Comment 6 for bug 1703091

Joao Luis (jpluis) wrote :

Answering first to note 5, yes.
All the times that kernel 4.4.0-83 crashed, the problem goes away if I boot my previous kernel 4.4.0-81.

kernel 4.4.0-83.106 does not crash during all the boots. Only some times. I could not figure out if they are related to cold/hot boots (in the morning always suceeds, thank goodness), CPU temperature, or any other factor.

Answering to note 4, regarding the 4.12.0-041200.201707022031 kernel it appears to solve the problem, but I can't be 100% sure.

Using the grub Advanced Ubuntu options, I performed 12 reboots, 6 reboots of each kernel 4.4.0-83.106
and 4.12.0-041200.201707022031 interleaved (boot 4.4.0-83.106, then reboot 4.12.0-041200, then reboot 4.4.0-83.106, etc...)

Kernel 4.4.0-83.106 crashed on boot (with the same error) 3 out of 6 times.

Kernel 4.12.0-041200 crashed 0 (zero) out of 6 times.

Not all reboots took a 5 second power-off interval. (The ones that kernel 4.4.0-83 crashed, did wait a few seconds before booting again).

So, the fix looks promising.

Meanwhile, I go back to kernel 4.4.0-83 or 4.4.0-81 because I need virtualbox (but I leave 4.12.0-041200 installed for now).

For my own record, the procedure I followed to install the 4.12.0-041200 was:




    sha1sum *.deb # sums looked ok
    # as root (su -) from now on

    dpkg -i *.deb
    # ...
    # ERROR (dkms apport): kernel package linux-headers-4.12.0-041200-generic is not supported (twice)
    # ...
    # lots of W: Possible missing firmware /lib/firmware/nvidia/...

But it did boot fine afterwards. (I did not attempt to remove or run virtualbox).