Re: PROBLEM : Device Re-enumeration not happening after reset

On Fri, 26 Oct 2007, Jithu Joseph wrote:

> Dear All,
> 
> I am trying to implement a firmware upgrade for my usb smart card
> reader. (device firmware update procedure : I issue a switch to DFU
> control message and issue a usb_reset which makes the device
> re-enumerate as a DFU device with different VID/PID, after which I
> proceed with the f/w download)
> 
> I am having a kernel level usb driver for my device (USB Smart card
> reader). When a user level  application triggers a firm ware update :
> 1.  i get the control in my kernel driver.
> 2.  I send the switch to DFU control message to my device  (which is
> ack ed by the device)
> 3.  i issue a reset using usb_reset_device (defined in hub.c)
> 
> The problem i am facing is that after issuing the  "usb_reset_device
> (struct usb_device *dev) ", though i am able to see a reset in the
> CATC( USB protocol Analyzer) , the disconnect of the current usb
> client driver  is not getting called  and also the probe of the DFU
> driver is not getting called (the driver has support for both the
> VIDs/PIDs)
> 
> Am i missing something - should i somehow notify the subystem to
> reenumerate the port or am i using the wrong call to reset
> 
> At the CATC level i am seeing a reset, and the same address is
> reassigned to the device and also get descriptors are issued.(and the
> device returns descriptors with the new VID/PID). The set address and
> get decriptors are called from usb_reset_device . -
> 
> But why is the USB subsystem not detecting the device removal and
> unloading the current driver and loading the DFU driver ?
> 
> // --- i have a similar DFU implementation for the same device using
> libusb which works  - ie without any kernel level usb drivers.
> 
> There the sequence of events is
> 
> The dfu app issues usb_reset(usb_dev_handle *dev)  (libusb call)
> 
> The usb reset occurs on bus, reenumeration occours and the next higer
> usb address is assigned.
>  i loop for my DFU VID/PID and open the handle and proceed. I couldn't
> make out which kernel api's libusb is using to trigger the
> reenumeration sequence
> 
> My environment Linux 2.4.27 on AT91 boards
> I hope I am posting to the right forum. If not, please re-direct me to
> the appropriate one.

The problem is very simple: Reenumeration after a reset wasn't 
implemented in 2.4!  It's present only in the 2.6 kernel.

Alan Stern


-------------------------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc.
Still grepping through log files to find problems?  Stop.
Now Search log events and configuration files using AJAX and a browser.
Download your FREE copy of Splunk now >> http://get.splunk.com/
_______________________________________________
[email protected]
To unsubscribe, use the last form field at:
https://lists.sourceforge.net/lists/listinfo/linux-usb-devel