lkml.org 
[lkml]   [2018]   [Feb]   [21]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
SubjectRe: [PATCH 2/2] selftests/gpio: Fix OUTPUT directory in Makefile
From
Date
On 02/21/2018 04:04 PM, Daniel Díaz wrote:
> Hello!
>
>
> On 02/21/2018 04:09 PM, Andy Shevchenko wrote:
>> On Wed, Feb 21, 2018 at 11:52 PM, Daniel Díaz <daniel.diaz@linaro.org> wrote:
>>> When simply running `make' from the selftests top dir, this
>>> error shows up:
>>>
>>> cc -O2 -g -std=gnu99 -Wall -I../../../../usr/include/ -I/usr/include/libmount -I/usr/include/blkid -I/usr/include/uuid gpio-mockup-chardev.c ../../../gpio/gpio-utils.o -lmount -o gpio-mockup-chardev
>>> cc: error: ../../../gpio/gpio-utils.o: No such file or directory
>>> <builtin>: recipe for target 'gpio-mockup-chardev' failed
>>> make[1]: *** [gpio-mockup-chardev] Error 1
>>>
>>> because the output directory is set to "selftests/gpio" and
>>> all binaries built from ../../../gpio/ end up there. In fact,
>>> they appear as, exempli gratia:
>>> * gpiogpio-event-mon
>>> * gpiogpio-hammer
>>> * gpioinclude/
>>> * gpiolsgpio
>>> which is wrong, as it's missing directory separator somewhere.
>>>
>>> This patch sets straight the output directory when building
>>> ../../../gpio/ so that binaries don't cross paths.
>>
>> This patch doesn't sound right like previous one.
>> Does selftest infrastructure have it's own build system like tools?
>
> Yes. See linux/Documentation/dev-tools/kselftest.rst § "Contributing new
> tests (details)" or linux/tools/testing/selftests/lib.mk itself, which
> starts saying:
> # This mimics the top-level Makefile. We do it explicitly here so that this
> # Makefile can operate with or without the kbuild infrastructure.
>
>
>> Does it use tools' one?
>
> No, as far as I can tell. If anything, it uses it only to build inside
> tools/gpio/ whatever is needed.
>
>
>> What's wrong with the current approach?
>
> When building from the tools/testing/selftests/ directory, binaries end
> up in the current directory (and with an improper name), which is
> undesirable. The gpio/ selftest fails to build when make'ing from there.
>
> This is with current master:
> /linux/tools/testing/selftests$ make
> make[1]: Entering directory '/linux/tools/testing/selftests/gpio'
> [...]
> gcc -O2 -g -std=gnu99 -Wall -I../../../../usr/include/
> gpio-mockup-chardev.c ../../../gpio/gpio-utils.o
> ../../../../usr/include/linux/gpio.h -lmount -I/usr/include/libmount -o
> gpio-mockup-chardev
> gcc: error: ../../../gpio/gpio-utils.o: No such file or directory
> <builtin>: recipe for target 'gpio-mockup-chardev' failed
> make[1]: *** [gpio-mockup-chardev] Error 1
> make[1]: Leaving directory '/linux/tools/testing/selftests/gpio'
> Makefile:32: recipe for target 'all' failed
> make: *** [all] Error 2
>
>
>> Why do you need tools in selftests?
>
> I'll leave that to the test author to answer, as this is only a fix in
> the Makefile, but the only part of tools needed here is tools/gpio/*,
> for make'ing the object files this test will be built against.
>

This test depends on tools/libs built in tools/gpio. The original
commit has details on the benefits of adding this test. Adding author
to the thread for further comments.

22f6592b23ef8a0c09283bcb13087340721e1154
commit 22f6592b23ef8a0c09283bcb13087340721e1154
Author: Bamvor Jian Zhang <bamvor.zhangjian@linaro.org>

thanks,
-- Shuah

\
 
 \ /
  Last update: 2018-02-22 00:12    [W:0.040 / U:0.056 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site