From mboxrd@z Thu Jan  1 00:00:00 1970
From: Petr Vorel <pvorel@suse.cz>
Date: Mon, 5 Feb 2018 14:28:57 +0100
Subject: [LTP] [RFC,
	COMMITTED 1/1] configure: add knob to control numa support
In-Reply-To: <20180205132441.19130-1-pvorel@suse.cz>
References: <20180205132441.19130-1-pvorel@suse.cz>
Message-ID: <20180205132857.GA20417@x230>
List-Id: <ltp.lists.linux.it>
MIME-Version: 1.0
Content-Type: text/plain; charset="us-ascii"
Content-Transfer-Encoding: 7bit
To: ltp@lists.linux.it

Hi,

> From: Anders Roxell <anders.roxell@linaro.org>

> Allow to disable numa from the top level.

> Based on patch:
> http://cgit.openembedded.org/openembedded-core/commit/meta/recipes-extended/ltp?id=4c7873552e13dfdba96afca7562c398d2966ca71

> Reported-by: Roy Li <rongqing.li@windriver.com>
> Signed-off-by: Anders Roxell <anders.roxell@linaro.org>
> Signed-off-by: Petr Vorel <pvorel@suse.cz>
> ---
>  configure.ac | 10 ++++++++++
>  1 file changed, 10 insertions(+)

> diff --git a/configure.ac b/configure.ac
> index 5c8d4eaf8..b39a31ddb 100644
> --- a/configure.ac
> +++ b/configure.ac
> @@ -83,6 +83,14 @@ else
>      AC_SUBST([WITH_EXPECT],["no"])
>  fi

> +# Numa
> +AC_ARG_WITH([numa],
> +  AC_HELP_STRING([--without-numa],
> +    [without numa support (default=yes)]),
> +  [with_numa=no],
> +  [with_numa=yes]
> +)
> +
>  # Perl
>  AC_ARG_WITH([perl],
>    [AC_HELP_STRING([--with-perl],
> @@ -157,7 +165,9 @@ LTP_CHECK_SELINUX
>  LTP_CHECK_SIGNAL
>  LTP_CHECK_SYSCALL_EVENTFD
>  LTP_CHECK_SYSCALL_MODIFY_LDT
> +if test "x$with_numa" = xyes; then
>  LTP_CHECK_SYSCALL_NUMA
> +fi
>  LTP_CHECK_SYSCALL_QUOTACTL
>  LTP_CHECK_SYSCALL_SIGNALFD
>  LTP_CHECK_SYSCALL_UNSHARE

Although the autodetection was fixed in commit bf3441268 ("numa: Check for API >= v2 with
autoconf") (the patch is for ltp released 4 years ago), I committed this as user should be
able to decide whether to compile tests or not.

I'm sorry, I've introduced two errors with this commit:

1) there should be default=no as we want to have
it default:
-    [without numa support (default=yes)]),
+    [without numa support (default=no)]),

So, this is easily to be fixed.

2) I haven't realized that, the TCONF message is misleading when
configured with '--without-numa':
tst_test.c:712: CONF: test requires libnuma >= 2 and it's development packages

So I guess the whole idea of our functionality compiling with error
message "you lack support" doesn't work well with enabling and disabling
features by user. There should be a way to define in config variable
WITHOUT_NUMA and print correct message "Compiled without NUMA support".

I'm going to fix this if you like the idea allowing users to configure
what is being build (there might be requests in the future to add more
switches). Otherwise I'll revert it.


Kind regards,
Petr