On Thu, Apr 13, 2017 at 02:20:10PM +0200, Johannes Thumshirn wrote: > On Thu, Apr 13, 2017 at 08:11:40PM +0800, Ming Lei wrote: > > Ok, could you apply the attached debug patch and collect the > > ftrace log? (ftrace_dump_on_oops need to be passed to kernel cmd line). > Thanks Johannes very much for collecting the log. The root cause should be the following: - if 'offset' of one segment isn't zero, we can't make that segment at full size, otherwise the segment ends in the unaligned virt boundary. Give me a while, I will figure out one patch to address the issue. Thanks, Ming
