[PATCH] cdc-acm: prevent infinite loop when parsing CDC headers.

[PATCH] cdc-acm: prevent infinite loop when parsing CDC headers.

[Quentin Casasnovas]

Phil and I found out a problem with commit:

  7e860a6e ("cdc-acm: add sanity checks")

It added some sanity checks to ignore potential garbage in CDC headers but
also introduced a potential infinite loop.  This can happen at the first
loop iteration (elength = 0 in that case) if the description isn't a
DT_CS_INTERFACE or later if 'buffer[0]' is zero.

It should also be noted that the wrong length was being added to 'buffer'
in case 'buffer[1]' was not a DT_CS_INTERFACE descriptor, since elength was
assigned after that check in the loop.

A specially crafted USB device could be used to trigger this infinite loop.

Fixes: 7e860a6e ("cdc-acm: add sanity checks")
Signed-off-by: Phil Turnbull <phil.turnbull@oracle.com>
Signed-off-by: Quentin Casasnovas <quentin.casasnovas@oracle.com>
CC: Oliver Neukum <oneukum@suse.de>
CC: Adam Lee <adam8157@gmail.com>
---
 drivers/usb/class/cdc-acm.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/drivers/usb/class/cdc-acm.c b/drivers/usb/class/cdc-acm.c
index 6836177..1ac4587 100644
--- a/drivers/usb/class/cdc-acm.c
+++ b/drivers/usb/class/cdc-acm.c
@@ -1133,11 +1133,12 @@ static int acm_probe(struct usb_interface *intf,
 	}
 
 	while (buflen > 0) {
+		if ((elength = buffer[0]) == 0)
+			break;
 		if (buffer[1] != USB_DT_CS_INTERFACE) {
 			dev_err(&intf->dev, "skipping garbage\n");
 			goto next_desc;
 		}
-		elength = buffer[0];
 
 		switch (buffer[2]) {
 		case USB_CDC_UNION_TYPE: /* we've found it */
-- 
2.0.5

Re: [PATCH] cdc-acm: prevent infinite loop when parsing CDC headers.

[Adam Lee]

On Mon, Apr 13, 2015 at 05:24:04PM +0200, Quentin Casasnovas wrote:
> Phil and I found out a problem with commit:
> 
>   7e860a6e ("cdc-acm: add sanity checks")
> 
> It added some sanity checks to ignore potential garbage in CDC headers but
> also introduced a potential infinite loop.  This can happen at the first
> loop iteration (elength = 0 in that case) if the description isn't a
> DT_CS_INTERFACE or later if 'buffer[0]' is zero.
> 
> It should also be noted that the wrong length was being added to 'buffer'
> in case 'buffer[1]' was not a DT_CS_INTERFACE descriptor, since elength was
> assigned after that check in the loop.
> 
> A specially crafted USB device could be used to trigger this infinite loop.

Yes, "elength = buffer[0]" should be moved to the front of
USB_DT_CS_INTERFACE check, ACK this part.

But I don't know in what case the buffer[0] could be zero, if it
happens, better to set elength 1 then goto next_desc? (I'm not a
maintainer, pleas also consider others' opinion)

-- 
Adam Lee
http://adam8157.info

Re: [PATCH] cdc-acm: prevent infinite loop when parsing CDC headers.

[Quentin Casasnovas]

Adding Greg on CC as suggested by Oliver.

On Mon, Apr 13, 2015 at 05:24:04PM +0200, Quentin Casasnovas wrote:
> Phil and I found out a problem with commit:
> 
>   7e860a6e ("cdc-acm: add sanity checks")
> 
> It added some sanity checks to ignore potential garbage in CDC headers but
> also introduced a potential infinite loop.  This can happen at the first
> loop iteration (elength = 0 in that case) if the description isn't a
> DT_CS_INTERFACE or later if 'buffer[0]' is zero.
> 
> It should also be noted that the wrong length was being added to 'buffer'
> in case 'buffer[1]' was not a DT_CS_INTERFACE descriptor, since elength was
> assigned after that check in the loop.
> 
> A specially crafted USB device could be used to trigger this infinite loop.
> 
> Fixes: 7e860a6e ("cdc-acm: add sanity checks")
> Signed-off-by: Phil Turnbull <phil.turnbull@oracle.com>
> Signed-off-by: Quentin Casasnovas <quentin.casasnovas@oracle.com>
> CC: Oliver Neukum <oneukum@suse.de>
> CC: Adam Lee <adam8157@gmail.com>
> ---
>  drivers/usb/class/cdc-acm.c | 3 ++-
>  1 file changed, 2 insertions(+), 1 deletion(-)
> 
> diff --git a/drivers/usb/class/cdc-acm.c b/drivers/usb/class/cdc-acm.c
> index 6836177..1ac4587 100644
> --- a/drivers/usb/class/cdc-acm.c
> +++ b/drivers/usb/class/cdc-acm.c
> @@ -1133,11 +1133,12 @@ static int acm_probe(struct usb_interface *intf,
>  	}
>  
>  	while (buflen > 0) {
> +		if ((elength = buffer[0]) == 0)
> +			break;
>  		if (buffer[1] != USB_DT_CS_INTERFACE) {
>  			dev_err(&intf->dev, "skipping garbage\n");
>  			goto next_desc;
>  		}
> -		elength = buffer[0];
>  
>  		switch (buffer[2]) {
>  		case USB_CDC_UNION_TYPE: /* we've found it */
> -- 
> 2.0.5
> 

Re: [PATCH] cdc-acm: prevent infinite loop when parsing CDC headers.

[Quentin Casasnovas]

On Mon, Apr 13, 2015 at 11:48:27PM +0800, Adam Lee wrote:
> On Mon, Apr 13, 2015 at 05:24:04PM +0200, Quentin Casasnovas wrote:
> > Phil and I found out a problem with commit:
> > 
> >   7e860a6e ("cdc-acm: add sanity checks")
> > 
> > It added some sanity checks to ignore potential garbage in CDC headers but
> > also introduced a potential infinite loop.  This can happen at the first
> > loop iteration (elength = 0 in that case) if the description isn't a
> > DT_CS_INTERFACE or later if 'buffer[0]' is zero.
> > 
> > It should also be noted that the wrong length was being added to 'buffer'
> > in case 'buffer[1]' was not a DT_CS_INTERFACE descriptor, since elength was
> > assigned after that check in the loop.
> > 
> > A specially crafted USB device could be used to trigger this infinite loop.
> 
> Yes, "elength = buffer[0]" should be moved to the front of
> USB_DT_CS_INTERFACE check, ACK this part.
> 
> But I don't know in what case the buffer[0] could be zero, if it
> happens, better to set elength 1 then goto next_desc? (I'm not a
> maintainer, pleas also consider others' opinion)
> 

Hi Adam,

I'm happy to change it as you suggest, though at that point we'll probably
be trying to parse garbage anyway.

If nobody had a different opinion in the meantime, I'll send a v2 tomorrow.

Thanks for the review :)

Quentin

Re: [PATCH] cdc-acm: prevent infinite loop when parsing CDC headers.

[Sergei Shtylyov]

Hello.

On 04/13/2015 06:24 PM, Quentin Casasnovas wrote:

> Phil and I found out a problem with commit:

>    7e860a6e ("cdc-acm: add sanity checks")

> It added some sanity checks to ignore potential garbage in CDC headers but
> also introduced a potential infinite loop.  This can happen at the first
> loop iteration (elength = 0 in that case) if the description isn't a
> DT_CS_INTERFACE or later if 'buffer[0]' is zero.

> It should also be noted that the wrong length was being added to 'buffer'
> in case 'buffer[1]' was not a DT_CS_INTERFACE descriptor, since elength was
> assigned after that check in the loop.

> A specially crafted USB device could be used to trigger this infinite loop.

> Fixes: 7e860a6e ("cdc-acm: add sanity checks")

    12-digit SHA1 hash is required here.

> Signed-off-by: Phil Turnbull <phil.turnbull@oracle.com>
> Signed-off-by: Quentin Casasnovas <quentin.casasnovas@oracle.com>
> CC: Oliver Neukum <oneukum@suse.de>
> CC: Adam Lee <adam8157@gmail.com>
> ---
>   drivers/usb/class/cdc-acm.c | 3 ++-
>   1 file changed, 2 insertions(+), 1 deletion(-)

> diff --git a/drivers/usb/class/cdc-acm.c b/drivers/usb/class/cdc-acm.c
> index 6836177..1ac4587 100644
> --- a/drivers/usb/class/cdc-acm.c
> +++ b/drivers/usb/class/cdc-acm.c
> @@ -1133,11 +1133,12 @@ static int acm_probe(struct usb_interface *intf,
>   	}
>
>   	while (buflen > 0) {
> +		if ((elength = buffer[0]) == 0)

    Please run your patches thru scripts/checkpatch.pl. Assignments in the 
*if* operator are not allowed.

[...]

WBR, Sergei

[PATCH v2] cdc-acm: prevent infinite loop when parsing CDC headers.

[Quentin Casasnovas]

Phil and I found out a problem with commit:

  7e860a6e7aa6 ("cdc-acm: add sanity checks")

It added some sanity checks to ignore potential garbage in CDC headers but
also introduced a potential infinite loop.  This can happen at the first
loop iteration (elength = 0 in that case) if the description isn't a
DT_CS_INTERFACE or later if 'buffer[0]' is zero.

It should also be noted that the wrong length was being added to 'buffer'
in case 'buffer[1]' was not a DT_CS_INTERFACE descriptor, since elength was
assigned after that check in the loop.

A specially crafted USB device could be used to trigger this infinite loop.

v2:
 - Use 12-digits sha1 to reference the offending commit.
 - Do not break from the loop and try next byte instead.
 - Move the assignment outside the 'if'.
 - Add a debug print.

Fixes: 7e860a6e7aa6 ("cdc-acm: add sanity checks")
Signed-off-by: Phil Turnbull <phil.turnbull@oracle.com>
Signed-off-by: Quentin Casasnovas <quentin.casasnovas@oracle.com>
CC: Sergei Shtylyov <sergei.shtylyov@cogentembedded.com>
CC: Oliver Neukum <oneukum@suse.de>
CC: Adam Lee <adam8157@gmail.com>
CC: <stable@vger.kernel.org>
---
 drivers/usb/class/cdc-acm.c | 7 ++++++-
 1 file changed, 6 insertions(+), 1 deletion(-)

diff --git a/drivers/usb/class/cdc-acm.c b/drivers/usb/class/cdc-acm.c
index 6836177..220c0fd 100644
--- a/drivers/usb/class/cdc-acm.c
+++ b/drivers/usb/class/cdc-acm.c
@@ -1133,11 +1133,16 @@ static int acm_probe(struct usb_interface *intf,
 	}
 
 	while (buflen > 0) {
+		elength = buffer[0];
+		if (!elength) {
+			dev_err(&intf->dev, "skipping garbage byte\n");
+			elength = 1;
+			goto next_desc;
+		}
 		if (buffer[1] != USB_DT_CS_INTERFACE) {
 			dev_err(&intf->dev, "skipping garbage\n");
 			goto next_desc;
 		}
-		elength = buffer[0];
 
 		switch (buffer[2]) {
 		case USB_CDC_UNION_TYPE: /* we've found it */
-- 
2.0.5

Re: [PATCH] cdc-acm: prevent infinite loop when parsing CDC headers.

[Sergei Shtylyov]

Hello.

On 4/13/2015 7:20 PM, Sergei Shtylyov wrote:

>> Phil and I found out a problem with commit:

>>    7e860a6e ("cdc-acm: add sanity checks")

>> It added some sanity checks to ignore potential garbage in CDC headers but
>> also introduced a potential infinite loop.  This can happen at the first
>> loop iteration (elength = 0 in that case) if the description isn't a
>> DT_CS_INTERFACE or later if 'buffer[0]' is zero.
>
>> It should also be noted that the wrong length was being added to 'buffer'
>> in case 'buffer[1]' was not a DT_CS_INTERFACE descriptor, since elength was
>> assigned after that check in the loop.

>> A specially crafted USB device could be used to trigger this infinite loop.

>> Fixes: 7e860a6e ("cdc-acm: add sanity checks")

>     12-digit SHA1 hash is required here.

>> Signed-off-by: Phil Turnbull <phil.turnbull@oracle.com>
>> Signed-off-by: Quentin Casasnovas <quentin.casasnovas@oracle.com>
>> CC: Oliver Neukum <oneukum@suse.de>
>> CC: Adam Lee <adam8157@gmail.com>
>> ---
>>   drivers/usb/class/cdc-acm.c | 3 ++-
>>   1 file changed, 2 insertions(+), 1 deletion(-)

>> diff --git a/drivers/usb/class/cdc-acm.c b/drivers/usb/class/cdc-acm.c
>> index 6836177..1ac4587 100644
>> --- a/drivers/usb/class/cdc-acm.c
>> +++ b/drivers/usb/class/cdc-acm.c
>> @@ -1133,11 +1133,12 @@ static int acm_probe(struct usb_interface *intf,
>>       }
>>
>>       while (buflen > 0) {
>> +        if ((elength = buffer[0]) == 0)

>     Please run your patches thru scripts/checkpatch.pl. Assignments in the
> *if* operator are not allowed.

    s/operator/statement/, of course. :-)

> [...]

WBR, Sergei

Re: [PATCH v2] cdc-acm: prevent infinite loop when parsing CDC headers.

[Quentin Casasnovas]

On Tue, Apr 14, 2015 at 11:25:43AM +0200, Quentin Casasnovas wrote:
> Phil and I found out a problem with commit:
> 
>   7e860a6e7aa6 ("cdc-acm: add sanity checks")
> 

Any comment on v2?

Thanks,
Quentin

Re: [PATCH v2] cdc-acm: prevent infinite loop when parsing CDC headers.

[Greg Kroah-Hartman]

On Mon, Apr 20, 2015 at 01:54:56PM +0200, Quentin Casasnovas wrote:
> On Tue, Apr 14, 2015 at 11:25:43AM +0200, Quentin Casasnovas wrote:
> > Phil and I found out a problem with commit:
> > 
> >   7e860a6e7aa6 ("cdc-acm: add sanity checks")
> > 
> 
> Any comment on v2?

I can't do anything with any patches until 4.1-rc1 is released so please
just be patient.