Hi,
W dniu 30.06.2020 o 20:33, Daniel Lezcano pisze:
On 30/06/2020 18:56, Andrzej Pietrasiewicz wrote:
Hi,
W dniu 30.06.2020 o 17:53, Daniel Lezcano pisze:
On 30/06/2020 17:29, Andrzej Pietrasiewicz wrote:
Hi Daniel,
W dniu 30.06.2020 o 16:53, Daniel Lezcano pisze:
On 30/06/2020 15:43, Andrzej Pietrasiewicz wrote:
Hi Daniel,
I am reading the logs and can't find anything specific to thermal.
What I can see is
"random: crng init done"
with large times (~200s) and then e.g.
'auto-login-action timed out after 283 seconds'
I'm looking at e.g.
https://storage.kernelci.org/thermal/testing/v5.8-rc3-11-gf5e50bf4d3ef/arm/multi_v7_defconfig/gcc-8/lab-baylibre/baseline-imx6q-sabrelite.html
f5e50bf4d3ef is PATCH 11/11. Does the problem happen at PATCH 1-10/11?
PATCH 11/11 renames a method and the code compiles, so it seems
unlikely that this is causing problems. One should never say never,
though ;)
The sha1 is just the HEAD for the kernel reference. The regression
happens with your series, somewhere.
The reported failure is not due to some test failing but rather due
to timeout logging into the test system. Could it be that there is
some other problem?
I did reproduce:
v5.8-rc3 + series => imx6 hang at boot time
v5.8-rc3 => imx6 boots correctly
What did you reproduce? Timeout logging in to the test system or a "real"
failure of a test?
I kindly ask for a bisect.
I will give a try but it is a very long process as the board is running
on kernelci.
I was not able to reproduce it on imx7 despite it is the same sensor :/
Could it be that the thermal sensors somehow contribute to entropy and after
the series is applied on some machines it takes more time to gather enough
entropy?
Andrzej
