Fake USB devices can send their own report descriptors for which the
input_mapping() hook does not get called.  In this case, pm->input_ep82 stays
NULL, which leads to a crash later.

This does not happen with the real device, but can be provoked by imposing as
one.

Cc: stable@vger.kernel.org
Signed-off-by: Günther Noack <gnoack@google.com>
---
 drivers/hid/hid-prodikeys.c | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/drivers/hid/hid-prodikeys.c b/drivers/hid/hid-prodikeys.c
index 74bddb2c3e82..6e413df38358 100644
--- a/drivers/hid/hid-prodikeys.c
+++ b/drivers/hid/hid-prodikeys.c
@@ -378,6 +378,10 @@ static int pcmidi_handle_report4(struct pcmidi_snd *pm, u8 *data)
 	bit_mask = (bit_mask << 8) | data[2];
 	bit_mask = (bit_mask << 8) | data[3];
 
+	/* robustness in case input_mapping hook does not get called */
+	if (!pm->input_ep82)
+		return 0;
+
 	/* break keys */
 	for (bit_index = 0; bit_index < 24; bit_index++) {
 		if (!((0x01 << bit_index) & bit_mask)) {
-- 
2.52.0.457.g6b5491de43-goog

On Fri, 9 Jan 2026, Günther Noack wrote:

> Fake USB devices can send their own report descriptors for which the 
> input_mapping() hook does not get called.  In this case, pm->input_ep82 
> sta= ys NULL, which leads to a crash later.
> 
> This does not happen with the real device, but can be provoked by 
> imposing = as one.

Applied. thanks.

-- 
Jiri Kosina
SUSE Labs