On Tue, 9 Jan 2007, Phil Endecott wrote: > Gadgetfs had a mode in which endpoint descriptors were written by > the user program before connection. This mode had some bugs, and > may never have been used. This patch removes that mode. The only > mode of operation now is for the user program to wait for endpoint > 0 to report a SET_CONFIGURATION, and then write the endpoint > descriptors. > > With this change, testusb works equally well as before for me: test > 10 fails. I am now wondering if this is a host problem, and would > appreciate feedback about whether testusb works for other people > after applying this patch.
Phil: I haven't tried testing your patch. But I have tried using the existing gadgetfs with net2280 and uhci-hcd. There was no trouble running test #10 with the usb program or with g_zero instead of gadgetfs. I can't imagine what's going wrong on your system. On the other hand, I did find (another) bug in the async I/O handling in inode.c. If you aren't using the AIO code in usb.c then it shouldn't affect you, but you might want to give it a try anyway. Alan Stern Index: usb-2.6/drivers/usb/gadget/inode.c =================================================================== --- usb-2.6.orig/drivers/usb/gadget/inode.c +++ usb-2.6/drivers/usb/gadget/inode.c @@ -586,7 +586,6 @@ static ssize_t ep_aio_read_retry(struct } kfree(priv->buf); kfree(priv); - aio_put_req(iocb); return len; } ------------------------------------------------------------------------- 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 _______________________________________________ linux-usb-devel@lists.sourceforge.net To unsubscribe, use the last form field at: https://lists.sourceforge.net/lists/listinfo/linux-usb-devel