[oe] RFC: remove old kernels

[Khem Raj]

On Fri, Oct 15, 2010 at 1:39 AM, Marcin Juszkiewicz
<marcin at juszkiewicz.com.pl> wrote:
>
> Hi
>
> recipes/linux/ contains 235 recipes. I would like to propose cleaning:
>
> 1. drop 2.4 kernels - if someone still needs them then .oz354x and .stable
>   branches have them.
>
> This may affect:
>
> - zauruses (which have 2.6 kernel)
> - ipaq h36/38/39/54/55 - some have 2.6 kernel
> - ipod - but does someone really uses 2.4.24 on own ipod?
> - ramses - but this is maintained out of tree
> - wrt54 - OpenWRT distro works better there and lets keep that way
> - xxs1500, mtx, sun4cdm - if someone maintain it at all then please raise your
>  hand
> - colinux - no one will cry
> - geodegx - 2.6 mainline works on it
>
>
> 2. Drop kernels older then 2.6.20 - it was over 3 years ago so if someone uses
>   older one then probably also uses 3 years old OE.
>
> This may affect (checked recipes):
>
> - ep93xx - machines with this cpu are mainline supported, patches from ts72xx
>  and sim.one works
> - chumby
> - compulab-pxa270
> - devkitidp-pxa255
> - fsg3
> - gumstix
> - amsdelta
> - bd-neon
> - colinux
> - epia
> - geodelx
> - h1940
> - h6300
> - ks8695
> - magicbox
> - mtx-3
> - neuros
> - sun4cdm
> - tornado-omap2
> - wrap-geode
> - x86
> - acern30
> - logicpd-pxa270
> - mainstone
> - mx21ads
> - netbook-pro
> - triton
> - blueangel
> - universal
>
> Some of those machines are not maintained anymore, many of them should work
> with current mainline kernels. If someone cares about anything on a list -
> raise your hand.

good plan. Go ahead with the removal.

>
> Why 2.6.20? Because NHK-15 uses it and does not work with mainline kernel.
>
> Regards,
> --
> JID:      hrw at jabber.org
> Website:  http://marcin.juszkiewicz.com.pl/
> LinkedIn: http://www.linkedin.com/in/marcinjuszkiewicz
>
>
>
> _______________________________________________
> Openembedded-devel mailing list
> Openembedded-devel at lists.openembedded.org
> http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-devel
>