linux-kselftest.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
From: Shuah Khan <skhan@linuxfoundation.org>
To: Michael Ellerman <mpe@ellerman.id.au>,
	yamada.masahiro@socionext.com, michal.lkml@markovi.net,
	shuah@kernel.org
Cc: linux-kbuild@vger.kernel.org, linux-kselftest@vger.kernel.org,
	linux-kernel@vger.kernel.org,
	"skh >> Shuah Khan" <skhan@linuxfoundation.org>
Subject: Re: [PATCH v2] selftests: Add kselftest-all and kselftest-install targets
Date: Wed, 16 Oct 2019 10:08:50 -0600	[thread overview]
Message-ID: <adcfcda4-c36e-c222-4964-f83b5f3d0097@linuxfoundation.org> (raw)
In-Reply-To: <87sgnttpoq.fsf@mpe.ellerman.id.au>

On 10/15/19 8:00 PM, Michael Ellerman wrote:
> Hi Shuah,
> 
> I know this has been merged already, so this is just FYI and in case it
> helps anyone else who's tracking down build failures.
> 

Thanks for letting me know. I have been sending updates about
breakages. Will send an update with this info.

> Sorry I didn't reply before you merged it, I was on leave.
> 
> Shuah Khan <skhan@linuxfoundation.org> writes:
>> Add kselftest-all target to build tests from the top level
>> Makefile. This is to simplify kselftest use-cases for CI and
>> distributions where build and test systems are different.
>>
>> Current kselftest target builds and runs tests on a development
>> system which is a developer use-case.
>>
>> Add kselftest-install target to install tests from the top level
>> Makefile. This is to simplify kselftest use-cases for CI and
>> distributions where build and test systems are different.
>>
>> This change addresses requests from developers and testers to add
>> support for installing kselftest from the main Makefile.
>>
>> In addition, make the install directory the same when install is
>> run using "make kselftest-install" or by running kselftest_install.sh.
>> Also fix the INSTALL_PATH variable conflict between main Makefile and
>> selftests Makefile.
> ...
>> diff --git a/tools/testing/selftests/Makefile b/tools/testing/selftests/Makefile
>> index c3feccb99ff5..bad18145ed1a 100644
>> --- a/tools/testing/selftests/Makefile
>> +++ b/tools/testing/selftests/Makefile
>> @@ -171,9 +171,12 @@ run_pstore_crash:
>>   # 1. output_dir=kernel_src
>>   # 2. a separate output directory is specified using O= KBUILD_OUTPUT
>>   # 3. a separate output directory is specified using KBUILD_OUTPUT
>> +# Avoid conflict with INSTALL_PATH set by the main Makefile
>>   #
>> -INSTALL_PATH ?= $(BUILD)/install
>> -INSTALL_PATH := $(abspath $(INSTALL_PATH))
>> +KSFT_INSTALL_PATH ?= $(BUILD)/kselftest_install
> 
> This change broke all my CI, because the tests no longer install in the
> place it's expecting them :/
> 

Sorry about that.

> I can fix it by explicitly specifying the install path in my CI scripts.
> 
>> +KSFT_INSTALL_PATH := $(abspath $(KSFT_INSTALL_PATH))
>> +# Avoid changing the rest of the logic here and lib.mk.
>> +INSTALL_PATH := $(KSFT_INSTALL_PATH)
>


I searched all the selftests Makefiles for it and convinced myself that,
the above would take care of it for these cases. I searched powerpc
Makefiles so this doesn't break it. Didn't think about the CI.

android/Makefile:	mkdir -p $(INSTALL_PATH)
android/Makefile:	install -t $(INSTALL_PATH) $(TEST_PROGS) 
$(TEST_PROGS_EXTENDED) $(TEST_FILES)
android/Makefile:		$(MAKE) OUTPUT=$$BUILD_TARGET -C $$SUBDIR 
INSTALL_PATH=$(INSTALL_PATH)/$$SUBDIR install; \
futex/Makefile:	mkdir -p $(INSTALL_PATH)
futex/Makefile:	install -t $(INSTALL_PATH) $(TEST_PROGS) 
$(TEST_PROGS_EXTENDED) $(TEST_FILES)
futex/Makefile:		$(MAKE) OUTPUT=$$BUILD_TARGET -C $$SUBDIR 
INSTALL_PATH=$(INSTALL_PATH)/$$SUBDIR install; \
sparc64/Makefile:	mkdir -p $(INSTALL_PATH)
sparc64/Makefile:	install -t $(INSTALL_PATH) $(TEST_PROGS) 
$(TEST_PROGS_EXTENDED) $(TEST_FILES)
sparc64/Makefile:		$(MAKE) OUTPUT=$$BUILD_TARGET -C $$SUBDIR 
INSTALL_PATH=$(INSTALL_PATH)/$$SUBDIR install; \


> But because the over-rideable variable changed from INSTALL_PATH to
> KSFT_INSTALL_PATH I will need to export both of them in order for my CI
> to work with old and new kernels.

My mistake. I overlooked that this could be overridden and could
be in used in CI scripts.

> 
> So to emulate the old behaviour for old & new kernels you need to do:
> 
> # export KSFT_INSTALL_PATH=install
> # export INSTALL_PATH=install

Can we avoid exporting both if we updated INSTALL_PATH with 
KSFT_INSTALL_PATH in all the Makefiles instead of my shortcut?
Would that make it easier for you?

> # make -C tools/testing/selftests install
> 
> 

I am looking to simplify the use-case and ran into the INSTALL_PATH
variable conflict between main Makefile and selftests Makefile.

I think this should have been KSFT_INSTALL_PATH to begin with.

thanks,
-- Shuah


  reply	other threads:[~2019-10-16 16:09 UTC|newest]

Thread overview: 5+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2019-09-26 22:40 [PATCH v2] selftests: Add kselftest-all and kselftest-install targets Shuah Khan
2019-09-28 17:53 ` Masahiro Yamada
2019-10-16  2:00 ` Michael Ellerman
2019-10-16 16:08   ` Shuah Khan [this message]
2019-10-17 22:51     ` Shuah Khan

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=adcfcda4-c36e-c222-4964-f83b5f3d0097@linuxfoundation.org \
    --to=skhan@linuxfoundation.org \
    --cc=linux-kbuild@vger.kernel.org \
    --cc=linux-kernel@vger.kernel.org \
    --cc=linux-kselftest@vger.kernel.org \
    --cc=michal.lkml@markovi.net \
    --cc=mpe@ellerman.id.au \
    --cc=shuah@kernel.org \
    --cc=yamada.masahiro@socionext.com \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).