Latest NXP Downstream U-Boot imx_v2020.04_5.4.70_2.3.0 Fails Regular GitLab CI/CD Pipeline

[marcelziswiler]

I find it rather disappointing that NXP's latest downstream U-Boot

https://source.codeaurora.org/external/imx/uboot-imx/log/?h=imx_v2020.04_5.4.70_2.3.0

does not even pass the regular GitLab CI/CD pipeline (see attached screenshot). Could we not please agree on at least a very bare minimal quality target?

Thanks!

[Bio_TICFSL]

Hello Marcelziswiler,

That uboot is still on development, you can download the previous release, I will check with developers team this situation.

Regards

[marcelziswiler]

> That uboot is still on development

I don't get what you mean. L5.4.70_2.3.0 has been released as a GA across all i.MX SoCs not?

https://www.nxp.com/design/software/embedded-software/i-mx-software/embedded-linux-for-i-mx-applicat...

[Bio_TICFSL]

Hello,

Please get the detail failure from the gitlab CI/CD log. Otherwise how do I supposed to help?

Regards

[marcelziswiler]

Any progress on this?

[marcelziswiler]

Sorry, for not providing this earlier!

[Bio_TICFSL]

Hello

Please let me know which config is used for building when this error happen? 

WARNING: no maintainers for 'imx8qm_mek_androidauto_xen'
Building current source for 1 boards (1 thread, 40 jobs per thread)
Starting build... 0 0 0 /1 (starting) sandbox: + sandbox_spl
+common/image-android.c:18:10: fatal error: asm/bootm.h: No such file or directory
+ #include <asm/bootm.h>
+ ^~~~~~~~~~~~~
+compilation terminated.
+make[2]: *** [common/image-android.o] Error 1
+make[1]: *** [common] Error 2
+drivers/mailbox/sandbox-mbox-test.c: In function 'sandbox_mbox_test_send':
+drivers/mailbox/sandbox-mbox-test.c:27:9: error: too few arguments to function 'mbox_send'
+ return mbox_send(&sbmt->chan, &msg);
+ ^~~~~~~~~
+In file included from drivers/mailbox/sandbox-mbox-test.c:8:0:
+include/mailbox.h:125:5: note: declared here
+ int mbox_send(struct mbox_chan *chan, const void *data, ulong timeout_us);
+ ^~~~~~~~~
+ }
+ ^

Regards

[marcelziswiler]

I don't think any such "config" is relevant in this case. This is GitLab CI/CD automation we are talking about.

https://gitlab.denx.de/u-boot/u-boot/-/commit/1a62a722186ba5159813a5737de9894364ceede7

https://gitlab.denx.de/u-boot/u-boot/pipelines/charts

[Bio_TICFSL]

Hi,

From your word: "now fail such automated build tests" you also understand this is build test failure. Then we have to see which board configurations caused this build failure, this is why I'm asking you for which configurations. Otherwise how can we identify which one brokes.

Do you really understand????

[Bio_TICFSL]

Hi,

From your word: "now fail such automated build tests" you also understand this is build test failure. Then we have to see which board configurations caused this build failure, this is why I'm asking you for which configurations. Otherwise how can we identify which one brokes.

Do you really understand????

[marcelziswiler]

As you really do not understand I will recommend not to use any of NXP's broken downstream forks in the future. Thanks!

[Bio_TICFSL]

Please read carefully and understand the GitLab CI/CD, it's an integration/build tool for all of the projects. Which means automation build for any merged patches and check. Please look at the log you sent me, it's build failure, so that's why I'm asking you which config for build failure.

[marcelziswiler]

I really doubt you understand what we are talking about here:

• The upstream U-Boot project agreed on using GitLab CI/CD e.g. to do such automated build tests.
• NXP forked off from upstream U-Boot and added patches on top which now fail such automated build tests.

So it is NXP's sole responsibility, nothing to do with any of our configurations. NXP's fork of U-Boot just fails such automated build tests! Either NXP plans to fix it or admits that their stuff is broken...