From: Jiri Slaby <jirislaby@kernel.org>
To: zjzhao <zjzhao@edatec.cn>,
Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: Rob Herring <robh@kernel.org>,
Krzysztof Kozlowski <krzk+dt@kernel.org>,
Conor Dooley <conor+dt@kernel.org>,
linux-kernel@vger.kernel.org, linux-serial@vger.kernel.org,
devicetree@vger.kernel.org
Subject: Re: [PATCH 1/2] serial: wk2xxx: Add WK2xxx SPI UART driver
Date: Fri, 4 Sep 2026 10:17:02 +0200 [thread overview]
Message-ID: <8f8a4946-e8e2-4683-b837-c783c40b81c1@kernel.org> (raw)
In-Reply-To: <20260904072027.67473-2-zjzhao@edatec.cn>
Hi,
On 04. 09. 26, 9:20, zjzhao wrote:
> Add a driver for the WK2xxx SPI to UART bridge ICs (WK2124, WK2132,
> WK2168, WK2202 and WK2204) from WKmic (Chengdu Weikai Microelectronics).
> Each IC exposes two or four full-duplex UART channels with 256-byte
> RX/TX FIFOs through a single SPI slave interface and one interrupt line.
>
> The driver is a rework of the vendor driver (https://github.com/britus/
> wk2xxx) modeled after the NXP sc16is7xx driver. It registers ttyWK0..N
> lines (IDA-allocated), uses devm_request_threaded_irq with a kthread
> worker for register access, falls back to polling when the interrupt
> line is not described, and supports hardware flow control and RS485
> where the IC provides them.
Have you checked that there is no similar driver which could be only
extended?
> Also allocate PORT_WK2XXX (124) and add the SERIAL_WK2XXX Kconfig
> option.
>
> Tested on Raspberry Pi 5 boards (EDATEC IPC1200 with WK2132 on SPI0 and
> SBC2300 with WK2204 on SPI1); loopback TX/RX tests pass.
>
> Signed-off-by: zjzhao <zjzhao@edatec.cn>
...
> --- /dev/null
> +++ b/drivers/tty/serial/wk2xxx.c
> @@ -0,0 +1,1245 @@
...
> +static int wk2xxx_port_reg_write(struct wk2xxx_port *s, unsigned int portno,
> + u8 reg, u8 val)
> +{
> + guard(mutex)(&s->reg_lock);
> + return wk2xxx_raw_port_write(s, portno, reg, val);
> +}
> +
> +static void wk2xxx_port_reg_update(struct wk2xxx_port *s, unsigned int portno,
> + u8 reg, u8 mask, u8 val)
> +{
> + u8 r = 0;
> +
> + scoped_guard(mutex, &s->reg_lock) {
why is this one scoped?
> + if (wk2xxx_raw_port_read(s, portno, reg, &r))
> + return;
> + wk2xxx_raw_port_write(s, portno, reg, (r & ~mask) | val);
> + }
> +}
...
> +static void wk2xxx_throttle(struct uart_port *port)
> +{
> + unsigned long flags;
> +
> + /* Stop draining the RX FIFO to apply back-pressure. */
> + uart_port_lock_irqsave(port, &flags);
This can be a guard too, right?
> + wk2xxx_ier_clear(port, WK2XXX_SIER_RFTRIG_IEN_BIT);
> + uart_port_unlock_irqrestore(port, flags);
> +}
> +
> +static void wk2xxx_unthrottle(struct uart_port *port)
> +{
> + unsigned long flags;
> +
> + uart_port_lock_irqsave(port, &flags);
Same here.
> + wk2xxx_ier_set(port, WK2XXX_SIER_RFTRIG_IEN_BIT);
> + uart_port_unlock_irqrestore(port, flags);
> +}
> +
> +static void wk2xxx_handle_tx(struct uart_port *port)
> +{
> + struct wk2xxx_one *one = to_wk2xxx_one(port, port);
> + struct wk2xxx_port *s = dev_get_drvdata(port->dev);
> + struct tty_port *tport = &port->state->port;
> + unsigned long flags;
> + unsigned int portno = port->iobase;
> + unsigned int txlen, to_send, sent;
> + const unsigned char *tail;
> + u8 fsr, tfcnt;
What's the reason not to use any of the uart_port_tx* helpers?
> + mutex_lock(&one->tx_lock);
Why not guard?
> +
> + if (unlikely(port->x_char)) {
> + wk2xxx_port_reg_write(s, portno, WK2XXX_FDAT_REG, port->x_char);
> + port->icount.tx++;
> + port->x_char = 0;
> + goto out;
And kill the goto then.
> + }
> +
> + if (kfifo_is_empty(&tport->xmit_fifo) || uart_tx_stopped(port)) {
> + uart_port_lock_irqsave(port, &flags);
> + wk2xxx_stop_tx(port);
> + uart_port_unlock_irqrestore(port, flags);
> + goto out;
> + }
> +
> + /* Limit to the free space available in the TX FIFO. */
> + wk2xxx_port_reg_read(s, portno, WK2XXX_TFCNT_REG, &tfcnt);
> + if (tfcnt == 0) {
> + wk2xxx_port_reg_read(s, portno, WK2XXX_FSR_REG, &fsr);
> + txlen = (fsr & WK2XXX_FSR_TFULL_BIT) ? 0 : WK2XXX_FIFO_SIZE;
> + } else {
> + txlen = WK2XXX_FIFO_SIZE - tfcnt;
> + }
> + if (txlen > WK2XXX_MAX_TX_CHARS)
> + txlen = WK2XXX_MAX_TX_CHARS;
> +
> + to_send = kfifo_out_linear_ptr(&tport->xmit_fifo, &tail, txlen);
> + sent = to_send;
> + while (to_send) {
> + unsigned int chunk = min_t(unsigned int, to_send,
> + WK2XXX_MAX_SPI_LEN);
> +
> + wk2xxx_fifo_write(s, portno, tail, chunk);
> + tail += chunk;
> + to_send -= chunk;
> + }
> + uart_xmit_advance(port, sent);
> +
> + uart_port_lock_irqsave(port, &flags);
> + if (kfifo_len(&tport->xmit_fifo) < WAKEUP_CHARS)
> + uart_write_wakeup(port);
> +
> + if (kfifo_is_empty(&tport->xmit_fifo))
> + wk2xxx_stop_tx(port);
> + else
> + wk2xxx_ier_set(port, WK2XXX_SIER_TFTRIG_IEN_BIT);
> + uart_port_unlock_irqrestore(port, flags);
> +
> +out:
> + mutex_unlock(&one->tx_lock);
> +}
...
> +static irqreturn_t wk2xxx_irq(int irq, void *dev_id)
> +{
> + struct wk2xxx_port *s = dev_id;
> + bool keep_polling;
> +
> + do {
> + u8 gifr;
> + int i;
> +
> + keep_polling = false;
> +
> + if (wk2xxx_reg_read(s, WK2XXX_GIFR_REG, &gifr))
> + return IRQ_HANDLED; /* Bus error; give up this pass. */
> +
> + for (i = 0; i < s->devtype->nr_uart; ++i)
> + if (gifr & BIT(i))
> + keep_polling |= wk2xxx_port_irq(s, i);
> + } while (keep_polling);
Should you perhaps cap the loop count as well?
> +
> + return IRQ_HANDLED;
Pointed out by sashiko, this is indeed bad for shared irqs.
> +}
...
> +static void wk2xxx_conf_port(struct uart_port *port, u8 lcr, u8 fwcr,
> + u8 baud0, u8 baud1, u8 pres)
> +{
> + struct wk2xxx_port *s = dev_get_drvdata(port->dev);
> + unsigned int portno = port->iobase;
> + u8 sier, scr, fsr;
> + int count = 200;
> +
> + scoped_guard(mutex, &s->reg_lock) {
Why is this scoped again?
> + /* Disable all sub-UART interrupts. */
> + wk2xxx_raw_port_read(s, portno, WK2XXX_SIER_REG, &sier);
> + wk2xxx_raw_port_write(s, portno, WK2XXX_SIER_REG, 0);
> +
> + /* Wait for the transmitter to become idle. */
> + do {
> + wk2xxx_raw_port_read(s, portno, WK2XXX_FSR_REG, &fsr);
> + } while ((fsr & WK2XXX_FSR_TBUSY_BIT) && count--);
> +
> + /* Disable the transmitter and receiver. */
> + wk2xxx_raw_port_read(s, portno, WK2XXX_SCR_REG, &scr);
> + wk2xxx_raw_port_write(s, portno, WK2XXX_SCR_REG,
> + scr & ~(WK2XXX_SCR_TXEN_BIT |
> + WK2XXX_SCR_RXEN_BIT));
> +
> + /* Program the line control register. */
> + wk2xxx_raw_port_write(s, portno, WK2XXX_LCR_REG, lcr);
> +
> + /* Configure hardware flow control levels. */
> + if (fwcr) {
> + wk2xxx_raw_port_write(s, portno, WK2XXX_FWCR_REG, fwcr);
> + wk2xxx_raw_port_write(s, portno, WK2XXX_SPAGE_REG, 1);
> + wk2xxx_raw_port_write(s, portno, WK2XXX_FWTH_REG, 0xf0);
> + wk2xxx_raw_port_write(s, portno, WK2XXX_FWTL_REG, 0x80);
> + wk2xxx_raw_port_write(s, portno, WK2XXX_SPAGE_REG, 0);
> + }
> +
> + /* Program the baud rate generator (page 1 registers). */
> + wk2xxx_raw_port_write(s, portno, WK2XXX_SPAGE_REG, 1);
> + wk2xxx_raw_port_write(s, portno, WK2XXX_BAUD0_REG, baud0);
> + wk2xxx_raw_port_write(s, portno, WK2XXX_BAUD1_REG, baud1);
> + wk2xxx_raw_port_write(s, portno, WK2XXX_PRES_REG, pres);
> + wk2xxx_raw_port_write(s, portno, WK2XXX_SPAGE_REG, 0);
> +
> + /* Re-enable the transmitter and receiver. */
> + wk2xxx_raw_port_write(s, portno, WK2XXX_SCR_REG,
> + scr | (WK2XXX_SCR_TXEN_BIT |
> + WK2XXX_SCR_RXEN_BIT));
> +
> + /* Restore the interrupt enable register. */
> + wk2xxx_raw_port_write(s, portno, WK2XXX_SIER_REG, sier);
> + }
> +}
> +
> +static void wk2xxx_calc_divisor(unsigned long clk, unsigned int baud,
> + u8 *baud0, u8 *baud1, u8 *pres)
> +{
> + unsigned int div;
> +
> + div = clk / (baud * 16);
> + if (div == 0)
> + div = 1;
> + div--;
> + *baud0 = div & 0xff;
> + *baud1 = (div >> 8) & 0xff;
> + *pres = ((unsigned long long)(clk % (baud * 16)) * 100 / baud + 50) / 100;
Did you mean to use explicit u64?
> +}
> +
> +static void wk2xxx_set_termios(struct uart_port *port, struct ktermios *termios,
> + const struct ktermios *old)
> +{
> + unsigned int baud;
> + unsigned long flags;
> + u8 lcr = 0, fwcr = 0;
> + u8 baud0, baud1, pres;
> +
> + /* The WK2xxx supports 8 data bits only. */
> + termios->c_cflag &= ~CSIZE;
> + termios->c_cflag |= CS8;
> +
> + /* Parity. */
> + if (termios->c_cflag & PARENB) {
> + lcr |= WK2XXX_LCR_PAEN_BIT;
> + switch (termios->c_cflag & (PARODD | CMSPAR)) {
> + case 0:
> + lcr |= WK2XXX_LCR_PAM1_BIT; /* even */
> + break;
> + case PARODD:
> + lcr |= WK2XXX_LCR_PAM0_BIT; /* odd */
> + break;
> + case CMSPAR:
> + break; /* space */
> + case PARODD | CMSPAR:
> + lcr |= WK2XXX_LCR_PAM1_BIT |
> + WK2XXX_LCR_PAM0_BIT; /* mark */
> + break;
> + }
> + }
> +
> + /* Stop bits. */
> + if (termios->c_cflag & CSTOPB)
> + lcr |= WK2XXX_LCR_STPL_BIT;
> +
> + /* Set read status mask. */
> + port->read_status_mask = WK2XXX_LSR_OE_BIT;
> + if (termios->c_iflag & INPCK)
> + port->read_status_mask |= WK2XXX_LSR_PE_BIT |
> + WK2XXX_LSR_FE_BIT;
> + if (termios->c_iflag & (BRKINT | PARMRK))
> + port->read_status_mask |= WK2XXX_LSR_BI_BIT;
> +
> + /* Set status ignore mask. */
> + port->ignore_status_mask = 0;
> + if (termios->c_iflag & IGNBRK)
> + port->ignore_status_mask |= WK2XXX_LSR_BI_BIT;
> + if (!(termios->c_cflag & CREAD))
> + port->ignore_status_mask |= WK2XXX_LSR_BRK_ERROR_MASK;
> +
> + /* Configure flow control. */
> + port->status &= ~(UPSTAT_AUTOCTS | UPSTAT_AUTORTS);
> + if (termios->c_cflag & CRTSCTS) {
> + fwcr = WK2XXX_FWCR_FWM_RTS_CTS;
> + port->status |= UPSTAT_AUTOCTS | UPSTAT_AUTORTS;
> + }
> +
> + /* Get the baud rate generator configuration. */
> + baud = uart_get_baud_rate(port, termios, old,
> + port->uartclk / 16 / 0xffff,
> + port->uartclk / 16);
> +
> + wk2xxx_calc_divisor(port->uartclk, baud, &baud0, &baud1, &pres);
> + wk2xxx_conf_port(port, lcr, fwcr, baud0, baud1, pres);
> +
> + uart_port_lock_irqsave(port, &flags);
guard
> + uart_update_timeout(port, termios->c_cflag, baud);
> + uart_port_unlock_irqrestore(port, flags);
> +}
...
> +static const struct uart_ops wk2xxx_ops = {
> + .tx_empty = wk2xxx_tx_empty,
> + .set_mctrl = wk2xxx_set_mctrl,
> + .get_mctrl = wk2xxx_get_mctrl,
> + .stop_tx = wk2xxx_stop_tx,
> + .start_tx = wk2xxx_start_tx,
> + .throttle = wk2xxx_throttle,
> + .unthrottle = wk2xxx_unthrottle,
> + .stop_rx = wk2xxx_stop_rx,
> + .enable_ms = wk2xxx_enable_ms,
> + .break_ctl = wk2xxx_break_ctl,
> + .startup = wk2xxx_startup,
> + .shutdown = wk2xxx_shutdown,
> + .set_termios = wk2xxx_set_termios,
> + .type = wk2xxx_type,
> + .request_port = wk2xxx_request_port,
> + .release_port = wk2xxx_null_void,
req + rel are optional. Drop them.
> + .config_port = wk2xxx_config_port,
> + .verify_port = wk2xxx_verify_port,
> +};
...
> +static int wk2xxx_probe(struct spi_device *spi)
> +{
> + const struct wk2xxx_devtype *devtype;
> + struct device *dev = &spi->dev;
> + struct wk2xxx_port *s;
> + unsigned long uartclk;
> + u32 clock_freq = 0;
> + bool port_registered[WK2XXX_MAX_PORTS];
> + u8 val;
> + int i, ret;
> +
> + /* Setup SPI bus. The SPI mode follows the device tree (spi-cpha,
> + * spi-cpol); it defaults to SPI mode 0 when unspecified.
> + */
> + spi->bits_per_word = 8;
> + spi->max_speed_hz = spi->max_speed_hz ? : 10 * HZ_PER_MHZ;
> + ret = spi_setup(spi);
> + if (ret)
> + return ret;
> +
> + devtype = spi_get_device_match_data(spi);
> + if (!devtype)
> + return dev_err_probe(dev, -ENODEV, "Failed to match device\n");
> +
> + /* Allocate port structure. */
> + s = devm_kzalloc(dev, struct_size(s, p, devtype->nr_uart), GFP_KERNEL);
> + if (!s)
> + return dev_err_probe(dev, -ENOMEM,
> + "Error allocating port structure\n");
> +
> + s->devtype = devtype;
> + s->spi = spi;
> + mutex_init(&s->reg_lock);
> + dev_set_drvdata(dev, s);
> +
> + /*
> + * The WK2xxx has no identification register, so the best we can do
> + * is to check that communication is at all possible.
> + */
> + ret = wk2xxx_reg_read(s, WK2XXX_GENA_REG, &val);
> + if (ret)
> + return dev_err_probe(dev, ret, "Failed to read GENA register\n");
> +
> + /* Crystal clock; allow an optional DT override. */
> + uartclk = devtype->crystal_freq;
> + if (device_property_read_u32(dev, "clock-frequency", &clock_freq) == 0)
> + uartclk = clock_freq;
> +
> + /* Mark each port line and status as uninitialized. */
> + for (i = 0; i < devtype->nr_uart; ++i) {
> + s->p[i].port.line = WK2XXX_MAX_DEVS;
> + port_registered[i] = false;
> + }
> +
> + kthread_init_worker(&s->kworker);
> + s->kworker_task = kthread_run(kthread_worker_fn, &s->kworker,
> + "wk2xxx");
> + if (IS_ERR(s->kworker_task)) {
> + ret = PTR_ERR(s->kworker_task);
> + goto out_ports;
> + }
> + sched_set_fifo(s->kworker_task);
> +
> + for (i = 0; i < devtype->nr_uart; ++i) {
> + ret = ida_alloc_max(&wk2xxx_lines, WK2XXX_MAX_DEVS - 1,
> + GFP_KERNEL);
> + if (ret < 0)
> + goto out_ports;
> +
> + s->p[i].port.line = ret;
> +
> + /* Initialize port data. */
> + s->p[i].port.dev = dev;
> + s->p[i].port.irq = spi->irq;
> + s->p[i].port.type = PORT_WK2XXX;
> + s->p[i].port.fifosize = WK2XXX_FIFO_SIZE;
> + s->p[i].port.flags = UPF_FIXED_TYPE | UPF_LOW_LATENCY;
> + s->p[i].port.iobase = i;
> + /*
> + * Use all ones as membase so that uart_configure_port() in
> + * serial_core.c does not abort for SPI devices.
> + */
> + s->p[i].port.membase = (void __iomem *)~0;
0 is int, does the compiler not complain about cast to pointer? Or
sparse? Maybe not, I am not sure anymore...
> + s->p[i].port.iotype = UPIO_PORT;
> + s->p[i].port.uartclk = uartclk;
> + s->p[i].port.rs485_config = wk2xxx_config_rs485;
> + s->p[i].port.rs485_supported = wk2xxx_rs485_supported;
> + s->p[i].port.ops = &wk2xxx_ops;
thanks,
--
js
suse labs
next prev parent reply other threads:[~2026-09-04 8:17 UTC|newest]
Thread overview: 20+ messages / expand[flat|nested] mbox.gz Atom feed top
2026-09-04 7:20 [PATCH 0/2] WK2xxx SPI to UART bridge driver zjzhao
2026-09-04 7:20 ` [PATCH 1/2] serial: wk2xxx: Add WK2xxx SPI UART driver zjzhao
2026-09-04 7:35 ` sashiko-bot
2026-09-04 8:17 ` Jiri Slaby [this message]
2026-09-04 9:32 ` zjzhao
2026-09-04 7:20 ` [PATCH 2/2] dt-bindings: serial: Document WK2xxx SPI UART bindings zjzhao
2026-09-04 7:28 ` sashiko-bot
2026-09-04 9:33 ` [PATCH v2 0/2] WK2xxx SPI to UART bridge driver zjzhao
2026-09-04 9:33 ` [PATCH v2 1/2] serial: wk2xxx: Add WK2xxx SPI UART driver zjzhao
2026-09-04 9:53 ` sashiko-bot
2026-09-04 9:33 ` [PATCH v2 2/2] dt-bindings: serial: Document WK2xxx SPI UART bindings zjzhao
2026-09-04 9:40 ` sashiko-bot
2026-09-05 7:31 ` Krzysztof Kozlowski
2026-09-04 10:52 ` [PATCH v3 0/2] WK2xxx SPI to UART bridge driver zjzhao
2026-09-04 10:52 ` [PATCH v3 1/2] serial: wk2xxx: Add WK2xxx SPI UART driver zjzhao
2026-09-04 11:10 ` sashiko-bot
2026-09-04 10:52 ` [PATCH v3 2/2] dt-bindings: serial: Document WK2xxx SPI UART bindings zjzhao
2026-09-04 11:12 ` sashiko-bot
2026-09-04 15:28 ` Conor Dooley
2026-09-05 7:32 ` Krzysztof Kozlowski
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=8f8a4946-e8e2-4683-b837-c783c40b81c1@kernel.org \
--to=jirislaby@kernel.org \
--cc=conor+dt@kernel.org \
--cc=devicetree@vger.kernel.org \
--cc=gregkh@linuxfoundation.org \
--cc=krzk+dt@kernel.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-serial@vger.kernel.org \
--cc=robh@kernel.org \
--cc=zjzhao@edatec.cn \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox