On 28/02/2023 20:07, Saravana Kannan wrote:
On Tue, Feb 28, 2023 at 9:40 AM Jean-Jacques Hiblot
<jjhiblot@xxxxxxxxxxxxxxx> wrote:
From: Jean Jacques Hiblot <jjhiblot@xxxxxxxxxxxxxxx>
As all the other parsers do, parse_interrupt() must increase the refcount
of the device_node. Otherwise the refcount is decremented every time
parse_interrupt() is called on this node, leading to a potential
use-after-free.
This is a regression introduced by commit f265f06af194 ("of: property:
Fix fw_devlink handling of interrupts/interrupts-extended"). The reason is
that of_irq_parse_one() does not increase the refcount while the previously
used of_irq_find_parent() does.
Thanks for catching the issue Jean!
This feels like a bug in of_irq_parse_one() to me. It's returning a
reference to a node without doing a of_node_get() on it.
Rob, Marc, Do you agree?
Sarvana,
it looks like you're right. The bug seems to be in of_irq_parse_one().
It doesn't behave in the same way for "interrupts-extended" where it
does a get() and 'interrupts" where it doesn't.
So please ignore this patch.
Thanks
Jean,
If they agree, can you please fix of_irq_parse_one() and add a
of_node_put() to existing callers (if they aren't already doing a
put()).
Thanks,
Saravana
Fixes: f265f06af194 ("of: property: Fix fw_devlink handling of interrupts/interrupts-extended")
Signed-off-by: Jean Jacques Hiblot <jjhiblot@xxxxxxxxxxxxxxx>
---
drivers/of/property.c | 5 ++++-
1 file changed, 4 insertions(+), 1 deletion(-)
diff --git a/drivers/of/property.c b/drivers/of/property.c
index 134cfc980b70b..1f23bcb765c4e 100644
--- a/drivers/of/property.c
+++ b/drivers/of/property.c
@@ -1380,7 +1380,10 @@ static struct device_node *parse_interrupts(struct device_node *np,
strcmp(prop_name, "interrupts-extended"))
return NULL;
- return of_irq_parse_one(np, index, &sup_args) ? NULL : sup_args.np;
+ if (of_irq_parse_one(np, index, &sup_args))
+ return NULL;
+
+ return of_node_get(sup_args.np);
}
static const struct supplier_bindings of_supplier_bindings[] = {
--
2.25.1
