[Buildroot] problem building openswan using svn-buildroot

[Buildroot] problem building openswan using svn-buildroot

[Robert Urban]

Hi Folks,

I'm getting the following error when trying to build openswan, as part of a
'make all':

/big/home/urban/Build/Buildroot/svn/buildroot/build_i686/staging_dir/usr/bin/i686-linux-uclibc-gcc
 -I/big/home/urban/Build/Buildroot/svn/buildroot/build_i686/openswan-2.4.9/programs/pluto/linux26
-I/big/home/urban/Build/Buildroot/svn/buildroot/build_i686/openswan-2.4.9/include
-I/big/home/urban/Build/Buildroot/svn/buildroot/build_i686/openswan-2.4.9/lib/libcrypto
-I/big/home/urban/Build/Buildroot/svn/buildroot/build_i686/openswan-2.4.9/linux/include
   -DPLUTO_SENDS_VENDORID -DUSE_KEYRR   -DKERNEL26_SUPPORT
-DKERNEL26_HAS_KAME_DUPLICATES  -DX509     -DPLUTO -DKLIPS
-DBUILDER=\"urban at noisy\" -DDEBUG -DGCC_LINT -DUSE_AES -DUSE_3DES -DIKE_ALG
-DKERNEL_ALG -DDB_CONTEXT -DAGGRESSIVE -DXAUTH -DMODECFG  -DNAT_TRAVERSAL
-DVIRTUAL_IP  -DI_KNOW_TRANSPORT_MODE_HAS_SECURITY_CONCERN_BUT_I_WANT_IT
 -DSHARED_SECRETS_FILE=\"/etc/ipsec.secrets\"
-DPOLICYGROUPSDIR=\"/etc/ipsec.d/policies\"
-DPERPEERLOGDIR=\"/var/log/pluto/peer\" -g -Wall -W -Wmissing-prototypes
-Wpointer-arith -Wbad-function-cast -Wcast-qual -Wmissing-declarations
-Wwrite-strings -Wstrict-prototypes  -UDEBUG -DNO_DEBUG -ULEAK_DETECTIVE -Os
-pipe -Os
-I/big/home/urban/Build/Buildroot/svn/buildroot/build_i686/staging_dir/usr/include
-I/big/home/urban/Build/Buildroot/svn/buildroot/build_i686/staging_dir/include
--sysroot=/big/home/urban/Build/Buildroot/svn/buildroot/build_i686/staging_dir/
-isysroot
/big/home/urban/Build/Buildroot/svn/buildroot/build_i686/staging_dir
-mtune=i686 -march=i686 -D_LARGEFILE_SOURCE -D_LARGEFILE64_SOURCE
-D_FILE_OFFSET_BITS=64 -isysroot
/big/home/urban/Build/Buildroot/svn/buildroot/build_i686/staging_dir
-I/big/home/urban/Build/Buildroot/svn/buildroot/project_build_i686/uclibc/root/usr/include
-c connections.c
In file included from connections.c:48:
connections.h:199: error: expected specifier-qualifier-list before
?CIRCLEQ_ENTRY?
connections.c: In function ?host_pair_enqueue_pending?:
connections.c:118: error: ?const struct connection? has no member named
?host_pair?
connections.c:119: error: ?const struct connection? has no member named
?host_pair?
connections.c: In function ?host_pair_first_pending?:

I pulled the svn copy a few days ago...

My config can be found at:

	http://www.spielwiese.de/rob/Buildroot/config

and an excerpt of the build logfile (starting just after openswan sources
are copied) can be found here:

	http://www.spielwiese.de/rob/Buildroot/openswan.log

My build host is x86 Fedora Core 10 box, with all current updates.

anyone have any idea what might be wrong?

thanks,

Rob Urban

[Buildroot] problem building openswan using svn-buildroot

[Peter Korsgaard]

>>>>> "Robert" == Robert Urban <urban@unix-beratung.de> writes:

 Robert> Hi Folks,
 Robert> I'm getting the following error when trying to build openswan, as part of a
 Robert> 'make all':

Our openswan support is very old (from ~2006), and that one apparently
doesn't like current kernel headers.

I tried updating to openswan-2.6.20, but that also has various build
problems. I don't know anything about openswan (and frankly, I don't
care much about it), so unless someone will step up and support/fix it
up, I'll mark it as broken and remove it after the next release.

Bernhard, you were the one adding it in the first place - What do you
say?

-- 
Bye, Peter Korsgaard

[Buildroot] problem building openswan using svn-buildroot

[Peter Korsgaard]

>>>>> "Robert" == Robert Urban <urban@unix-beratung.de> writes:

 Robert> Hi Peter,

 Robert> thanks for your reply.  I was beginning to get the impression
 Robert> that nobody cares about openswan :) Is there a better
 Robert> alternative?

Sorry, I don't know - Anyone else?

 Robert> In the meantime I managed to get openswan to compile after a
 Robert> simplistic one-line-correction to
 Robert> .../programs/pluto/connection.h.  The diff follows.  I dunno
 Robert> if my fix suffices for the general population...

Thanks, but that ofcourse doesn't change the general situation that
the openswan package hasn't seen any development since early 2007,
which probably isn't a good thing for something security related.

-- 
Bye, Peter Korsgaard

[Buildroot] problem building openswan using svn-buildroot

[Steffen Schulz]

On 090222 at 19:15, Peter Korsgaard wrote:
> >>>>> "Robert" == Robert Urban <urban@unix-beratung.de> writes:
> 
>  Robert> Hi Peter,
> 
>  Robert> thanks for your reply.  I was beginning to get the impression
>  Robert> that nobody cares about openswan :) Is there a better
>  Robert> alternative?
> 
> Sorry, I don't know - Anyone else?

strongSwan is, to my knowledge, the only free IKEv1/IKEv2 server
that is in active development and ready for productive use.

It's far from small though.


/steffen
-- 
  ,''`,
 : :' :                                                +49/1781384223
 `. `'                                        gpg --recv-key A04D7875
   `-  www.debian.org                     mailto: pepe at cbg.dyndns.org

[Buildroot] problem building openswan using svn-buildroot

[Peter Korsgaard]

>>>>> "Steffen" == Steffen Schulz <pepe_ml@gmx.net> writes:

Hi,

 Robert> thanks for your reply.  I was beginning to get the impression
 Robert> that nobody cares about openswan :) Is there a better
 Robert> alternative?
 >> 
 >> Sorry, I don't know - Anyone else?

 Steffen> strongSwan is, to my knowledge, the only free IKEv1/IKEv2 server
 Steffen> that is in active development and ready for productive use.

 Steffen> It's far from small though.

Ok, anyone wants to step up and support openswan/strongswan? Otherwise
I think our users are better off if we remove the package.

-- 
Bye, Peter Korsgaard

[Buildroot] problem building openswan using svn-buildroot

[Steffen Schulz]

On 090223 at 12:15, Peter Korsgaard wrote:
> Steffen Schulz wrote:
>  Steffen> It's far from small though.

On the bright side, it seems to be very modular..

> Ok, anyone wants to step up and support openswan/strongswan? Otherwise
> I think our users are better off if we remove the package.

I just updated a package I made a few months ago. It compiles and
binaries seem to work, but I can not test it right now.


BTW, it happened the second time now that I could not use the most
recent version of a program because the libtool patch won't apply.
I used the most recent strongswan version that seems to work.
The other package this happened with is gnutls > 2.4.2.


/steffen
-- 
Chair for System Security               mail: steffen.schulz at rub.de
Ruhr-University Bochum                  phone:  +49 (0)234 32 29177
Key fingerprint = B805 57BE E4AF 0104 CC51 77A1 CE6F 8D46 A04D 7875
-------------- next part --------------
A non-text attachment was scrubbed...
Name: strongswan4.2.9.patch
Type: text/x-diff
Size: 6865 bytes
Desc: not available
URL: <http://lists.busybox.net/pipermail/buildroot/attachments/20090223/c0fc6124/attachment-0001.bin>

[Buildroot] problem building openswan using svn-buildroot

[Peter Korsgaard]

>>>>> "Steffen" == Steffen Schulz <pepe_ml@gmx.net> writes:

 >> Ok, anyone wants to step up and support openswan/strongswan?
 >> Otherwise I think our users are better off if we remove the
 >> package.

 Steffen> I just updated a package I made a few months ago. It
 Steffen> compiles and binaries seem to work, but I can not test it
 Steffen> right now.

Ok.

 Steffen> BTW, it happened the second time now that I could not use the most
 Steffen> recent version of a program because the libtool patch won't apply.
 Steffen> I used the most recent strongswan version that seems to work.
 Steffen> The other package this happened with is gnutls > 2.4.2.

Yes, I've seen that as well. I think we need to come up with a more
robust way of handling the libtool stuff than blindly trying to apply
a patch to whatever version the package includes.

-- 
Bye, Peter Korsgaard

[Buildroot] problem building openswan using svn-buildroot

[Ulf Samuelsson]

m?n 2009-02-23 klockan 12:14 +0100 skrev Peter Korsgaard:
> >>>>> "Steffen" == Steffen Schulz <pepe_ml@gmx.net> writes:
> 
> Hi,
> 
>  Robert> thanks for your reply.  I was beginning to get the impression
>  Robert> that nobody cares about openswan :) Is there a better
>  Robert> alternative?
>  >> 
>  >> Sorry, I don't know - Anyone else?
> 
>  Steffen> strongSwan is, to my knowledge, the only free IKEv1/IKEv2 server
>  Steffen> that is in active development and ready for productive use.
> 
>  Steffen> It's far from small though.
> 

Why not declare it as BROKEN?
I would not be surprised if someone would need it
at some point even if it has little attention at the moment.

> Ok, anyone wants to step up and support openswan/strongswan? Otherwise
> I think our users are better off if we remove the package.
> 
BR
Ulf Samuelsson

[Buildroot] problem building openswan using svn-buildroot

[Peter Korsgaard]

>>>>> "Ulf" == Ulf Samuelsson <ulf.samuelsson@atmel.com> writes:

Hi,

 Steffen> It's far from small though.

 Ulf> Why not declare it as BROKEN?
 Ulf> I would not be surprised if someone would need it
 Ulf> at some point even if it has little attention at the moment.

Well - It's HUGE (>7M), but OK, I'll mark it as BROKEN and remove if
after the 2009.05 release unless someone steps up to fix it. I'll also
remove the kernel patches as we remove old kernel versions.
-- 
Bye, Peter Korsgaard

[Buildroot] problem building openswan using svn-buildroot

[Ulf Samuelsson]

>>>>>> "Ulf" == Ulf Samuelsson <ulf.samuelsson@atmel.com> writes:
>
> Hi,
>
> Steffen> It's far from small though.
>
> Ulf> Why not declare it as BROKEN?
> Ulf> I would not be surprised if someone would need it
> Ulf> at some point even if it has little attention at the moment.
>
> Well - It's HUGE (>7M), but OK, I'll mark it as BROKEN and remove if
> after the 2009.05 release unless someone steps up to fix it. I'll also
> remove the kernel patches as we remove old kernel versions.
> -- 
> Bye, Peter Korsgaard
>

If you feel that the stuff should not be in the svn, then
make the patches loadable and store the patches in the mirror.

As pointed out to you earlier, buildroot users are using linux-2.6.22 and 
later
so there is no reason to remove things which is active use.
Again, it makes sense to move the patches to the mirror,
but yuu are the only one to have write acdess as far as I know.

Best Regards
Ulf Samuelsson

[Buildroot] problem building openswan using svn-buildroot

[Peter Korsgaard]

>>>>> "Ulf" == Ulf Samuelsson <ulf.samuelsson@atmel.com> writes:

Hi,

 >> Well - It's HUGE (>7M), but OK, I'll mark it as BROKEN and remove if
 >> after the 2009.05 release unless someone steps up to fix it. I'll also
 >> remove the kernel patches as we remove old kernel versions.

 Ulf> If you feel that the stuff should not be in the svn, then
 Ulf> make the patches loadable and store the patches in the mirror.

What will that change if it's broken?

 Ulf> As pointed out to you earlier, buildroot users are using
 Ulf> linux-2.6.22 and later so there is no reason to remove things
 Ulf> which is active use.  Again, it makes sense to move the patches
 Ulf> to the mirror, but yuu are the only one to have write acdess as
 Ulf> far as I know.

Active openswan users? How come we haven't seen any bugreports or
patches since 2007 then?

As said earlier, I think it's absolutely critical to trim the package
versions between releases, otherwise we'll never be able to get any
significant testing coverage.

-- 
Bye, Peter Korsgaard

[Buildroot] problem building openswan using svn-buildroot

[Ulf Samuelsson]

>>>>>> "Ulf" == Ulf Samuelsson <ulf.samuelsson@atmel.com> writes:
> 
> Hi,
> 
> >> Well - It's HUGE (>7M), but OK, I'll mark it as BROKEN and remove if
> >> after the 2009.05 release unless someone steps up to fix it. I'll also
> >> remove the kernel patches as we remove old kernel versions.
> 
> Ulf> If you feel that the stuff should not be in the svn, then
> Ulf> make the patches loadable and store the patches in the mirror.
> 
> What will that change if it's broken?
> 
> Ulf> As pointed out to you earlier, buildroot users are using
> Ulf> linux-2.6.22 and later so there is no reason to remove things
> Ulf> which is active use.  Again, it makes sense to move the patches
> Ulf> to the mirror, but yuu are the only one to have write acdess as
> Ulf> far as I know.
> 
> Active openswan users? How come we haven't seen any bugreports or
> patches since 2007 then?
> 

No active kernel users.
You indicate that you want to "remove old kernel versions",
and at the moment, I think that there is no need to do this.
It is desirable to make older patches downloadable from the mirror server.

I believe we should do the following for openswan.
* Mark the package as broken in Config
* Move the latest kernel patches to the mirror server.
* Add a comment about this in the Config
* Remove all openswan kernel patches from svn.
Even if someone does not use openswan today, it seems
to be useful and I would like to have some encouragement 
left for people to fix it.


> As said earlier, I think it's absolutely critical to trim the package
> versions between releases, otherwise we'll never be able to get any
> significant testing coverage.
> 
> -- 
> Bye, Peter Korsgaard
>


Best Regards
Ulf Samuelsson