From mboxrd@z Thu Jan 1 00:00:00 1970 From: Thomas Petazzoni Date: Sat, 3 Aug 2019 08:24:34 +0200 Subject: [Buildroot] [PATCH 0/4] new builtin kernel with entropy for the test infra In-Reply-To: <20190803052216.22292-1-ricardo.martincoski@gmail.com> References: <20190803052216.22292-1-ricardo.martincoski@gmail.com> Message-ID: <20190803082434.0080779f@windsurf.home> List-Id: MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: buildroot@busybox.net Hello, On Sat, 3 Aug 2019 02:22:12 -0300 Ricardo Martincoski wrote: > This small series adds a new builtin kernel image to the test infra. > It contains the VirtIORNG driver so the target will have enough entropy at > startup, a requirement for more and more packages being tested. > > It also uses this new builtin kernel to speed-up some test cases (atop, > syslog-ng, python-crossbar, python-nacl) and fix others (python-txtorcon, > python-treq). > > The binaries I used can be found here [1] and have this version string: > >Linux version 4.19.16 (br-user at runner-ed2dce3a-project-3755559-concurrent-0) (gcc version 4.9.4 (Buildroot 2017.08-git-01078-g95b1dae)) #1 Fri Aug 2 23:05:09 UTC 2019 > > One can see a complete run of "-runtime-tests" before [2] and after [3] this > series in the URLs below. > > In order to provide a v1 of this series, I created a naming convention for the > artifacts and also for the string passed to emulator.py in parameter "kernel". > It was the first naming that came to mind, so of course it is open to > suggestions/improvements. > > [1] https://gitlab.com/RicardoMartincoski/buildroot/commit/2ed44c7471cf5db246b592e24e9b2cfd69702f0f > [2] https://gitlab.com/RicardoMartincoski/buildroot/pipelines/74501223 > [3] https://gitlab.com/RicardoMartincoski/buildroot/pipelines/74516103 Thanks a lot for working on this. However, instead of having this new kernel as a separate "version" from the current ARMv5 pre-built kernel, why don't we simply update the existing one ? One issue is that downloads get cached, so if the filename is the same, it won't be redownloaded. Perhaps we need to store hashes of the download artifacts, so that we can detect if they need to be re-downloaded ? Best regards, Thomas -- Thomas Petazzoni, CTO, Bootlin Embedded Linux and Kernel engineering https://bootlin.com