On enabling CONFIG_OF_SELFTEST which enables CONFIG_OF_DYNAMIC,
we found out that while getting PHYs for the controller we were
doing an extra of_node_put on the child node in our routines -
exynos_e/ohci_get_phy().
This child is however already put by of_get_next_available_child()
which does a of_node_put() on the "prev" node. So there's no point
in putting the same node again in our routine.
Reported-by: Daniel Drake <drake@endlessm.com>
Signed-off-by: Vivek Gautam <gautam.vivek@samsung.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
phy = devm_of_phy_get(dev, child, NULL);
exynos_ehci->phy[phy_number] = phy;
- of_node_put(child);
if (IS_ERR(phy)) {
ret = PTR_ERR(phy);
if (ret == -EPROBE_DEFER) {
phy = devm_of_phy_get(dev, child, NULL);
exynos_ohci->phy[phy_number] = phy;
- of_node_put(child);
if (IS_ERR(phy)) {
ret = PTR_ERR(phy);
if (ret == -EPROBE_DEFER) {