Comment 10 for bug 598664

Revision history for this message
Andy Whitcroft (apw) wrote : Re: resume fails in development kernels

@nnutter -- the testing you have reported seems to be with (what is now) pretty early Maverick kernels. I wonder if you would be able to test with the very latest version (based on the released v2.6.35 + stable updates) and confirm if that is affected; we have reports of known suspend/resume issues being resolved for other Lenovo models with the most recent upstream updates. You also mention that you had no issues running Lucid on this machine, It might be helpful to have the dmesg output from the machine following a boot/suspend/resume on a Lucid kernel (please label this dmesg very clearly when attaching), you should be able to install a Lucid kernel into your Maverick userspace and boot it for that. For the v2.6.34/35 testing you report above, which kernel builds were these, was the v2.6.35 the official release version. Also where did you get the builds were they the ones from the Mainline Builds archive? Could we also get a dmesg from the working kernel you have, again label this dmesg very clearly.

    https://wiki.ubuntu.com/Kernel/MainlineBuilds

I have spent some time looking at your working configuration but it is so divorced from the Ubuntu config as to be almost impossible to use as a comparison source, nothing obvious jumps out from a visual inspection. The diffstat between your working one and our current kernel configuration shows it to be something like 90% different:

     i386-config.flavour.generic | 5606 +++++++++++++++++++++++++++++++++-----------
     1 file changed, 4283 insertions(+), 1323 deletions(-)

Looking at the output of your debugging run, the Hash is pointing to the PNP Timer, it might be worth testing the broken kernels with each of 'nohpet' and 'nolapic' (individually) to see if they have any effect.

If you could perform any of the testing above and report back here that would be helpful. Thanks.