6.12-stable review patch. If anyone has any objections, please let me know.
------------------
From: Zijun Hu <quic_zijuhu@xxxxxxxxxxx>
commit 0f7ca6f69354e0c3923bbc28c92d0ecab4d50a3e upstream.
of_irq_parse_one() may use uninitialized variable @addr_len as shown below:
// @addr_len is uninitialized
int addr_len;
// This operation does not touch @addr_len if it fails.
addr = of_get_property(device, "reg", &addr_len);
// Use uninitialized @addr_len if the operation fails.
if (addr_len > sizeof(addr_buf))
addr_len = sizeof(addr_buf);
// Check the operation result here.
if (addr)
memcpy(addr_buf, addr, addr_len);
Fix by initializing @addr_len before the operation.
Fixes: b739dffa5d57 ("of/irq: Prevent device address out-of-bounds read in interrupt map walk")
Cc: stable@xxxxxxxxxxxxxxx
Signed-off-by: Zijun Hu <quic_zijuhu@xxxxxxxxxxx>
Link: https://lore.kernel.org/r/20241209-of_irq_fix-v1-4-782f1419c8a1@xxxxxxxxxxx
Signed-off-by: Rob Herring (Arm) <robh@xxxxxxxxxx>
Signed-off-by: Greg Kroah-Hartman <gregkh@xxxxxxxxxxxxxxxxxxx>
---
drivers/of/irq.c | 1 +
1 file changed, 1 insertion(+)
--- a/drivers/of/irq.c
+++ b/drivers/of/irq.c
@@ -355,6 +355,7 @@ int of_irq_parse_one(struct device_node
return of_irq_parse_oldworld(device, index, out_irq);
/* Get the reg property (if any) */
+ addr_len = 0;
addr = of_get_property(device, "reg", &addr_len);
/* Prevent out-of-bounds read in case of longer interrupt parent address size */
