Re: [PATCH] 2.6.17-rt1 : fix x86_64 oops

["Paul E. McKenney" <paulmck@us.ibm.com>]

On Thu, Jun 29, 2006 at 04:24:42PM +0200, Ingo Molnar wrote:
> 
> * Dipankar Sarma <dipankar@in.ibm.com> wrote:
> 
> > OK, I need to catch up, but I see a lot of oops while running 
> > rcutorture in my box (rt1). I am investigating this atm.
> 
> fyi, 2.6.17-mm4 throws tons of these:
> 
>  BUG: scheduling while atomic: rcu_torture_rea/0x00010000/1471
>   [<c0106123>] show_trace+0xd/0x10
>   [<c010613d>] dump_stack+0x17/0x1a
>   [<c123b4e2>] schedule+0x61/0xc61
>   [<c015f380>] rcu_torture_reader+0x12e/0x17e
>   [<c014101f>] kthread+0xc4/0xf0
>   [<c0102005>] kernel_thread_helper+0x5/0xb

Probably the fault of my new ops-ization of rcutorture.c.  :-/
(For whatever it is worth, Dipankar would still be using the older
non-ops-ized version of rcutorture.c.)

Did you get these oopses with default setting of torture_type, or did you
specify torture_type=rcu_bh or torture_type=srcu to the modprobe command?

This was on i386, x86_64, or on something else?

Ah!  This would have been a CONFIG_PREEMPT build, right?

						Thanx, Paul