Re: Problems with 2.6.17-rt8

[Steven Rostedt <rostedt@goodmis.org>]

On Thu, 2006-08-03 at 13:22 -0700, Bill Huey wrote:
> On Thu, Aug 03, 2006 at 10:27:41AM -0400, Steven Rostedt wrote:
> > The rest was probably caused as a side effect from above.  The above is
> > already broken!
> > 
> > You have NUMA configured too, so this is also something to look at.
> > 
> > I still wouldn't ignore the first bug message you got:
> > 
> > ----
> > BUG: scheduling while atomic: udev_run_devd/0x00000001/1568
> > 
> > Call Trace:
> >        <ffffffff8045c693>{__schedule+155}
> >        <ffffffff8045f156>{_raw_spin_unlock_irqrestore+53}
> >        <ffffffff80242241>{task_blocks_on_rt_mutex+518}
> >        <ffffffff80252da0>{free_pages_bulk+39}
> >        <ffffffff80252da0>{free_pages_bulk+39}
> > ...
> > ----
> > 
> > This could also have a side effect that messes things up.
> > 
> > Unfortunately, right now I'm assigned to other tasks and I cant spend
> > much more time on this at the moment.  So hopefully, Ingo, Thomas or
> > Bill, or someone else can help you find the reason for this problem.
> 
> free_pages_bulk is definitely being called inside of an atomic.
> I force this stack trace when the in_atomic() test is true at the
> beginning of the function.
> 
> 
> [   29.362863] Call Trace:
> [   29.367107]        <ffffffff802a82ac>{free_pages_bulk+86}
> [   29.373122]        <ffffffff80261726>{_raw_spin_unlock_irqrestore+44}
> [   29.380233]        <ffffffff802a8778>{__free_pages_ok+428}
> [   29.386336]        <ffffffff8024f101>{free_hot_page+25}
> [   29.392165]        <ffffffff8022e298>{__free_pages+41}
> [   29.397898]        <ffffffff806b604d>{__free_pages_bootmem+174}
> [   29.404457]        <ffffffff806b5266>{free_all_bootmem_core+253}
> [   29.411112]        <ffffffff806b5340>{free_all_bootmem_node+9}
> [   29.417574]        <ffffffff806b254e>{numa_free_all_bootmem+61}
> [   29.424122]        <ffffffff8046e96e>{_etext+0}
> [   29.429224]        <ffffffff806b1392>{mem_init+128}
> [   29.434691]        <ffffffff806a17ab>{start_kernel+377}
> [   29.440520]        <ffffffff806a129b>{_sinittext+667}
> [   29.446669] ---------------------------
> [   29.450963] | preempt count: 00000001 ]
> [   29.455257] | 1-level deep critical section nesting:
> [   29.460732] ----------------------------------------
> [   29.466212] .. [<ffffffff806a169a>] .... start_kernel+0x68/0x221
> [   29.472815] .....[<ffffffff806a129b>] ..   ( <= _sinittext+0x29b/0x2a2)
> [   29.480056]

Perhaps you could put in that in_atomic check at the start of each of
these functions and point to where it is a problem.  Perhaps a spinlock
is taken that was real and not a mutex.

-- Steve


-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/