Skip to content

[BUG] check-kmod-load-unload fails on MTL #1037

Closed

Description

@plbossart

TPLG=/lib/firmware/intel/sof-ace-tplg/sof-hda-generic-4ch.tplg MODEL=MTLP_RVP_HDA ~/sof-test/test-case/check-kmod-load-unload.sh
Polling 'sof_firmware_boot_complete --since=@1684343454' every 1s for 10s
Command 'sof_firmware_boot_complete --since=@1684343454' timed out after 10s and 11 attempts
2023-05-17 17:11:08 UTC [INFO] Test Result: FAIL!

However running the kmod scripts works fine.

It's likely a timeout value that's not long enough for MTL. Can we increase it to 70s as done for the boot?

@fredoh9 @marc-hb @keqiaozhang this is a gap in our coverage, we do need to start testing module load/unload - possibly with a reduced number of cycles since we lose one minute with the i915 issue

cc:

Activity

  1. marc-hb commented on May 17, 2023

    @marc-hb
    Collaborator

    This is not a sof-test bug, this is new, platform-specific timeout feature (not questioning the priority)

    This i915 timeout/bug has been a royal pain all across the board, @kv2019i is there really no way to avoid it?

  2. added
    type:enhancementNew framework feature or request
    and removed on May 17, 2023
  3. plbossart commented on May 17, 2023

    @plbossart
    MemberAuthor

    we already have a 70s timeout in other parts of sof-test, so I will assert that it's a bug. This should just work with the delay being handled.

    We have the same issue for all FPGA tests btw, the tests need to deal with the i915 60s wait.

  4. plbossart commented on May 17, 2023

    @plbossart
    MemberAuthor

    case-lib/config.sh:MAX_WAIT_FW_LOADING=${MAX_WAIT_FW_LOADING:-70}

    # Set MAX Polling time to check FW Loading. If FW is already loaded, it will
    # return immediately. Default value is set to 60 seconds, because i915 driver
    # timeout is 60 seconds.
    MAX_WAIT_FW_LOADING=${MAX_WAIT_FW_LOADING:-70}
    

    why can't we use this for kmod tests as well?

  5. added
    P1Blocker bugs or important features
    on Jun 22, 2023
  6. marc-hb commented on Jun 27, 2023

    @marc-hb
    Collaborator

    Re-opening because this is not in any test plan yet, will close once this is running daily.

    EDIT: blocked by bug 4442 below.

  7. keqiaozhang commented on Jun 28, 2023

    @keqiaozhang
    Contributor
  8. marc-hb commented on Oct 9, 2023

    @marc-hb
    Collaborator

    I checked the last daily (32811) and the test is now passing on MTL, probably has for a while. However it's still not enabled on PR testing: https://sof-ci.01.org/softestpr/PR1110/build783/devicetest/index.html (#1110).

  9. plbossart commented on Oct 9, 2023

    @plbossart
    MemberAuthor

    probably because of the 60s graphics timeout, but this one will be removed "soon"

  10. marc-hb commented on Oct 9, 2023

    @marc-hb
    Collaborator

    Pretty sure we did the work for this test to pass even with the i915 60s timeout.

    Moreover the timeout happens only once right after boot AFAIK.

    For sure daily tests haven't been affected one way or the other.

    So I think we should just add it back.

  11. keqiaozhang commented on Oct 10, 2023

    @keqiaozhang
    Contributor

    We have been enabling the kmod related tests in our mtl-daily test for weeks now. I just enabled them in PR testing.

  12. marc-hb commented on Oct 11, 2023

    @marc-hb
    Collaborator

    It's passing now, thanks @keqiaozhang and everyone else!

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

No one assigned

    Labels

    P1Blocker bugs or important featurestype:enhancementNew framework feature or requesttype:test coverage gapThis requires a new test case, not just fixing one

    Type

    No type

    Projects

    No projects

      Milestone

      No milestone

      Relationships

      None yet

      Development

      No branches or pull requests

      Issue actions