> > When we load the File backed storage gadget (FBSG) driver (with a good > > backing file) and connect the USB cable to a system (we've tested Mac > > and Windows) the OS is never fully able to install the drive and gives > > up after 3-5 minutes of trying. Eventually, I found through debug > > prints if we have a single general INFO print in the main while loop > > in fsg_main_thread(), the host OS is able to correctly install the > > drive (in < 10 seconds) and things work ok. What's more, if I add a > > check make the INFO prints in fsg_main_thread() stop printing after > > 400 prints, things keep working fine after that (suggesting there is > > some sort of problem with connection initialization only). > > > > Have you ever heard of anything similar? Any ideas on what the > > problem could be? > > What gadget controller driver is this, and what kernel version? > > Have you looked at the USB packets to see if there is some odd timing > issue going on here? What's the difference in them between a working > and not working run? ... > Sorry for not replying sooner. We're using the OMAP 3530 > musb port/controller (the phy is in the tps power controller) > in device mode (not OTG). The kernel we're using is 2.6.32 > (from rowboat android éclair). What parameters do you pass to g_file_storage when you load it? In particular, do you use stall=n? I'm not sure why we need to pass this, but if it fixes things for you, maybe we can go look at that code and see why. - Anand -- To unsubscribe from this list: send the line "unsubscribe linux-usb" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html
