On Wed, 6 Mar 2024 09:50:03 +0100 Herve Codina herve.codina@bootlin.com wrote:
In the following sequence:
- of_platform_depopulate()
- of_overlay_remove()
During the step 1, devices are destroyed and devlinks are removed. During the step 2, OF nodes are destroyed but __of_changeset_entry_destroy() can raise warnings related to missing of_node_put(): ERROR: memory leak, expected refcount 1 instead of 2 ...
Indeed, during the devlink removals performed at step 1, the removal itself releasing the device (and the attached of_node) is done by a job queued in a workqueue and so, it is done asynchronously with respect to function calls. When the warning is present, of_node_put() will be called but wrongly too late from the workqueue job.
In order to be sure that any ongoing devlink removals are done before the of_node destruction, synchronize the of_changeset_destroy() with the devlink removals.
Fixes: 80dd33cf72d1 ("drivers: base: Fix device link removal") Cc: stable@vger.kernel.org Signed-off-by: Herve Codina herve.codina@bootlin.com
drivers/of/dynamic.c | 7 +++++++ 1 file changed, 7 insertions(+)
diff --git a/drivers/of/dynamic.c b/drivers/of/dynamic.c index 3bf27052832f..169e2a9ae22f 100644 --- a/drivers/of/dynamic.c +++ b/drivers/of/dynamic.c @@ -9,6 +9,7 @@ #define pr_fmt(fmt) "OF: " fmt +#include <linux/device.h> #include <linux/of.h> #include <linux/spinlock.h> #include <linux/slab.h> @@ -667,6 +668,12 @@ void of_changeset_destroy(struct of_changeset *ocs) { struct of_changeset_entry *ce, *cen;
- /*
* Wait for any ongoing device link removals before destroying some of* nodes.*/- device_link_wait_removal();
Tested-by: Luca Ceresoli luca.ceresoli@bootlin.com
And no problem appeared in my tests due to the removed unlock/lock around device_link_wait_removal().
Luca