Re: advice on reading a call trace

2010-10-07 Thread Benjamin Herrenschmidt
On Tue, 2010-10-05 at 07:21 +0200, Jean-Mickael Guerin wrote: > SLUB is turned on, and the oops does not seem to happen when SLAB replaces > SLUB. > I just got lucky with SLAB, or does it sound familiar on ppc? Nope, they should both work fine. Try enabling SLAB_DEBUG or SLUB_DEBUG maybe ? You mi

Re: advice on reading a call trace

2010-10-07 Thread Benjamin Herrenschmidt
On Mon, 2010-10-04 at 13:24 +0200, Jean-Mickael Guerin wrote: > Hello, > I'm stepping into ppc world and I'd like to know how to read this call trace, > I enabled debug options but I'm not able to track the origin of this bug, I > mean > what happens before handle_page_fault(): > > Unable to hand

Re: advice on reading a call trace

2010-10-04 Thread Jean-Mickael Guerin
SLUB is turned on, and the oops does not seem to happen when SLAB replaces SLUB. I just got lucky with SLAB, or does it sound familiar on ppc? Regards, Jean-Mickael On 10/4/2010 1:24 PM, Jean-Mickael Guerin wrote: > Hello, > I'm stepping into ppc world and I'd like to know how to read this call t

advice on reading a call trace

2010-10-04 Thread Jean-Mickael Guerin
Hello, I'm stepping into ppc world and I'd like to know how to read this call trace, I enabled debug options but I'm not able to track the origin of this bug, I mean what happens before handle_page_fault(): Unable to handle kernel paging request for data at address 0x0008 Faulting instruction