Re: kmemleak report on isp1763 and sierra MC8705

[Richard Retanubun <richardretanubun@ruggedcom.com>]

On 26/10/12 07:35 PM, Greg KH wrote:
> On Fri, Oct 26, 2012 at 05:57:23PM -0400, Richard Retanubun wrote:
>> Hi Guys,
>>
>> I am debugging a reported kmemleak involving a sierra wireless MC8705 connected
>> through isp1763 on powerpc linux-3.0.22
>
> Does this also happen on 3.6.3?
>
> thanks,
>
> greg k-h

Hi Greg,

Unfortunately, it is not trivial for us to update the kernel on this platform,
Is there a specific experiment/patch I should look at for 3.0.22?

I will be attempting to use 3.6.3 on another platform, but this may take some time.

I am thinking there may be an action that can be done at /sysfs or /procfs to
do the disconnect without actually removing the power to the device.
I tried "echo 1 > /sys/bus/usb/drivers/usb/2-1.2/remove" and then take down the power
but this produced the same leak signature even before I take down the power.

Update on trigger to problem
============================
This will happen as the modem is powered down and /dev/ttyUSB from sierra is teared down
either by powering it off/removing it, or sending at!reset.

It does not happen when the same thing is done using a simple usb to serial converter (pl2303)

Focusing down on one of the dumps:

unreferenced object 0xd3849740 (size 8):
   comm "khubd", pid 1026, jiffies 232553037 (age 506.597s)
   hex dump (first 8 bytes):
     4d 43 38 37 30 35 00 00                          MC8705..
   backtrace:
     [<e30efd74>] usb_cache_string+0x74/0xac [usbcore]
     [<e30e77bc>] usb_enumerate_device+0x44/0xf8 [usbcore]
     [<e30e7aa0>] usb_new_device+0x3c/0x13c [usbcore]
     [<e30e9824>] hub_thread+0xc8c/0x1544 [usbcore]
     [<c0043aa8>] kthread+0x7c/0x80
     [<c000ed48>] kernel_thread+0x4c/0x68

I have a small question. How does the memory kmalloc-ed() in usb_cache_string is supposed to be released?
(during usb_serial_disconnect()?) Is the sierra driver is supposed to participate
in the tear down process (in sierra_release() maybe) and not doing something that is expected?
I am still missing the link between the actions done by the hub_thread() for the caching the stings
and the sierra driver code.

Thanks a lot for your time.