[16:30] <jodh> xnox: Still having problems with my ->valid branch but it's demonstrated that current upstream is still iterating through job_change_state() too many times; try applying http://paste.ubuntu.com/7588618/ and booting!
[16:32] <xnox> jodh: we don't gurantee clean process_data[process] at that point.
[16:32] <xnox> jodh: we only reset it to null after a job is starting from stopped.
[16:33] <jodh> xnox: the point is that we're calling job_process_data_new() multiple times for the *same* job process
[16:33] <xnox> true.
[16:34] <xnox> but old one might be invalid at that point already. hm, need to transverse through this again.