> I am not sure about this, but I wonder can we do the reset in > cp210x_open() instead of in cp210x_startup()? Possibly, but I'd rather perform the reset just once, right when the device is detected and bound to the driver. After doing some more reading, I've found that calling `usb_lock_device_for_reset` just before `usb_reset_device` seems to work. Here's a possible patch: --- cp210x.c +++ cp210x.c.new.c @@ -856,7 +856,8 @@ struct cp210x_serial_private *spriv; /* cp210x buffers behave strangely unless device is reset */ - usb_reset_device(serial->dev); + if(usb_lock_device_for_reset(serial->dev, NULL) == 0) + usb_reset_device(serial->dev); spriv = kzalloc(sizeof(*spriv), GFP_KERNEL); if (!spriv) The downside here is that the reset takes a few hundred extra milliseconds, and the reset is issued for each interface. But, at least both interfaces are properly detected the first time. :) Any thoughts? Thanks, Blake On Fri, May 16, 2014 at 2:47 AM, Jidong Xiao <jidong.xiao@xxxxxxxxx> wrote: > On Thu, May 15, 2014 at 6:53 PM, Miner, Blake <bminer@xxxxxxxxxxxxxxxx> wrote: >> After debugging further, I figured out what was causing the problem. >> >> In the driver for the cp210x, there is a line of code in the `.attach` >> function that calls `usb_reset_device(serial->dev);`. See the source >> code here: https://github.com/torvalds/linux/blob/master/drivers/usb/serial/cp210x.c#L859 >> >> This call to `usb_reset_device` causes the device to be reset just >> before the second interface is detected. For whatever reason, the >> reset prevents the second interface from being created in `/dev/`. >> >> By simply removing this line, both interfaces are properly detected >> the first time. Unfortunately, there is a comment above this line >> that reads "/* cp210x buffers behave strangely unless device is reset >> */" Therefore, this reset might actually be necessary in order to >> clear the buffers of the CP210x. In this case, when is the >> appropriate time to perform a device reset so that both interfaces are >> properly detected? >> >> Thanks, > > I am not sure about this, but I wonder can we do the reset in > cp210x_open() instead of in cp210x_startup()? > > -Jidong -- 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
