On Sat, Dec 17, 2016 at 03:27:43AM -0800, Russell Senior wrote:
> All testing is with minicom.
Thanks for this through report.
> Starting with 00013-gc510871:
>
> In 8-bit mode, interoperates correctly with pl2303. Switching to
> 5-bit mode on both sides (I get unicode boxes of indeterminate
> values). Switching back to 8-bit mode on both sides, I get correct
> text out both sides. Switching just the ch341 side back to 5-bit
> mode, I get 0xff's out on the pl2303 side (still in 8-bit mode). This
> would seem to imply that byte-size changes are working on this
> version. Changing baud rate also works.
Can you verify also what comes through in 5-bit mode is indeed what's
expected (e.g. 'a' -> 0x01, 'b' -> 0x02, ...)?
You can enable usb-serial debugging to get a log of what is received:
modprobe usbserial dyndbg==p
> 00014-g79e475a:
>
> dmesg shows on insert:
>
> [ 1430.504116] usb 6-2: new full-speed USB device number 3 using uhci_hcd
> [ 1430.682143] usb 6-2: New USB device found, idVendor=1a86, idProduct=7523
> [ 1430.682148] usb 6-2: New USB device strings: Mfr=0, Product=2, SerialNumber=0
> [ 1430.682152] usb 6-2: Product: USB2.0-Ser!
> [ 1430.686294] ch341 6-2:1.0: ch341-uart converter detected
> [ 1430.689122] usb 6-2: ch341_dbg - [0x2518] = c3 00, [0x1213] = 00 00
> (pre version)
> [ 1430.693121] usb 6-2: ch341_dbg - [0x2518] = c3 00, [0x1213] = 00 00
> (post init-0)
> [ 1430.693124] usb 6-2: ch341_set_baudrate_lcr - speed = 9600, lcr =
> c3, a = b202
> [ 1430.696120] usb 6-2: ch341_dbg - [0x2518] = c3 00, [0x1213] = 00 00
> (post divisor)
> [ 1430.699119] usb 6-2: ch341_dbg - [0x2518] = c3 00, [0x1213] = 00 00
> (post lcr)
> [ 1430.700501] usb 6-2: ch341-uart converter now attached to ttyUSB0
>
> Changing baud rate and then byte-size:
>
> [ 1621.189700] usb 6-2: ch341_set_baudrate_lcr - speed = 9600, lcr =
> c3, a = b202
> [ 1621.193556] usb 6-2: ch341_dbg - [0x2518] = c3 00, [0x1213] = 00 00
> (post divisor)
> [ 1621.196557] usb 6-2: ch341_dbg - [0x2518] = c3 00, [0x1213] = 00 00
> (post lcr)
> [ 1621.199565] usb 6-2: ch341_set_baudrate_lcr - speed = 115200, lcr =
> c3, a = cc03
> [ 1621.202571] usb 6-2: ch341_dbg - [0x2518] = c3 00, [0x1213] = 00 00
> (post divisor)
> [ 1621.205569] usb 6-2: ch341_dbg - [0x2518] = c3 00, [0x1213] = 00 00
> (post lcr)
> [ 1621.207572] usb 6-2: ch341_set_baudrate_lcr - speed = 115200, lcr =
> c3, a = cc03
> [ 1621.210570] usb 6-2: ch341_dbg - [0x2518] = c3 00, [0x1213] = 00 00
> (post divisor)
> [ 1621.213573] usb 6-2: ch341_dbg - [0x2518] = c3 00, [0x1213] = 00 00
> (post lcr)
> [ 1644.117396] usb 6-2: ch341_set_baudrate_lcr - speed = 9600, lcr =
> c3, a = b202
> [ 1644.120933] usb 6-2: ch341_dbg - [0x2518] = c3 00, [0x1213] = 00 00
> (post divisor)
> [ 1644.123952] usb 6-2: ch341_dbg - [0x2518] = c3 00, [0x1213] = 00 00
> (post lcr)
> [ 1666.950341] usb 6-2: ch341_set_baudrate_lcr - speed = 9600, lcr =
> c0, a = b202
> [ 1666.954320] usb 6-2: ch341_dbg - [0x2518] = c3 00, [0x1213] = 00 00
> (post divisor)
> [ 1666.957316] usb 6-2: ch341_dbg - [0x2518] = c0 00, [0x1213] = 00 00
> (post lcr)
> [ 1700.500318] usb 6-2: ch341_set_baudrate_lcr - speed = 9600, lcr =
> c3, a = b202
> [ 1700.503845] usb 6-2: ch341_dbg - [0x2518] = c0 00, [0x1213] = 00 00
> (post divisor)
> [ 1700.506866] usb 6-2: ch341_dbg - [0x2518] = c3 00, [0x1213] = 00 00
> (post lcr)
>
> It works in 8-bit mode, I can change to 5-bit mode and then back, as
> before. And I can change baud rate.
Note that you only need to test this commit, which simply adds the
debugging printks on top of the previous.
The log indicates that the LCR is updated correctly, but also reveals
that the divisor registers always read back as zero even if changing
speeds appear to work. Both my devices read back the values set either
through direct register write or using the init commmand.
We may be able to use this to detect these devices and fall back to
direct register writes.
> 00015-gee5a27b:
> I can switch baud rate and byte size back and forth and it seems to
> work. In 5-bit mode, I guess minicom just doesn't know how to render
> the values, so that makes sense, switching back to 8-bit mode and I
> get readable text again. So, this one is working too, afaict.
Good, this one only differs from the previous in that the 7th bit of the
divisor register is set, something which is required for CH341A. Good to
know it does not seem to have any negative effects for this device.
> 00016-ga06b45d:
>
> dmesg on insert:
>
> [ 2588.016052] usb 6-2: new full-speed USB device number 5 using uhci_hcd
> [ 2588.189182] usb 6-2: New USB device found, idVendor=1a86, idProduct=7523
> [ 2588.189187] usb 6-2: New USB device strings: Mfr=0, Product=2, SerialNumber=0
> [ 2588.189192] usb 6-2: Product: USB2.0-Ser!
> [ 2588.193366] ch341 6-2:1.0: ch341-uart converter detected
> [ 2588.196146] usb 6-2: ch341_dbg - [0x2518] = c3 00, [0x1213] = 00 00
> (pre version)
> [ 2588.200162] usb 6-2: ch341_dbg - [0x2518] = c3 00, [0x1213] = 00 00
> (post init-0)
> [ 2588.200165] usb 6-2: ch341_set_baudrate_lcr - speed = 9600, lcr =
> c3, a = b202
> [ 2588.203182] usb 6-2: ch341_dbg - [0x2518] = c3 00, [0x1213] = 00 00
> (post init)
> [ 2588.204488] usb 6-2: ch341-uart converter now attached to ttyUSB0
>
> This doesn't work. Even when matching baud rate and byte size, the
> best I get is 0xff's out on the ch341. Nothing decodes as printable
> anything at the pl2303 side.
Ok, so the init command cannot be used to switch baud rate for these
devices.
Could you see if you can communicate using 19200 baud (when using this
or the next commit), which is the default baud rate for the devices I
have (i.e. default settings are 19200 8N1).
> 00017-g6edc283:
> This doesn't work either. At 9600 8N1, I get 0xff's on the pl2303,
> and some unicode on the ch341.
If the baud rate cannot be set using the init command, then this should
behave as the previous commit. 19200 baud might work.
> Let me know if I can test anything else!
Apart from the two additional tests mentioned above, can you also
provide a log from when connecting the device using the following commit
that I just pushed to the ch341 branch:
f341ee36198d ("dbg: ch341: add register dumps to probe")
which provides dumps of the register settings during initialisation.
(Make sure ch341 dynamic debugging is not enabled to avoid cluttering
the log.)
Thanks,
Johan
--
To unsubscribe from this list: send the line "unsubscribe stable" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html