spi: add ch341a usb2spi driver
PATCH 1/1 spi: add ch341a usb2spi driver
Add a driver for the QiHeng Electronics ch341a USB-to-SPI adapter.
This driver is loosly based on the ch341a module from the flashrom project.
--- drivers/spi/Kconfig | 6 + drivers/spi/Makefile | 1 + drivers/spi/spi-ch341.c | 236 ++++++++++++++++++++++++++++++++++++++++ 3 files changed, 243 insertions(+) create mode 100644 drivers/spi/spi-ch341.c
drivers/spi/Kconfig +6 -0
@@ -1198,6 +1198,12 @@ config SPI_AMD help Enables SPI controller driver for AMD SoC. +config SPI_CH341 + tristate "CH341 USB2SPI adapter" + depends on SPI_MASTER && USB + help + Enables the SPI controller on the CH341a USB to serial chip + # # Add new SPI master controllers in alphabetical order above this line #
drivers/spi/Makefile +1 -0
@@ -154,6 +154,7 @@ obj-$(CONFIG_SPI_XTENSA_XTFPGA) += spi-xtensa-xtfpga.o obj-$(CONFIG_SPI_ZYNQ_QSPI) += spi-zynq-qspi.o obj-$(CONFIG_SPI_ZYNQMP_GQSPI) += spi-zynqmp-gqspi.o obj-$(CONFIG_SPI_AMD) += spi-amd.o +obj-$(CONFIG_SPI_CH341) += spi-ch341.o # SPI slave protocol handlers obj-$(CONFIG_SPI_SLAVE_TIME) += spi-slave-time.o
drivers/spi/spi-ch341.c new file +236 -0
236 lines not on this page · raw message
Signature · 1 line
2.43.0
Mark Brown
maintainerQuoted 5 lines · Johannes Thumshirn
drivers/spi/Kconfig +0
-0@@ -1198,6 +1198,12 @@ config SPI_AMD help Enables SPI controller driver for AMD SoC.
drivers/spi/Kconfig config SPI_AMD · line 1209
Quoted 10 lines
drivers/spi/Kconfig +6
-0+config SPI_CH341 + tristate "CH341 USB2SPI adapter" + depends on SPI_MASTER && USB + help + Enables the SPI controller on the CH341a USB to serial chip + # # Add new SPI master controllers in alphabetical order above this line #
Like the comment says please keep these files sorted (I appreciate there's some things been missed).
drivers/spi/Kconfig line 5
Quoted 7 lines
+++
b/drivers/spi/spi-ch341.cdrivers/spi/Kconfig +5
-0@@ -0,0 +1,236 @@ +// SPDX-License-Identifier: GPL-2.0 +/* + * QiHeng Electronics ch341a USB-to-SPI adapter driver + * + * Copyright (C) 2024 Johannes Thumshirn <[email protected]>
Please make the entire comment block a C++ one so things look clearer.
Quoted 8 lines
drivers/spi/Kconfig +8
-0+static void ch341_set_cs(struct spi_device *spi, bool is_high) +{ + struct ch341_spi_dev *ch341 = + spi_controller_get_devdata(spi->controller); + + memset(ch341->tx_buf, 0, CH341_PACKET_LENGTH); + ch341->tx_buf[0] = CH341A_CMD_UIO_STREAM; + ch341->tx_buf[1] = CH341A_CMD_UIO_STM_OUT | (is_high ? 0x36 : 0x37);
drivers/spi/Kconfig line 23
Quoted 10 lines
drivers/spi/Kconfig +10
-0+ + if (is_high) { + ch341->tx_buf[2] = CH341A_CMD_UIO_STM_DIR | 0x3f; + ch341->tx_buf[3] = CH341A_CMD_UIO_STM_END; + } else { + ch341->tx_buf[2] = CH341A_CMD_UIO_STM_END; + } + + (void)usb_bulk_msg(ch341->udev, ch341->write_pipe, ch341->tx_buf, + (is_high ? 4 : 3), NULL, CH341_DEFAULT_TIMEOUT);
The cast to void here is very suspicious, what's it doing?
drivers/spi/Kconfig line 32
Quoted 9 lines
drivers/spi/Kconfig +9
-0+static int ch341_config_stream(struct ch341_spi_dev *ch341, int speed) +{ + memset(ch341->tx_buf, 0, CH341_PACKET_LENGTH); + ch341->tx_buf[0] = CH341A_CMD_I2C_STREAM; + ch341->tx_buf[1] = CH341A_CMD_I2C_STM_SET | (speed & 0x7); + ch341->tx_buf[2] = CH341A_CMD_I2C_STM_END; + + return usb_bulk_msg(ch341->udev, ch341->write_pipe, ch341->tx_buf, 3, + NULL, CH341_DEFAULT_TIMEOUT);
No validation of speed?
Quoted 4 lines · Mark Brown
Quoted 4 lines
drivers/spi/Kconfig +4
-0+config SPI_CH341 + tristate "CH341 USB2SPI adapter" + depends on SPI_MASTER && USB + help
Quoted 8 lines
drivers/spi/Kconfig config SPI_AMD · line 1209
Quoted 5 lines
drivers/spi/Kconfig +2
-0+ Enables the SPI controller on the CH341a USB to serial chip + # # Add new SPI master controllers in alphabetical order above this line #Like the comment says please keep these files sorted (I appreciate there's some things been missed).
Oops. Fixed, sorry.
Quoted 10 lines
drivers/spi/spi-ch341.c line 5
Quoted 8 lines
+++
b/drivers/spi/spi-ch341.cdrivers/spi/spi-ch341.c new file +5
-0@@ -0,0 +1,236 @@ +// SPDX-License-Identifier: GPL-2.0 +/* +* QiHeng Electronics ch341a USB-to-SPI adapter driver +* +* Copyright (C) 2024 Johannes Thumshirn <[email protected]> Please make the entire comment block a C++ one so things look clearer.
I've never seen that in kernel code, but sure changed it.
Quoted 13 lines
Quoted 13 lines
drivers/spi/spi-ch341.c new file +12
-0+static void ch341_set_cs (struct spi_device *spi,bool is_high) +{ +struct ch341_spi_dev *ch341 = +spi_controller_get_devdata (spi->controller); + +memset (ch341->tx_buf,0 ,CH341_PACKET_LENGTH ); + ch341->tx_buf[0 ] =CH341A_CMD_UIO_STREAM ; + ch341->tx_buf[1 ] =CH341A_CMD_UIO_STM_OUT | (is_high ?0x36 :0x37 ); + +if (is_high) { + ch341->tx_buf[2 ] =CH341A_CMD_UIO_STM_DIR |0x3f ; + ch341->tx_buf[3 ] =CH341A_CMD_UIO_STM_END ;
Quoted 8 lines
drivers/spi/spi-ch341.c line 59
Quoted 6 lines
drivers/spi/spi-ch341.c new file +6
-0+ }else { + ch341->tx_buf[2 ] =CH341A_CMD_UIO_STM_END ; + } + + (void )usb_bulk_msg (ch341->udev, ch341->write_pipe, ch341->tx_buf, + (is_high ?4 :3 ),NULL ,CH341_DEFAULT_TIMEOUT );The cast to void here is very suspicious, what's it doing?
Explicitly ignoring errors from usb_bulk_msg() as ->set_cs() can't really handle error returns.
I've changed it to an error print.
Quoted 12 lines
drivers/spi/spi-ch341.c line 118
Quoted 10 lines
drivers/spi/spi-ch341.c new file +9
-0+static int ch341_config_stream (struct ch341_spi_dev *ch341,int speed) +{ +memset (ch341->tx_buf,0 ,CH341_PACKET_LENGTH ); + ch341->tx_buf[0 ] =CH341A_CMD_I2C_STREAM ; + ch341->tx_buf[1 ] =CH341A_CMD_I2C_STM_SET | (speed &0x7 ); + ch341->tx_buf[2 ] =CH341A_CMD_I2C_STM_END ; + +return usb_bulk_msg (ch341->udev, ch341->write_pipe, ch341->tx_buf,3 , +NULL ,CH341_DEFAULT_TIMEOUT );No validation of speed?
TBH I haven't found a command that reads the current settings of the device.
Mark Brown
maintainerQuoted 5 lines · Johannes Thumshirn
Quoted 1 line · Mark Brown
No validation of speed?
TBH I haven't found a command that reads the current settings of the device.
Why would reading things from the device be relevant to validating what we are writing?
Quoted 9 lines · Mark Brown
Quoted 5 lines · Johannes Thumshirn
Quoted 1 line · Mark Brown
No validation of speed?
TBH I haven't found a command that reads the current settings of the device.
Why would reading things from the device be relevant to validating what we are writing?
Then I don't really understand your comment.
Why would I need to validate a constant? But indeed it doesn't make any sense at all to pass in the speed parameter.
Will fix in v2.