Alan Stern wrote: > On Tue, 24 Aug 2010, Murphy, Robert wrote: > > > > What happens when you plug your gadget into a host running Linux? Does > > > it behave any differently? > > No, we see the same problem if we plug it into a Linux host. > > Then you should capture a usbmon trace on the Linux host and post it > here. It will give us a good hint as to where the problem lies. > Instructions are in the kernel source file > Documentation/usb/usbmon.txt. > I quickly tried this with v2.6.36-rc1 on an OMAP4 board (similar MUSB controller as the OMAP3 that Robert is running). I see that with stall=y, I have repeated messages like this on the host, a little after enumeration: [2091018.128077] usb 1-5: reset high speed USB device using ehci_hcd and address 71 Assuming Robert's problem is due to the same cause, here're a couple of usbmon traces from my test - one with stall=n passed to g_file_storage, and one without. (I've added 2 blank lines in the stall=y log, just before I got the reset message). The corresponding dmesg logs, are at the top of the same traces. - Anand
Attachment:
usbmon-stall-y.log
Description: usbmon-stall-y.log
Attachment:
usbmon-stall-n.log
Description: usbmon-stall-n.log
