u-boot-utils badness

u-boot-utils badness

[Frans Meulenbroeks]

Incidentally I stumbled upon u-boot-utils.
This recipe did not seem too good.

We only have a version for 1.2.
Guess u-boot tools do not change that much, but still it seems somewhat odd.

It's do_compile reads:

do_compile () {
        oe_runmake Sandpoint8240_config
        oe_runmake tools
}

Which is also somewhat odd. Probably the u-boot env is fairlly standard, but
still it seems odd to have Sandpoint8240 there (especially since I could not
even relate it to a machine in oe).

Furthermore the recipe seems mostly useless anyway as u-boot itself also
generates the tools.
From u-boot.inc:

do_compile () {
        unset LDFLAGS
        unset CFLAGS
        unset CPPFLAGS
        oe_runmake ${UBOOT_MACHINE}
        oe_runmake all
        oe_runmake tools env
}

So there seems to be enough reason to nuke this recipe, but ...

A quick grep learned that there are several machines that recommend or
depend on this.

frans@frans-desktop:~/workspace/gtd/openembedded.git$ grep u-boot-utils -r
conf
conf/machine/lsmipsel.conf:MACHINE_EXTRA_RRECOMMENDS    = "u-boot-utils"
conf/machine/include/lsppc.inc:MACHINE_EXTRA_RRECOMMENDS = "u-boot-utils"
conf/machine/n1200.conf:MACHINE_EXTRA_RRECOMMENDS       = "temper
u-boot-utils"
conf/machine/kuropro.conf:MACHINE_EXTRA_RDEPENDS          = "micro-evtd
u-boot-utils"
conf/machine/openrd-base.conf:MACHINE_EXTRA_RDEPENDS          =
"u-boot-utils"
conf/machine/sheevaplug.conf:MACHINE_EXTRA_RDEPENDS          =
"u-boot-utils"
conf/machine/storcenter.conf:MACHINE_EXTRA_RDEPENDS = "u-boot-utils xfsprogs
sccd"
conf/machine/lspro.conf:MACHINE_EXTRA_RDEPENDS          = "micro-evtd
u-boot-utils"
conf/machine/mv2120.conf:MACHINE_EXTRA_RDEPENDS          = "u-boot-utils"
conf/machine/tsx09.conf:MACHINE_EXTRA_RDEPENDS                          =
"u-boot-utils piccontrol"
conf/machine/dns323.conf:MACHINE_EXTRA_RDEPENDS          = "u-boot-utils"
conf/machine/openrd-client.conf:MACHINE_EXTRA_RDEPENDS          =
"u-boot-utils"
conf/distro/include/preferred-om-2008-versions.inc:PREFERRED_VERSION_u-boot-utils
?= "1.2.0"

and in recipes:
frans@frans-desktop:~/workspace/gtd/openembedded.git$ grep -r u-boot-utils
recipes
recipes/tasks/task-openprotium.bb:DISTRO_EXTRA_RDEPENDS += "u-boot-utils"
recipes/linux/linux-gumstix_2.6.15.bb:DEPENDS = "u-boot-utils-native"
recipes/linux/linux-turbostation_2.6.20.2.bb:DEPENDS = "u-boot-utils-native"
recipes/linux/linux-bd-neon-2.6_2.6.22.bb:DEPENDS += "u-boot-utils-native

How to deal with this???

Frans

Re: u-boot-utils badness

[Philip Balister]

On 07/15/2010 02:41 AM, Frans Meulenbroeks wrote:
> Incidentally I stumbled upon u-boot-utils.
> This recipe did not seem too good.
>
> We only have a version for 1.2.
> Guess u-boot tools do not change that much, but still it seems somewhat odd.

.....

>
> How to deal with this???

Why not just leave it alone? It is entirely possible the recipe is being 
used soley to generate mkimage for machines that have working u-boots 
from another source.

Much better to have a u-boot recipe that generates the tools than have 
everyone who just needs mkimage hacking together a different recipe to 
get it.

Philip



>
> Frans
> _______________________________________________
> Openembedded-devel mailing list
> Openembedded-devel@lists.openembedded.org
> http://lists.linuxtogo.org/cgi-bin/mailman/listinfo/openembedded-devel
>

Re: u-boot-utils badness

[Frans Meulenbroeks]

2010/7/16 Philip Balister <philip@balister.org>

> On 07/15/2010 02:41 AM, Frans Meulenbroeks wrote:
>
>> Incidentally I stumbled upon u-boot-utils.
>> This recipe did not seem too good.
>>
>> We only have a version for 1.2.
>> Guess u-boot tools do not change that much, but still it seems somewhat
>> odd.
>>
>
> .....
>
>
>
>> How to deal with this???
>>
>
> Why not just leave it alone? It is entirely possible the recipe is being
> used soley to generate mkimage for machines that have working u-boots from
> another source.
>
> Much better to have a u-boot recipe that generates the tools than have
> everyone who just needs mkimage hacking together a different recipe to get
> it.
>
> Philip
>
>
As you wish :-)

Note that u-boot-tools not only provides mkimage but also things like fw_env
Also, to further mess up things: we also have u-boot-mkimage_1.3.2.bb
There seems to be room for some quality improvement.
Oh well....

Frans

Re: u-boot-utils badness

[Philip Balister]

On 07/16/2010 07:33 AM, Frans Meulenbroeks wrote:
> 2010/7/16 Philip Balister<philip@balister.org>
>
>> On 07/15/2010 02:41 AM, Frans Meulenbroeks wrote:
>>
>>
>>
> As you wish :-)
>
> Note that u-boot-tools not only provides mkimage but also things like fw_env
> Also, to further mess up things: we also have u-boot-mkimage_1.3.2.bb
> There seems to be room for some quality improvement.
> Oh well....

It might be worth consolidating these two recipes. But you'll need to 
figure out why they both exist.

Philip

Re: u-boot-utils badness

[Frans Meulenbroeks]

2010/7/16 Philip Balister <philip@balister.org>

>
>
> On 07/16/2010 07:33 AM, Frans Meulenbroeks wrote:
>
>> 2010/7/16 Philip Balister<philip@balister.org>
>>
>>  On 07/15/2010 02:41 AM, Frans Meulenbroeks wrote:
>>>
>>>
>>>
>>>  As you wish :-)
>>
>> Note that u-boot-tools not only provides mkimage but also things like
>> fw_env
>> Also, to further mess up things: we also have u-boot-mkimage_1.3.2.bb
>> There seems to be room for some quality improvement.
>> Oh well....
>>
>
> It might be worth consolidating these two recipes. But you'll need to
> figure out why they both exist.
>

This info is probably hidden in time. They seem to be recipes from the pre
git era.

Wrt the machines using it:
No idea why Tim added it for sheevaplug.
Guess Leon just used sheevaplug as starting point for openrd base.
Wrt openrd client: I just copied over from openrd base, but openrd client
does not need it.

(btw there is also something called u-boot-env.bb
This is apparently a dummy recipe to get u-boot env into deploy and is only
used by neuros)

Frans.