Re: [PATCH v2 2/3] linux-module: honor modules Makefile

[Jan Kiszka <jan.kiszka@siemens.com>]

On 01.09.22 19:17, Henning Schild wrote:
> Am Thu, 1 Sep 2022 18:08:08 +0200
> schrieb Jan Kiszka <jan.kiszka@siemens.com>:
> 
>> On 27.08.22 10:59, Henning Schild wrote:
>>> External modules might have their own appends and target
>>> definitions in their Makefile. All we need to give them is the
>>> target name and KDIR, not dive into KDIR right away.  
>>
>> My observation with external modules makefile having their own rules
>> is that they generally also have their own KDIR variables. There is no
>> standard way of telling those pick up the kernel from a standard place
>> and install it at the standard location.
>>
>>>
>>> Signed-off-by: Henning Schild <henning.schild@siemens.com>
>>> ---
>>>  meta/recipes-kernel/linux-module/files/debian/rules.tmpl | 6 +++---
>>>  1 file changed, 3 insertions(+), 3 deletions(-)
>>>
>>> diff --git
>>> a/meta/recipes-kernel/linux-module/files/debian/rules.tmpl
>>> b/meta/recipes-kernel/linux-module/files/debian/rules.tmpl index
>>> d3bd7dc30f21..0d16186b5ff3 100755 ---
>>> a/meta/recipes-kernel/linux-module/files/debian/rules.tmpl +++
>>> b/meta/recipes-kernel/linux-module/files/debian/rules.tmpl @@
>>> -48,13 +48,13 @@ KDIR := $(shell dpkg -L $(KERNEL_DEP) | grep
>>> "/lib/modules/.*/build") endif override_dh_auto_clean:
>>> -	$(MAKE) -C $(KDIR) M=$(PWD) clean
>>> +	$(MAKE) KDIR=$(KDIR) clean
>>>  
>>>  override_dh_auto_build:
>>> -	$(MAKE) -C $(KDIR) M=$(PWD) modules
>>> +	$(MAKE) KDIR=$(KDIR) modules
>>>  
>>>  override_dh_auto_install:
>>> -	$(MAKE) -C $(KDIR) M=$(PWD)
>>> INSTALL_MOD_PATH=$(PWD)/debian/${PN} modules_install
>>> +	$(MAKE) KDIR=$(KDIR) INSTALL_MOD_PATH=$(PWD)/debian/${PN}
>>> modules_install 
>>>  %:
>>>  	CFLAGS= LDFLAGS= dh $@ --parallel  
>>
>> This breaks existing users with makefile that are unusable or do not
>> have own rules. I would refrain from that breakage.
>>
>> If you have a more complex case than what we support, you can simply
>> provide your own rules and can then even account for the other
>> specialties that usually come with such module makefiles.
>>
>> Same arguing applies to patch 3.
> 
> I am not sure i would agree that the example Makefile is indeed valid,
> since it does not allow self sustained building with a simple "make
> modules_install". But yes for modules with such Makefiles this would be
> a breaking change.
> 
> Will go and carry the special rules entries downstream, while i still
> think Isar should take that and rather make those weirdos carry their
> own rules ... accepting a breakage.
> 
> But i have no clue how "weird" vs "common" it would be to have such a
> super-minimal Makefile. And my case is probably also not "common"
> enough to ague for too long.

Isar implements to official way of writing external module makefiles and
invoking them, see
https://www.kernel.org/doc/html/latest/kbuild/modules.html#command-syntax.
Anything beyond in its generic code would be guesswork about how
makefiles were written, what they expect as input etc.

Jan

-- 
Siemens AG, Technology
Competence Center Embedded Linux