USB: cdc-acm: handle broken union descriptors

PATCH In review Johan Hovold 3 participants 22 messages Sep 21, 2020 – Sep 22, 2020

Subsystems (3)

Stable branch USB USB ACM driver

lore

Share & tools

lore

Johan Hovold

author

2020-09-21 14:00 UTC #0 lore raw

v2 cover USB: cdc-acm: handle broken union descriptors

This series adds support for handling broken union descriptors by falling back to "combined-interface" probing.

The first patch drops some bogus altsetting sanity checks which would otherwise have had to be needlessly reproduced for consistency. The third patch drops the driver specific data class define in favour of the common one. The last one, cleans up the no-union-descriptor handling by probing for a "combined-interface" before falling back to the call-management descriptor.

Note that I changed my mind on the stable tag; we can't be overly paranoid about a theoretical risk of breaking some quirky devices. And if we do, we still want to know about it, right?

Daniel, would you mind giving these a spin as well?

Johan

v2
- add stable tag to 2/2 as it enables a new class of devices
- demote a broken-union warning to dev_dbg
- replace the fourth RFC patch with a clean up of the no-union-descriptor case only

Johan Hovold (4):
Revert "cdc-acm: hardening against malicious devices"
USB: cdc-acm: handle broken union descriptors
USB: cdc-acm: use common data-class define
USB: cdc-acm: clean up no-union-descriptor handling

 drivers/usb/class/cdc-acm.c | 55 ++++++++++++++++---------------------
 drivers/usb/class/cdc-acm.h | 13 ++++-----
 2 files changed, 29 insertions(+), 39 deletions(-)

Signature · 1 line

2.26.2

Johan Hovold

author

2020-09-21 14:00 UTC #1 lore raw

PATCH v2 1/4 Revert "cdc-acm: hardening against malicious devices"

This reverts commit 2ad9d544f2497a7bf239c34bd2b86fd19683dbb5.

Drop bogus sanity check; an interface in the active configuration will always have a current altsetting assigned by USB core.

Signed-off-by Johan Hovold <[email protected]>

---
 drivers/usb/class/cdc-acm.c | 5 -----
 1 file changed, 5 deletions(-)

drivers/usb/class/cdc-acm.c +0 -5

index 7f6f3ab5b8a6..e28ac640ff9c 100644
@@ -1196,9 +1196,6 @@ static int acm_probe(struct usb_interface *intf,
 		return -EINVAL;
 	}
 
-	if (!intf->cur_altsetting)
-		return -EINVAL;
-
 	if (!buflen) {
 		if (intf->cur_altsetting->endpoint &&
 				intf->cur_altsetting->endpoint->extralen &&

@@ -1252,8 +1249,6 @@ static int acm_probe(struct usb_interface *intf,
 		dev_dbg(&intf->dev, "no interfaces\n");
 		return -ENODEV;
 	}
-	if (!data_interface->cur_altsetting || !control_interface->cur_altsetting)
-		return -ENODEV;
 
 	if (data_intf_num != call_intf_num)
 		dev_dbg(&intf->dev, "Separate call control interface. That is not fully supported.\n");

Signature · 1 line

2.26.2

Johan Hovold

author

2020-09-21 14:00 UTC #2 lore raw

PATCH v2 2/4 USB: cdc-acm: handle broken union descriptors

In reply to Johan Hovold

Handle broken union functional descriptors where the master-interface doesn't exist or where its class is of neither Communication or Data type (as required by the specification) by falling back to "combined-interface" probing.

Note that this still allows for handling union descriptors with switched interfaces.

This specifically makes the Whistler radio scanners TRX series devices work with the driver without adding further quirks to the device-id table.

Link https://lore.kernel.org/r/[email protected] open here

Reported-by Daniel Caujolle-Bert <[email protected]>

Cc [email protected]

Signed-off-by Johan Hovold <[email protected]>

---
 drivers/usb/class/cdc-acm.c | 12 ++++++++++++
 1 file changed, 12 insertions(+)

drivers/usb/class/cdc-acm.c +12 -0

index e28ac640ff9c..f42ade505569 100644
@@ -1240,9 +1240,21 @@ static int acm_probe(struct usb_interface *intf,
 			}
 		}
 	} else {
+		int class = -1;
+
 		data_intf_num = union_header->bSlaveInterface0;
 		control_interface = usb_ifnum_to_if(usb_dev, union_header->bMasterInterface0);
 		data_interface = usb_ifnum_to_if(usb_dev, data_intf_num);
+
+		if (control_interface)
+			class = control_interface->cur_altsetting->desc.bInterfaceClass;
+
+		if (class != USB_CLASS_COMM && class != USB_CLASS_CDC_DATA) {
+			dev_dbg(&intf->dev, "Broken union descriptor, assuming single interface\n");
+			combined_interfaces = 1;
+			control_interface = data_interface = intf;
+			goto look_for_collapsed_interface;
+		}
 	}
 
 	if (!control_interface || !data_interface) {

Signature · 1 line

2.26.2

<Daniel Caujolle-Bert>

Tested-by on v2 2/4

2020-09-22 09:53 UTC #3 lore raw

Tested-by Daniel Caujolle-Bert <[email protected]>

Johan Hovold

author

2020-09-21 14:00 UTC #4 lore raw

PATCH v2 3/4 USB: cdc-acm: use common data-class define

In reply to Johan Hovold

Use the data-class define provided by USB core and drop the driver-specific one.

Signed-off-by Johan Hovold <[email protected]>

---
 drivers/usb/class/cdc-acm.c | 6 ++----
 drivers/usb/class/cdc-acm.h | 2 --
 2 files changed, 2 insertions(+), 6 deletions(-)

drivers/usb/class/cdc-acm.c +2 -4

index f42ade505569..a361b937684a 100644
@@ -1287,10 +1287,8 @@ static int acm_probe(struct usb_interface *intf,
 skip_normal_probe:
 
 	/*workaround for switched interfaces */
-	if (data_interface->cur_altsetting->desc.bInterfaceClass
-						!= CDC_DATA_INTERFACE_TYPE) {
-		if (control_interface->cur_altsetting->desc.bInterfaceClass
-						== CDC_DATA_INTERFACE_TYPE) {
+	if (data_interface->cur_altsetting->desc.bInterfaceClass != USB_CLASS_CDC_DATA) {
+		if (control_interface->cur_altsetting->desc.bInterfaceClass == USB_CLASS_CDC_DATA) {
 			dev_dbg(&intf->dev,
 				"Your device has switched interfaces.\n");
 			swap(control_interface, data_interface);

drivers/usb/class/cdc-acm.h +0 -2

index cd5e9d8ab237..b7174a0098a5 100644
@@ -131,8 +131,6 @@ struct acm {
 	unsigned long quirks;
 };
 
-#define CDC_DATA_INTERFACE_TYPE	0x0a
-
 /* constants describing various quirks and errors */
 #define NO_UNION_NORMAL			BIT(0)
 #define SINGLE_RX_URB			BIT(1)

Signature · 1 line

2.26.2

Johan Hovold

author

2020-09-21 14:00 UTC #5 lore raw

PATCH v2 4/4 USB: cdc-acm: clean up no-union-descriptor handling

In reply to Johan Hovold

For interfaces that lack a union descriptor, probe for a "combined-interface" before falling back to the call-management descriptor instead of the other way round.

This allows for the removal of the NO_DATA_INTERFACE quirk and makes the probe algorithm somewhat easier to follow.

Signed-off-by Johan Hovold <[email protected]>

---
 drivers/usb/class/cdc-acm.c | 32 ++++++++++----------------------
 drivers/usb/class/cdc-acm.h | 11 +++++------
 2 files changed, 15 insertions(+), 28 deletions(-)

drivers/usb/class/cdc-acm.c +10 ~~-22~~

index a361b937684a..357e896a4fc0 100644
@@ -1218,26 +1218,19 @@ static int acm_probe(struct usb_interface *intf,
 		call_intf_num = cmgmd->bDataInterface;
 
 	if (!union_header) {
-		if (call_intf_num > 0) {
+		if (intf->cur_altsetting->desc.bNumEndpoints == 3) {
+			dev_dbg(&intf->dev, "No union descriptor, assuming single interface\n");
+			combined_interfaces = 1;
+			control_interface = data_interface = intf;
+			goto look_for_collapsed_interface;
+		} else if (call_intf_num > 0) {
 			dev_dbg(&intf->dev, "No union descriptor, using call management descriptor\n");
-			/* quirks for Droids MuIn LCD */
-			if (quirks & NO_DATA_INTERFACE) {
-				data_interface = usb_ifnum_to_if(usb_dev, 0);
-			} else {
-				data_intf_num = call_intf_num;
-				data_interface = usb_ifnum_to_if(usb_dev, data_intf_num);
-			}
+			data_intf_num = call_intf_num;
+			data_interface = usb_ifnum_to_if(usb_dev, data_intf_num);
 			control_interface = intf;
 		} else {
-			if (intf->cur_altsetting->desc.bNumEndpoints != 3) {
-				dev_dbg(&intf->dev,"No union descriptor, giving up\n");
-				return -ENODEV;
-			} else {
-				dev_warn(&intf->dev,"No union descriptor, testing for castrated device\n");
-				combined_interfaces = 1;
-				control_interface = data_interface = intf;
-				goto look_for_collapsed_interface;
-			}
+			dev_dbg(&intf->dev, "No union descriptor, giving up\n");
+			return -ENODEV;
 		}
 	} else {
 		int class = -1;

@@ -1881,11 +1874,6 @@ static const struct usb_device_id acm_ids[] = {
 
 	/* NOTE: non-Nokia COMM/ACM/0xff is likely MSFT RNDIS... NOT a modem! */
 
-	/* Support for Droids MuIn LCD */
-	{ USB_DEVICE(0x04d8, 0x000b),
-	.driver_info = NO_DATA_INTERFACE,
-	},
-
 #if IS_ENABLED(CONFIG_INPUT_IMS_PCU)
 	{ USB_DEVICE(0x04d8, 0x0082),	/* Application mode */
 	.driver_info = IGNORE_DEVICE,

drivers/usb/class/cdc-acm.h +5 -6

index b7174a0098a5..b2135095898f 100644
@@ -135,9 +135,8 @@ struct acm {
 #define NO_UNION_NORMAL			BIT(0)
 #define SINGLE_RX_URB			BIT(1)
 #define NO_CAP_LINE			BIT(2)
-#define NO_DATA_INTERFACE		BIT(4)
-#define IGNORE_DEVICE			BIT(5)
-#define QUIRK_CONTROL_LINE_STATE	BIT(6)
-#define CLEAR_HALT_CONDITIONS		BIT(7)
-#define SEND_ZERO_PACKET		BIT(8)
-#define DISABLE_ECHO			BIT(9)
+#define IGNORE_DEVICE			BIT(3)
+#define QUIRK_CONTROL_LINE_STATE	BIT(4)
+#define CLEAR_HALT_CONDITIONS		BIT(5)
+#define SEND_ZERO_PACKET		BIT(6)
+#define DISABLE_ECHO			BIT(7)

Signature · 1 line

2.26.2

Oliver Neukum

maintainer

2020-09-21 14:17 UTC #6 lore raw

Re: [PATCH v2 4/4] USB: cdc-acm: clean up no-union-descriptor handling

Am Montag, den 21.09.2020, 15:59 +0200 schrieb Johan Hovold:

Quoted 3 lines

For interfaces that lack a union descriptor, probe for a "combined-interface" before falling back to the call-management descriptor instead of the other way round.

Hi,

the more I look at this the more it seems to me like the device that has the quirk does NOT have a collapsed interface but two interfaces and just a lack of a union descriptor. I am taking the original author into CC (hoping it still workd)

Johan, would just taking the first three patches of the series work for now?

Regards
Oliver

Johan Hovold

author

2020-09-21 14:28 UTC #7 lore raw

Re: [PATCH v2 4/4] USB: cdc-acm: clean up no-union-descriptor handling

Quoted 10 lines · Oliver Neukum

Am Montag, den 21.09.2020, 15:59 +0200 schrieb Johan Hovold:
Quoted 3 lines
For interfaces that lack a union descriptor, probe for a "combined-interface" before falling back to the call-management descriptor instead of the other way round.
Hi,
the more I look at this the more it seems to me like the device that has the quirk does NOT have a collapsed interface but two interfaces and just a lack of a union descriptor.

But then why name the quirk NO_DATA_INTERFACE if it has a data interface? By hardcoding the data-interface number to be the one and only interface, you'd end up probing for a "combined" interface also with a broken call-management descriptor.

Side note: I really think we should start mandating lsusb output to go along with any patch for quirky devices.

Quoted 3 lines

I am taking the original author into CC (hoping it still workd) Johan, would just taking the first three patches of the series work for now?

Of course, no problem holding back the last one for a bit.

Johan

Oliver Neukum

maintainer

2020-09-21 15:04 UTC #8 lore raw

Re: [PATCH v2 4/4] USB: cdc-acm: clean up no-union-descriptor handling

Am Montag, den 21.09.2020, 16:28 +0200 schrieb Johan Hovold:

Quoted 7 lines · Oliver Neukum

Quoted 6 lines
Am Montag, den 21.09.2020, 15:59 +0200 schrieb Johan Hovold:
Quoted 3 lines
For interfaces that lack a union descriptor, probe for a "combined-interface" before falling back to the call-management descriptor instead of the other way round.
Hi,

Hi,

Quoted 6 lines

Quoted 4 lines
the more I look at this the more it seems to me like the device that has the quirk does NOT have a collapsed interface but two interfaces and just a lack of a union descriptor.
But then why name the quirk NO_DATA_INTERFACE if it has a data

In hindsight that seems not the best name.

Quoted 3 lines

interface? By hardcoding the data-interface number to be the one and only interface, you'd end up probing for a "combined" interface also with a broken call-management descriptor.

Well, by the changelog assuming a combined interface caused an oops. Thence I am forced to conclude that the davices _has_ a separate data interface, but no union descriptor.

Quoted 2 lines

Side note: I really think we should start mandating lsusb output to go along with any patch for quirky devices.

Good idea. Convince Greg.

Regards
Oliver

Johan Hovold

author

2020-09-21 15:16 UTC #9 lore raw

Re: [PATCH v2 4/4] USB: cdc-acm: clean up no-union-descriptor handling

Quoted 20 lines · Oliver Neukum

Am Montag, den 21.09.2020, 16:28 +0200 schrieb Johan Hovold:
Quoted 7 lines · Oliver Neukum
Quoted 6 lines
Am Montag, den 21.09.2020, 15:59 +0200 schrieb Johan Hovold:
Quoted 3 lines
For interfaces that lack a union descriptor, probe for a "combined-interface" before falling back to the call-management descriptor instead of the other way round.
Hi,
Hi,
Quoted 6 lines
Quoted 4 lines
the more I look at this the more it seems to me like the device that has the quirk does NOT have a collapsed interface but two interfaces and just a lack of a union descriptor.
But then why name the quirk NO_DATA_INTERFACE if it has a data
In hindsight that seems not the best name.

Quoted 7 lines

Quoted 3 lines
interface? By hardcoding the data-interface number to be the one and only interface, you'd end up probing for a "combined" interface also with a broken call-management descriptor.
Well, by the changelog assuming a combined interface caused an oops. Thence I am forced to conclude that the davices _has_ a separate data interface, but no union descriptor.

No, the oops was probably due to the missing sanity check later added by 403dff4e2c94 ("USB: cdc-acm: check for valid interfaces").

With a broken call-management descriptor pointing to a non-existent interface we'd oops before that commit.

Quoted 4 lines

Quoted 2 lines
Side note: I really think we should start mandating lsusb output to go along with any patch for quirky devices.
Good idea. Convince Greg.

Heh. I'm sure he can be convinced. :)

Johan

Oliver Neukum

maintainer

2020-09-21 17:17 UTC #10 lore raw

Re: [PATCH v2 4/4] USB: cdc-acm: clean up no-union-descriptor handling

Am Montag, den 21.09.2020, 17:16 +0200 schrieb Johan Hovold:

Quoted 8 lines

Quoted 7 lines
Quoted 3 lines
interface? By hardcoding the data-interface number to be the one and only interface, you'd end up probing for a "combined" interface also with a broken call-management descriptor.
Well, by the changelog assuming a combined interface caused an oops. Thence I am forced to conclude that the davices _has_ a separate data interface, but no union descriptor.

Quoted 5 lines

No, the oops was probably due to the missing sanity check later added by 403dff4e2c94 ("USB: cdc-acm: check for valid interfaces").
With a broken call-management descriptor pointing to a non-existent interface we'd oops before that commit.

Hi,

maybe I am dense, but a patch that comes after a patch that is said to fix something? Furthermore that patch would not come it work, it would merely make probe() fail cleanly. If I read the changelog correctly, the change makes the device work.

Regards
Oliver

Johan Hovold

author

2020-09-22 07:05 UTC #11 lore raw

Re: [PATCH v2 4/4] USB: cdc-acm: clean up no-union-descriptor handling

Quoted 15 lines · Oliver Neukum

Am Montag, den 21.09.2020, 17:16 +0200 schrieb Johan Hovold:
Quoted 8 lines
Quoted 7 lines
Quoted 3 lines
interface? By hardcoding the data-interface number to be the one and only interface, you'd end up probing for a "combined" interface also with a broken call-management descriptor.
Well, by the changelog assuming a combined interface caused an oops. Thence I am forced to conclude that the davices _has_ a separate data interface, but no union descriptor.
Quoted 5 lines
No, the oops was probably due to the missing sanity check later added by 403dff4e2c94 ("USB: cdc-acm: check for valid interfaces").
With a broken call-management descriptor pointing to a non-existent interface we'd oops before that commit.

Quoted 6 lines

Hi,
maybe I am dense, but a patch that comes after a patch that is said to fix something? Furthermore that patch would not come it work, it would merely make probe() fail cleanly. If I read the changelog correctly, the change makes the device work.

The relevant commits are

  a2bfb4a346d2 ("USB: support for cdc-acm of single interface devices")      (2009)
  fd5054c169d2 ("USB: cdc_acm: Fix oops when Droids MuIn LCD is connected")  (2011)
  403dff4e2c94 ("USB: cdc-acm: check for valid interfaces")                  (2014)

Before Greg added the sanity checks in that last commit, a broken call-management descriptor referring to a non-existent interface would cause a NULL-pointer dereference.

The second commit, adding support for a specific device, didn't fix that problem generally and only worked around it for one device by hardcoding the data interface to match the control interface, thereby falling back to the "combined-interface" probing you added in that first commit.

See? The second commit fixed the oops *and* made the device probe successfully. If we'd had the sanity check in place by then it would only have make it probe successfully.

Johan

Oliver Neukum

maintainer

2020-09-22 10:41 UTC #12 lore raw

Re: [PATCH v2 4/4] USB: cdc-acm: clean up no-union-descriptor handling

Am Dienstag, den 22.09.2020, 09:05 +0200 schrieb Johan Hovold:

Quoted 9 lines

The relevant commits are
  a2bfb4a346d2 ("USB: support for cdc-acm of single interface devices")      (2009)
  fd5054c169d2 ("USB: cdc_acm: Fix oops when Droids MuIn LCD is connected")  (2011)
  403dff4e2c94 ("USB: cdc-acm: check for valid interfaces")                  (2014) 
Before Greg added the sanity checks in that last commit, a broken call-management descriptor referring to a non-existent interface would cause a NULL-pointer dereference.

Yes.

Quoted 2 lines

The second commit, adding support for a specific device, didn't fix that problem generally

Yes

Quoted 2 lines

and only worked around it for one device by hardcoding
the data interface to match the control interface,

How do you know. It hardcoded the data interface. That it matches the control interface is a guess.

Quoted 2 lines

thereby falling back
to the "combined-interface" probing you added in that first commit.

How do you know? They may or may not match.

Regards
Oliver

Johan Hovold

author

2020-09-22 10:54 UTC #13 lore raw

Re: [PATCH v2 4/4] USB: cdc-acm: clean up no-union-descriptor handling

Quoted 17 lines · Oliver Neukum

Am Dienstag, den 22.09.2020, 09:05 +0200 schrieb Johan Hovold:
Quoted 9 lines
The relevant commits are
  a2bfb4a346d2 ("USB: support for cdc-acm of single interface devices")      (2009)
  fd5054c169d2 ("USB: cdc_acm: Fix oops when Droids MuIn LCD is connected")  (2011)
  403dff4e2c94 ("USB: cdc-acm: check for valid interfaces")                  (2014) 
Before Greg added the sanity checks in that last commit, a broken call-management descriptor referring to a non-existent interface would cause a NULL-pointer dereference.
Yes.
Quoted 2 lines
The second commit, adding support for a specific device, didn't fix that problem generally

Quoted 7 lines

Yes
Quoted 2 lines
and only worked around it for one device by hardcoding
the data interface to match the control interface,
How do you know. It hardcoded the data interface. That it matches the control interface is a guess.

No, see below.

Quoted 4 lines

Quoted 2 lines
thereby falling back
to the "combined-interface" probing you added in that first commit.
How do you know? They may or may not match.

Heh. Did you actually read the commit message?

"Add NO_DATA_INTERFACE quirk to tell the driver that "control"
and "data" interfaces are not separated for this device, which prevents dereferencing a null pointer in the device probe code."

Convinced yet?

Johan

Oliver Neukum

maintainer

2020-09-22 11:42 UTC #14 lore raw

Re: [PATCH v2 4/4] USB: cdc-acm: clean up no-union-descriptor handling

Am Dienstag, den 22.09.2020, 12:54 +0200 schrieb Johan Hovold:

Quoted 8 lines

Heh. Did you actually read the commit message?
"Add NO_DATA_INTERFACE quirk to tell the driver that "control"
and "data" interfaces are not separated for this device, which prevents dereferencing a null pointer in the device probe code."
Convinced yet?

This patch does not fully match its description. Very well. Proceed. Could you resubmit and I'll ack?

Regards
Oliver

Johan Hovold

author

2020-09-22 11:49 UTC #15 lore raw

Re: [PATCH v2 4/4] USB: cdc-acm: clean up no-union-descriptor handling

Quoted 9 lines · Oliver Neukum

Am Dienstag, den 22.09.2020, 12:54 +0200 schrieb Johan Hovold:
Quoted 7 lines
Heh. Did you actually read the commit message?
"Add NO_DATA_INTERFACE quirk to tell the driver that "control"
and "data" interfaces are not separated for this device, which prevents dereferencing a null pointer in the device probe code."

Quoted 4 lines

Quoted 1 line
Convinced yet?
This patch does not fully match its description. Very well. Proceed. Could you resubmit and I'll ack?

I think you can just ack whole series by replying to 0/4. If Greg wants it resubmitted I'll include that ack when resending. Works for you?

Johan

<Daniel Caujolle-Bert>

2020-09-21 14:21 UTC #16 lore raw

In reply to Johan Hovold

Hi Johan,

Yes, sure. Patching....

Cheers.
---
Daniel

<Daniel Caujolle-Bert>

2020-09-21 16:19 UTC #17 lore raw

In reply to Johan Hovold

Hi Johan, Oliver,

I just compiled and tested, everything still works.

Cheers.
---
Daniel

Johan Hovold

author

2020-09-22 07:08 UTC #18 lore raw

Quoted 3 lines · On Mon, Sep 21, 2020 at 06:19:12PM +0200, <Daniel Caujolle-Bert> wrote:

Hi Johan, Oliver,
I just compiled and tested, everything still works.

Thanks for testing, Daniel.

If you want to you can reply to the second patch with a Tested-by tag so that it gets added by Greg's tooling (or reply to the cover letter if you want to have that tag added to all the patches in the series).

Johan

<Daniel Caujolle-Bert>

2020-09-22 09:56 UTC #19 lore raw

Hi Johan,

Okay, I replied to the second patch, hope I didn't make any mistake.

Cheers.
---
Daniel

Johan Hovold

author

2020-09-22 10:07 UTC #20 lore raw

Quoted 3 lines · On Tue, Sep 22, 2020 at 11:56:02AM +0200, <Daniel Caujolle-Bert> wrote:

Hi Johan,
Okay, I replied to the second patch, hope I didn't make any mistake.

Looks good. Thanks again!

Johan

Oliver Neukum

maintainer Acked-by on the v2 series

2020-09-22 12:11 UTC #21 lore raw

In reply to Johan Hovold

Am Montag, den 21.09.2020, 15:59 +0200 schrieb Johan Hovold:

Quoted 10 lines

This series adds support for handling broken union descriptors by falling back to "combined-interface" probing.
The first patch drops some bogus altsetting sanity checks which would otherwise have had to be needlessly reproduced for consistency. The third patch drops the driver specific data class define in favour of the common one. The last one, cleans up the no-union-descriptor handling by probing for a "combined-interface" before falling back to the call-management descriptor.

Quoted 3 lines

Note that I changed my mind on the stable tag; we can't be overly paranoid about a theoretical risk of breaking some quirky devices. And if we do, we still want to know about it, right?

Acked-by Oliver Neukum <[email protected]>