On Wed, Oct 15, 2025 at 10:03:44AM +0200, Lorenzo Pieralisi wrote:
On Tue, Oct 14, 2025 at 06:20:55PM -0400, Frank Li wrote:
quoted
On Tue, Oct 14, 2025 at 11:58:43AM +0200, Lorenzo Pieralisi wrote:
quoted
In of_msi_get_domain() if the iterator loop stops early because an
irq_domain match is detected, an of_node_put() on the iterator node is
needed to keep the OF node refcount in sync.
Add it.
Signed-off-by: Lorenzo Pieralisi <lpieralisi@kernel.org>
Cc: Rob Herring <robh@kernel.org>
---
Reviewed-by: Frank Li <Frank.Li@nxp.com>
After go though of_for_each_phandle, I understand why need of_node_put()
at break branch.
It will be nice if add of_for_each_phandle_scoped() help macro.
Yes because this fix is not the end of it AFAICS.
Please review and test patch (4) as well since I slightly change
the existing logic there, I don't want to break the EP mapping code you
added in f1680d9081e1 (by the way, I don't get that commit logic, if the
msi-parent loop would match it could just return and we could have
removed the
if (ret)
guarding of_map_id(), correct ?).
That's what I did in patch (4), please have a look.
It looks correct. PCIE-EP use msi-map. There are comments from Marc Zyngier.
Suppose you will update new version. I will test after new version posted.
Frank
Thanks,
Lorenzo
quoted
quoted
drivers/of/irq.c | 4 +++-
1 file changed, 3 insertions(+), 1 deletion(-)
diff --git a/drivers/of/irq.c b/drivers/of/irq.c
index e67b2041e73b..9f6cd5abba76 100644
--- a/drivers/of/irq.c
+++ b/drivers/of/irq.c
@@ -773,8 +773,10 @@ struct irq_domain *of_msi_get_domain(struct device *dev,
of_for_each_phandle(&it, err, np, "msi-parent", "#msi-cells", 0) {
d = irq_find_matching_host(it.node, token);
- if (d)
+ if (d) {
+ of_node_put(it.node);
return d;
+ }
}
return NULL;
--2.50.1