On 22.02.21 10:12, Asherah Connor wrote: > Hi Tom, Simon; > > On 21/02/20 05:02:p, Tom Rini wrote: >> On Sat, Feb 20, 2021 at 04:54:41AM -0700, Simon Glass wrote: >>> Probably create a uclass. I expect there are operations and some >>> private data. We also need to think about testing, since all uclasses >>> need a sandbox test. >> >> Well, need a test. Since this is something for qemu, which is already a >> virtual platform, I'd be quite happy to see this tested in the qemu >> platforms that will support it and then not sandbox. > > I've completed the rewrite to uclass and will be sending the next > version of the patch series along shortly. I'm now looking to complete > the tests. > > Tom, what do you have in mind when you say "see this tested in the qemu > platforms that will support it"? It looks like we don't have any > existing tests that use qemu, with the exception of > test/nokia_rx51_test.sh, which doesn't look like anything to pattern > from.
If you look into a Gitlab job log like https://gitlab.denx.de/u-boot/u-boot/-/jobs/225922 you will see a lot of tests executed.'on QEMU. Have a look into file .gitlab-ci.yml for variables TEST_PY* and at https://gitlab.denx.de/u-boot/u-boot-test-hooks. In https://github.com/swarren/uboot-test-hooks you will have to change bin/travis-ci/conf.qemu_arm64_na, bin/travis-ci/conf.qemu_arm64_na to enable the framebuffer device. Best regards Heinrich > > A simple sandbox test in test/dm looks like it shouldn't be much work, > if it's ideal. > > Best, > > Asherah >