Skip to contentSkip to conversation

HID: hid-logitech-dj: add some range checks

PATCH In review Dan Carpenter 3 participants 4 messages Mar 5, 2013

Subsystems (3)
lore
Share & tools
lore
#0 lore raw

PATCH 1/1 HID: hid-logitech-dj: add some range checks

We can't trust dj_report->device_index because it comes from the user and hasn't been range checked. It is used as an offset into the djrcv_dev->paired_dj_devices[] array which has 7 elements. There is one range check already in logi_dj_recv_add_djhid_device() and I have copy and pasted that here.

DJ_DEVICE_INDEX_MIN is 1.
DJ_DEVICE_INDEX_MAX is 6.

---
This is a static checker fix and I'm not certain it's correct, please look it over carefully.

I do not know this code well, so I don't know why a zero index is invalid.

drivers/hid/hid-logitech-dj.c +7 -0
index 9500f2f..c01cd25 100644
@@ -701,6 +701,13 @@ static int logi_dj_raw_event(struct hid_device *hdev,
 	 * anything else with it.
 	 */
 
+	if ((dj_report->device_index < DJ_DEVICE_INDEX_MIN) ||
+	    (dj_report->device_index > DJ_DEVICE_INDEX_MAX)) {
+		dev_err(&hdev->dev, "%s: invalid device index:%d\n", __func__,
+			dj_report->device_index);
+		return true;
+	}
+
 	spin_lock_irqsave(&djrcv_dev->lock, flags);
 	if (dj_report->report_id = REPORT_ID_DJ_SHORT) {
 		switch (dj_report->report_type) {
#1 lore raw

Hi Dan,

Quoted 22 lines · Dan Carpenter

We can't trust dj_report->device_index because it comes from the user and hasn't been range checked. It is used as an offset into the djrcv_dev->paired_dj_devices[] array which has 7 elements. There is one range check already in logi_dj_recv_add_djhid_device() and I have copy and pasted that here.

DJ_DEVICE_INDEX_MIN is 1.
DJ_DEVICE_INDEX_MAX is 6.

---
This is a static checker fix and I'm not certain it's correct, please look it over carefully.

I do not know this code well, so I don't know why a zero index is invalid.

drivers/hid/hid-logitech-dj.c +0 -0
index 9500f2f..c01cd25 100644

drivers/hid/hid-logitech-dj.c static int logi_dj_raw_event(struct hid_device *hdev, · line 710

Quoted 10 lines
drivers/hid/hid-logitech-dj.c +7 -0
@@ -701,6 +701,13 @@ static int logi_dj_raw_event(struct hid_device *hdev,
         * anything else with it.
         */
 
+       if ((dj_report->device_index < DJ_DEVICE_INDEX_MIN) ||
+           (dj_report->device_index > DJ_DEVICE_INDEX_MAX)) {
+               dev_err(&hdev->dev, "%s: invalid device index:%d\n", __func__,
+                       dj_report->device_index);
+               return true;
+       }
+

Sorry, but it's a NACK in its current form:
- device_index = 0 is the receiver. IIRC, the receiver sends the different REPORT_TYPE_NOTIF_*. So I would say that this patch blocks events from the receiver.
- the DJ protocol specifies that the device index is between 1 and 6 and that 0 means the receiver. So unless there is a problem in the USB line from the receiver to the driver, device index will be between 1 and 6 for input events. This is only true as long as with stay with real firmware from Logitech, not from events from uhid.

If you want to add a check, it need to be in logi_dj_recv_forward_report(). The current access to djrcv_dev->paired_dj_devices[] in delayedwork_callback() has been removed in latest HID tree:
https://git.kernel.org/cgit/linux/kernel/git/jikos/hid.git/log/?h=for-3.10/logitech

Cheers,
Benjamin

Quoted 7 lines
drivers/hid/hid-logitech-dj.c +0 -1
        spin_lock_irqsave(&djrcv_dev->lock, flags);
        if (dj_report->report_id = REPORT_ID_DJ_SHORT) {
                switch (dj_report->report_type) {
--

To unsubscribe from this list: send the line "unsubscribe linux-input" in the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html

#2 lore raw

Hi Dan,

Quoted 28 lines · Benjamin Tissoires

Hi Dan,

Quoted 25 lines · Dan Carpenter

We can't trust dj_report->device_index because it comes from the user and hasn't been range checked. It is used as an offset into the djrcv_dev->paired_dj_devices[] array which has 7 elements. There is one range check already in logi_dj_recv_add_djhid_device() and I have copy and pasted that here.

DJ_DEVICE_INDEX_MIN is 1.
DJ_DEVICE_INDEX_MAX is 6.

---
This is a static checker fix and I'm not certain it's correct, please look it over carefully.

I do not know this code well, so I don't know why a zero index is invalid.

drivers/hid/hid-logitech-dj.c +0 -0
index 9500f2f..c01cd25 100644
@@ -701,6 +701,13 @@ static int logi_dj_raw_event(struct hid_device *hdev,
         * anything else with it.
         */
 
Quoted 12 lines

drivers/hid/hid-logitech-dj.c static int logi_dj_raw_event(struct hid_device *hdev, · line 710

Quoted 7 lines
drivers/hid/hid-logitech-dj.c +7 -0
+       if ((dj_report->device_index < DJ_DEVICE_INDEX_MIN) ||
+           (dj_report->device_index > DJ_DEVICE_INDEX_MAX)) {
+               dev_err(&hdev->dev, "%s: invalid device index:%d\n", __func__,
+                       dj_report->device_index);
+               return true;
+       }
+

Sorry, but it's a NACK in its current form:
- device_index = 0 is the receiver. IIRC, the receiver sends the different REPORT_TYPE_NOTIF_*. So I would say that this patch blocks events from the receiver.

Correct. This would block notifications from the receiver.

Quoted 10 lines

- the DJ protocol specifies that the device index is between 1 and 6 and that 0 means the receiver. So unless there is a problem in the USB line from the receiver to the driver, device index will be between 1 and 6 for input events. This is only true as long as with stay with real firmware from Logitech, not from events from uhid.

If you want to add a check, it need to be in logi_dj_recv_forward_report(). The current access to djrcv_dev->paired_dj_devices[] in delayedwork_callback() has been removed in latest HID tree:
https://git.kernel.org/cgit/linux/kernel/git/jikos/hid.git/log/?h=for-3.10/logitech

The test could be added there. It would be quite defensive though, as we should always get the right device indexes.

Quoted 15 lines

Cheers,
Benjamin

Quoted 7 lines
drivers/hid/hid-logitech-dj.c +0 -1
        spin_lock_irqsave(&djrcv_dev->lock, flags);
        if (dj_report->report_id = REPORT_ID_DJ_SHORT) {
                switch (dj_report->report_type) {
--

To unsubscribe from this list: send the line "unsubscribe linux-input" in the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html

--
To unsubscribe from this list: send the line "unsubscribe linux-input" in the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html

#3 lore raw
Quoted 6 lines · Nestor Lopez Casado
Quoted 4 lines

If you want to add a check, it need to be in logi_dj_recv_forward_report(). The current access to djrcv_dev->paired_dj_devices[] in delayedwork_callback() has been removed in latest HID tree:
https://git.kernel.org/cgit/linux/kernel/git/jikos/hid.git/log/?h=for-3.10/logitech

The test could be added there. It would be quite defensive though, as we should always get the right device indexes.

Ah... I appologize for the noise. I misunderstood how hid_input_report() works. I thought it could be called with user supplied parameters.

Never mind.

regards,
dan carpenter