knecht]

Daniel,
   See my other post on this thread but as for the kernel it is updated:

mark@godzilla ~ $ uname -a
Linux godzilla 2.6.12-gentoo-r4 #1 Sat Jul 16 09:29:59 PDT 2005 i686
AMD Athlon(TM) XP1600+ AuthenticAMD GNU/Linux
mark@godzilla ~ $

Thanks,
Mark

On 7/17/05, Daniel Drake <dsd@gentoo.org> wrote:
> Hi,
> 
> Mark Knecht wrote:
> > which would have made the device. Is the fact that the device isn't
> > created a problem with the ebuild? Even if it was, how should this
> > really be created for a udev system? A rule somewhere in some
> > udev.conf file?
> 
> Its a kernel "bug" which is fixed in 2.6.12. If you upgrade, it will start
> working automatically.
> 
> Daniel
> --
> gentoo-user@gentoo.org mailing list
> 
>

-- 
gentoo-user@gentoo.org mailing list

Re: [gentoo-user] make dev / udev question (1394 related)

[Mark Knecht]

On 7/17/05, Daniel Drake <dsd@gentoo.org> wrote:
> Hi,
> 
> Mark Knecht wrote:
> > which would have made the device. Is the fact that the device isn't
> > created a problem with the ebuild? Even if it was, how should this
> > really be created for a udev system? A rule somewhere in some
> > udev.conf file?
> 
> Its a kernel "bug" which is fixed in 2.6.12. If you upgrade, it will start
> working automatically.
> 
> Daniel

Daniel,
   So I'm probably doing something wrong (as usual...no comments...)
;-) but the device having moved seems to break 1394commander:

mark@godzilla ~ $ 1394commander
1394commander 0.1.1
Copyright (C) 2002-2004 by Manfred Weihs <weihs@ict.tuwien.ac.at>
This software comes with absolutely no warranty.

No adapter specified!
couldn't get handle: No such file or directory
This probably means that you don't have raw1394 support in the kernel or that
you haven't loaded the raw1394 module.
mark@godzilla ~ $

Granted I've not used this program in quite awhile, but I think it
used to work under Gentoo. I'm guessing that it's hard wired to look
for /dev/raw1394, but that's just a guess.

thanks,
Mark

-- 
gentoo-user@gentoo.org mailing list

Re: [gentoo-user] make dev / udev question (1394 related)

[Daniel Drake]

Mark Knecht wrote:
> What's throwing me is that the libraw1394 emerge makes this limited complaint:
> 
> ********************************************
> Required /dev/raw1394 device file not found.
> Run 'make dev' to create it.
> ********************************************

This is more suited for the old-style 'static' /dev, where /dev was just a 
load of device nodes stored on disk.

With more modern solutions (devfs/udev), you don't ever have to worry about 
creating device nodes, they are created automatically when the appropriate 
driver is loaded. Its safe to ignore this warning on Gentoo.

> It seems to me that if the device was moved by either udev or the
> Gentoo developers then shouldn't this package have been modified to
> understand that and not throw a message like this? Or is there
> possibly supposed to be a link of some type from /dev/raw1394 to
> /dev/raw/raw1394?

I forgot about this one. This is a udev rule problem, which is supposed to 
deal with other raw devices, but not raw1394. It's been fixed for future udev 
versions, but for now, you can do this:

# echo 'KERNEL="raw1394", SYMLINK="%k"' > /etc/udev/rules.d/10-raw1394.rules
# udevstart

Daniel
-- 
gentoo-user@gentoo.org mailing list

Re: [gentoo-user] make dev / udev question (1394 related)

[Mark Knecht]

On 7/17/05, Daniel Drake <dsd@gentoo.org> wrote:
> Mark Knecht wrote:
> > What's throwing me is that the libraw1394 emerge makes this limited complaint:
> >
> > ********************************************
> > Required /dev/raw1394 device file not found.
> > Run 'make dev' to create it.
> > ********************************************
> 
> This is more suited for the old-style 'static' /dev, where /dev was just a
> load of device nodes stored on disk.
> 
> With more modern solutions (devfs/udev), you don't ever have to worry about
> creating device nodes, they are created automatically when the appropriate
> driver is loaded. Its safe to ignore this warning on Gentoo.
> 
> > It seems to me that if the device was moved by either udev or the
> > Gentoo developers then shouldn't this package have been modified to
> > understand that and not throw a message like this? Or is there
> > possibly supposed to be a link of some type from /dev/raw1394 to
> > /dev/raw/raw1394?
> 
> I forgot about this one. This is a udev rule problem, which is supposed to
> deal with other raw devices, but not raw1394. It's been fixed for future udev
> versions, but for now, you can do this:
> 
> # echo 'KERNEL="raw1394", SYMLINK="%k"' > /etc/udev/rules.d/10-raw1394.rules
> # udevstart
> 
> Daniel
> --
> gentoo-user@gentoo.org mailing list
> 
> 

Thank you Daniel. That's exactly what I was looking for:

godzilla ~ # echo 'KERNEL="raw1394", SYMLINK="%k"' >
/etc/udev/rules.d/10-raw1394.rules
godzilla ~ # udevstart
godzilla ~ # ls -al /dev/raw
raw/     raw1394
godzilla ~ # ls -al /dev/raw1394
lrwxrwxrwx  1 root root 11 Jul 17 15:50 /dev/raw1394 -> raw/raw1394
godzilla ~ # 1394commander
1394commander 0.1.1
Copyright (C) 2002-2004 by Manfred Weihs <weihs@ict.tuwien.ac.at>
This software comes with absolutely no warranty.

No adapter specified!
successfully got handle
current generation number (driver): 2
2 card(s) found
  nodes on bus:  2, card name: ohci1394
  nodes on bus:  1, card name: ohci1394
using adapter 0
found: 2 nodes on bus, local ID is 1, IRM is 1
current generation number (adapter): 1

entering command mode
Type 'help' for more information!
Command:

I appreciate your help very much!

Thanks,
Mark

-- 
gentoo-user@gentoo.org mailing list