On Wednesday 03 January 2007 3:30 am, Phil Endecott wrote:
> 
> > Something to be aware of:  not all gadget controller drivers seem to
> > handle control-OUT transfers correctly ... especially in the "deferred
> > response" mode needed by gadgetfs.  You can test control-OUT transfers
> > using "usbtest" test case #14; and g_file_storage uses deferred response
> > mode, but ISTR not for control-OUT transfers.  You might be hitting a
> > controller driver (or controller!) bug specific to that combination...
> 
> I'm using a Net2280 PCI card. 

That "should" be a pretty solid combo, but I don't know that anyone's
run the full test battery against it for a few years now; which likely
means a few odd behaviors have snuck in, along with support for newer
hardware and bugfixes.


> With g_zero, test 14 passes when I run  
> it as test.sh does:
> 
> # ./testusb -D /proc/bus/usb/001/005 -t 14 -c 15000 -s 256 -v 1
> unknown speed /proc/bus/usb/001/005
> /proc/bus/usb/001/005 test 14,    6.164763 secs
> 
> Tests 0-13 pass (with g_zero) when run with default settings.

That's all good.


> With gadgetfs and usb.c, tests 0-9 pass, and it hangs on test 10.  Is 
> that expected?

No.  It used to work.  (And quite a PITA it was to get it working...
a mix of requests handled in-kernel, and handed out to userspace.)

Where does it hang, and why?  What kernel release with net2280.c?
(Suggest a separate thread for such discussion.)  I think using the
"verbose" option of "usb.c" on the peripheral side, and dmesg on
the host side (plus maybe alt-sysrq-t to see specifically what
usbtest and usb.c are waiting for) would be appropriately informative.

 
> Are you suggesting that I should run testusb against g_file_storage in 
> order to prove whether the system would support control OUT if my 
> gadgetfs app needed them?

No, just pointing out things you can test; if you ran into a problem
with those, it could explain the other issues you're seeing.  That
issue with test #10, for example, presumably means some change to the
net2280 or gadgetfs code misbehaves in some race scenario ... it'd be
slightly strange to have it interact with control-OUT transfers, but
I've certainly seen far stranger bug interactions.

- Dave


-------------------------------------------------------------------------
Take Surveys. Earn Cash. Influence the Future of IT
Join SourceForge.net's Techsay panel and you'll get the chance to share your
opinions on IT & business topics through brief surveys - and earn cash
http://www.techsay.com/default.php?page=join.php&p=sourceforge&CID=DEVDEV
_______________________________________________
[email protected]
To unsubscribe, use the last form field at:
https://lists.sourceforge.net/lists/listinfo/linux-usb-devel

Reply via email to