Re: [PATCH] clocksource: Do not drop unheld reference on device node

["Uwe Kleine-König" <u.kleine-koenig@pengutronix.de>]

On Fri, Nov 22, 2013 at 05:31:46PM +0100, Daniel Lezcano wrote:
> On 11/22/2013 05:16 PM, Thierry Reding wrote:
> >On Sat, Oct 19, 2013 at 12:49:48AM +0200, Thierry Reding wrote:
> >>When booting a recent kernel on ARM with OF_DYNAMIC enabled, the kernel
> >>warns about the following:
> >>
> >>	[    0.000000] ERROR: Bad of_node_put() on /timer@50004600
> >>	[    0.000000] CPU: 0 PID: 0 Comm: swapper/0 Not tainted 3.12.0-rc5-next-20131017-00077-gedfd827-dirty #406
> >>	[    0.000000] [<c0015b68>] (unwind_backtrace+0x0/0xf4) from [<c00117e4>] (show_stack+0x10/0x14)
> >>	[    0.000000] [<c00117e4>] (show_stack+0x10/0x14) from [<c055f734>] (dump_stack+0x9c/0xc8)
> >>	[    0.000000] [<c055f734>] (dump_stack+0x9c/0xc8) from [<c03b47d4>] (of_node_release+0x90/0x9c)
> >>	[    0.000000] [<c03b47d4>] (of_node_release+0x90/0x9c) from [<c03b5084>] (of_find_matching_node_and_match+0x78/0xb4)
> >>	[    0.000000] [<c03b5084>] (of_find_matching_node_and_match+0x78/0xb4) from [<c07887c8>] (clocksource_of_init+0x60/0x70)
> >>	[    0.000000] [<c07887c8>] (clocksource_of_init+0x60/0x70) from [<c076e99c>] (start_kernel+0x1f4/0x33c)
> >>	[    0.000000] [<c076e99c>] (start_kernel+0x1f4/0x33c) from [<80008074>] (0x80008074)
> >>
> >>This is caused by clocksource_of_init() dropping a reference on the
> >>device node that it never took. The reference taken by the loop is
> >>implicitly dropped on subsequent iterations. See the implementation of
> >>and the comment on top of the of_find_matching_node_and_match()
> >>function for reference (no pun intended).
I only saw the end of the comment saying:

	Returns a node pointer with refcount incremented, use
	of_node_put() on it when done.

Maybe that could be improved to something like:

/**
 * of_find_matching_node_and_match - Find a node based on an of_device_id
 *                                   match table
 * @from: The node to start searching from or NULL, the node you pass
 *        will not be searched, only the next one will; typically, you
 *        pass what the previous call returned.
 * @matches: ...
 * @match: ...
 *
 * Reference counting: Puts a reference on @from and increases the
 * reference count of the returned node.
 */

> >>
> >>Signed-off-by: Thierry Reding <treding@nvidia.com>
> >>---
> >>  drivers/clocksource/clksrc-of.c | 1 -
> >>  1 file changed, 1 deletion(-)
> >
> >Can someone look at this?
> 
> Yes. Sounds like I missed it.
> 
> This regression has been introduced by:
> 
> commit 326e31eebe61dc838e031ea16968b2cfb43443e3
> Author: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
> Date:   Tue Oct 1 11:00:53 2013 +0200
> 
>     clocksource: Put nodes passed to CLOCKSOURCE_OF_DECLARE
> callbacks centrally
> 
>     Instead of letting each driver call of_node_put do it centrally in the
>     loop that also calls the CLOCKSOURCE_OF_DECLARE callbacks. This is less
>     prone to error and also moves getting and putting the references
> into the
>     same function.
> 
>     Consequently all respective of_node_put calls in drivers are removed.
> 
>     Signed-off-by: Uwe Kleine-König <u.kleine-koenig@pengutronix.de>
>     Signed-off-by: Daniel Lezcano <daniel.lezcano@linaro.org>
>     Acked-by: David Brown <davidb@codeaurora.org>
Still all but the hook in clocksource_of_init of this commit was
correct, right? (Well, but this buggy hunk makes the commit log wrong.)

Best regards
Uwe

-- 
Pengutronix e.K.                           | Uwe Kleine-König            |
Industrial Linux Solutions                 | http://www.pengutronix.de/  |