Building uml with an allyesconfig and a few changes in UML only fails with:
...
CC arch/um/drivers/vde_user.o
/home/roel/dnld/src/kernel/git/linux-2.6/arch/um/drivers/vde_user.c:8:24: error: libvdeplug.h: No such file or directory
/home/roel/dnld/src/kernel/git/linux-2.6/arch/um/drivers/vde_user.c: In function 'vde_user_init':
/home/roel/dnld/src/kernel/git/linux-2.6/arch/um/drivers/vde_user.c:18: error: 'VDECONN' undeclared (first use in this function)
...
the file libvdeplug.h is missing and I can't find it neither in the tree or on
the net. 'git-grep VDECONN' shows that VDECONN isn't defined anywhere.
this code was added in this commit:
--
commit ad43c3565bebada7e5a13288e37542fd940369e8
Author: Jeff Dike <[email protected]>
Date: Tue Oct 16 01:26:48 2007 -0700
uml: add VDE networking support
Added vde network backend in uml to introduce native Virtual Distributed
Ethernet support (using libvdeplug).
Signed-off-by: Luca Bigliardi <[email protected]>
Signed-off-by: Jeff Dike <[email protected]>
Signed-off-by: Andrew Morton <[email protected]>
Signed-off-by: Linus Torvalds <[email protected]>
--
There is a reference to http://wiki.virtualsquare.org/index.php/Main_Page in
the Kconfig (also added by this commit).
After reverting this commit the build continues. possibly the header
libvdeplug.h was forgotten?
Roel
Roel Kluin wrote:
> the file libvdeplug.h is missing
>
> There is a reference to http://wiki.virtualsquare.org/index.php/Main_Page in
> the Kconfig (also added by this commit).
> After reverting this commit the build continues. possibly the header
> libvdeplug.h was forgotten?
Sorry for answering myself, but apparently it can be found here:
http://vde.sourceforge.net/
I think, however that there is a 'depends on' missing in kconfig - to prevent build
failure?
Roel Kluin wrote:
[ added some peoples to Cc ]
> Roel Kluin wrote:
>> the file libvdeplug.h is missing
>
>> There is a reference to http://wiki.virtualsquare.org/index.php/Main_Page in
>> the Kconfig (also added by this commit).
>> After reverting this commit the build continues. possibly the header
>> libvdeplug.h was forgotten?
>
> Sorry for answering myself, but apparently it can be found here:
>
> http://vde.sourceforge.net/
>
> I think, however that there is a 'depends on' missing in kconfig - to prevent build
> failure?
On Mon, Oct 29, 2007 at 04:24:54PM +0100, Roel Kluin wrote:
> CC arch/um/drivers/vde_user.o
> /home/roel/dnld/src/kernel/git/linux-2.6/arch/um/drivers/vde_user.c:8:24: error: libvdeplug.h: No such file or directory
> /home/roel/dnld/src/kernel/git/linux-2.6/arch/um/drivers/vde_user.c: In function 'vde_user_init':
> /home/roel/dnld/src/kernel/git/linux-2.6/arch/um/drivers/vde_user.c:18: error: 'VDECONN' undeclared (first use in this function)
> ...
>
> the file libvdeplug.h is missing and I can't find it neither in the tree or on
> the net. 'git-grep VDECONN' shows that VDECONN isn't defined anywhere.
This is an option which requires support on the host (in the form of
vde being installed - pcap is another example of this).
Jeff
--
Work email - jdike at linux dot intel dot com
Hello!
I tried 2.6.24-rc1-git11 with "allmodconfig ARCH=um" today.
this gave at least two errors - i needed to set CONFIG_SMP=n and CONFIG_UML_NET_VDE=n to make it compile.
not sure if options should be set to "n" automatically or being fixed differently, but if there is something broken or some header file missing, from an "end-user`s" point of view errors like this may be hard to get around.
it`s nothing against UML, it`s just the fact that i would expect "allmodconfig" or "allyesconfig" to work out of the box.
one should be able to compile a kernel without the need to use google, imho.
many automated cross-compile tests i have seen till now mostly show more broken than successful builds, as you can see at http://l4x.org/k , and issues like this don`t help making this better.
minor issue, though......
regards
roland
On Mon, Oct 29, 2007 at 04:24:54PM +0100, Roel Kluin wrote:
> CC arch/um/drivers/vde_user.o
> /home/roel/dnld/src/kernel/git/linux-2.6/arch/um/drivers/vde_user.c:8:24: error: \
> libvdeplug.h: No such file or directory
> /home/roel/dnld/src/kernel/git/linux-2.6/arch/um/drivers/vde_user.c: In function \
> 'vde_user_init':
> /home/roel/dnld/src/kernel/git/linux-2.6/arch/um/drivers/vde_user.c:18: error: \
> 'VDECONN' undeclared (first use in this function)
> ...
>
> the file libvdeplug.h is missing and I can't find it neither in the tree or on
> the net. 'git-grep VDECONN' shows that VDECONN isn't defined anywhere.
This is an option which requires support on the host (in the form of
vde being installed - pcap is another example of this).
Jeff
> Roel Kluin wrote:
>> the file libvdeplug.h is missing
>
>> There is a reference to http://wiki.virtualsquare.org/index.php/Main_Page in
>> the Kconfig (also added by this commit).
>> After reverting this commit the build continues. possibly the header
>> libvdeplug.h was forgotten?
>
> Sorry for answering myself, but apparently it can be found here:
>
> http://vde.sourceforge.net/
>
> I think, however that there is a 'depends on' missing in kconfig - to prevent build
> failure?
______________________________________________________________________
XXL-Speicher, PC-Virenschutz, Spartarife & mehr: Nur im WEB.DE Club!
Jetzt testen! http://produkte.web.de/club/?mc=021130
On Fri, Nov 02, 2007 at 01:05:59PM +0100, [email protected] wrote:
> I tried 2.6.24-rc1-git11 with "allmodconfig ARCH=um" today.
>
> this gave at least two errors - i needed to set CONFIG_SMP=n
SMP should depend on BROKEN for now, and I thought it did.
> and CONFIG_UML_NET_VDE=n to make it compile.
For this, my immediate reaction is that if you're going to build
something, the dependencies should be there. However, I guess we
should have some sympathy for the automatic allyesconfig builders.
I'm wondering whether, instead of failing the build if headers or
libraries aren't available, compile in some sort of stub which will
produce a run-time error instead.
The violates my preference to have error detected as early as
possible, but maybe it's preferable to the current situation.
Jeff
--
Work email - jdike at linux dot intel dot com