From: Thomas Petazzoni Date: Mon, 13 Jul 2020 21:54:02 +0200 Subject: [Buildroot] [PATCH 1/1] package/dhcpcd: fix build without fork In-Reply-To: <20200713090140.154594-1-fontaine.fabrice@gmail.com> References: <20200713090140.154594-1-fontaine.fabrice@gmail.com> Message-ID: <20200713215402.29a5770b@windsurf.home> List-Id: MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: buildroot@busybox.net Hello Fabrice, On Mon, 13 Jul 2020 11:01:40 +0200 Fabrice Fontaine wrote: > - Add a patch to fix build without fork in src/dhcpcd. > - Disable privsep as it unconditionally uses fork (privsep has been > enabled by default since version 9.0.0 and > https://github.com/rsmarples/dhcpcd/commit/3a4c2e5604d72151b06ed365aa71493740a3ad75) > > Fixes: > - http://autobuild.buildroot.org/results/9fcc88abedcb8a02946f37837dcf4fff02f66c23 > > Signed-off-by: Fabrice Fontaine I have applied, but your explanations are still too short, and they require additional effort from the maintainers to do research and have a proper commit log. Indeed, where there is a build failure like this, just "fixing" it is not enough. We need to understand since when it is happening, and why, so that we can understand if the fix should be backported to a LTS or stable release. So, I've done that research here, and extended the commit log as follows: package/dhcpcd: fix build without fork - Add a patch to fix build without fork in src/dhcpcd.c. This regression was introduced in upstream commit 3063ebb6c8ac7c96196fa923cdd5f7c0384de23b, which was merged in dhcpcd 9.0.0. Therefore, Buildroot is affected since we bumped from 8.0.3 to 9.1.4 in commit 809f548e79c6c099f1fa3e3728d90842be7059a7, which was applied after 2020.05 - Disable privsep as it unconditionally uses fork (privsep has been enabled by default since version 9.0.0 and https://github.com/rsmarples/dhcpcd/commit/3a4c2e5604d72151b06ed365aa71493740a3ad75) Fixes: - http://autobuild.buildroot.org/results/9fcc88abedcb8a02946f37837dcf4fff02f66c23 Could you try to do that research in the future, and extend the commit log as appropriate ? This would really help reducing the effort on the maintainer shoulders. Thanks a lot! Thomas -- Thomas Petazzoni, CTO, Bootlin Embedded Linux and Kernel engineering https://bootlin.com