Re: [PATCH v5 1/3] mfd: ch341: add core driver for the WCH CH341 in I2C/SPI/GPIO mode

[Lee Jones <lee.jones@xxxxxxxxxx>]

On Thu, 31 Mar 2022, frank zago wrote:

> The CH341 is a multifunction chip, presenting 3 different USB PID. One
> of these functions is for I2C/SPI/GPIO. This new set of drivers will
> manage I2C and GPIO.
> 
> Signed-off-by: frank zago <frank@xxxxxxxx>
> ---
>  Documentation/misc-devices/ch341.rst | 114 +++++++++++++++++++++++++++
>  Documentation/misc-devices/index.rst |   1 +

Not sure I've seen this bundled with an MFD driver before.

Please separate them into their own patches.

>  MAINTAINERS                          |   7 ++
>  drivers/mfd/Kconfig                  |  10 +++
>  drivers/mfd/Makefile                 |   1 +
>  drivers/mfd/ch341-core.c             |  99 +++++++++++++++++++++++
>  include/linux/mfd/ch341.h            |  26 ++++++
>  7 files changed, 258 insertions(+)
>  create mode 100644 Documentation/misc-devices/ch341.rst
>  create mode 100644 drivers/mfd/ch341-core.c
>  create mode 100644 include/linux/mfd/ch341.h
> 
> diff --git a/Documentation/misc-devices/ch341.rst b/Documentation/misc-devices/ch341.rst
> new file mode 100644
> index 000000000000..bf0b83f2eb85
> --- /dev/null
> +++ b/Documentation/misc-devices/ch341.rst
> @@ -0,0 +1,114 @@
> +.. SPDX-License-Identifier: GPL-2.0-or-later
> +
> +===========================================================
> +WinChipHead (沁恒) CH341 linux driver for I2C and GPIO mode
> +===========================================================
> +
> +The CH341 is declined in several flavors, and may support one or more
> +of UART, SPI, I2C and GPIO, but not always simultaneously:
> +
> +  - CH341 A/B/F: UART, Printer, SPI, I2C and GPIO
> +  - CH341 C/T: UART and I2C
> +  - CH341 H: SPI
> +
> +They work in 3 different modes, with only one being presented
> +depending on the USB PID:
> +
> +  - 0x5523: UART mode, covered by the USB `ch341` serial driver
> +  - 0x5512: SPI/I2C/GPIO mode, covered by the ch341 MFD drivers
> +  - 0x5584: Parallel printer mode, covered by the USB `usblp` driver
> +
> +Mode selection is done at the hardware level by tying some
> +pins. Breakout boards with one of the CH341 chip usually have one or
> +more jumpers to select which mode they work on. At least one model
> +(CJMCU-341) appears to need bridging some solder pads to select a
> +different default. Breakout boards also don't usually offer an option
> +to configure the chip into printer mode; for that case, connect the
> +SCL and SDA lines directly together.
> +
> +The various CH341 appear to be indistinguishable from the
> +software. For instance the ch341 MFD driver will present a GPIO
> +interface for the CH341T although physical pins are not present, and
> +the device will accept GPIO commands.
> +
> +The ch341 MFD driver has been tested with a CH341A, CH341B and
> +CH341T.
> +
> +Some breakout boards work in 3.3V and 5V depending on some jumpers.
> +
> +The black chip programmer with a ZIF socket will power the CH341 at
> +3.3V if a jumper is set, but will only output 5V to the chips to be
> +programmed, which is not always desirable. A hardware hack to use 3.3V
> +everywhere, involving some soldering, is available at
> +https://eevblog.com/forum/repair/ch341a-serial-memory-programmer-power-supply-fix/
> +
> +Some sample code for the CH341 is available at the manufacturer
> +website, at http://wch-ic.com/products/CH341.html
> +
> +The repository at https://github.com/boseji/CH341-Store contains a lot
> +of information on these chips, including datasheets.
> +
> +This driver is based on the pre-existing work at
> +https://github.com/gschorcht/i2c-ch341-usb
> +
> +
> +I2C Caveats
> +-----------
> +
> +The ch341 doesn't work with a Wii nunchuk, possibly because the
> +pull-up value is too low (1500 ohms).
> +
> +i2c AT24 eeproms can be read but not programmed properly because the
> +at24 linux driver tries to write a byte at a time, and doesn't wait at
> +all (or enough) between writes. Data corruption on writes does occur.
> +
> +The driver doesn't support detection of I2C devices present on the
> +bus. Apparently when a device is not present at a given address, the
> +CH341 will return an extra byte of data, but the driver doesn't
> +support that. This may be addressed in the future.
> +
> +
> +The GPIOs
> +---------
> +
> +16 GPIOs are available on the CH341 A/B/F. The first 6 are input/output,
> +and the last 10 are input only.
> +
> +Pinout and their names as they appear on some breakout boards::
> +
> +  CH341A/B/F     GPIO  Names                    Mode
> +    pin          line
> +
> +   15             0     D0, CS0                  input/output
> +   16             1     D1, CS1                  input/output
> +   17             2     D2, CS2                  input/output
> +   18             3     D3, SCK, DCK             input/output
> +   19             4     D4, DOUT2, CS3           input/output
> +   20             5     D5, MOSI, DOUT, SDO      input/output
> +   21             6     D6, DIN2                 input
> +   22             7     D7, MISO, DIN            input
> +    5             8     ERR                      input
> +    6             9     PEMP                     input
> +    7            10     INT                      input
> +    8            11     SLCT (SELECT)            input
> +   26            12     RST# (?)                 input
> +   27            13     WT (WAIT)                input
> +    4            14     DS (Data Select?)        input
> +    3            15     AS (Address Select?)     input
> +
> +
> +GPIO interrupt
> +~~~~~~~~~~~~~~
> +
> +The INT pin, corresponding to GPIO 10 is an input pin that can trigger
> +an interrupt on a rising edge. Only that pin is able to generate an
> +interrupt, and only on a rising edge. Trying to monitor events on
> +another GPIO, or that GPIO on something other than a rising edge, will
> +be rejected.
> +
> +
> +SPI
> +---
> +
> +This driver doesn't offer an SPI interface (yet) due to the
> +impossibility of declaring an SPI device like I2C does.
> diff --git a/Documentation/misc-devices/index.rst b/Documentation/misc-devices/index.rst
> index 30ac58f81901..52d03715601e 100644
> --- a/Documentation/misc-devices/index.rst
> +++ b/Documentation/misc-devices/index.rst
> @@ -19,6 +19,7 @@ fit into other categories.
>     bh1770glc
>     eeprom
>     c2port
> +   ch341
>     dw-xdata-pcie
>     ibmvmc
>     ics932s401
> diff --git a/MAINTAINERS b/MAINTAINERS
> index cbbd3ce7e0c2..b61af813fb9f 100644
> --- a/MAINTAINERS
> +++ b/MAINTAINERS
> @@ -21211,6 +21211,13 @@ M:	David Härdeman <david@xxxxxxxxxxx>
>  S:	Maintained
>  F:	drivers/media/rc/winbond-cir.c
>  
> +WINCHIPHEAD CH341 I2C/GPIO MFD DRIVER
> +M:	Frank Zago <frank@xxxxxxxx>
> +L:	linux-usb@xxxxxxxxxxxxxxx
> +S:	Maintained
> +F:	drivers/mfd/ch341-core.c
> +F:	include/linux/mfd/ch341.h
> +
>  WINSYSTEMS EBC-C384 WATCHDOG DRIVER
>  M:	William Breathitt Gray <vilhelm.gray@xxxxxxxxx>
>  L:	linux-watchdog@xxxxxxxxxxxxxxx
> diff --git a/drivers/mfd/Kconfig b/drivers/mfd/Kconfig
> index 3b59456f5545..893acc821a42 100644
> --- a/drivers/mfd/Kconfig
> +++ b/drivers/mfd/Kconfig
> @@ -1784,6 +1784,16 @@ config MFD_LOCHNAGAR
>  	help
>  	  Support for Cirrus Logic Lochnagar audio development board.
>  
> +config MFD_CH341
> +	tristate "WinChipHead CH341 in I2C/SPI/GPIO mode"
> +	depends on USB
> +	help
> +	  If you say yes to this option, support for the CH341 series
> +	  of chips, running in I2C/SPI/GPIO mode will be included.
> +
> +	  This driver can also be built as a module.  If so, the
> +	  module will be called ch341-core.
> +
>  config MFD_ARIZONA
>  	select REGMAP
>  	select REGMAP_IRQ
> diff --git a/drivers/mfd/Makefile b/drivers/mfd/Makefile
> index 858cacf659d6..fd615ab3929f 100644
> --- a/drivers/mfd/Makefile
> +++ b/drivers/mfd/Makefile
> @@ -13,6 +13,7 @@ obj-$(CONFIG_MFD_ASIC3)		+= asic3.o tmio_core.o
>  obj-$(CONFIG_ARCH_BCM2835)	+= bcm2835-pm.o
>  obj-$(CONFIG_MFD_BCM590XX)	+= bcm590xx.o
>  obj-$(CONFIG_MFD_BD9571MWV)	+= bd9571mwv.o
> +obj-$(CONFIG_MFD_CH341)		+= ch341-core.o
>  obj-$(CONFIG_MFD_CROS_EC_DEV)	+= cros_ec_dev.o
>  obj-$(CONFIG_MFD_ENE_KB3930)	+= ene-kb3930.o
>  obj-$(CONFIG_MFD_EXYNOS_LPASS)	+= exynos-lpass.o
> diff --git a/drivers/mfd/ch341-core.c b/drivers/mfd/ch341-core.c
> new file mode 100644
> index 000000000000..0bb6eb8057e9
> --- /dev/null
> +++ b/drivers/mfd/ch341-core.c
> @@ -0,0 +1,99 @@
> +// SPDX-License-Identifier: GPL-2.0
> +/*
> + * Core driver for the CH341A, CH341B and CH341T in I2C/SPI/GPIO
> + * mode. There are cell drivers available for I2C and GPIO. SPI is not
> + * yet supported.
> + *
> + * Copyright 2022, Frank Zago
> + * Copyright (c) 2017 Gunar Schorcht (gunar@xxxxxxxxxxxx)
> + * Copyright (c) 2016 Tse Lun Bien
> + * Copyright (c) 2014 Marco Gittler
> + * Copyright (c) 2006-2007 Till Harbaum (Till@xxxxxxxxxxx)
> + */
> +
> +#include <linux/kernel.h>
> +#include <linux/mfd/ch341.h>
> +#include <linux/mfd/core.h>
> +#include <linux/module.h>
> +#include <linux/slab.h>
> +#include <linux/usb.h>
> +
> +static const struct mfd_cell ch341_devs[] = {
> +};
> +
> +static int ch341_usb_probe(struct usb_interface *iface,
> +			   const struct usb_device_id *usb_id)
> +{
> +	struct usb_host_endpoint *endpoints;
> +	struct ch341_device *dev;

Please rename this struct 'ch341_ddata' and s/dev/ddata/.

> +	int rc;

'ret' please.

> +	dev = devm_kzalloc(&iface->dev, sizeof(*dev), GFP_KERNEL);
> +	if (!dev)
> +		return -ENOMEM;
> +
> +	dev->usb_dev = usb_get_dev(interface_to_usbdev(iface));

Can this come later?

It would save on the goto dance.

> +	dev->iface = iface;
> +	mutex_init(&dev->usb_lock);
> +
> +	if (iface->cur_altsetting->desc.bNumEndpoints != 3) {

Why 3?

> +		rc = -ENODEV;
> +		goto free_dev;
> +	}
> +
> +	endpoints = iface->cur_altsetting->endpoint;
> +	if (!usb_endpoint_is_bulk_in(&endpoints[0].desc) ||
> +	    !usb_endpoint_is_bulk_out(&endpoints[1].desc) ||
> +	    !usb_endpoint_xfer_int(&endpoints[2].desc)) {

What has happened if we get here?

Perhaps a comment would be useful?

> +		rc = -ENODEV;
> +		goto free_dev;
> +	}
> +
> +	dev->ep_in = endpoints[0].desc.bEndpointAddress;
> +	dev->ep_out = endpoints[1].desc.bEndpointAddress;
> +	dev->ep_intr = endpoints[2].desc.bEndpointAddress;
> +	dev->ep_intr_interval = endpoints[2].desc.bInterval;
> +
> +	usb_set_intfdata(iface, dev);
> +
> +	rc = mfd_add_hotplug_devices(&iface->dev, ch341_devs,

Why are you using 'hotplug' here?

ch341_devs is empty right?

So no child devices are registered.

In which case this is not (yet) an MFD and cannot be accepted.

Please add the children.

> +				     ARRAY_SIZE(ch341_devs));
> +	if (rc) {
> +		rc = dev_err_probe(&iface->dev, rc,
> +				   "Failed to add mfd devices to core\n");

I'm not even sure what this means.  Should be:

"Failed to register child devices\n"

> +		goto free_dev;
> +	}
> +
> +	return 0;
> +
> +free_dev:
> +	usb_put_dev(dev->usb_dev);
> +
> +	return rc;
> +}
> +
> +static void ch341_usb_disconnect(struct usb_interface *usb_if)
> +{
> +	struct ch341_device *dev = usb_get_intfdata(usb_if);
> +
> +	mfd_remove_devices(&usb_if->dev);
> +	usb_put_dev(dev->usb_dev);
> +}
> +
> +static const struct usb_device_id ch341_usb_table[] = {
> +	{ USB_DEVICE(0x1a86, 0x5512) },
> +	{ }
> +};
> +MODULE_DEVICE_TABLE(usb, ch341_usb_table);
> +
> +static struct usb_driver ch341_usb_driver = {
> +	.name       = "ch341-mfd",
> +	.id_table   = ch341_usb_table,
> +	.probe      = ch341_usb_probe,
> +	.disconnect = ch341_usb_disconnect,
> +};
> +module_usb_driver(ch341_usb_driver);
> +
> +MODULE_AUTHOR("Various");

This does not look valid.  Please drop it.

> +MODULE_DESCRIPTION("CH341 USB to I2C/SPI/GPIO adapter");

Is it?  What makes it one of those?

> +MODULE_LICENSE("GPL");
> diff --git a/include/linux/mfd/ch341.h b/include/linux/mfd/ch341.h
> new file mode 100644
> index 000000000000..a87b23e30123
> --- /dev/null
> +++ b/include/linux/mfd/ch341.h
> @@ -0,0 +1,26 @@
> +/* SPDX-License-Identifier: GPL-2.0 */
> +/* Definitions for the CH341 MFD driver */

Drop the term MFD please.

> +#include <linux/mutex.h>
> +#include <linux/types.h>
> +
> +#define DEFAULT_TIMEOUT_MS 1000	/* 1s USB requests timeout */

Where is this used?

> +/* All commands fit inside a 32-byte segment. There may be several
> + * of these segments in a USB command.
> + */

This is not a properly formatted multi-line comment.

> +#define SEG_SIZE 32
> +
> +struct usb_device;
> +struct usb_interface;
> +
> +struct ch341_device {
> +	struct usb_device *usb_dev;
> +	struct usb_interface *iface;
> +	struct mutex usb_lock;
> +
> +	int ep_in;
> +	int ep_out;
> +	int ep_intr;
> +	u8 ep_intr_interval;
> +};

-- 
Lee Jones [李琼斯]
Principal Technical Lead - Developer Services
Linaro.org │ Open source software for Arm SoCs
Follow Linaro: Facebook | Twitter | Blog