Re: debian installer images for LS-XHL/LS-CHLv2

2014-04-14 Thread Michael Walle

Am 2014-04-02 18:24, schrieb Martin Michlmayr:

* Michael Walle  [2014-04-02 17:20]:

any preferences here? i'd prefer a preseed file as long as i don't
need any conditional behaviour. Could you point me to a board which
already uses a preseed file? Eg. in which repository should this file
be added?


Take a look at e.g.
debian-installer/build/config/armel/iop32x/network-console/ss4000e.cfg

At the moment, all Kirkwood images are generated from one file
(kirkwood/network-console.cfg), which doesn't work if you want to
include different preseed files for different machines.  However, I
don't see a problem with adding a new .cfg file under armel/kirkwood
for your device, which can then include a PRESEED directive.


FYI, i've filed bugs #744714 (oldsys-preseed), #744715 (network-console)
and #744716 (debian-installer).

--
michael


--
To UNSUBSCRIBE, email to debian-arm-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Archive: https://lists.debian.org/2d6d9760e397f5965c6a4941ee8a4...@walle.cc



Re: Dropping armel/ixp4xx

2014-04-14 Thread Ben Hutchings
On Mon, 2014-04-14 at 12:28 +0200, G.K. wrote:
> -BEGIN PGP SIGNED MESSAGE-
> Hash: SHA1
> 
> On 14/04/2014 03:02, Ben Hutchings wrote:
> > The ixp4xx kernel is now too big to fit in the flash partitions:
> > https://buildd.debian.org/status/fetch.php?pkg=linux&arch=armel&ver=3.14-1~exp1&stamp=1397151242
> 
> As fot IOP32X, why stopping support as we can install out from the flash (USB 
> stick is a good alternative) ?

Show me the code.

Ben.

> > I intend to disable this flavour for the next upload of 3.14 and remove
> > it for 3.15 if no-one provides a configuration fix.
> 
> Then we only need to change flashkernel ?

-- 
Ben Hutchings
I say we take off; nuke the site from orbit.  It's the only way to be sure.


signature.asc
Description: This is a digitally signed message part


Re: Dropping armel/ixp4xx

2014-04-14 Thread G.K.
-BEGIN PGP SIGNED MESSAGE-
Hash: SHA1

On 14/04/2014 03:02, Ben Hutchings wrote:
> The ixp4xx kernel is now too big to fit in the flash partitions:
> https://buildd.debian.org/status/fetch.php?pkg=linux&arch=armel&ver=3.14-1~exp1&stamp=1397151242

As fot IOP32X, why stopping support as we can install out from the flash (USB 
stick is a good alternative) ?

> I intend to disable this flavour for the next upload of 3.14 and remove
> it for 3.15 if no-one provides a configuration fix.

Then we only need to change flashkernel ?


-- 
To UNSUBSCRIBE, email to debian-arm-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Archive: https://lists.debian.org/534bb838.4010...@gk2.fr



Re: Dropping armel/ixp4xx

2014-04-14 Thread Jérémy Lal
Le lundi 14 avril 2014 à 02:02 +0100, Ben Hutchings a écrit :
> The ixp4xx kernel is now too big to fit in the flash partitions:
> https://buildd.debian.org/status/fetch.php?pkg=linux&arch=armel&ver=3.14-1~exp1&stamp=1397151242
> 
> I intend to disable this flavour for the next upload of 3.14 and remove
> it for 3.15 if no-one provides a configuration fix.

Would this change arch minimum requirement (-march=armv4t) ?

Jérémy.



-- 
To UNSUBSCRIBE, email to debian-arm-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Archive: https://lists.debian.org/1397467217.29197.2.camel@imac.chaumes



Re: ConcurrencyKit (libck0 package) on ARM?

2014-04-14 Thread Riku Voipio
Hi,

On Thu, Apr 10, 2014 at 03:56:41PM +0200, Daniel Pocock wrote:
> Upstream says that armel can't be supported:
> https://groups.google.com/d/msg/concurrencykit/I34MWr_mK8Q/sKeh1Ro3eAYJ
 
> Should I go back to the Ganglia community and suggest moving away from
> ConcurrencyKit?

In case ConcurrencyKit developers are not interested in portablity, I suggest
to Ganglia community to add a portable fallback in case ck is not available. 

> Could anybody with more ARM expertise than myself provide any more
> feedback to upstream about how this package could work in Debian?

Looking at ck sources briefly, ck could possibly have a backend
based on gcc atomic builtins[0]. It would not be very fast, but
should be guaranteed to work on debian archs. 

> As a compromise, we could ensure the Ganglia agent (gmond) builds
> without this dependency but the data collection server (gmetad package)
> would only build on i386 or amd64

I think anyone seriously interested in ganglia would be using armhf 
(or in future arm64), which already works with ck. So this compromise
could indeed work, at least as long as ganglia upstream keeps gmond
portable.

Riku

[0] http://gcc.gnu.org/onlinedocs/gcc/_005f_005fatomic-Builtins.html

> 
> -- 
> To UNSUBSCRIBE, email to debian-arm-requ...@lists.debian.org
> with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
> Archive: https://lists.debian.org/5346a319.60...@pocock.pro


-- 
To UNSUBSCRIBE, email to debian-arm-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Archive: https://lists.debian.org/20140414065949.ga32...@afflict.kos.to