Re: [PATCH 4.5 00/88] 4.5.4-stable review

[Kevin Hilman]

Greg Kroah-Hartman <gregkh@xxxxxxxxxxxxxxxxxxx> writes:

> On Mon, May 09, 2016 at 12:49:49AM -0700, kernelci.org bot wrote:
>> stable-queue boot: 269 boots: 6 failed, 247 passed with 16 offline (v4.5.3-88-gdb49c07fbfff)
>> 
>> Full Boot Summary: https://kernelci.org/boot/all/job/stable-queue/kernel/v4.5.3-88-gdb49c07fbfff/
>> Full Build Summary: https://kernelci.org/build/stable-queue/kernel/v4.5.3-88-gdb49c07fbfff/
>> 
>> Tree: stable-queue
>> Branch: local/linux-4.5.y.queue
>> Git Describe: v4.5.3-88-gdb49c07fbfff
>> Git Commit: db49c07fbffff689c30ce13696cdd8326d4f80a8
>> Git URL: git://server.roeck-us.net/git/linux-stable.git
>> Tested: 65 unique boards, 21 SoC families, 22 builds out of 108
>> 
>> Boot Failures Detected: https://kernelci.org/boot/?v4.5.3-88-gdb49c07fbfff&fail
>> 
>> arm:
>> 
>>     multi_v7_defconfig+CONFIG_PROVE_LOCKING=y:
>>         vexpress-v2p-ca9: 1 failed lab
>> 
>>     sunxi_defconfig:
>>         sun7i-a20-cubieboard2_rootfs:nfs: 1 failed lab
>> 
>>     socfpga_defconfig:
>>         socfpga_cyclone5_socrates: 1 failed lab
>> 
>>     multi_v7_defconfig+CONFIG_EFI=y:
>>         imx53-qsrb: 1 failed lab
>>         imx6dl-riotboard: 1 failed lab
>>         socfpga_cyclone5_socrates: 1 failed lab
>
> I still don't know how to intrepret these, sorry.
>
> What does "1 failed lab" mean?  What can I do with this?
>
> I looked at one of these "failures":
> 	https://kernelci.org/boot/vexpress-v2p-ca9/job/stable-queue/kernel/v4.5.3-88-gdb49c07fbfff/defconfig/multi_v7_defconfig+CONFIG_PROVE_LOCKING=y/lab/lab-baylibre-seattle/?_id=572fdd1e59b51453c3af381d
>
> and I can't really understand what is going on here, what should I be
> looking for?

Normally, this should be very easy for you: any non-zero number of
failures indicates a regression.

However, we've recently added a few more board farms to kernel CI which
have boards that work with newer kernels, but not older/stable kernels.

We need to blacklist these new boards from stable testing until they are
known to work, so they don't show up as false positives.

Sorry for the noise.  Until we sort out the blacklisting, I'll try to
follow-up on these with an interpretation of the restuls.

Kevin