From: Denys Dmytriyenko <denis@denix.org>
To: Anuj Mittal <anuj.mittal@intel.com>
Cc: Bruce Ashfield <bruce.ashfield@windriver.com>,
openembedded-core@lists.openembedded.org
Subject: Re: [rocko][PATCH 2/2] make-mod-scripts: add build requirements for external modules
Date: Wed, 15 Aug 2018 20:26:36 -0400 [thread overview]
Message-ID: <20180816002636.GW19965@denix.org> (raw)
In-Reply-To: <20180728084950.13242-2-anuj.mittal@intel.com>
On Sat, Jul 28, 2018 at 04:49:50PM +0800, Anuj Mittal wrote:
> From: Bruce Ashfield <bruce.ashfield@windriver.com>
>
> Newer kernels (4.14/v4.15+) have dependencies for the build of
> modules (and hence external modules). Without these dependencies
> explicitly in the build chain, you can end up with build failures like:
>
> work-shared/qemux86/kernel-source/scripts/extract-cert.c:21:25: fatal
> error: openssl/bio.h: No such file or directory
> | #include <openssl/bio.h>
> | ^
> | compilation terminated.
> | make[2]: *** [scripts/extract-cert] Error 1
> | make[1]: *** [scripts] Error 2
>
> To ensure that these headers are in place, and that the scripts use
> our build environment flags, we add a dependency on openssl-native
> and use oe_make to invoke the build.
>
> Older kernels have no issues with the extra dependency, so there's no
> need to make this conditional.
>
> (From OE-Core rev: 916cb2029d3c97bf12ebf03832b9ba980451dbcf)
>
> Fixes [YOCTO #12860]
>
> Signed-off-by: Bruce Ashfield <bruce.ashfield@windriver.com>
> Signed-off-by: Richard Purdie <richard.purdie@linuxfoundation.org>
> [Tweaked to have changes in module*.bbclass instead from where
> make-mod-scripts was split in sumo]
> Signed-off-by: Anuj Mittal <anuj.mittal@intel.com>
> ---
> meta/classes/module-base.bbclass | 5 ++++-
> meta/classes/module.bbclass | 2 +-
> 2 files changed, 5 insertions(+), 2 deletions(-)
>
> diff --git a/meta/classes/module-base.bbclass b/meta/classes/module-base.bbclass
> index 64729770c7..9f3465eb9d 100644
> --- a/meta/classes/module-base.bbclass
> +++ b/meta/classes/module-base.bbclass
> @@ -12,6 +12,9 @@ export CROSS_COMPILE = "${TARGET_PREFIX}"
> # we didn't pick the name.
> export KBUILD_OUTPUT = "${STAGING_KERNEL_BUILDDIR}"
>
> +DEPENDS += "bc-native"
> +EXTRA_OEMAKE += " HOSTCC="${BUILD_CC} ${BUILD_CFLAGS} ${BUILD_LDFLAGS}" HOSTCPP="${BUILD_CPP}""
> +
> export KERNEL_VERSION = "${@base_read_file('${STAGING_KERNEL_BUILDDIR}/kernel-abiversion')}"
> KERNEL_OBJECT_SUFFIX = ".ko"
>
> @@ -22,6 +25,6 @@ PACKAGE_ARCH = "${MACHINE_ARCH}"
> # be called before do_compile. See module.bbclass for an example.
> do_make_scripts() {
> unset CFLAGS CPPFLAGS CXXFLAGS LDFLAGS
> - make CC="${KERNEL_CC}" LD="${KERNEL_LD}" AR="${KERNEL_AR}" \
> + oe_runmake CC="${KERNEL_CC}" LD="${KERNEL_LD}" AR="${KERNEL_AR}" \
> -C ${STAGING_KERNEL_DIR} O=${STAGING_KERNEL_BUILDDIR} scripts prepare
This now breaks out-of-tree kernel modules that provide own non-standard
Makefiles with something like "-f build.mak" added to EXTRA_OEMAKE.
That is because do_make_scripts() gets executed inside kernel-source directory
(-C ${STAGING_KERNEL_DIR} above) and supposed to use kernel's Makefile, while
do_compile()/do_install() are executed within module's sources and use
whatever Makefile it provides. That is why it historically was using "make"
here and not "oe_runmake".
When Bruce separated this code into make-mod-scripts, it was possible to also
use oe_runmake there, as it now had own context. But the backport to rocko
moved it back into module-base.bbclass and it gets executed within module's
environment, including custom EXTRA_OEMAKE, which often breaks (i.e. few of my
modules are now broken).
Should this backport only touch DEPENDS and do_make_scripts[depends], as the
changelog claims, and leave make/oe_runmake and EXTRA_OEMAKE alone?
> }
> diff --git a/meta/classes/module.bbclass b/meta/classes/module.bbclass
> index 78d1b21dbd..282900dc6d 100644
> --- a/meta/classes/module.bbclass
> +++ b/meta/classes/module.bbclass
> @@ -2,7 +2,7 @@ inherit module-base kernel-module-split pkgconfig
>
> addtask make_scripts after do_prepare_recipe_sysroot before do_configure
> do_make_scripts[lockfiles] = "${TMPDIR}/kernel-scripts.lock"
> -do_make_scripts[depends] += "virtual/kernel:do_shared_workdir"
> +do_make_scripts[depends] += "virtual/kernel:do_shared_workdir openssl-native:do_populate_sysroot"
>
> EXTRA_OEMAKE += "KERNEL_SRC=${STAGING_KERNEL_DIR}"
>
> --
> 2.17.1
>
> --
> _______________________________________________
> Openembedded-core mailing list
> Openembedded-core@lists.openembedded.org
> http://lists.openembedded.org/mailman/listinfo/openembedded-core
next prev parent reply other threads:[~2018-08-16 0:32 UTC|newest]
Thread overview: 4+ messages / expand[flat|nested] mbox.gz Atom feed top
2018-07-28 8:49 [rocko][PATCH 1/2] make-mod-scripts/kernel-devsrc: Fix objtool issues Anuj Mittal
2018-07-28 8:49 ` [rocko][PATCH 2/2] make-mod-scripts: add build requirements for external modules Anuj Mittal
2018-08-16 0:26 ` Denys Dmytriyenko [this message]
2018-08-16 1:25 ` Denys Dmytriyenko
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=20180816002636.GW19965@denix.org \
--to=denis@denix.org \
--cc=anuj.mittal@intel.com \
--cc=bruce.ashfield@windriver.com \
--cc=openembedded-core@lists.openembedded.org \
/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