[Buildroot] equivalent of a 'make modules'

[Buildroot] equivalent of a 'make modules'

[Jeff Krasky]

Hello,

 

I am working with another program that builds itself against the kernel.  I
am using BuildRoot for making the kernel and filesystem.  It is believed
that the next step would be to do something like a "make modules" command so
that the kernel and filesystem can be rebuilt with this other program's
information in it.

 

In Buildroot, where I normally do the 'make' command, I tried 'make modules'
and got this message:

 

"No rule to make target 'modules'. Stop."

 

Is there an equivalent?

 

Thanks.

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.busybox.net/pipermail/buildroot/attachments/20111122/44bd57ee/attachment.html>

[Buildroot] equivalent of a 'make modules'

[Thomas Petazzoni]

Le Tue, 22 Nov 2011 14:51:53 -0600,
"Jeff Krasky" <jeff.krasky@dspcg.com> a ?crit :

> I am working with another program that builds itself against the
> kernel.  I am using BuildRoot for making the kernel and filesystem.
> It is believed that the next step would be to do something like a
> "make modules" command so that the kernel and filesystem can be
> rebuilt with this other program's information in it.

Can you expand on what this program is ?

With your information, it's a bit hard to understand what you're trying
to achieve, and therefore to give you possible solutions.

> In Buildroot, where I normally do the 'make' command, I tried 'make
> modules' and got this message:

'modules' is a target of the Linux build system. Buildroot has
different targets, because Buildroot does a different job than the
Linux build system.

We really need more information about what your program is, how it
relates to the kernel build process, etc.

Thomas
-- 
Thomas Petazzoni, Free Electrons
Kernel, drivers, real-time and embedded Linux
development, consulting, training and support.
http://free-electrons.com

[Buildroot] equivalent of a 'make modules'

[Jeff Krasky]

>> I am working with another program that builds itself against the
>> kernel.  I am using BuildRoot for making the kernel and filesystem.
>> It is believed that the next step would be to do something like a
>> "make modules" command so that the kernel and filesystem can be
>> rebuilt with this other program's information in it.
>
>Can you expand on what this program is ?

EtherCAT.  I think it should have some kernel modules, be built against the
kernel source, etc.  A command like 'make modules_install install' can be
part of building for Linux PC.

Generally what I am doing is building EtherCAT and pointing it to the
buildroot directory for the kernel sources.  Then when EtherCAT build
finishes, I go back into BuildRoot and run make and the filesystem has been
updated.

But when I try and run, I get a modules not found error.  So the thought is
I need to do a make install type of command.

Thanks,



Jeff

[Buildroot] equivalent of a 'make modules'

[Thomas Petazzoni]

Le Tue, 22 Nov 2011 16:11:45 -0600,
"Jeff Krasky" <jeff.krasky@dspcg.com> a ?crit :

> EtherCAT.  I think it should have some kernel modules, be built
> against the kernel source, etc.  A command like 'make modules_install
> install' can be part of building for Linux PC.
> 
> Generally what I am doing is building EtherCAT and pointing it to the
> buildroot directory for the kernel sources.  Then when EtherCAT build
> finishes, I go back into BuildRoot and run make and the filesystem
> has been updated.
> 
> But when I try and run, I get a modules not found error.  So the
> thought is I need to do a make install type of command.

Ok. I guess the EtherCAT stuff is not open-source, unfortunately.

What you should do is create a Buildroot package that builds EtherCAT
against the Linux kernel built by Buildroot. Within a package, you can
access the location of the built kernel tree using the $(LINUX_DIR)
variable. Remember to depend on the "linux" package to make sure it is
built before your "ethercat" package. I would happily help you with
more details, but without access to the EtherCAT source code, it's
going to be a bit difficult.

Best regards,

Thomas
-- 
Thomas Petazzoni, Free Electrons
Kernel, drivers, real-time and embedded Linux
development, consulting, training and support.
http://free-electrons.com

[Buildroot] equivalent of a 'make modules'

[Jeff Krasky]

> Ok. I guess the EtherCAT stuff is not open-source, unfortunately.

This is what I am working with:

http://www.etherlab.org/en/ethercat/index.php


> What you should do is create a Buildroot package

Ok, this is where it becomes new for me.  Is this the "only way" or the
"most convenient" way, or the "simplest way" for someone not that familiar
with BuildRoot?

Thanks,


Jeff

[Buildroot] equivalent of a 'make modules'

[Thomas De Schampheleire]

On Tue, Nov 22, 2011 at 11:35 PM, Jeff Krasky <jeff.krasky@dspcg.com> wrote:
>> Ok. I guess the EtherCAT stuff is not open-source, unfortunately.
>
> This is what I am working with:
>
> http://www.etherlab.org/en/ethercat/index.php
>
>
>> What you should do is create a Buildroot package
>
> Ok, this is where it becomes new for me. ?Is this the "only way" or the
> "most convenient" way, or the "simplest way" for someone not that familiar
> with BuildRoot?

I would say it's both.

Below is an example package .mk file for a kernel module. It uses the
GENTARGETS infrastructure to explicitly specify the commands that
should be executed when building and installing.
Depending on the organization of EtherCAT, it may be a good idea to
create two packages: one for the application and one for the kernel
modules. It will probably make the .mk files a tad simpler.

FOOMODULE_VERSION = 0.1.2
FOOMODULE_SOURCE = foomodule-$(FOOMODULE_VERSION).tar.gz
FOOMODULE_SITE = http://yoursite

FOOMODULE_DEPENDENCIES = linux

FOOMODULE_INSTALL_MOD_DIR = foo

define FOOMODULE_BUILD_CMDS
       $(MAKE) $(LINUX_MAKE_FLAGS) -C $(LINUX_DIR) M=$(@D) modules
endef

define FOOMODULE_INSTALL_TARGET_CMDS
       $(MAKE) $(LINUX_MAKE_FLAGS) -C $(LINUX_DIR) M=$(@D) \
               INSTALL_MOD_STRIP=1 \
               INSTALL_MOD_DIR=$(FOOMODULE_INSTALL_MOD_DIR) \
               modules_install
endef

define FOOMODULE_UNINSTALL_TARGET_CMDS
       rm -rf $(TARGET_DIR)/lib/modules/$(LINUX_VERSION_PROBED)/$(FOOMODULE_INSTALL_MOD_DIR)
endef

$(eval $(call GENTARGETS))


Documentation on adding packages is available here:
http://buildroot.uclibc.org/buildroot.html#add_packages
If the EtherCAT application uses autotools (i.e. when it has files
like configure.in, configure.ac, Makefile.in, Makefile.am) then you
can use the AUTOTOOLS infrastructure for that part. The modules should
be done with GENTARGETS.

Best regards,
Thomas

[Buildroot] equivalent of a 'make modules'

[Thomas Petazzoni]

Le Tue, 22 Nov 2011 16:35:20 -0600,
"Jeff Krasky" <jeff.krasky@dspcg.com> a ?crit :

> Ok, this is where it becomes new for me.  Is this the "only way" or
> the "most convenient" way, or the "simplest way" for someone not that
> familiar with BuildRoot?

Basically you have the choice of :

 * Generating your system with Buildroot, then manually build EtherCAT
   outside of Buildroot using the Buildroot cross-compiler and the
   kernel built by Buildroot. Advantages: I don't see any, except maybe
   the fact that you don't need to learn out to create packages in
   Buildroot. Drawbacks: it might actually be *harder* to achieve than
   to create a package in Buildroot, and your build process is no
   longer nicely integrated.

 * Add a package in Buildroot. Requires learning a bit about how to
   create packages (but our documentation on this is quite good I
   think, and the Buildroot source tree is full of examples). In the
   end, it might be simpler to package your software component than
   building it manually, and you have the benefit of having it nicely
   integrated.

Regards,

Thomas
-- 
Thomas Petazzoni, Free Electrons
Kernel, drivers, real-time and embedded Linux
development, consulting, training and support.
http://free-electrons.com

[Buildroot] equivalent of a 'make modules'

[Jeff Krasky]

>Basically you have the choice of :
>
> * Generating your system with Buildroot, then manually build EtherCAT
>   outside of Buildroot using the Buildroot cross-compiler and the
>   kernel built by Buildroot. Advantages: I don't see any, except maybe
>   the fact that you don't need to learn out to create packages in
>   Buildroot.

That, and I am already trying this approach.  If there are just one or two
small things to do along this approach, it might be easier to stick with it.
For example, if there was just an easy way to do something like 'make
modules' then this approach might be quickest.

> Drawbacks: it might actually be *harder* to achieve than
>   to create a package in Buildroot, and your build process is no
>   longer nicely integrated.

Is the 'make modules' idea considered hard?

> * Add a package in Buildroot. Requires learning a bit about how to
>   create packages (but our documentation on this is quite good I
>   think, and the Buildroot source tree is full of examples).

I believe another poster sent a link to read about this, so I will look at
that.  In the meantime, can you comment on the first part of this email?

Thanks.

[Buildroot] equivalent of a 'make modules'

[Thomas Petazzoni]

Le Wed, 23 Nov 2011 08:17:24 -0600,
"Jeff Krasky" <jeff.krasky@dspcg.com> a ?crit :

> > Drawbacks: it might actually be *harder* to achieve than
> >   to create a package in Buildroot, and your build process is no
> >   longer nicely integrated.
> 
> Is the 'make modules' idea considered hard?

It doesn't make any sense for Buildroot to provide "make modules". This
is a *Linux* build target, not a *Buildroot* build target. You
certainly don't need Buildroot to provide "make modules" to build
external kernel modules outside of Buildroot.

I would need to look at how EtherCAT works to provide more details here.

Regards,

Thomas
-- 
Thomas Petazzoni, Free Electrons
Kernel, drivers, real-time and embedded Linux
development, consulting, training and support.
http://free-electrons.com

[Buildroot] LIBFOO_SITE in .mk file

[Jeff Krasky]

According to the documentation for making the .mk file
(http://buildroot.uclibc.org/buildroot.html#autotools-tutorial), what is the
syntax for specifying a local (hard drive location) of the .bz2 file?

LIBFOO_SITE = http://www.foosoftware.org/download

Is the example given.  Can I just do something like

LIBFOO_SITE = /home/libfoo.tar.bz2
?

Thanks.

[Buildroot] LIBFOO_SITE in .mk file

[Jeff Krasky]

In addition to the previous email (about LIBFOO) I followed the instructions
for making a package subdirectory (I called it EtherCAT), made a Config.in
file for it, added 'source "package/EtherCAT/Config.in"' to
package/Config.in and made the EtherCAT.mk file.

I then ran 'make menuconfig' again and selected ethercat.  Now when I
'make', I get this error.  I will post the Config.in and EtherCAT.mk below.

user1 at linux-0mmn:~/buildroot-2011.05> make
/usr/bin/make -j2  HOSTCC="/usr/bin/gcc" HOSTCXX="/usr/bin/g++"
silentoldconfig
make[1]: Entering directory `/home/user1/buildroot-2011.05'
KCONFIG_AUTOCONFIG=/home/user1/buildroot-2011.05/output/build/buildroot-conf
ig/auto.conf
KCONFIG_AUTOHEADER=/home/user1/buildroot-2011.05/output/build/buildroot-conf
ig/autoconf.h
KCONFIG_TRISTATE=/home/user1/buildroot-2011.05/output/build/buildroot-config
/tristate.config BUILDROOT_CONFIG=/home/user1/buildroot-2011.05/.config
/home/user1/buildroot-2011.05/output/build/buildroot-config/conf
--silentoldconfig Config.in
#
# make dependencies written to .auto.deps
# ATTENTION buildroot devels!
# See top of this file before playing with this auto-preprequisites!
#
make[1]: Leaving directory `/home/user1/buildroot-2011.05'
>>> ethercat 1.0 Downloading
/home/user1/Desktop/EtherCAT4Atmel/etherlabmaster-4adb56c4198d.tar.bz2:
Scheme missing.
>>> ethercat 1.0 Extracting
bzcat: Can't open input file
/home/user1/buildroot-2011.05/dl/etherlabmaster-4adb56c4198d.tar.bz2: No
such file or directory.
tar: This does not look like a tar archive
tar: Exiting with failure status due to previous errors
make: ***
[/home/user1/buildroot-2011.05/output/build/ethercat-1.0/.stamp_extracted]
Error 2
user1 at linux-0mmn:~/buildroot-2011.05>

The code originally came in as a .bz2 file.  I unzipped it into a .tar file,
then extracted it.  I then used the .tar file to convert it back to bz2
using bzip2 shell command.  Now it is not recognized?  I get the same type
of error if I put it back as a tar file and use gzip to make a .gz
extension.

Thoughts?

Here are the files:

Config.in
--------------------------------------------------------
config BR2_PACKAGE_ETHERCAT
	bool "ethercat"
	help
	  Tries to build EtherCAT into the Linux image.





EtherCAT.mk
----------------------------------------------------------
#############################################################
#
# EtherCAT
#
#############################################################
ETHERCAT_VERSION = 1.0
ETHERCAT_SOURCE = etherlabmaster-4adb56c4198d.tar.gz
ETHERCAT_SITE = /home/user1/Desktop/EtherCAT4Atmel
ETHERCAT_INSTALL_STAGING = YES
ETHERCAT_INSTALL_TARGET = YES
ETHERCAT_CONF_OPT = --enable-generic
#####LIBFOO_DEPENDENCIES = libglib2 host-pkg-config

$(eval $(call AUTOTARGETS,package,ethercat))



-----Original Message-----
From: buildroot-bounces@busybox.net [mailto:buildroot-bounces at busybox.net]
On Behalf Of Jeff Krasky
Sent: Wednesday, November 23, 2011 3:39 PM
To: buildroot at busybox.net
Subject: [Buildroot] LIBFOO_SITE in .mk file

According to the documentation for making the .mk file
(http://buildroot.uclibc.org/buildroot.html#autotools-tutorial), what is the
syntax for specifying a local (hard drive location) of the .bz2 file?

LIBFOO_SITE = http://www.foosoftware.org/download

Is the example given.  Can I just do something like

LIBFOO_SITE = /home/libfoo.tar.bz2
?

Thanks.

[Buildroot] LIBFOO_SITE in .mk file

[Thomas De Schampheleire]

On Wed, Nov 23, 2011 at 10:38 PM, Jeff Krasky <jeff.krasky@dspcg.com> wrote:
> According to the documentation for making the .mk file
> (http://buildroot.uclibc.org/buildroot.html#autotools-tutorial), what is the
> syntax for specifying a local (hard drive location) of the .bz2 file?
>
> LIBFOO_SITE = http://www.foosoftware.org/download
>
> Is the example given. ?Can I just do something like
>
> LIBFOO_SITE = /home/libfoo.tar.bz2

You can use

LIBFOO_SITE = file:///home/libfoo.tar.bz2

(note the three slashes; two from the scheme, the third one for an
absolute path)

Note that using file:// will not really help other people, except if
they have access to the same filesystem. You may consider using an
http/ftp server, or an scp url.

Best regards,
Thomas

[Buildroot] LIBFOO_SITE in .mk file

[Jeff Krasky]

> You can use
> 
> LIBFOO_SITE = file:///home/libfoo.tar.bz2
> 
> (note the three slashes; two from the scheme, the third one for an
> absolute path)

One issue now is that 'make' in Buildroot doesn't seem to find this file and
put it in the download directory.  Here is EtherCAT.mk:

#############################################################
#
# EtherCAT
#
#############################################################
ETHERCAT_VERSION = 1.0
ETHERCAT_SOURCE = etherlabmaster.tar.bz2
ETHERCAT_SITE = file:///home/user1/Desktop/EtherCAT4Atmel
ETHERCAT_INSTALL_STAGING = YES
ETHERCAT_INSTALL_TARGET = YES
ETHERCAT_CONF_OPT = --enable-generic
#####LIBFOO_DEPENDENCIES = libglib2 host-pkg-config

$(eval $(call AUTOTARGETS,package,ethercat))



So you can see that the file is at /home/user1/Desktop/EtherCAT4Atmel/
etherlabmaster.tar.bz2

Now this is the message from 'make' in Buildroot:

user1 at linux-0mmn:~/buildroot-2011.05> make
>>> ethercat 1.0 Extracting
bzcat: Can't open input file
/home/user1/buildroot-2011.05/dl/etherlabmaster.tar.bz2: No such file or
directory.
tar: This does not look like a tar archive
tar: Exiting with failure status due to previous errors
make: ***
[/home/user1/buildroot-2011.05/output/build/ethercat-1.0/.stamp_extracted]
Error 2
user1 at linux-0mmn:~/buildroot-2011.05>


So it looks like it never copies /home/user1/Desktop/EtherCAT4Atmel/
etherlabmaster.tar.bz2 over to the buildroot-2011.05/dl/ directory.  When I
manually inspect buildroot-2011.05/dl/ I do not see the file there.

Why isn't it getting copied over to dl/ ?


Thanks.

[Buildroot] LIBFOO_SITE in .mk file

[Thomas Petazzoni]

Le Mon, 28 Nov 2011 09:07:54 -0600,
"Jeff Krasky" <jeff.krasky@dspcg.com> a ?crit :

> ETHERCAT_SITE = file:///home/user1/Desktop/EtherCAT4Atmel

This looks like a directory name and not a file. You have two choices
for building from local sources :

 * The 'file' method, where ETHERCAT_SITE needs to be a tarball location

 * The 'local' method, where ETHERCAT_SITE needs to be a directory

But since ETHERCAT is available online, you really should be doing
something like :

ETHERCAT_SITE = http://somesite.com/sometarball.tar.gz

Thomas
-- 
Thomas Petazzoni, Free Electrons
Kernel, drivers, real-time and embedded Linux
development, consulting, training and support.
http://free-electrons.com

[Buildroot] LIBFOO_SITE in .mk file

[Jeff Krasky]

> > ETHERCAT_SITE = file:///home/user1/Desktop/EtherCAT4Atmel
> 
> This looks like a directory name and not a file.

My understanding from the tutorial at
http://buildroot.uclibc.org/buildroot.html#add_packages was that

ETHERCAT_SITE = file:///home/user1/Desktop/EtherCAT4Atmel

specifies the path (whether it is http or filesystem)

and

ETHERCAT_SOURCE = etherlabmaster.tar.bz2

specifies the file name.

Then I thought Buildroot would put ETHERCAT_SITE + ETHERCAT_SOURCE to come
up with the location of the file.  This is not correct?


I edited the .mk file to put the entire path in ETHERCAT_SITE, like

ETHERCAT_SITE = file:///home/user1/Desktop/EtherCAT4Atmel/
etherlabmaster.tar.bz2

but it still didn't find it.

[Buildroot] LIBFOO_SITE in .mk file

[Thomas Petazzoni]

Le Mon, 28 Nov 2011 09:19:29 -0600,
"Jeff Krasky" <jeff.krasky@dspcg.com> a ?crit :

> My understanding from the tutorial at
> http://buildroot.uclibc.org/buildroot.html#add_packages was that
> 
> ETHERCAT_SITE = file:///home/user1/Desktop/EtherCAT4Atmel
> 
> specifies the path (whether it is http or filesystem)
> 
> and
> 
> ETHERCAT_SOURCE = etherlabmaster.tar.bz2
> 
> specifies the file name.
> 
> Then I thought Buildroot would put ETHERCAT_SITE + ETHERCAT_SOURCE to come
> up with the location of the file.  This is not correct?

Woops, you're right, my bad. The manual is obvisouly correct.

You should use :

ETHERCAT_VERSION = 3.3
ETHERCAT_SITE = http://foobar.com/somewhere/
ETHERCAT_SOURCE = ethercat-$(ETHERCAT_VERSION).tar.bz2

> ETHERCAT_SITE = file:///home/user1/Desktop/EtherCAT4Atmel/
> etherlabmaster.tar.bz2

Why are you using a local directory ? You *really* should be using a
tarball available from http or ftp so that you can later contribute
your package to Buildroot.

Thanks,

Thomas
-- 
Thomas Petazzoni, Free Electrons
Kernel, drivers, real-time and embedded Linux
development, consulting, training and support.
http://free-electrons.com

[Buildroot] LIBFOO_SITE in .mk file

[Jeff Krasky]

> Woops, you're right, my bad. The manual is obvisouly correct.
> 
> You should use :
>
> ETHERCAT_VERSION = 3.3
> ETHERCAT_SITE = http://foobar.com/somewhere/
> ETHERCAT_SOURCE = ethercat-$(ETHERCAT_VERSION).tar.bz2
> 
> Why are you using a local directory ? You *really* should be using a
> tarball available from http or ftp so that you can later contribute
> your package to Buildroot.

Because I don't want just the default source of EtherCAT built into the
Linux image.  Lets say I want to modify something in the examples directory
of EtherCAT.  For instance, I had to modify the sample code and rebuild.  So
what I want is this "rebuilt" version of EtherCAT, with customized example
code, as part of the Linux image.  If I point at the EtherCAT website, that
won't be the correct version.  The EtherCAT examples development will occur
on the same machine that Buildroot is on, hence why I like the local
filesystem solution.

So any thoughts on why it is not finding the file?

Thanks.

[Buildroot] LIBFOO_SITE in .mk file

[Michael S. Zick]

On Mon November 28 2011, Jeff Krasky wrote:
> > Woops, you're right, my bad. The manual is obvisouly correct.
> > 
> > You should use :
> >
> > ETHERCAT_VERSION = 3.3
> > ETHERCAT_SITE = http://foobar.com/somewhere/
> > ETHERCAT_SOURCE = ethercat-$(ETHERCAT_VERSION).tar.bz2
> > 
> > Why are you using a local directory ? You *really* should be using a
> > tarball available from http or ftp so that you can later contribute
> > your package to Buildroot.
> 
> Because I don't want just the default source of EtherCAT built into the
> Linux image.  Lets say I want to modify something in the examples directory
> of EtherCAT.  For instance, I had to modify the sample code and rebuild.  So
> what I want is this "rebuilt" version of EtherCAT, with customized example
> code, as part of the Linux image.  If I point at the EtherCAT website, that
> won't be the correct version.  The EtherCAT examples development will occur
> on the same machine that Buildroot is on, hence why I like the local
> filesystem solution.
>

The typical way to handle that is to create a patch of the difference
between your local version and the download version;
Then let the build system apply the patch.
 
> So any thoughts on why it is not finding the file?
> 

Try a "make clean" then try again.
The missing file was reported to be the downloaded (*/dl/* directory)
the system may have that marked as already downloaded.

Mike
> Thanks.
> 
> _______________________________________________
> buildroot mailing list
> buildroot at busybox.net
> http://lists.busybox.net/mailman/listinfo/buildroot
> 
> 

[Buildroot] LIBFOO_SITE in .mk file

[Jeff Krasky]

> Try a "make clean" then try again.
> The missing file was reported to be the downloaded (*/dl/* directory)
> the system may have that marked as already downloaded.

>>> ethercat 1.0 Downloading
file:///home/user1/Desktop/EtherCAT4Atmel//etherlabmaster.tar.bz2:
Unsupported scheme "file".
>>> ethercat 1.0 Extracting
bzcat: Can't open input file
/home/user1/buildroot-2011.05/dl/etherlabmaster.tar.bz2: No such file or
directory.
tar: This does not look like a tar archive
tar: Exiting with failure status due to previous errors
make: ***
[/home/user1/buildroot-2011.05/output/build/ethercat-1.0/.stamp_extracted]
Error 2
user1 at linux-0mmn:~/buildroot-2011.05>


I did the 'make clean' and now it appears like it doesn't like the "file://"
part of things.  Has anyone done it like this?  The website for EtherCAT
doesn't point directly at a .bz2 file and so I can't do the http:// method.

Thanks.

[Buildroot] LIBFOO_SITE in .mk file

[Quotient Remainder]

On 28 Samh 2011 16:15, "Jeff Krasky" <jeff.krasky@dspcg.com> wrote:
...
>
> So any thoughts on why it is not finding the file?
>

Try "make ethercat-dirclean", then "rm dl/ethercat*" and finally, "make
ethercat". The stamp files or a partial download may be confusing matters.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.busybox.net/pipermail/buildroot/attachments/20111128/8be27223/attachment.html>

[Buildroot] LIBFOO_SITE in .mk file

[Jeff Krasky]

> Try "make ethercat-dirclean"

This removed output/build/ethercat-1.0

> "rm dl/ethercat*"

No files like this existed.

> "make ethercat"

Same problem still.  What makes Buildroot think it has the file in dl
already?

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.busybox.net/pipermail/buildroot/attachments/20111128/a78bddbf/attachment.html>

[Buildroot] LIBFOO_SITE in .mk file

[Michael S. Zick]

On Mon November 28 2011, Jeff Krasky wrote:
> > Try a "make clean" then try again.
> > The missing file was reported to be the downloaded (*/dl/* directory)
> > the system may have that marked as already downloaded.
> 
> >>> ethercat 1.0 Downloading
> file:///home/user1/Desktop/EtherCAT4Atmel//etherlabmaster.tar.bz2:
> Unsupported scheme "file".
> >>> ethercat 1.0 Extracting
> bzcat: Can't open input file
> /home/user1/buildroot-2011.05/dl/etherlabmaster.tar.bz2: No such file or
> directory.
> tar: This does not look like a tar archive
> tar: Exiting with failure status due to previous errors
> make: ***
> [/home/user1/buildroot-2011.05/output/build/ethercat-1.0/.stamp_extracted]
> Error 2
> user1 at linux-0mmn:~/buildroot-2011.05>
> 
> 
> I did the 'make clean' and now it appears like it doesn't like the "file://"

Try:
"file:///"

> part of things.  Has anyone done it like this?  The website for EtherCAT
> doesn't point directly at a .bz2 file and so I can't do the http:// method.
> 

Check me on this - see if I am following your description -
The site:
http://etherlab.org/en/components.php

Has a multiple component download file:
http://etherlab.org/download/bundle/etherlab-bundle-1.4.tar.bz2

That should be your download site and download package name.

Inside the bundle, it has directory trees of (among other things):
etherlab-bundle-1.6/2_rtal/linux-2.6.24.7.tar.bz2
etherlab-bundle-1.6/2_rtal/rtal-3.6.2.tar.bz2
etherlab-bundle-1.6/3_ethercat-1.4.0.tar.bz2
etherlab-bundle-1.6/5_etherlab/etherlab-1.2-rc4.tar.gz
etherlab-bundle-1.6/8_examples/example.mdl
etherlab-bundle-1.6/8_examples/example.sml
......

and you want to extract, patch, and build __some__ of the contained
archives (or, for examples, some of the files)....

Although I haven't done this myself, I do recall reading on the M.L.
here that there is a way in Buildroot to define a "custom extract"
(or download) step (or script?) -
That should be able to handle such an ugly package structure.

I think all it has to do is download the bundle of archives,
open just the outer bundle and mark each of the contained archives
as having been downloaded.  I __think__.

With the contained archives now downloaded (by extracting from the
containing package) the BR system should be able to return to running
the "extract,..." steps of each of the make files for each contained archive.

If that is what you need to do - one of the experts on the Buildroot
infrastructure needs to chime in with a "how to" here.  ;-)

Mike
> Thanks.
> 
> _______________________________________________
> buildroot mailing list
> buildroot at busybox.net
> http://lists.busybox.net/mailman/listinfo/buildroot
> 
> 

[Buildroot] LIBFOO_SITE in .mk file

[Jeff Krasky]

> Try:
> "file:///"

Sorry, I meant to imply I had the 3 slashes, with the 3rd one being for the
path.  So I've been using "file:///" all along, I just wrote "file://" since
that is the "scheme" part.

> Check me on this - see if I am following your description -
> The site:
> http://etherlab.org/en/components.php
>
> Has a multiple component download file:
> http://etherlab.org/download/bundle/etherlab-bundle-1.4.tar.bz2

That's the correct site, but not the version.  The latest tries to take you
to sourceforge for a download.  And navigating there didn't work too well.

If the "file" versus "http" was the problem, I could try working on that.
But it seems like something else is the issue for now.  Somehow this thinks
it already put the file in dl/ and it hasn't.

[Buildroot] LIBFOO_SITE in .mk file

[Jeff Krasky]

It looks like the possible methods for getting it are:

wget
svn
git
bzr

and none of those seem to apply to wanting to use a local copy.  So instead
I just did this:

cp etherlabmaster-1.0.tar.bz2 /home/user1/buildroot-2011.05/dl/

to get it in the dl/ directory.

Can anyone see a problem with this?

[Buildroot] LIBFOO_SITE in .mk file

[Jeff Krasky]

> cp etherlabmaster-1.0.tar.bz2 /home/user1/buildroot-2011.05/dl/
>
> to get it in the dl/ directory.

In addition to anyone seeing a problem with this, now when I run the
program:

/etc/init.d/ethercat start

I get a message like this:

Starting EtherCAT master devel modprobe: module ec_master not found in
modules.d
ep
 failed



This error message is what got me started down the path of adding the
package in Buildroot.  I used to build EtherCAT outside of Buildroot and
then that build process would put EtherCAT in the linux image.  But I'd get
this error, so the suggestion was made to make the EtherCAT build part of
Buildroot.  Somehow the kernel needs to be told of this module.  The
Buildroot process must not be doing a 'make install'?  Or is there a command
I should run on the board itself once Linux is running?

[Buildroot] LIBFOO_SITE in .mk file

[Thomas Petazzoni]

Le Mon, 28 Nov 2011 10:14:58 -0600,
"Jeff Krasky" <jeff.krasky@dspcg.com> a ?crit :

> Because I don't want just the default source of EtherCAT built into
> the Linux image.  Lets say I want to modify something in the examples
> directory of EtherCAT.  For instance, I had to modify the sample code
> and rebuild.  So what I want is this "rebuilt" version of EtherCAT,
> with customized example code, as part of the Linux image.  If I point
> at the EtherCAT website, that won't be the correct version.  The
> EtherCAT examples development will occur on the same machine that
> Buildroot is on, hence why I like the local filesystem solution.

I am not sure what you are trying to achieve here. If the upstream
package has issues that need to be fixed, then you can put patches in
the package/ethercat/ directory, which Buildroot will apply after
extracting the upstream tarball.

In addition, if you want to do quick development on EtherCAT, you can
also use the new "source override" feature that is part of the
upcoming Buildroot 2011.11. With this feature, you can tell Buildroot
to use a local directory for the source code instead of an upstream
tarball without changing the package .mk file itself, so that the .mk
file remains usable for upstreaming in Buildroot.

To do this, just create a local.mk file in the top directory of
Buildroot with the following contents :

ETHERCAT_OVERRIDE_SRCDIR = /my/local/place/

and then reference this local.mk file in the BR2_PACKAGE_OVERRIDE_FILE
configuration option.

If you really want to use a local directory directly in the package .mk
file, then you can use

ETHERCAT_SITE = /my/local/place
ETHERCAT_SITE_METHOD = local

But this is not great as it won't allow you to submit Ethercat to the
Buildroot developers, and will also force all users (including in your
company) to have Ethercat available at this location.

Regards,

Thomas
-- 
Thomas Petazzoni, Free Electrons
Kernel, drivers, real-time and embedded Linux
development, consulting, training and support.
http://free-electrons.com

[Buildroot] LIBFOO_SITE in .mk file

[Thomas Petazzoni]

Le Mon, 28 Nov 2011 12:50:22 -0600,
"Jeff Krasky" <jeff.krasky@dspcg.com> a ?crit :

> >>> ethercat 1.0 Downloading
> file:///home/user1/Desktop/EtherCAT4Atmel//etherlabmaster.tar.bz2:
> Unsupported scheme "file".

Support for "file" has been added in 2011.11, and you're using 2011.05,
so it's not supported.

Thomas
-- 
Thomas Petazzoni, Free Electrons
Kernel, drivers, real-time and embedded Linux
development, consulting, training and support.
http://free-electrons.com

[Buildroot] LIBFOO_SITE in .mk file

[Thomas Petazzoni]

Le Mon, 28 Nov 2011 16:15:02 -0600,
"Jeff Krasky" <jeff.krasky@dspcg.com> a ?crit :

> This error message is what got me started down the path of adding the
> package in Buildroot.  I used to build EtherCAT outside of Buildroot
> and then that build process would put EtherCAT in the linux image.
> But I'd get this error, so the suggestion was made to make the
> EtherCAT build part of Buildroot.  Somehow the kernel needs to be
> told of this module.  The Buildroot process must not be doing a 'make
> install'?  Or is there a command I should run on the board itself
> once Linux is running?

It depends on :

 * what your ethercat.mk file looks like (could you post it on the
   list ?)
 * how the EtherCAT build system works

Regards,

Thomas
-- 
Thomas Petazzoni, Free Electrons
Kernel, drivers, real-time and embedded Linux
development, consulting, training and support.
http://free-electrons.com

[Buildroot] LIBFOO_SITE in .mk file

[Jeff Krasky]

> > This error message is what got me started down the path of adding the
> > package in Buildroot.  I used to build EtherCAT outside of Buildroot
> > and then that build process would put EtherCAT in the linux image.
> > But I'd get this error, so the suggestion was made to make the
> > EtherCAT build part of Buildroot.  Somehow the kernel needs to be
> > told of this module.  The Buildroot process must not be doing a 'make
> > install'?  Or is there a command I should run on the board itself
> > once Linux is running?
>
> It depends on :
> 
> * what your ethercat.mk file looks like (could you post it on the
>   list ?)
> * how the EtherCAT build system works

EtherCAT.mk:

#############################################################
#
# EtherCAT
#
#############################################################
ETHERCAT_VERSION = 1.0
ETHERCAT_SOURCE = etherlabmaster-1.0.tar.bz2
###ETHERCAT_SITE = file:///home/user1/Desktop/EtherCAT4Atmel/
ETHERCAT_SITE = http://127.0.0.1/home/user1/Desktop/EtherCAT4Atmel/
ETHERCAT_INSTALL_STAGING = YES
ETHERCAT_INSTALL_TARGET = YES
ETHERCAT_CONF_OPT = --enable-generic
#####LIBFOO_DEPENDENCIES = libglib2 host-pkg-config

$(eval $(call AUTOTARGETS,package,ethercat))



Regarding how the EtherCAT build system works, how do I answer this for you?
Are there some Makefiles that would explain it?  Some other file I can post
that will explain it?

[Buildroot] LIBFOO_SITE in .mk file

[Thomas Petazzoni]

Le Tue, 29 Nov 2011 08:28:56 -0600,
"Jeff Krasky" <jeff.krasky@dspcg.com> a ?crit :

> #############################################################
> #
> # EtherCAT
> #
> #############################################################
> ETHERCAT_VERSION = 1.0
> ETHERCAT_SOURCE = etherlabmaster-1.0.tar.bz2
> ###ETHERCAT_SITE = file:///home/user1/Desktop/EtherCAT4Atmel/
> ETHERCAT_SITE = http://127.0.0.1/home/user1/Desktop/EtherCAT4Atmel/

That will work, but you still seem to not understand how to do this
properly. I've already explained in the past e-mails. You should do 

ETHERCAT_SITE = http://www.ethercat.org/somedirectory/

And then if you really need to hack EtherCAT locally, use the source
override feature.

> Regarding how the EtherCAT build system works, how do I answer this for you?
> Are there some Makefiles that would explain it?  Some other file I can post
> that will explain it?

I'll try to look at the EtherCAT build system, but I'm quite busy at
the moment. The INSTALL file has some details, but to build within
Buildroot, probably quite a few environment variables have to be passed.

Regards,

Thomas
-- 
Thomas Petazzoni, Free Electrons
Kernel, drivers, real-time and embedded Linux
development, consulting, training and support.
http://free-electrons.com

[Buildroot] LIBFOO_SITE in .mk file

[Jeff Krasky]

> I'll try to look at the EtherCAT build system, but I'm quite busy at
> the moment. The INSTALL file has some details, but to build within
> Buildroot, probably quite a few environment variables have to be passed.

Here are the contents of the INSTALL file:

----------------------------------------------------------------------------
---

This is the INSTALL file of the IgH EtherCAT Master.

$Id$

vim: set spelllang=en spell tw=78

----------------------------------------------------------------------------
---

Building and installing
=======================

The complete build and installation procedure is described in the respective
section of the documentation available from http://etherlab.org/en/ethercat.

----------------------------------------------------------------------------
---

For the impatient: The procedure mainly consists of calling

$ ./boostrap # to create the configure script, if downloaded from the repo

$ ./configure
$ make all modules

... and as root:

# make modules_install install
# depmod

... and linking the init script and copying the sysconfig file from
$PREFIX/etc
to the appropriate locations and customizing the sysconfig file.

# ln -s ${PREFIX}/etc/init.d/ethercat /etc/init.d/ethercat
# cp ${PREFIX}/etc/sysconfig/ethercat /etc/sysconfig/ethercat
# vi /etc/sysconfig/ethercat

Make sure, that the 'udev' package is installed, to automatically create the
EtherCAT character devices. The character devices will be created with mode
0660 and group root by default. If you want to give normal users reading
access, create a udev rule like this:

# echo KERNEL==\"EtherCAT[0-9]*\", MODE=\"0664\" >
/etc/udev/rules.d/99-EtherCAT.rules

Now you can start the EtherCAT master:

# /etc/init.d/ethercat start

Have a look at the examples/ subdirectory for some application examples.

Have fun!

----------------------------------------------------------------------------
---

[Buildroot] LIBFOO_SITE in .mk file

[Michael S. Zick]

On Tue November 29 2011, Thomas Petazzoni wrote:
> Le Tue, 29 Nov 2011 08:28:56 -0600,
> "Jeff Krasky" <jeff.krasky@dspcg.com> a ?crit :
> 
> > #############################################################
> > #
> > # EtherCAT
> > #
> > #############################################################
> > ETHERCAT_VERSION = 1.0
> > ETHERCAT_SOURCE = etherlabmaster-1.0.tar.bz2
> > ###ETHERCAT_SITE = file:///home/user1/Desktop/EtherCAT4Atmel/
> > ETHERCAT_SITE = http://127.0.0.1/home/user1/Desktop/EtherCAT4Atmel/
> 
> That will work, but you still seem to not understand how to do this
> properly. I've already explained in the past e-mails. You should do 
> 
> ETHERCAT_SITE = http://www.ethercat.org/somedirectory/
> 

And if you want more recent than version 1.4, either
switch to Buildroot 2011.11(-rc?) and use the hg methods for download
or
use the direct hg link to the tarball of the stable version 1.5 tag
that I posted.

The hg setup installed at SourceForge is set to provide a *.tar.bz2
of any tag in the repository, down to a specific file level.
Other hg repositories may be setup differently (the one I run provides
zip and *.tar.gz).

The Buildroot hg method will clone and tar-ball the entire repository,
which may take more than a few cpu cycles.  ;-)
But that is the more general method.

Hardcoding the URL displayed by the "tar-ball button(s)" when browsing
an hg repository leaves you with a strangely named file but it may
be a much smaller download.
Although not as general, that only needs the http methods available
in the older Buildroot versions.

Mike
> And then if you really need to hack EtherCAT locally, use the source
> override feature.
> 
> > Regarding how the EtherCAT build system works, how do I answer this for you?
> > Are there some Makefiles that would explain it?  Some other file I can post
> > that will explain it?
> 
> I'll try to look at the EtherCAT build system, but I'm quite busy at
> the moment. The INSTALL file has some details, but to build within
> Buildroot, probably quite a few environment variables have to be passed.
> 
> Regards,
> 
> Thomas

[Buildroot] LIBFOO_SITE in .mk file

[Jeff Krasky]

The message I get is "Starting EtherCAT master devel modprobe: module
ec_master not found in modules.d ep  failed" and I think the problem is with
module dependencies.

On the Linux PC, in /lib/modules/<source> there is an 'ethercat'
subdirectory.  On the Linux built for the board, there is not.  This is
probably the problem.

Now when installing on Linux PC, the instructions mention doing:

make modules_install install
depmod


and so perhaps I need to do the 'depmod' in Buildroot somehow?

[Buildroot] LIBFOO_SITE in .mk file

[Jeff Krasky]

> On the Linux PC, in /lib/modules/<source> there is an 'ethercat'
> subdirectory.  On the Linux built for the board, there is not.  This is
> probably the problem.

Also, since this directory isn't there, it means that 

/lib/modules/<source>/ethercat/master/ec_master.ko

isn't present.  So perhaps the 'make modules_install install' command is
what would do this?  If so, how do I run it?

[Buildroot] LIBFOO_SITE in .mk file

[Jeff Krasky]

> /lib/modules/<source>/ethercat/master/ec_master.ko
> 
> isn't present.

If anyone can just help for a second, I have narrowed this down more.  I
found out that "ethercat's 'make modules_install' just calls the kernel's
make -C <kernel_dir> modules_install which in turn installs the modules
where it seems fit".

Since this is what needs to be done, can someone mention how to do it in
Buildroot?  Maybe I need the Makefile for Generic Packages, not the
autotools-based packages makefile?

[Buildroot] LIBFOO_SITE in .mk file

[Thomas Petazzoni]

Le Wed, 30 Nov 2011 09:59:24 -0600,
"Jeff Krasky" <jeff.krasky@dspcg.com> a ?crit :

> Since this is what needs to be done, can someone mention how to do it in
> Buildroot?  Maybe I need the Makefile for Generic Packages, not the
> autotools-based packages makefile?

You'll need to pass the INSTALL_MOD_PATH environment variable, but also
the ARCH= and CROSS_COMPILE variable. The easiest way is probably to use
$(LINUX_MAKE_FLAGS) variable defined in linux/linux.mk.

Thomas
-- 
Thomas Petazzoni, Free Electrons
Kernel, drivers, real-time and embedded Linux
development, consulting, training and support.
http://free-electrons.com

[Buildroot] LIBFOO_SITE in .mk file

[Jeff Krasky]

> > Since this is what needs to be done, can someone mention how to do it in
> > Buildroot?  Maybe I need the Makefile for Generic Packages, not the
> > autotools-based packages makefile?
>
> You'll need to pass the INSTALL_MOD_PATH environment variable, but also
> the ARCH= and CROSS_COMPILE variable. The easiest way is probably to use
> $(LINUX_MAKE_FLAGS) variable defined in linux/linux.mk.

Thank you so much for the reply.  I just checked and I don't see
"LINUX_MAKE_FLAGS" in buildroot-2011.05/linux/linux.mk.  Maybe that's in the
2011.11 version?  But, are you saying that I could still use the
autotools-based makefile and just add INSTALL_MOD_PATH, ARCH, and
CROSS_COMPILE?  I have it running now, but the build takes a while, so if
this is wrong can you tell me?

#############################################################
#
# EtherCAT
#
#############################################################
ETHERCAT_VERSION = 1.0
ETHERCAT_SOURCE = etherlabmaster-1.0.tar.bz2
###ETHERCAT_SITE = file:///home/user1/Desktop/EtherCAT4Atmel/
ETHERCAT_SITE = http://127.0.0.1/home/user1/Desktop/EtherCAT4Atmel/
ETHERCAT_INSTALL_STAGING = YES
ETHERCAT_INSTALL_TARGET = YES
ETHERCAT_CONF_OPT = --enable-generic
#####LIBFOO_DEPENDENCIES = libglib2 host-pkg-config
#####ETHERCAT_MAKE_OPT = INSTALL_MOD_DIR="ethercat" modules_install
ETHERCAT_INSTALL_MOD_PATH = /lib/modules/2.6.33/ethercat
ETHERCAT_ARCH = arm
ETHERCAT_CROSS_COMPILE = arm-unknown-linux-uclibcgnueabi-

$(eval $(call AUTOTARGETS,package,ethercat))

[Buildroot] LIBFOO_SITE in .mk file

[Thomas Petazzoni]

Le Wed, 30 Nov 2011 10:20:41 -0600,
"Jeff Krasky" <jeff.krasky@dspcg.com> a ?crit :

> Thank you so much for the reply.  I just checked and I don't see
> "LINUX_MAKE_FLAGS" in buildroot-2011.05/linux/linux.mk.  Maybe that's in the
> 2011.11 version?  But, are you saying that I could still use the
> autotools-based makefile and just add INSTALL_MOD_PATH, ARCH, and
> CROSS_COMPILE?  I have it running now, but the build takes a while, so if
> this is wrong can you tell me?

It is wrong.

> #############################################################
> #
> # EtherCAT
> #
> #############################################################
> ETHERCAT_VERSION = 1.0
> ETHERCAT_SOURCE = etherlabmaster-1.0.tar.bz2
> ###ETHERCAT_SITE = file:///home/user1/Desktop/EtherCAT4Atmel/
> ETHERCAT_SITE = http://127.0.0.1/home/user1/Desktop/EtherCAT4Atmel/
> ETHERCAT_INSTALL_STAGING = YES
> ETHERCAT_INSTALL_TARGET = YES
> ETHERCAT_CONF_OPT = --enable-generic
> #####LIBFOO_DEPENDENCIES = libglib2 host-pkg-config
> #####ETHERCAT_MAKE_OPT = INSTALL_MOD_DIR="ethercat" modules_install
> ETHERCAT_INSTALL_MOD_PATH = /lib/modules/2.6.33/ethercat
> ETHERCAT_ARCH = arm
> ETHERCAT_CROSS_COMPILE = arm-unknown-linux-uclibcgnueabi-

Those are not Buildroot variables, but Kbuild variables, that you must
pass when running the modules_install. The fact that EtherCAT contains
kernel modules makes it not completely straightforward to package.

Do you need it right now ? I might be interested in helping package
EtherCAT, but I don't have much time at the moment.

Best regards,

Thomas
-- 
Thomas Petazzoni, Free Electrons
Kernel, drivers, real-time and embedded Linux
development, consulting, training and support.
http://free-electrons.com

[Buildroot] LIBFOO_SITE in .mk file

[Jeff Krasky]

> Those are not Buildroot variables, but Kbuild variables, that you must
> pass when running the modules_install. The fact that EtherCAT contains
> kernel modules makes it not completely straightforward to package.

At what point is modules_install being run?

> Do you need it right now ? I might be interested in helping package
> EtherCAT, but I don't have much time at the moment.

Time is very short, yes.  Anything you can do to help is definitely
appreciated.  Like, for now, can you say more about running modules_install
and passing parameters to it?

Thank you very much, again.

[Buildroot] LIBFOO_SITE in .mk file

[Michael S. Zick]

On Wed November 30 2011, Jeff Krasky wrote:
> > Those are not Buildroot variables, but Kbuild variables, that you must
> > pass when running the modules_install. The fact that EtherCAT contains
> > kernel modules makes it not completely straightforward to package.
> 
> At what point is modules_install being run?
> 
> > Do you need it right now ? I might be interested in helping package
> > EtherCAT, but I don't have much time at the moment.
> 
> Time is very short, yes.  Anything you can do to help is definitely
> appreciated.  Like, for now, can you say more about running modules_install
> and passing parameters to it?
>

Would a worked example help?
This package patches the kernel and builds the related software:
http://git.buildroot.net/buildroot/commit/?id=956d4ab0f501c44f977297d8faac95afa43f3082

Mike 
> Thank you very much, again.
> 
> _______________________________________________
> buildroot mailing list
> buildroot at busybox.net
> http://lists.busybox.net/mailman/listinfo/buildroot
> 
> 

[Buildroot] LIBFOO_SITE in .mk file

[Jeff Krasky]

> Would a worked example help?
> This package patches the kernel and builds the related software:
>http://git.buildroot.net/buildroot/commit/?id=956d4ab0f501c44f977297d8faac9
5afa43f3082


Thanks.  I'm going over that now.  I see two .mk files.  Is that what needs
to be changed?  In other words, this helps and I can try things, but I'm not
completely sure what to look at in this patch.

Thanks again.

[Buildroot] LIBFOO_SITE in .mk file

[Michael S. Zick]

On Wed November 30 2011, Jeff Krasky wrote:
> > Would a worked example help?
> > This package patches the kernel and builds the related software:
> >http://git.buildroot.net/buildroot/commit/?id=956d4ab0f501c44f977297d8faac9
> 5afa43f3082
> 
> 
> Thanks.  I'm going over that now.  I see two .mk files.  Is that what needs
> to be changed?  In other words, this helps and I can try things, but I'm not
> completely sure what to look at in this patch.
>

I do not know either.

There where a number of changes made to that package since it
was first added at the above link.

Better if you downloaded 2011.11-rc3 tar-ball and look at that
package as it exists today in the Buildroot sources.

Mike 
> Thanks again.
> 
> _______________________________________________
> buildroot mailing list
> buildroot at busybox.net
> http://lists.busybox.net/mailman/listinfo/buildroot
> 
> 

[Buildroot] LIBFOO_SITE in .mk file

[Thomas De Schampheleire]

Jeff,

On Wed, Nov 30, 2011 at 5:38 PM, Jeff Krasky <jeff.krasky@dspcg.com> wrote:
>> Those are not Buildroot variables, but Kbuild variables, that you must
>> pass when running the modules_install. The fact that EtherCAT contains
>> kernel modules makes it not completely straightforward to package.
>
> At what point is modules_install being run?
>
>> Do you need it right now ? I might be interested in helping package
>> EtherCAT, but I don't have much time at the moment.
>
> Time is very short, yes. ?Anything you can do to help is definitely
> appreciated. ?Like, for now, can you say more about running modules_install
> and passing parameters to it?
>

Since your ethercat is actually containing two things, being an
application *and* a kernel module, you should treat it as such. In my
opinion, you should either:

* create two different packages, one for ethercat and one for the
ethercat driver. Each would have a .mk file in a package/* directory.
* create one .mk file, but with two different sets of build rules, one
for the application and one for the driver. You can make sure that one
depends on the other so that they're always built together. If the
application is autotools-based, you can use the AUTOTOOLS
infrastructure. For the driver I suggest the GENTARGETS one (see
below).
* create one .mk file, with one set of build rules, that combine the
build steps for the application and for the kernel.

With regards to the kernel driver part, I have already given you an
example set of build rules in one of your other threads, see
http://permalink.gmane.org/gmane.comp.lib.uclibc.buildroot/35273
Do you understand how these build rules work? Did you try them?

Building this with older versions of buildroot is perfectly possible,
if you have to stick with older versions for some reason. (if you are
free to update, I'd suggest you do). Do keep in mind that in previous
versions of buildroot, the LINUX_ prefix was called LINUX26_, so for
example LINUX_MAKE_FLAGS needs to be renamed to LINUX26_MAKE_FLAGS,
likewise with LINUX_DIR.

I hope this helps.

If you still have questions, I suggest you provide us with your
current.mk file and the problems (e.g. build errors) you still
experience with it. Please take some time to provide detailed
information, this will allow you to get more relevant responses from
the list.

Best regards,
Thomas

[Buildroot] LIBFOO_SITE in .mk file

[Thomas Petazzoni]

Le Wed, 30 Nov 2011 18:48:13 +0100,
Thomas De Schampheleire <patrickdepinguin+buildroot@gmail.com> a ?crit :

> * create two different packages, one for ethercat and one for the
> ethercat driver. Each would have a .mk file in a package/* directory.

I don't agree here. This will make it hard to keep both things in sync.

The build system of EtherCAT seems to integrate both the build of the
kernel-space parts and the user-space parts, and it's probably doable
inside a single package.

> * create one .mk file, but with two different sets of build rules, one
> for the application and one for the driver. You can make sure that one
> depends on the other so that they're always built together. If the
> application is autotools-based, you can use the AUTOTOOLS
> infrastructure. For the driver I suggest the GENTARGETS one (see
> below).

I haven't had the time to look closely, but it seems like the
autotools-based build system triggers the build of the kernel parts. At
least, the kernel part is described by a Kbuild.in file that needs to
be processed by autoconf before being able to build the kernel side
part. So definitely, the userspace parts and kernelspace parts are
related in terms of build system.

Best regards,

Thomas
-- 
Thomas Petazzoni, Free Electrons
Kernel, drivers, real-time and embedded Linux
development, consulting, training and support.
http://free-electrons.com

[Buildroot] LIBFOO_SITE in .mk file

[Jeff Krasky]

> * create one .mk file, with one set of build rules, that combine the
> build steps for the application and for the kernel.

I'll try this one first.  Here is the latest EtherCAT.mk:

#############################################################
#
# EtherCAT
#
#############################################################
ETHERCAT_VERSION = 1.0
ETHERCAT_SOURCE = etherlabmaster-1.0.tar.bz2
###ETHERCAT_SITE = file:///home/user1/Desktop/EtherCAT4Atmel/
ETHERCAT_SITE = http://127.0.0.1/home/user1/Desktop/EtherCAT4Atmel/
ETHERCAT_INSTALL_STAGING = YES
ETHERCAT_INSTALL_TARGET = YES
ETHERCAT_CONF_OPT = --enable-generic
ETHERCAT_DEPENDENCIES = linux
ETHERCAT_INSTALL_MOD_DIR = ethercat
#####ETHERCAT_INSTALL_MOD_PATH

define ETHERCAT_BUILD_CMDS
       $(MAKE) $(LINUX26_MAKE_FLAGS) -C $(LINUX26_DIR) M=$(@D) modules
endef

define ETHERCAT_INSTALL_TARGET_CMDS
       $(MAKE) $(LINUX26_MAKE_FLAGS) -C $(LINUX26_DIR) M=$(@D) \
               INSTALL_MOD_STRIP=1 \
               INSTALL_MOD_DIR=$(ETHERCAT_INSTALL_MOD_DIR) \
               modules_install
endef

define ETHERCAT_UNINSTALL_TARGET_CMDS
       rm -rf
$(TARGET_DIR)/lib/modules/$(LINUX26_VERSION_PROBED)/$(ETHERCAT_INSTALL_MOD_D
IR)
endef

$(eval $(call GENTARGETS))

#####$(eval $(call AUTOTARGETS,package,ethercat))


> Do keep in mind that in previous
> versions of buildroot, the LINUX_ prefix was called LINUX26_, so for
> example LINUX_MAKE_FLAGS needs to be renamed to LINUX26_MAKE_FLAGS,
> likewise with LINUX_DIR.

Thanks.  I changed them, but I left

ETHERCAT_DEPENDENCIES = linux

as just "linux".  Should it be "linux26"?  The build is going on now, but it
takes a while, so if it should be "linux26" can you tell me (or if anything
else is wrong in .mk) so I can stop the build and fix it?

> If you still have questions, I suggest you provide us with your
> current.mk file and the problems (e.g. build errors) you still
> experience with it.

I'll see how this latest build goes, but I think the result needs to be that


/lib/modules/2.6.33

needs to have an 'ethercat' subdirectory containing ec_master.ko, among
other files.  So far, no build has done this.

Thanks so much, time is short and this is really helping.

[Buildroot] LIBFOO_SITE in .mk file

[Jeff Krasky]

> $(eval $(call GENTARGETS))

This needed to be changed to $(eval $(call GENTARGETS,package,ethercat))

and then it built the modules!  Now there is /lib/modules/2.6.33/ethercat
with the correct kernel modules.

However, there is no longer

/etc/init.d/ethercat

executable file.  So in changing the EtherCAT.mk file now the "application"
part isn't getting built.  Do I need to put back this line:

$(eval $(call AUTOTARGETS,package,ethercat))

If not, what should I do?  It seems I can make application OR kernel module,
just not both.


Thanks again.

[Buildroot] LIBFOO_SITE in .mk file

[Thomas De Schampheleire]

Hi,

On Wed, Nov 30, 2011 at 7:11 PM, Jeff Krasky <jeff.krasky@dspcg.com> wrote:
>> * create one .mk file, with one set of build rules, that combine the
>> build steps for the application and for the kernel.
>
> I'll try this one first. ?Here is the latest EtherCAT.mk:
>
> #############################################################
> #
> # EtherCAT
> #
> #############################################################
> ETHERCAT_VERSION = 1.0
> ETHERCAT_SOURCE = etherlabmaster-1.0.tar.bz2
> ###ETHERCAT_SITE = file:///home/user1/Desktop/EtherCAT4Atmel/
> ETHERCAT_SITE = http://127.0.0.1/home/user1/Desktop/EtherCAT4Atmel/
> ETHERCAT_INSTALL_STAGING = YES
> ETHERCAT_INSTALL_TARGET = YES
> ETHERCAT_CONF_OPT = --enable-generic
> ETHERCAT_DEPENDENCIES = linux
> ETHERCAT_INSTALL_MOD_DIR = ethercat
> #####ETHERCAT_INSTALL_MOD_PATH
>
> define ETHERCAT_BUILD_CMDS
> ? ? ? $(MAKE) $(LINUX26_MAKE_FLAGS) -C $(LINUX26_DIR) M=$(@D) modules
> endef
>
> define ETHERCAT_INSTALL_TARGET_CMDS
> ? ? ? $(MAKE) $(LINUX26_MAKE_FLAGS) -C $(LINUX26_DIR) M=$(@D) \
> ? ? ? ? ? ? ? INSTALL_MOD_STRIP=1 \
> ? ? ? ? ? ? ? INSTALL_MOD_DIR=$(ETHERCAT_INSTALL_MOD_DIR) \
> ? ? ? ? ? ? ? modules_install
> endef
>
> define ETHERCAT_UNINSTALL_TARGET_CMDS
> ? ? ? rm -rf
> $(TARGET_DIR)/lib/modules/$(LINUX26_VERSION_PROBED)/$(ETHERCAT_INSTALL_MOD_D
> IR)
> endef
>
> $(eval $(call GENTARGETS))
>
> #####$(eval $(call AUTOTARGETS,package,ethercat))

You do realize that these rules are purely intended for the module, right.

Based on what ThomasP says, things may be a little more complicated
than I originally thought.

>
>
>> Do keep in mind that in previous
>> versions of buildroot, the LINUX_ prefix was called LINUX26_, so for
>> example LINUX_MAKE_FLAGS needs to be renamed to LINUX26_MAKE_FLAGS,
>> likewise with LINUX_DIR.
>
> Thanks. ?I changed them, but I left
>
> ETHERCAT_DEPENDENCIES = linux
>
> as just "linux". ?Should it be "linux26"? ?The build is going on now, but it
> takes a while, so if it should be "linux26" can you tell me (or if anything
> else is wrong in .mk) so I can stop the build and fix it?

In buildroot 2010.05, it was possible to use both linux as linux26, so
you should be fine.

>
>> If you still have questions, I suggest you provide us with your
>> current.mk file and the problems (e.g. build errors) you still
>> experience with it.
>
> I'll see how this latest build goes, but I think the result needs to be that
>
>
> /lib/modules/2.6.33
>
> needs to have an 'ethercat' subdirectory containing ec_master.ko, among
> other files. ?So far, no build has done this.

Just checking: you are looking into output/target/lib/modules or
/lib/modules on the target right, not /lib/modules on the host?
If the modules_install step in your .mk file goes well, this should
put the file there.
Seeing the build log of that part should hint you what is wrong, if anything.

>
> Thanks so much, time is short and this is really helping.
>

[Buildroot] LIBFOO_SITE in .mk file

[Jeff Krasky]

> You do realize that these rules are purely intended for the module, right.

No, I thought it would do both.  I have a second build running with both
$evals in it.  Hopefully that will do it.

> Based on what ThomasP says, things may be a little more complicated
> than I originally thought.

Do I need to separate sets of build rules in the .mk?

> Just checking: you are looking into output/target/lib/modules or
> /lib/modules on the target right, not /lib/modules on the host?

Correct, I am looking at /lib/modules when running the Linux image on the
target.

> If the modules_install step in your .mk file goes well, this should
> put the file there.

It did.  I just wrote a post about how it's there but the application isn't.

Thanks again.

[Buildroot] LIBFOO_SITE in .mk file

[Thomas Petazzoni]

Le Wed, 30 Nov 2011 14:01:10 -0600,
"Jeff Krasky" <jeff.krasky@dspcg.com> a ?crit :

> It did.  I just wrote a post about how it's there but the application isn't.

I have just posted a patch that adds the igh-ethercat package to
Buildroot. It applies on the latest Buildroot Git, which should become
in a few hours the 2011.11 release.

The patch has only been compile tested, I don't have the EtherCAT
slaves hardware to test this. Note that you must use a kernel version
compatible with the EtherCAT network drivers.

Best regards,

Thomas
-- 
Thomas Petazzoni, Free Electrons
Kernel, drivers, real-time and embedded Linux
development, consulting, training and support.
http://free-electrons.com

[Buildroot] LIBFOO_SITE in .mk file

[Thomas Petazzoni]

Hello,

Please don't reply to me directly, but use the list instead.

Le Wed, 30 Nov 2011 16:34:59 -0600,
"Jeff Krasky" <jeff.krasky@dspcg.com> a ?crit :

> > I have just posted a patch that adds the igh-ethercat package to
> > Buildroot. It applies on the latest Buildroot Git, which should become
> > in a few hours the 2011.11 release.
> 
> So if I:
> 
> wget http://buildroot.uclibc.org/downloads/buildroot-2011.11.tar.bz2
> 
> I can un-tar it, go into 'make menuconfig' and select EtherCAT and then it
> will build both the application and kernel module as part of the Linux
> image/filesystem?

After untarring the tarball, you have to apply the patch. The
igh-ethercat package is *not* part of the release.

> Right now I am building with 2.6.33, will I still be able to select that in
> buildroot-2011.11?

Buildroot allows you to select whatever kernel version you want.
However, none of the EtherCAT drivers seem to be supported for 2.6.33.

Thomas
-- 
Thomas Petazzoni, Free Electrons
Kernel, drivers, real-time and embedded Linux
development, consulting, training and support.
http://free-electrons.com

[Buildroot] LIBFOO_SITE in .mk file

[Jeff Krasky]

> After untarring the tarball, you have to apply the patch. The
> igh-ethercat package is *not* part of the release.

Oh, ok.  So the patch will be a .patch file on 

http://buildroot.uclibc.org/downloads/

and I just have to

1. download it
2. go into buildroot-2011.11 directory
3. run 'patch filename.patch'
4. then do the 'make menuconfig', 'make' etc.

is that correct?

Thanks again.

[Buildroot] LIBFOO_SITE in .mk file

[Thomas Petazzoni]

Le Wed, 30 Nov 2011 16:54:58 -0600,
"Jeff Krasky" <jeff.krasky@dspcg.com> a ?crit :

> Oh, ok.  So the patch will be a .patch file on 
> 
> http://buildroot.uclibc.org/downloads/

No, the patch is in an e-mail I have sent to the list yesterday evening.

> 1. download it
> 2. go into buildroot-2011.11 directory
> 3. run 'patch filename.patch'

No:

cat filename.patch | patch -p1

Thomas
-- 
Thomas Petazzoni, Free Electrons
Kernel, drivers, real-time and embedded Linux
development, consulting, training and support.
http://free-electrons.com

[Buildroot] LIBFOO_SITE in .mk file

[Jeff Krasky]

> No, the patch is in an e-mail I have sent to the list yesterday evening.

Ok, so I go to the email with a subject of "[Buildroot] [PATCH 1/1]
igh-ethercat: new package" and I copy the text from



---
 package/Config.in                    |    1 +
 package/igh-ethercat/Config.in       |   23 +++++++++++++++++++++++




all the way down to the text




+IGH_ETHERCAT_POST_INSTALL_TARGET_HOOKS += IGH_ETHERCAT_INSTALL_MODULES
+
+$(eval $(call AUTOTARGETS))
\ No newline at end of file
--
1.7.4.1



and place it in a text file called filename.patch.  Then I go into the
directory buildroot-2011.11 and I run 'cat filename.patch | patch -p1'

Then I can just run 'make menuconfig', 'make', etc?

Thanks so much.

[Buildroot] LIBFOO_SITE in .mk file

[Michael S. Zick]

On Thu December 1 2011, Jeff Krasky wrote:
> > No, the patch is in an e-mail I have sent to the list yesterday evening.
> 
> Ok, so I go to the email with a subject of "[Buildroot] [PATCH 1/1]
> igh-ethercat: new package" and I copy the text from
> 

You don't even have to do that, just save the e-mail as a plain text
file using your e-mail client.

Patch will ignore the leading, non-patch, discussion.

It was designed that way, back in the days when the only way to
pass patches was by e-mail.  ;-)

Mike
> 
> 
> ---
>  package/Config.in                    |    1 +
>  package/igh-ethercat/Config.in       |   23 +++++++++++++++++++++++
> 
> 
> 
> 
> all the way down to the text
> 
> 
> 
> 
> +IGH_ETHERCAT_POST_INSTALL_TARGET_HOOKS += IGH_ETHERCAT_INSTALL_MODULES
> +
> +$(eval $(call AUTOTARGETS))
> \ No newline at end of file
> --
> 1.7.4.1
> 
> 
> 
> and place it in a text file called filename.patch.  Then I go into the
> directory buildroot-2011.11 and I run 'cat filename.patch | patch -p1'
> 
> Then I can just run 'make menuconfig', 'make', etc?
> 
> Thanks so much.
> 
> _______________________________________________
> buildroot mailing list
> buildroot at busybox.net
> http://lists.busybox.net/mailman/listinfo/buildroot
> 
> 

[Buildroot] LIBFOO_SITE in .mk file

[Jeff Krasky]

> You don't even have to do that, just save the e-mail as a plain text
> file using your e-mail client.
> 
> Patch will ignore the leading, non-patch, discussion.
> 
> It was designed that way, back in the days when the only way to
> pass patches was by e-mail.  ;-)

Ok, I saved the entire email in notepad as an ANSI text document, then
renamed it to have the .patch extension.

I went to run 'make linux-menuconfig' and I got this message:

package/igh-ethercat/igh-ethercat.mk:11: *** missing separator.  Stop.

When using Buildroot-2011.05 I needed to run 'make linux-menuconfig'.  Here
is the .mk file:

IGH_ETHERCAT_VERSION = stable-1.5
IGH_ETHERCAT_SITE    =
http://etherlabmaster.hg.sourceforge.net/hgweb/etherlabmaster/etherlabmaster
/archive/
IGH_ETHERCAT_SOURCE  = $(IGH_ETHERCAT_VERSION).tar.bz2 
IGH_ETHERCAT_AUTORECONF = YES IGH_ETHERCAT_DEPENDENCIES = linux 
IGH_ETHERCAT_INSTALL_STAGING = YES

IGH_ETHERCAT_CONF_OPT = \
	--with-linux-dir=$(LINUX_DIR)

IGH_ETHERCAT_CONF_OPT += $(if 
$(BR2_PACKAGE_IGH_ETHERCAT_8139TOO),--enable-8139too,--disable-8139too)
IGH_ETHERCAT_CONF_OPT += $(if 
$(BR2_PACKAGE_IGH_ETHERCAT_E100),--enable-e100,--disable-e100)
IGH_ETHERCAT_CONF_OPT += $(if 
$(BR2_PACKAGE_IGH_ETHERCAT_E1000),--enable-e1000,--disable-e1000)
IGH_ETHERCAT_CONF_OPT += $(if 
$(BR2_PACKAGE_IGH_ETHERCAT_R8169),--enable-r8169,--disable-r8169)

define IGH_ETHERCAT_CREATE_CHANGELOG
	touch $(@D)/ChangeLog
endef

IGH_ETHERCAT_POST_PATCH_HOOKS += IGH_ETHERCAT_CREATE_CHANGELOG

define IGH_ETHERCAT_BUILD_MODULES
	$(MAKE) -C $(@D) $(LINUX_MAKE_FLAGS) modules endef

IGH_ETHERCAT_POST_BUILD_HOOKS += IGH_ETHERCAT_BUILD_MODULES

define IGH_ETHERCAT_INSTALL_MODULES
	$(MAKE) -C $(@D) $(LINUX_MAKE_FLAGS) modules_install endef

IGH_ETHERCAT_POST_INSTALL_TARGET_HOOKS += IGH_ETHERCAT_INSTALL_MODULES



Line 11 appears to be
"$(BR2_PACKAGE_IGH_ETHERCAT_8139TOO),--enable-8139too,--disable-8139too)".
Does this need to be tabbed in or something?  The build takes about 30
minutes and so I'd like to correct in advance if I could.

Thanks again.

[Buildroot] LIBFOO_SITE in .mk file

[Jeff Krasky]

> Ok, I saved the entire email in notepad as an ANSI text document, then
> renamed it to have the .patch extension.

Now I even re-newlined each line to make sure it wasn't a windows newline
versus linux newline issue.

> I went to run 'make linux-menuconfig' and I got this message:
> 
> package/igh-ethercat/igh-ethercat.mk:11: *** missing separator.  Stop.

I have tried running just 'make' and that gives the same error.

Please help.

[Buildroot] LIBFOO_SITE in .mk file

[Jeff Krasky]

> > I went to run 'make linux-menuconfig' and I got this message:
> > 
> > package/igh-ethercat/igh-ethercat.mk:11: *** missing separator.  Stop.
>
> I have tried running just 'make' and that gives the same error.

Looks like those lines had to be on the same line as the 'if' that proceeded
them:

IGH_ETHERCAT_CONF_OPT += $(if
$(BR2_PACKAGE_IGH_ETHERCAT_8139TOO),--enable-8139too,--disable-8139too)

all one line.

Also, the endef's appearing later on had to be placed on their own line, at
the beginning.

[Buildroot] LIBFOO_SITE in .mk file

[Michael S. Zick]

On Thu December 1 2011, Jeff Krasky wrote:
> > > I went to run 'make linux-menuconfig' and I got this message:
> > > 
> > > package/igh-ethercat/igh-ethercat.mk:11: *** missing separator.  Stop.
> >
> > I have tried running just 'make' and that gives the same error.
> 
> Looks like those lines had to be on the same line as the 'if' that proceeded
> them:
> 
> IGH_ETHERCAT_CONF_OPT += $(if
> $(BR2_PACKAGE_IGH_ETHERCAT_8139TOO),--enable-8139too,--disable-8139too)
> 
> all one line.
> 
> Also, the endef's appearing later on had to be placed on their own line, at
> the beginning.
>
Ah, the joys of:
X-Mailer: Microsoft Office Outlook 11
In-Reply-To: <001101ccb044$956d1340$0400a8c0@dspcgrnzks9p98>

Have you ever considered running Linux on that box?
Or at least a sane e-mail client?

Mike 
> _______________________________________________
> buildroot mailing list
> buildroot at busybox.net
> http://lists.busybox.net/mailman/listinfo/buildroot
> 
> 

[Buildroot] LIBFOO_SITE in .mk file

[Jeff Krasky]

> Looks like those lines had to be on the same line as the 'if' that
> proceeded them:
> 
> IGH_ETHERCAT_CONF_OPT += $(if
$(BR2_PACKAGE_IGH_ETHERCAT_8139TOO),--enable-8139too,--disable-8139too)
> 
> all one line.
> 
> Also, the endef's appearing later on had to be placed on their own line, 
> at the beginning.

Ok, it built the image and filesystem now.  I put them in Nandflash where
the previous kernel image and jffs2 filesystem would go (0xA0000 and
0x400000 respectively).

Now I see this error:

Freeing init memory: 124K
JFFS2 error: (1) jffs2_do_read_inode_internal: Argh. Special inode #95 with
mode
 0x21b6 had more than one node
iget() failed for ino #95
Warning: unable to open an initial console.


Any ideas?  Is something different with 2011.11?  I am using the same
bootstrap and uboot that I have been using.

Two paths:

1) Can I just look at the new igh-ethercat.mk file and put that stuff in the
EtherCAT.mk file I was working on yesterday?  I'd like to see something
work.

2) Anyone have a clue about this latest error?


Thanks.

[Buildroot] LIBFOO_SITE in .mk file

[Jeff Krasky]

The following is a message from yesterday.  It was the closest I have gotten
to both ethercat application and kernel modules being built.  As can be seen
from the last email I sent today, I can't build either one with the new .mk
file.  Something in that file only applies to buildroot-2011.11?

Can someone help out with what I had going yesterday?  The suggestions from
Thomas De Schampheleire helped to build the kernel modules, but in so doing
the application was no longer being built.  I will show both .mk files.  It
would get REALLY appreciated if someone would merge these for me somehow.

.mk file for building the application:


#############################################################
#
# EtherCAT
#
#############################################################
ETHERCAT_VERSION = 1.0
ETHERCAT_SOURCE = etherlabmaster-1.0.tar.bz2
###ETHERCAT_SITE = file:///home/user1/Desktop/EtherCAT4Atmel/
ETHERCAT_SITE = http://127.0.0.1/home/user1/Desktop/EtherCAT4Atmel/
ETHERCAT_AUTORECONF = YES
ETHERCAT_INSTALL_STAGING = YES

ETHERCAT_INSTALL_TARGET = YES
ETHERCAT_CONF_OPT = --enable-generic

$(eval $(call AUTOTARGETS,package,ethercat))




And here is the .mk for building the kernel module:

#############################################################
#
# EtherCAT
#
#############################################################
ETHERCAT_VERSION = 1.0
ETHERCAT_SOURCE = etherlabmaster-1.0.tar.bz2
###ETHERCAT_SITE = file:///home/user1/Desktop/EtherCAT4Atmel/
ETHERCAT_SITE = http://127.0.0.1/home/user1/Desktop/EtherCAT4Atmel/
ETHERCAT_AUTORECONF = YES
ETHERCAT_INSTALL_STAGING = YES

ETHERCAT_INSTALL_TARGET = YES
ETHERCAT_CONF_OPT = --enable-generic

###These next two lines were a part of building the kernel module.
ETHERCAT_DEPENDENCIES = linux
ETHERCAT_INSTALL_MOD_DIR = ethercat

# From here to the GENTARGETS line is part of building the kernel module.
define ETHERCAT_BUILD_CMDS
       $(MAKE) $(LINUX26_MAKE_FLAGS) -C $(LINUX26_DIR) M=$(@D) modules
endef

define ETHERCAT_INSTALL_TARGET_CMDS
       $(MAKE) $(LINUX26_MAKE_FLAGS) -C $(LINUX26_DIR) M=$(@D) \
               INSTALL_MOD_STRIP=1 \
               INSTALL_MOD_DIR=$(ETHERCAT_INSTALL_MOD_DIR) \
               modules_install
endef

define ETHERCAT_UNINSTALL_TARGET_CMDS
       rm -rf
$(TARGET_DIR)/lib/modules/$(LINUX26_VERSION_PROBED)/$(ETHERCAT_INSTALL_MOD_D
IR)
endef

$(eval $(call GENTARGETS,package,ethercat))











> You do realize that these rules are purely intended for the module, right.

No, I thought it would do both.  I have a second build running with both
$evals in it.  Hopefully that will do it.

> Based on what ThomasP says, things may be a little more complicated
> than I originally thought.

Do I need to separate sets of build rules in the .mk?

> Just checking: you are looking into output/target/lib/modules or
> /lib/modules on the target right, not /lib/modules on the host?

Correct, I am looking at /lib/modules when running the Linux image on the
target.

> If the modules_install step in your .mk file goes well, this should
> put the file there.

It did.  I just wrote a post about how it's there but the application isn't.

Thanks again.

[Buildroot] LIBFOO_SITE in .mk file

[Thomas De Schampheleire]

Jeff,

On Thu, Dec 1, 2011 at 11:56 PM, Jeff Krasky <jeff.krasky@dspcg.com> wrote:
> The following is a message from yesterday. ?It was the closest I have gotten
> to both ethercat application and kernel modules being built. ?As can be seen
> from the last email I sent today, I can't build either one with the new .mk
> file. ?Something in that file only applies to buildroot-2011.11?
>
> Can someone help out with what I had going yesterday? ?The suggestions from
> Thomas De Schampheleire helped to build the kernel modules, but in so doing
> the application was no longer being built. ?I will show both .mk files. ?It
> would get REALLY appreciated if someone would merge these for me somehow.
>

I have the impression you are missing some basic skills, and I really
think it will help you if you took the time to learn about them. For
example:
* how does make work and how are Makefiles organized
* what is a patch, how do you apply a patch

There is some information on patching here:
http://www.linuxchix.org/content/courses/kernel_hacking/lesson9
but take into account that some of it is specific to the linux kernel.


Regarding ethercat: what Thomas Petazzoni has provided you with is a
patch for the entire thing, application + drivers. It is based on a
clean 2011.11 release. Here is a step-by-step guide of things I think
you should do:

1. Start from a clean 2011.11 buildroot tarball (unpack it)
2. Move to the buildroot directory with cd
3. Apply the patch of ThomasP by saving the e-mail in text file format) and run:
patch -p1 < saved-email
You should see output like:
patching file package/Config.in
patching file package/igh-ethercat/Config.in
patching file package/igh-ethercat/igh-ethercat.mk

4. Configure buildroot properly (using make menuconfig) and don't
forget to enable the new igh-ethercat package, which is present under
'Networking applications'

5. Run 'make'.

That should be it. Buildroot will do all the rest for you and when the
build is done, you can test ethercat to see if everything works.

Best regards,
Thomas

[Buildroot] LIBFOO_SITE in .mk file

[Thomas Petazzoni]

Hello,

Le Fri, 2 Dec 2011 08:50:55 +0100,
Thomas De Schampheleire <patrickdepinguin+buildroot@gmail.com> a ?crit :

> I have the impression you are missing some basic skills, and I really
> think it will help you if you took the time to learn about them.

I am sorry to say that I agree here. You really should learn about the
basics before trying to build an embedded Linux system with a piece of
software as complicated to configure as EtherCAT. I don't think it is
possible to work on EtherCAT without having knowledge of makefiles,
build process, patching, etc.

Regards,

Thomas
-- 
Thomas Petazzoni, Free Electrons
Kernel, drivers, real-time and embedded Linux
development, consulting, training and support.
http://free-electrons.com

[Buildroot] LIBFOO_SITE in .mk file

[Michael S. Zick]

On Fri December 2 2011, Thomas De Schampheleire wrote:
> Jeff,
> 
> On Thu, Dec 1, 2011 at 11:56 PM, Jeff Krasky <jeff.krasky@dspcg.com> wrote:
> > The following is a message from yesterday. ?It was the closest I have gotten
> > to both ethercat application and kernel modules being built. ?As can be seen
> > from the last email I sent today, I can't build either one with the new .mk
> > file. ?Something in that file only applies to buildroot-2011.11?
> >
> > Can someone help out with what I had going yesterday? ?The suggestions from
> > Thomas De Schampheleire helped to build the kernel modules, but in so doing
> > the application was no longer being built. ?I will show both .mk files. ?It
> > would get REALLY appreciated if someone would merge these for me somehow.
> >
> 
> I have the impression you are missing some basic skills, and I really
> think it will help you if you took the time to learn about them. For
> example:
> * how does make work and how are Makefiles organized
> * what is a patch, how do you apply a patch
> 
> There is some information on patching here:
> http://www.linuxchix.org/content/courses/kernel_hacking/lesson9
> but take into account that some of it is specific to the linux kernel.
> 
> 
> Regarding ethercat: what Thomas Petazzoni has provided you with is a
> patch for the entire thing, application + drivers. It is based on a
> clean 2011.11 release. Here is a step-by-step guide of things I think
> you should do:
> 
> 1. Start from a clean 2011.11 buildroot tarball (unpack it)
> 2. Move to the buildroot directory with cd
> 3. Apply the patch of ThomasP by saving the e-mail in text file format) and run:
> patch -p1 < saved-email
> You should see output like:
> patching file package/Config.in
> patching file package/igh-ethercat/Config.in
> patching file package/igh-ethercat/igh-ethercat.mk
> 
> 4. Configure buildroot properly (using make menuconfig) and don't
> forget to enable the new igh-ethercat package, which is present under
> 'Networking applications'
>

4.5 Ensure that you have a kernel version selected that is supported
by the EtherCAT patches. 
I seem to recall the most recent supported kernel in stable-1.5 was
2.6.32, __not__ 2.6.33 which was mentioned in your posts earlier.

Mike 
> 5. Run 'make'.
> 
> That should be it. Buildroot will do all the rest for you and when the
> build is done, you can test ethercat to see if everything works.
> 
> Best regards,
> Thomas
> _______________________________________________
> buildroot mailing list
> buildroot at busybox.net
> http://lists.busybox.net/mailman/listinfo/buildroot
> 
> 

[Buildroot] LIBFOO_SITE in .mk file

[Jeff Krasky]

Ok, ok, I'm tryin again from the 2011.11 version.  Can someone at least say
if there is a way to just clean the ethercat part?  Like a

'make clean-igh-ethercat' or something?

[Buildroot] LIBFOO_SITE in .mk file

[Quotient Remainder]

On 2 Noll 2011 15:58, "Jeff Krasky" <jeff.krasky@dspcg.com> wrote:
>
> Ok, ok, I'm tryin again from the 2011.11 version.  Can someone at least
say
> if there is a way to just clean the ethercat part?  Like a
>
> 'make clean-igh-ethercat' or something?
>
>
>

make igh-ethercat-clean to run the clean commands for the package. (in the
..._CLEAN_CMDS variable in the .mk file)

make igh-ethercat-dirclean to start from blank build directory again.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.busybox.net/pipermail/buildroot/attachments/20111202/48a3b675/attachment.html>

[Buildroot] Kernel panic - not syncing: Attempted to kill init!

[Jeff Krasky]

I'm giving things a try with 2011.11 and when booting the board I see this:

 

VFS: Mounted root (jffs2 filesystem) on device 31:1.

Freeing init memory: 124K

Kernel panic - not syncing: Attempted to kill init!

 

If anyone has a hint it would be appreciated.  I think I've done the steps
correctly, so if I can just get it to boot..

 

 

Thanks.

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.busybox.net/pipermail/buildroot/attachments/20111202/27c2a065/attachment.html>

[Buildroot] Kernel panic - not syncing: Attempted to kill init!

[Jeff Krasky]

> I'm giving things a try with 2011.11 and when booting the board I see
this:

> 

> VFS: Mounted root (jffs2 filesystem) on device 31:1.

> Freeing init memory: 124K

> Kernel panic - not syncing: Attempted to kill init!

> 

 

If it helps any, I get this same error message whether I put the jffs2
filesystem at 0x400000 (four hundred thousand) or 0x2a0000 in Nandflash.  So
maybe the kernel doesn't know where to look?  When building kernel 2.6.33 in
2011.05 I would put kernel at 0xa0000 and filesystem at 0x400000 (four
hundred thousand).

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.busybox.net/pipermail/buildroot/attachments/20111202/8263d55f/attachment.html>

[Buildroot] Kernel panic - not syncing: Attempted to kill init!

[Tim Judd]

"Attempted to kill init" means the init process died.? Kernel must panic because init cannot terminate.
There's something preventing init from running, or it is dying during execution.

Is your userland correctly have an init?? normally /sbin/init is called.? If your init is somewhere else, use the init= paramater to kernel append

you may have a mounted jffs2 filesystem but not directly usable (think squashfs), is it loaded to ramdisk?? if so, root=/dev/ram0 init=/sbin/init

I'm trying to rework an OEM userland and have come up on many attempts to kill init.? It sucks for diagnostics.




?
If opportunity doesn't knock, build a door.
"I can" is a way of life.
More and Bigger is not always Better.
The road to success is always uphill.
Life isn't about finding yourself, it's about creating yourself.


________________________________
 From: Jeff Krasky <jeff.krasky@dspcg.com>
To: buildroot at busybox.net 
Sent: Friday, December 2, 2011 9:51 AM
Subject: [Buildroot] Kernel panic - not syncing: Attempted to kill init!
 

 
I?m giving things a try with 2011.11
and when booting the board I see this:
?
VFS: Mounted root (jffs2 filesystem) on
device 31:1.
Freeing init memory: 124K
Kernel panic - not syncing: Attempted to
kill init!
?
If anyone has a hint it would be
appreciated.? I think I?ve done the steps correctly, so if I can just get
it to boot?.
?
?
Thanks.
_______________________________________________
buildroot mailing list
buildroot at busybox.net
http://lists.busybox.net/mailman/listinfo/buildroot
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.busybox.net/pipermail/buildroot/attachments/20111202/329157bd/attachment.html>

[Buildroot] Kernel panic - not syncing: Attempted to kill init!

[Jeff Krasky]

> "Attempted to kill init" means the init process died.  Kernel must panic
because init cannot terminate.

> There's something preventing init from running, or it is dying during
execution.

> 

> Is your userland correctly have an init?  normally /sbin/init is called.
If your init is somewhere else, use

> the init= paramater to kernel append

 

How do I know where init is if I can't bring the board up and run any linux
commands?

 

>you may have a mounted jffs2 filesystem but not directly usable (think
squashfs), is it loaded to

> ramdisk?  if so, root=/dev/ram0 init=/sbin/init

 

When I successfully ran 2.6.33 built from Buildroot-2011.05, in U-boot the
setting for root was root=/dev/mtdblock1.  I am not sure what squashfs is.
How can the filesystem not be usable?

 

Thanks.

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.busybox.net/pipermail/buildroot/attachments/20111202/62bb2304/attachment-0001.html>

[Buildroot] Kernel panic - not syncing: Attempted to kill init!

[Aleksander Dutkowski]

afair this message could be because of lack of dev/console and/or dev/null.
You must run "mknod -m 660 dev/console c 5 1" and "mknod -m
660dev/null c 1 3" in root partition.

How to do it? I have my on usb stick, so i
simply mount usb drive, and do some changes. If youre using yaffs2 and
NAND memory, you can use

System configuration --> Custom script to run before creating filesystem images

to invoke those commands or just find unyaffs script (somewhere on the
Internet, on google code
afair), unyaffs your rootfs, make changes and make yaffs image one
more time.
You have output/target folder with your rootfs, so you can make changes there :)


On Fri, Dec 2, 2011 at 7:02 PM, Jeff Krasky <jeff.krasky@dspcg.com> wrote:
>> "Attempted to kill init" means the init process died.? Kernel must panic
>> because init cannot terminate.
>
>> There's something preventing init from running, or it is dying during
>> execution.
>
>>
>
>> Is your userland correctly have an init?? normally /sbin/init is called.
>> If your init is somewhere else, use
>
>> the init= paramater to kernel append
>
>
>
> How do I know where init is if I can?t bring the board up and run any linux
> commands?
>
>
>
>>you may have a mounted jffs2 filesystem but not directly usable (think
>> squashfs), is it loaded to
>
>> ramdisk?? if so, root=/dev/ram0 init=/sbin/init
>
>
>
> When I successfully ran 2.6.33 built from Buildroot-2011.05, in U-boot the
> setting for root was root=/dev/mtdblock1. ?I am not sure what squashfs is.
> How can the filesystem not be usable?
>
>
>
> Thanks.
>
>
> _______________________________________________
> buildroot mailing list
> buildroot at busybox.net
> http://lists.busybox.net/mailman/listinfo/buildroot



-- 
pozdrawiam
Aleksander Dutkowski

[Buildroot] Kernel panic - not syncing: Attempted to kill init!

[Tim Judd]

"usable" on the filesystem is maybe not the perfect word to use, it is more meant in terms of what is read-write and freely accessible on the root device.? A mtdblock is an embedded flash memory area who often times has a read-only compressed filesystem such as squashfs (google it).? squashfs is not directly usable like ext2 on a hard disk is.

You might have to hack 'n slash your way in (this is my limited linux knowledge.? I'm not claiming to be an expert).
use an additional append= argument: "init=/bin/sh" or "init=/bin/bash" or "init=/bin/ash" to try to get into a usable shell to see what's there.? This argument overrides the traditional init that runs so you can look in the directories trying to find where your init is supposed to be.? Remember that init itself is never supposed to terminate/end like you can close Firefox.? So if you leave the shell, linux kernel WILL panic (same error you started this thread with)

I'm not very confident mtdblock1 is supposed to be a read-write filesystem, but may have some kind of read-only squashfs kind of filesystem on it.


you're still welcome to try "root=/dev/ram0 init=/bin/ash" and related combinations trying to get a usable root filesystem and working shell.? This is where the headaches for me start.


I'm just knowledgeable enough to know basics, others here are guaranteed to be more experienced.





?
If opportunity doesn't knock, build a door.
"I can" is a way of life.
More and Bigger is not always Better.
The road to success is always uphill.
Life isn't about finding yourself, it's about creating yourself.


________________________________
 From: Jeff Krasky <jeff.krasky@dspcg.com>
To: buildroot at busybox.net 
Sent: Friday, December 2, 2011 11:02 AM
Subject: Re: [Buildroot] Kernel panic - not syncing: Attempted to kill init!
 

 
> "Attempted
to kill init" means the init process died.? Kernel must panic because
init cannot terminate.
> There's
something preventing init from running, or it is dying during execution.
> 
> Is
your userland correctly have an init?? normally /sbin/init is
called.? If your init is somewhere else, use
> the
init= paramater to kernel append
?
How do I know
where init is if I can?t bring the board up and run any linux commands?
?
>you
may have a mounted jffs2 filesystem but not directly usable (think squashfs),
is it loaded to
> ramdisk??
if so, root=/dev/ram0 init=/sbin/init
?
When I
successfully ran 2.6.33 built from Buildroot-2011.05, in U-boot the setting for
root was root=/dev/mtdblock1. ?I am not sure what squashfs
is.? How can the filesystem not be usable?
?
Thanks.
_______________________________________________
buildroot mailing list
buildroot at busybox.net
http://lists.busybox.net/mailman/listinfo/buildroot
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.busybox.net/pipermail/buildroot/attachments/20111202/c9e38a49/attachment.html>

[Buildroot] Kernel panic - not syncing: Attempted to kill init!

[Jeff Krasky]

> afair this message could be because of lack of dev/console and/or
dev/null.
> You must run "mknod -m 660 dev/console c 5 1" and "mknod -m
> 660dev/null c 1 3" in root partition.
>
> How to do it? I have my on usb stick, so i
> simply mount usb drive, and do some changes. If youre using yaffs2 and
> NAND memory, you can use
> 
> System configuration --> Custom script to run before creating filesystem
> images
> 
> to invoke those commands or just find unyaffs script (somewhere on the
> Internet, on google code
> afair), unyaffs your rootfs, make changes and make yaffs image one
> more time.

Ok, I ran 'make menuconfig' and went into "System configuration -> Custom
script to run before creating filesystem images" and I entered this string:

mknod -m 660 dev/console c 5 1; mknod -m 660 dev/null c 1 3

and I got this when running 'make':

echo 2011.11 > /home/user1/buildroot-2011.11/output/target/etc/br-version
"mknod -m 660 dev/console c 5 1; mknod -m 660 dev/null c 1 3"
/home/user1/buildroot-2011.11/output/target
/bin/sh: mknod -m 660 dev/console c 5 1; mknod -m 660 dev/null c 1 3: No
such file or directory
make: *** [target-finalize] Error 127
user1 at dhcppc15:~/buildroot-2011.11>


> You have output/target folder with your rootfs, so you can make changes
> there :)

How to do it to a already built .jffs2 file?  Can I just open a terminal
window on the linux PC that buildroot is on and run some commands?

[Buildroot] Kernel panic - not syncing: Attempted to kill init!

[Tim Judd]

lack of dev/console or /dev/null should give a message something like "can't allocate initial tty!"
I was having those errors too.? his is unable to run/survive init


having dev/console and dev/null (and if necessary dev/tty*) so the console can be opened.? It's odd the kernel can spit the bootup on ttyS0 but complains when trying to run init.? one of the great mysteries of linux.



?

If opportunity doesn't knock, build a door.
"I can" is a way of life.
More and Bigger is not always Better.
The road to success is always uphill.
Life isn't about finding yourself, it's about creating yourself.


________________________________
 From: Aleksander Dutkowski <adutkowski@gmail.com>
To: buildroot at busybox.net 
Sent: Friday, December 2, 2011 11:21 AM
Subject: Re: [Buildroot] Kernel panic - not syncing: Attempted to kill init!
 
afair this message could be because of lack of dev/console and/or dev/null.
You must run "mknod -m 660 dev/console c 5 1" and "mknod -m
660dev/null c 1 3" in root partition.

How to do it? I have my on usb stick, so i
simply mount usb drive, and do some changes. If youre using yaffs2 and
NAND memory, you can use

System configuration --> Custom script to run before creating filesystem images

to invoke those commands or just find unyaffs script (somewhere on the
Internet, on google code
afair), unyaffs your rootfs, make changes and make yaffs image one
more time.
You have output/target folder with your rootfs, so you can make changes there :)


On Fri, Dec 2, 2011 at 7:02 PM, Jeff Krasky <jeff.krasky@dspcg.com> wrote:
>> "Attempted to kill init" means the init process died.? Kernel must panic
>> because init cannot terminate.
>
>> There's something preventing init from running, or it is dying during
>> execution.
>
>>
>
>> Is your userland correctly have an init?? normally /sbin/init is called.
>> If your init is somewhere else, use
>
>> the init= paramater to kernel append
>
>
>
> How do I know where init is if I can?t bring the board up and run any linux
> commands?
>
>
>
>>you may have a mounted jffs2 filesystem but not directly usable (think
>> squashfs), is it loaded to
>
>> ramdisk?? if so, root=/dev/ram0 init=/sbin/init
>
>
>
> When I successfully ran 2.6.33 built from Buildroot-2011.05, in U-boot the
> setting for root was root=/dev/mtdblock1. ?I am not sure what squashfs is.
> How can the filesystem not be usable?
>
>
>
> Thanks.
>
>
> _______________________________________________
> buildroot mailing list
> buildroot at busybox.net
> http://lists.busybox.net/mailman/listinfo/buildroot



-- 
pozdrawiam
Aleksander Dutkowski
_______________________________________________
buildroot mailing list
buildroot at busybox.net
http://lists.busybox.net/mailman/listinfo/buildroot
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.busybox.net/pipermail/buildroot/attachments/20111202/b8eaca41/attachment.html>

[Buildroot] Kernel panic - not syncing: Attempted to kill init!

[Aleksander Dutkowski]

On Fri, Dec 2, 2011 at 7:30 PM, Jeff Krasky <jeff.krasky@dspcg.com> wrote:
>> afair this message could be because of lack of dev/console and/or
> dev/null.
>> You must run "mknod -m 660 dev/console c 5 1" and "mknod -m
>> 660dev/null c 1 3" in root partition.
>>
>> How to do it? I have my on usb stick, so i
>> simply mount usb drive, and do some changes. If youre using yaffs2 and
>> NAND memory, you can use
>>
>> System configuration --> Custom script to run before creating filesystem
>> images
>>
>> to invoke those commands or just find unyaffs script (somewhere on the
>> Internet, on google code
>> afair), unyaffs your rootfs, make changes and make yaffs image one
>> more time.
>
> Ok, I ran 'make menuconfig' and went into "System configuration -> Custom
> script to run before creating filesystem images" and I entered this string:
>
> mknod -m 660 dev/console c 5 1; mknod -m 660 dev/null c 1 3
>
> and I got this when running 'make':
>
> echo 2011.11 > /home/user1/buildroot-2011.11/output/target/etc/br-version
> "mknod -m 660 dev/console c 5 1; mknod -m 660 dev/null c 1 3"
> /home/user1/buildroot-2011.11/output/target
> /bin/sh: mknod -m 660 dev/console c 5 1; mknod -m 660 dev/null c 1 3: No
> such file or directory
> make: *** [target-finalize] Error 127
> user1 at dhcppc15:~/buildroot-2011.11>

"System configuration --> Custom script to run before creating
filesystem image" input is for specify a PATH to bash script.
You you must simply create script, fill with mknod commands (remember
- mknod can be ran only by root! so do 'sudo mknod')
and then put the path to the script in "Custom script to run before
creating filesystem image".
Remeber to add execute mode to the script (chmod +x file)
>
>
>> You have output/target folder with your rootfs, so you can make changes
>> there :)
>
> How to do it to a already built .jffs2 file? ?Can I just open a terminal
> window on the linux PC that buildroot is on and run some commands?

When buildroot is done compiling all packages, it installs all in
'output/target' directory, and then create image.
I am used to tell buildroot not to create any images, make my own
adjustements in output/target dir, and then create image by hand.
If you need to create yaffs, you can look for "mkyaffs2" script, which
you can find on code.google.com.

On Fri, Dec 2, 2011 at 7:31 PM, Tim Judd <tjudd2k@yahoo.com> wrote:
> lack of dev/console or /dev/null should give a message something like "can't
> allocate initial tty!"
> I was having those errors too.? his is unable to run/survive init
>
>
> having dev/console and dev/null (and if necessary dev/tty*) so the console
> can be opened.? It's odd the kernel can spit the bootup on ttyS0 but
> complains when trying to run init.? one of the great mysteries of linux.

I had another error message (I experienced problem with lack of
/dev/console some time ago), but you can be right. I'm just assuming,
that my solution is proper, and want Jeff to check it ;)

>
> _______________________________________________
> buildroot mailing list
> buildroot at busybox.net
> http://lists.busybox.net/mailman/listinfo/buildroot



-- 
pozdrawiam
Aleksander Dutkowski

[Buildroot] Device files and Buildroot

[Thomas Petazzoni]

Le Sat, 3 Dec 2011 01:25:41 +0100,
Aleksander Dutkowski <adutkowski@gmail.com> a ?crit :

> "System configuration --> Custom script to run before creating
> filesystem image" input is for specify a PATH to bash script.
> You you must simply create script, fill with mknod commands (remember
> - mknod can be ran only by root! so do 'sudo mknod')

No, no, no and no ! I'm sorry but this is totally horrible and is
definitely not the way of doing things with Buildroot.

With Buildroot, you have four ways of managing the device files
in /dev. The mechanism used to manage device files is configured from
System configuration -> /dev management. The four ways are :

 * Static using device table. In this case the "System configuration ->
   Path to the device tables" option gives a space-separated list of
   files, each of which containing a list of devices to create at build
   time in the root filesystem. By default, this list is defined to
   just the target/generic/device_table_dev.txt, which creates some
   basic device files. Those device files are created at *build* time
   and are statically present in the root filesystem image generated by
   Buildroot. All basic devices such as /dev/console, /dev/null and al.
   are already present in the default device table. If you are in this
   mode and want to add more device files, then you should add them to
   target/generic/device_table_dev.txt, or better, create your own
   additional device table in
   board/<yourcompany>/<yourproject>/device_table.txt, and add it to
   the space-separated list in "System configuration -> Path to the
   device tables".

 * Dynamic using devtmpfs only. Devtmpfs is a virtual filesystem
   implemented in the Linux kernel that can be mounted in /dev. The
   kernel will automatically create/remove device files from this
   filesystem as devices appear/disappear from the system. devtmpfs
   exists in the Linux kernel since 2.6.32. When this option is
   selected *and* Buildroot is responsible for building the kernel,
   then Buildroot ensures that the kernel is built with the appropriate
   options to make devtmpfs work. When Buildroot is *not* responsible
   for building the kernel (the user does it on its own), then the user
   is responsible for making sure that CONFIG_DEVTMPFS and
   CONFIG_DEVTMPFS_MOUNT are both enabled in the kernel configuration.
   When this mode is used, no static device files are created in the
   root filesystem: the device files are automatically created at boot
   time by the kernel.

 * Dynamic using mdev. This is exactly like with 'devtmpfs' (i.e,
   devtmpfs is required for this mode to work), but Buildroot adds the
   mdev utility into the mix. mdev is an utility bundled with Busybox
   which gets executed when the kernel notifies that a device has been
   added or removed from the system. Compared to a pure 'devtmpfs'
   solution, it allows to execute arbitrary applications or shell
   scripts when devices appear/disappear. mdev behaviour can be
   configured from /etc/mdev.conf, refer to the Busybox documentation
   for more details. Since this case relies on devtmpfs, there are no
   static device files created in the root filesystem, and no device
   table is used.

 * Dynamic using udev. This is also exactly like with 'devtmpfs' (i.e,
   devtmpfs is required for this mode to work), but Buildroot adds the
   udev daemon into the mix. udev is the "device event manager" used in
   all Linux desktop and server systems and can be seen as a
   "full-featured" mdev. It is more configurable, provides a library
   called libudev to allow applications to query for which devices are
   available, etc.

So, no, do *NOT* ever use the post-build script to create device files.
Just understand how those device files work and the different mode
Buildroot provides to handle them.

If you need precisions or have questions on this topic, don't hesitate
to ask. We have many people on the list ready to answer your questions.

Best regards,

Thomas
-- 
Thomas Petazzoni, Free Electrons
Kernel, drivers, real-time and embedded Linux
development, consulting, training and support.
http://free-electrons.com

[Buildroot] Device files and Buildroot

[Trevor Woerner]

Hi Thomas,

Excellent write-up. Does buildroot's /dev handling also tweak
/etc/fstab? I would assume /dev would only need to be mounted in the
non-static cases.

Best regards,
    Trevor

[Buildroot] Device files and Buildroot

[Thomas Petazzoni]

Le Sat, 3 Dec 2011 07:53:51 -0500,
Trevor Woerner <twoerner@gmail.com> a ?crit :

> Excellent write-up. Does buildroot's /dev handling also tweak
> /etc/fstab? I would assume /dev would only need to be mounted in the
> non-static cases.

Two cases here :

 * Static /dev. In this case /dev is part of the root filesystem
   itself, so there is no need for anything special in /etc/fstab.

 * Dynamic /dev (either with devtmpfs only, or with devtmpfs+mdev or
   devtmpfs+udev). In this case, we use the CONFIG_DEVTMPFS_MOUNT
   kernel option, which tells the kernel to automatically mount the
   devtmpfs filesystem at boot time. So the root filesystem simply
   contains a /dev directory, completely empty. Nothing is needed
   in /etc/fstab: the kernel will mount devtmpfs in /dev automatically.
   The only special case here is when the root filesystem is an
   initramfs: for some reason, the kernel does not mount devtmpfs
   automatically in this case, but Buildroot installs a special /init
   wrapper script, which mounts devtmpfs before starting the real init.
   But even in this case, nothing special is needed in /etc/fstab.

Does that answer your question?

Thomas
-- 
Thomas Petazzoni, Free Electrons
Kernel, drivers, real-time and embedded Linux
development, consulting, training and support.
http://free-electrons.com

[Buildroot] Device files and Buildroot

[Aleksander Dutkowski]

I've read your message, and I now know, what was the problem in my
case - I'am using kernel 2.6.30 with udev and it doesn't have
devtmpfs, so it is not creating dev/null and /dev/console. Thanks!

Best Regards,
Aleksander

On Sat, Dec 3, 2011 at 9:25 AM, Thomas Petazzoni
<thomas.petazzoni@free-electrons.com> wrote:
> Le Sat, 3 Dec 2011 01:25:41 +0100,
> Aleksander Dutkowski <adutkowski@gmail.com> a ?crit :
>
>> "System configuration --> Custom script to run before creating
>> filesystem image" input is for specify a PATH to bash script.
>> You you must simply create script, fill with mknod commands (remember
>> - mknod can be ran only by root! so do 'sudo mknod')
>
> No, no, no and no ! I'm sorry but this is totally horrible and is
> definitely not the way of doing things with Buildroot.
>
> With Buildroot, you have four ways of managing the device files
> in /dev. The mechanism used to manage device files is configured from
> System configuration -> /dev management. The four ways are :
>
> ?* Static using device table. In this case the "System configuration ->
> ? Path to the device tables" option gives a space-separated list of
> ? files, each of which containing a list of devices to create at build
> ? time in the root filesystem. By default, this list is defined to
> ? just the target/generic/device_table_dev.txt, which creates some
> ? basic device files. Those device files are created at *build* time
> ? and are statically present in the root filesystem image generated by
> ? Buildroot. All basic devices such as /dev/console, /dev/null and al.
> ? are already present in the default device table. If you are in this
> ? mode and want to add more device files, then you should add them to
> ? target/generic/device_table_dev.txt, or better, create your own
> ? additional device table in
> ? board/<yourcompany>/<yourproject>/device_table.txt, and add it to
> ? the space-separated list in "System configuration -> Path to the
> ? device tables".
>
> ?* Dynamic using devtmpfs only. Devtmpfs is a virtual filesystem
> ? implemented in the Linux kernel that can be mounted in /dev. The
> ? kernel will automatically create/remove device files from this
> ? filesystem as devices appear/disappear from the system. devtmpfs
> ? exists in the Linux kernel since 2.6.32. When this option is
> ? selected *and* Buildroot is responsible for building the kernel,
> ? then Buildroot ensures that the kernel is built with the appropriate
> ? options to make devtmpfs work. When Buildroot is *not* responsible
> ? for building the kernel (the user does it on its own), then the user
> ? is responsible for making sure that CONFIG_DEVTMPFS and
> ? CONFIG_DEVTMPFS_MOUNT are both enabled in the kernel configuration.
> ? When this mode is used, no static device files are created in the
> ? root filesystem: the device files are automatically created at boot
> ? time by the kernel.
>
> ?* Dynamic using mdev. This is exactly like with 'devtmpfs' (i.e,
> ? devtmpfs is required for this mode to work), but Buildroot adds the
> ? mdev utility into the mix. mdev is an utility bundled with Busybox
> ? which gets executed when the kernel notifies that a device has been
> ? added or removed from the system. Compared to a pure 'devtmpfs'
> ? solution, it allows to execute arbitrary applications or shell
> ? scripts when devices appear/disappear. mdev behaviour can be
> ? configured from /etc/mdev.conf, refer to the Busybox documentation
> ? for more details. Since this case relies on devtmpfs, there are no
> ? static device files created in the root filesystem, and no device
> ? table is used.
>
> ?* Dynamic using udev. This is also exactly like with 'devtmpfs' (i.e,
> ? devtmpfs is required for this mode to work), but Buildroot adds the
> ? udev daemon into the mix. udev is the "device event manager" used in
> ? all Linux desktop and server systems and can be seen as a
> ? "full-featured" mdev. It is more configurable, provides a library
> ? called libudev to allow applications to query for which devices are
> ? available, etc.
>
> So, no, do *NOT* ever use the post-build script to create device files.
> Just understand how those device files work and the different mode
> Buildroot provides to handle them.
>
> If you need precisions or have questions on this topic, don't hesitate
> to ask. We have many people on the list ready to answer your questions.
>
> Best regards,
>
> Thomas
> --
> Thomas Petazzoni, Free Electrons
> Kernel, drivers, real-time and embedded Linux
> development, consulting, training and support.
> http://free-electrons.com
> _______________________________________________
> buildroot mailing list
> buildroot at busybox.net
> http://lists.busybox.net/mailman/listinfo/buildroot

[Buildroot] Device files and Buildroot

[Felipe Contreras]

On Sat, Dec 3, 2011 at 10:25 AM, Thomas Petazzoni
<thomas.petazzoni@free-electrons.com> wrote:
> Le Sat, 3 Dec 2011 01:25:41 +0100,
> Aleksander Dutkowski <adutkowski@gmail.com> a ?crit :
>
>> "System configuration --> Custom script to run before creating
>> filesystem image" input is for specify a PATH to bash script.
>> You you must simply create script, fill with mknod commands (remember
>> - mknod can be ran only by root! so do 'sudo mknod')
>
> No, no, no and no ! I'm sorry but this is totally horrible and is
> definitely not the way of doing things with Buildroot.
>
> With Buildroot, you have four ways of managing the device files
> in /dev. The mechanism used to manage device files is configured from
> System configuration -> /dev management. The four ways are :

Why is this not described when one press '?' on the menu options?

Cheers.

-- 
Felipe Contreras

[Buildroot] Device files and Buildroot

[Michael S. Zick]

On Thu December 8 2011, Felipe Contreras wrote:
> On Sat, Dec 3, 2011 at 10:25 AM, Thomas Petazzoni
> <thomas.petazzoni@free-electrons.com> wrote:
> > Le Sat, 3 Dec 2011 01:25:41 +0100,
> > Aleksander Dutkowski <adutkowski@gmail.com> a ?crit :
> >
> >> "System configuration --> Custom script to run before creating
> >> filesystem image" input is for specify a PATH to bash script.
> >> You you must simply create script, fill with mknod commands (remember
> >> - mknod can be ran only by root! so do 'sudo mknod')
> >
> > No, no, no and no ! I'm sorry but this is totally horrible and is
> > definitely not the way of doing things with Buildroot.
> >
> > With Buildroot, you have four ways of managing the device files
> > in /dev. The mechanism used to manage device files is configured from
> > System configuration -> /dev management. The four ways are :
> 
> Why is this not described when one press '?' on the menu options?
>

Because you haven't sent the patch for the menu that does that yet.

Mike 
> Cheers.
> 

[Buildroot] Device files and Buildroot

[Thomas Petazzoni]

Le Thu, 8 Dec 2011 23:30:21 +0200,
Felipe Contreras <felipe.contreras@gmail.com> a ?crit :

> Why is this not described when one press '?' on the menu options?

Yeah, that's possible, but help options on kconfig "choices" is a bit
non-user friendly. There is no way to specify a global help for the
choice, it is only possible to provide a per-choice help.

Regards,

Thomas
-- 
Thomas Petazzoni, Free Electrons
Kernel, drivers, real-time and embedded Linux
development, consulting, training and support.
http://free-electrons.com