From mboxrd@z Thu Jan  1 00:00:00 1970
From: Mauro Condarelli <mc5686@mclink.it>
Date: Mon, 31 Aug 2015 14:48:27 +0200
Subject: [Buildroot] [PATCH 1/1] Fix Image name not configurable (issue
 #28).
In-Reply-To: <20150831081740.GA3596@free.fr>
References: <1440842823-1283-1-git-send-email-mc5686@mclink.it>
 <1440845031-3965-1-git-send-email-mc5686@mclink.it>
 <1440845031-3965-2-git-send-email-mc5686@mclink.it>
 <20150829115935.GC4857@free.fr> <55E1DDB9.6040101@mclink.it>
 <20150831093214.6ae89b8a@free-electrons.com> <20150831081740.GA3596@free.fr>
Message-ID: <55E44D1B.5050207@mclink.it>
List-Id: <buildroot.busybox.net>
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Content-Transfer-Encoding: 7bit
To: buildroot@busybox.net

Agreed.
I will press upstream.
In the meantime I will live with my local changes.
I'll keep the list posted on any (eventual) development.

Regards
Mauro

Il 31/08/2015 10:17, Yann E. MORIN ha scritto:
> Mauro, Thomas, All,
>
> On 2015-08-31 09:32 +0200, Thomas Petazzoni spake thusly:
>> On Sat, 29 Aug 2015 18:28:41 +0200, Mauro Condarelli wrote:
>>> After the previous discussion on IRC I came up with this very different patch which does not touch the upstream package at all (even though I still think allowing user to chose isn't evil).
>>> Before I try to submit it I would like a comment.
>>> The aim is to force the booted image name according to requested kernel target if (and only if) user contextually requests the image to be installed.
>>> Obvious downside is user has no feedback of the change (and I don't know how to provide this feedback).
>>> Would this be more acceptable?
>>>
>>> diff --git a/boot/at91bootstrap3/at91bootstrap3.mk b/boot/at91bootstrap3/at91bootstrap3.mk
>>> index fa67ea6..ccb5e95 100644
>>> --- a/boot/at91bootstrap3/at91bootstrap3.mk
>>> +++ b/boot/at91bootstrap3/at91bootstrap3.mk
>>> @@ -40,6 +40,23 @@ define AT91BOOTSTRAP3_INSTALL_IMAGES_CMDS
>>>           cp $(@D)/binaries/*.bin $(BINARIES_DIR)
>>>    endef
>>> +ifeq ($(BR2_LINUX_KERNEL_INSTALL_TARGET),y)
>>> +define AT91BOOTSTRAP3_KCONFIG_FIXUP_CMDS
>>> +       $(if $(BR2_LINUX_KERNEL_UIMAGE),
>>> +               $(call KCONFIG_SET_OPT,CONFIG_IMAGE_NAME,"uImage",$(@D)/.config))
>>> +       $(if $(BR2_LINUX_KERNEL_APPENDED_UIMAGE),
>>> +               $(call KCONFIG_SET_OPT,CONFIG_IMAGE_NAME,"uImage",$(@D)/.config))
>>> +       $(if $(BR2_LINUX_KERNEL_ZIMAGE),
>>> +               $(call KCONFIG_SET_OPT,CONFIG_IMAGE_NAME,"zImage",$(@D)/.config))
>>> +       $(if $(BR2_LINUX_KERNEL_APPENDED_ZIMAGE),
>>> +               $(call KCONFIG_SET_OPT,CONFIG_IMAGE_NAME,"zImage",$(@D)/.config))
>>> +       $(if $(BR2_LINUX_KERNEL_VMLINUX),
>>> +               $(call KCONFIG_SET_OPT,CONFIG_IMAGE_NAME,"vmlinux",$(@D)/.config))
>>> +       $(if $(BR2_LINUX_KERNEL_IMAGE_TARGET_CUSTOM),
>>> +               $(call KCONFIG_SET_OPT,CONFIG_IMAGE_NAME,$(BR2_LINUX_KERNEL_IMAGE_NAME),$(@D)/.config))
>>> +endef
>>> +endif
>> I think this is not correct because it assumes you want to boot the
>> Linux kernel directly from ATBootstrap, while many people usually do
>> AT91Bootstrap -> another bootloader (U-Boot or Barebox) -> Linux kernel.
> Besides, this is changing an option that is not user-configurable, so
> its lifetime in the .config is not guaranteed. I.e. if the buildsystem
> of at01bootstrap3 would do an "oldconfig" on it, it would get lost
> (well, at least that's the case on their master branch).
>
> Please, get the matter with upstream and see what they think.
>
> When upstream has agreed on a "fix", *then* we can see at backporting it
> until they do a new release.
>
> Regards,
> Yann E. MORIN.
>