Prabhav Kumar Vaish <pvkumar5749...@gmail.com> writes:
> `dev->of_node` has a pointer to device node, of_node_get call seems
> unnecessary.
Sorry but it is necessary.
> Signed-off-by: Prabhav Kumar Vaish <pvkumar5749...@gmail.com>
> ---
> arch/powerpc/platforms/cell/iommu.c | 9 +++------
> 1 file changed, 3 insertions(+), 6 deletions(-)
>
> diff --git a/arch/powerpc/platforms/cell/iommu.c
> b/arch/powerpc/platforms/cell/iommu.c
> index 4cd9c0de22c2..5b794ce08689 100644
> --- a/arch/powerpc/platforms/cell/iommu.c
> +++ b/arch/powerpc/platforms/cell/iommu.c
> @@ -780,14 +780,13 @@ static int __init cell_iommu_init_disabled(void)
> static u64 cell_iommu_get_fixed_address(struct device *dev)
> {
> u64 cpu_addr, size, best_size, dev_addr = OF_BAD_ADDR;
> - struct device_node *np;
> + struct device_node *np = dev->of_node;
> const u32 *ranges = NULL;
> int i, len, best, naddr, nsize, pna, range_size;
>
> /* We can be called for platform devices that have no of_node */
> - np = of_node_get(dev->of_node);
> if (!np)
> - goto out;
> + return dev_addr;
>
> while (1) {
> naddr = of_n_addr_cells(np);
nsize = of_n_size_cells(np);
np = of_get_next_parent(np);
if (!np)
break;
of_get_next_parent() drops the reference of the node passed to it (np).
So if you actually tested your patch you should see a recount underflow.
cheers
