Re: [Barry-devel] app downloading

2008-03-02 Thread Niels de Vos
Hello,

On 3/1/08, Brian Edginton [EMAIL PROTECTED] wrote:
  If/When you get useful USB captures, feel free to pass them along to me,
  and we can put them somewhere public... either in the sourceforge wiki
  or CVS.
 

 So, what have you found to be the best way to snoop the USB? Is windows
 better than Linux? Is there a tool that produces the same output as you have
 used in the past? Have you standardized anything?

At work we have a Ellisys USB Explorer. This is a hardware USB
Analyser and I could create a trace if you're interested. However the
software for viewing is only available for Windows (free download).

All I need is a URL for a little/cool java app for installing on my BB 8100.

Cheers,
Niels

-
This SF.net email is sponsored by: Microsoft
Defy all challenges. Microsoft(R) Visual Studio 2008.
http://clk.atdmt.com/MRT/go/vse012070mrt/direct/01/
___
Barry-devel mailing list
Barry-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/barry-devel


Re: [Barry-devel] app downloading

2008-03-01 Thread Chris Frey
On Sat, Mar 01, 2008 at 05:28:39AM -0700, Brian Edginton wrote:
 So, what have you found to be the best way to snoop the USB? Is windows
 better than Linux? Is there a tool that produces the same output as you have
 used in the past? Have you standardized anything?

I've used windows snooping and linux snooping.  They are both good, but
linux captures more of the protocol than windows... I had to use linux
snooping in order to find the special usb control messages for bcharge,
for example.  If the traffic you're trying to see only uss Bulk messages
(which is the case for most of what the Blackberry does, that I've seen),
then windows snooping is fine.

One drawback of linux snooping is that I've found on my system, sometimes
it seems like the kernel printk message buffering gets overwhelmed, and
somewhere along the line from kernel printk to syslog /var/log/kern.log
some data is dropped.  This may be a peculiarity of my system, but I haven't
tracked it down.

Both methods are documented under barry/doc/USB-capture.txt.

Another issue with linux snooping is that it tries to log the entire URB
packet, whether or not the data is being used at all... this makes following
the traffic a little confusing, since a data packet might be logged that
really has nothing to do with reality.  You can tell this by the context,
but you have to concentrate.

But the nice thing about linux snooping is that you have the source code
to everything. ;-)  I've attached a patch to kernel 2.6.24.3 that adds
endpoint logging and tries to limit the data packet size that is written to
the log.

- Chris

diff -ru linux-2.6.24.3-vanilla/drivers/usb/core/devio.c 
linux-2.6.24.3-cdfsnooplimit/drivers/usb/core/devio.c
--- linux-2.6.24.3-vanilla/drivers/usb/core/devio.c 2008-02-25 
19:20:20.0 -0500
+++ linux-2.6.24.3-cdfsnooplimit/drivers/usb/core/devio.c   2008-02-29 
17:12:24.0 -0500
@@ -284,7 +284,7 @@
 
 static void snoop_urb(struct urb *urb, void __user *userurb)
 {
-   int j;
+   int j, len;
unsigned char *data = urb-transfer_buffer;
 
if (!usbfs_snoop)
@@ -297,7 +297,12 @@
 urb-transfer_buffer_length);
dev_info(urb-dev-dev, actual_length=%d\n, urb-actual_length);
dev_info(urb-dev-dev, data: );
-   for (j = 0; j  urb-transfer_buffer_length; ++j)
+
+   len = urb-actual_length;
+   if( len  urb-transfer_buffer_length )
+   len = urb-transfer_buffer_length;
+
+   for (j = 0; j  len; ++j)
printk (%02x , data[j]);
printk(\n);
 }
@@ -726,8 +731,8 @@
kfree(tbuf);
return -EINVAL;
}
-   snoop(dev-dev, bulk read: len=0x%02x timeout=%04d\n,
-   bulk.len, bulk.timeout);
+   snoop(dev-dev, bulk read: ep=0x%02x, len=0x%02x 
timeout=%04d\n,
+   bulk.ep, bulk.len, bulk.timeout);
usb_unlock_device(dev);
i = usb_bulk_msg(dev, pipe, tbuf, len1, len2, tmo);
usb_lock_device(dev);
@@ -750,8 +755,8 @@
return -EFAULT;
}
}
-   snoop(dev-dev, bulk write: len=0x%02x timeout=%04d\n,
-   bulk.len, bulk.timeout);
+   snoop(dev-dev, bulk write: ep=0x%02x, len=0x%02x 
timeout=%04d\n,
+   bulk.ep, bulk.len, bulk.timeout);
if (usbfs_snoop) {
dev_info(dev-dev, bulk write: data: );
for (j = 0; j  len1; ++j)
@@ -1092,6 +1097,8 @@
return -EFAULT;
}
}
+   snoop(ps-dev-dev, submit urb: bEndpoint %02x\n,
+   uurb-endpoint);
snoop_urb(as-urb, as-userurb);
 async_newpending(as);
 if ((ret = usb_submit_urb(as-urb, GFP_KERNEL))) {
Only in linux-2.6.24.3-cdfsnooplimit: tags
-
This SF.net email is sponsored by: Microsoft
Defy all challenges. Microsoft(R) Visual Studio 2008.
http://clk.atdmt.com/MRT/go/vse012070mrt/direct/01/___
Barry-devel mailing list
Barry-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/barry-devel


[Barry-devel] app downloading

2008-02-29 Thread Brian Edginton
If no one else is too far into it, or if they would like some help, I'm
going to start looking at the java loader  protocol.

-edge
-
This SF.net email is sponsored by: Microsoft
Defy all challenges. Microsoft(R) Visual Studio 2008.
http://clk.atdmt.com/MRT/go/vse012070mrt/direct/01/___
Barry-devel mailing list
Barry-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/barry-devel