On 11/03/2020 21:11, Kevin Hilman wrote:

Hi Guillaume,

"kernelci.org bot" bot@kernelci.org writes:

...

stable/linux-4.19.y boot: 58 boots: 1 failed, 56 passed with 1 untried/unknown (v4.19.109)

Full Boot Summary: https://kernelci.org/boot/all/job/stable/branch/linux-4.19.y/kernel/v4.19.10... Full Build Summary: https://kernelci.org/build/stable/branch/linux-4.19.y/kernel/v4.19.109/

Tree: stable Branch: linux-4.19.y Git Describe: v4.19.109 Git Commit: 5692097116094a4a7045abcc1dbc172dbdc5657e Git URL: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux-stable.git Tested: 46 unique boards, 16 SoC families, 15 builds out of 206

Boot Regressions Detected:

arm:

versatile_defconfig:
    gcc-8:
      versatile-pb:
          lab-collabora: new failure (last pass: v4.19.108)

I think this is a false positive.

This is showing up as a regression in the email report, but looking at the web link above, it's actually a "conflict": lab-baylibre shows passing and lab-collaabora shows offline (infra error.)

I've stopped trusting boot email reports, they got broken when support for multiple compilers was added. The test email reports have a separate logic for detecting regressions and as far as I know it doesn't show this kind of false-positives. We'll be able to stop sending boot email reports when we stop running boot tests, see my other email about the plan for rolling that out.

Guillaume