On Thu, Sep 17, 2015 at 04:24:21PM +0530, Archit Taneja wrote: > > > On 9/17/2015 2:04 PM, Ingo Molnar wrote: > > > > > >* Ingo Molnar <mingo at kernel.org> wrote: > > > > > > > >>So this patch was whitespace damaged - I applied it by hand and made the > >>commit > > > >>below. This has solved the crash, thanks Archit! > > > > > > > >Spoke too soon - the attached (allyesconfig-ish) config still crashes, first > >there > > > >are a handful of kobject debug warnings, then: > > The error handling in the driver is bad. The main problem is that > the driver_load op calls mgag200_driver_unload if anything fails, > which doesn't work well if driver_load fails mid way. mgag200_driver_unload is trying to unload everything evenif that has not succeeded in initializing. Here the ttm failed to initialize but still mgag200_mm_fini was called to unload it.
regards sudip
