Hi, On Wed, 2011-07-13 at 10:37 +0200, Uwe Bonnes wrote: > Ionut> Now, one problem I see in ftdi_2232h_baud_base_to_divisor is that > Ionut> another bit is set (not MSB): > > Ionut> divisor |= 0x00020000; > > Ionut> I think it should be: > > Ionut> divisor |= 0x80000000; > > But at 0x0001000 you have the high bit of the encoded fractional divider. > > Until now, the bit at 0x0001000 and 0x00020000 are moved for devices with > several Uarts to 0x0100000 and 0x2000000 later, so you solution would impact > other devices. > > See my solution in an earlier post in this thread. > I saw your previous post now. Yes, there's 1 bit of the fractional divider in the upper 16-bit half of the 32-bit encoded divisor. The other two are in the lower 16-bit half so they will get into urb_value together with the integer divisor. I have to admit I'm a little confused about how the FTDI_SIO_SET_BAUDRATE_REQUEST command should work: * how do you send a 32 bit result with just one usb_control_message? * isn't encoding the first 16-bit half into urb_index wrong? Shouldn't urb_index contain the interface/channel number for multi-channel devices? * is there any documentation for this protocol (the control messages exchanged with the device over usb)? I couldn't find any ... I still think MSB needs to be set. If you look at their example in AN120, section 4.4 for 38400 bps: 38,41,00,80 => divisor = 312.5, rate = 38,400 The u32 divisor is encoded as 0x80004138. This decodes to div_frac = 001 => 0.5, div_int = 0x138 = 312. So the divisor is 312.5 from a clock of 12 Mhz. Doing the math 12000000/312.5 equals exactly 38400. Regards, Ionut. -- 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
