serial: 8250: Add driver for Broadcom UART

Add a UART driver for the new Broadcom 8250 based STB UART. The new UART is backward compatible with the standard 8250, but has some additional features. The new features include a high accuracy baud rate clock system and DMA support.

The driver will use the new optional BAUD MUX clock to select the best one of the four master clocks (81MHz, 108MHz, 64MHz and 48MHz) to feed the baud rate selection logic for any requested baud rate. This allows for more accurate BAUD rates when high speed baud rates are selected.

The driver will use the new UART DMA hardware if the UART DMA registers are specified in Device Tree "reg" property.

The driver also sets the UPSTAT_AUTOCTS flag when hardware flow control is enabled. This flag is needed for UARTs that don't assert a CTS changed interrupt when CTS changes and AFE (Hardware Flow Control) is enabled.

The driver also contains a workaround for a bug in the Synopsis 8250 core. The problem is that at high baud rates, the RX partial FIFO timeout interrupt can occur but there is no RX data (DR not set in the LSR register). In this case the driver will not read the Receive Buffer Register, which clears the interrupt, and the system will get continuous UART interrupts until the next RX character arrives. The fix originally suggested by Synopsis was to read the Receive Buffer Register and discard the character when the DR bit in the LSR was not set, to clear the interrupt. The problem was that occasionally a character would arrive just after the DR bit check and a valid character would be discarded. The fix that was added will clear receive interrupts to stop the interrupt, deassert RTS to insure that no new data can arrive, wait for 1.5 character times for the sender to react to RTS and then check for data and either do a dummy read or a valid read. Sysfs error counters were also added and were used to help create test software that would cause the error condition. The counters can be found at:
/sys/devices/platform/rdb/*serial/rx_bad_timeout_late_char /sys/devices/platform/rdb/*serial/rx_bad_timeout_no_char

Signed-off-by Al Cooper <[email protected]>

---
 MAINTAINERS                            |    8 +
 drivers/tty/serial/8250/8250_bcm7271.c | 1099 ++++++++++++++++++++++++
 drivers/tty/serial/8250/Kconfig        |   11 +
 drivers/tty/serial/8250/Makefile       |    1 +
 drivers/tty/serial/8250/bcm7271_uart.h |  158 ++++
 5 files changed, 1277 insertions(+)
 create mode 100644 drivers/tty/serial/8250/8250_bcm7271.c
 create mode 100644 drivers/tty/serial/8250/bcm7271_uart.h

MAINTAINERS +0 -0

index 64c7169db617..bb6ad2fc4376 100644

Quoted 57 lines · Al Cooper

Add a UART driver for the new Broadcom 8250 based STB UART. The new UART is backward compatible with the standard 8250, but has some additional features. The new features include a high accuracy baud rate clock system and DMA support.

The driver will use the new optional BAUD MUX clock to select the best one of the four master clocks (81MHz, 108MHz, 64MHz and 48MHz) to feed the baud rate selection logic for any requested baud rate. This allows for more accurate BAUD rates when high speed baud rates are selected.

The driver will use the new UART DMA hardware if the UART DMA registers are specified in Device Tree "reg" property.

The driver also sets the UPSTAT_AUTOCTS flag when hardware flow control is enabled. This flag is needed for UARTs that don't assert a CTS changed interrupt when CTS changes and AFE (Hardware Flow Control) is enabled.

The driver also contains a workaround for a bug in the Synopsis 8250 core. The problem is that at high baud rates, the RX partial FIFO timeout interrupt can occur but there is no RX data (DR not set in the LSR register). In this case the driver will not read the Receive Buffer Register, which clears the interrupt, and the system will get continuous UART interrupts until the next RX character arrives. The fix originally suggested by Synopsis was to read the Receive Buffer Register and discard the character when the DR bit in the LSR was not set, to clear the interrupt. The problem was that occasionally a character would arrive just after the DR bit check and a valid character would be discarded. The fix that was added will clear receive interrupts to stop the interrupt, deassert RTS to insure that no new data can arrive, wait for 1.5 character times for the sender to react to RTS and then check for data and either do a dummy read or a valid read. Sysfs error counters were also added and were used to help create test software that would cause the error condition. The counters can be found at:
/sys/devices/platform/rdb/*serial/rx_bad_timeout_late_char /sys/devices/platform/rdb/*serial/rx_bad_timeout_no_char

Signed-off-by Al Cooper <[email protected]>

---
 MAINTAINERS                            |    8 +
 drivers/tty/serial/8250/8250_bcm7271.c | 1099 ++++++++++++++++++++++++
 drivers/tty/serial/8250/Kconfig        |   11 +
 drivers/tty/serial/8250/Makefile       |    1 +
 drivers/tty/serial/8250/bcm7271_uart.h |  158 ++++
 5 files changed, 1277 insertions(+)
 create mode 100644 drivers/tty/serial/8250/8250_bcm7271.c
 create mode 100644 drivers/tty/serial/8250/bcm7271_uart.h

MAINTAINERS +0 -0

index 64c7169db617..bb6ad2fc4376 100644
@@ -3582,6 +3582,14 @@ S:	Supported
 F:	Documentation/devicetree/bindings/i2c/brcm,brcmstb-i2c.yaml
 F:	drivers/i2c/busses/i2c-brcmstb.c
 

MAINTAINERS S: Supported · line 3587

Quoted 3 lines

MAINTAINERS +3 -0

+BROADCOM BRCMSTB UART DRIVER
+M:	Al Cooper <[email protected]>
+L:	[email protected]

This should probably be linux-serial, copy pasted from the USB entry down below presumably.

Add a UART driver for the new Broadcom 8250 based STB UART. The new UART is backward compatible with the standard 8250, but has some additional features. The new features include a high accuracy baud rate clock system and DMA support.

The driver will use the new optional BAUD MUX clock to select the best one of the four master clocks (81MHz, 108MHz, 64MHz and 48MHz) to feed the baud rate selection logic for any requested baud rate. This allows for more accurate BAUD rates when high speed baud rates are selected.

drivers/tty/serial/8250/8250_bcm7271.c new file +0 -0

drivers/tty/serial/8250/8250_bcm7271.c new file +14 -0

+static void init_real_clk_rates(struct device *dev, struct brcmuart_priv *priv)
+{
+	int x;
+	int rc;
+
+	priv->default_mux_rate = clk_get_rate(priv->baud_mux_clk);
+	dev_dbg(dev, "Default BAUD MUX Clock rate is %lu\n",
+		priv->default_mux_rate);
+
+	for (x = 0; x < ARRAY_SIZE(priv->real_rates); x++) {
+		if (priv->rate_table[x] == 0) {
+			priv->real_rates[x] = 0;
+			continue;
+		}

drivers/tty/serial/8250/8250_bcm7271.c new file +65 -0

+}
+
+static void set_clock_mux(struct uart_port *up, struct brcmuart_priv *priv,
+			u32 baud)
+{
+	u32 percent;
+	u32 best_percent = UINT_MAX;
+	u32 quot;
+	u32 best_quot = 1;
+	u32 rate;
+	int best_index = -1;
+	u64 hires_rate;
+	u64 hires_baud;
+	u64 hires_err;
+	int rc;
+	int i;
+	int real_baud;
+
+	/* If the Baud Mux Clock was not specified, just return */
+	if (priv->baud_mux_clk == NULL)
+		return;
+
+	/* Find the closest match for specified baud */
+	for (i = 0; i < ARRAY_SIZE(priv->real_rates); i++) {
+		if (priv->real_rates[i] == 0)
+			continue;
+		rate = priv->real_rates[i] / 16;
+		quot = DIV_ROUND_CLOSEST(rate, baud);
+		if (!quot)
+			continue;
+
+		/* increase resolution to get xx.xx percent */
+		hires_rate = (u64)rate * 10000;
+		hires_baud = (u64)baud * 10000;
+
+		hires_err = div_u64(hires_rate, (u64)quot);
+
+		/* get the delta */
+		if (hires_err > hires_baud)
+			hires_err = (hires_err - hires_baud);
+		else
+			hires_err = (hires_baud - hires_err);
+
+		percent = (unsigned long)DIV_ROUND_CLOSEST_ULL(hires_err, baud);
+		dev_dbg(up->dev,
+			"Baud rate: %u, MUX Clk: %u, Error: %u.%u%%\n",
+			baud, priv->real_rates[i], percent / 100,
+			percent % 100);
+		if (percent < best_percent) {
+			best_percent = percent;
+			best_index = i;
+			best_quot = quot;
+		}
+	}
+	if (best_index == -1) {
+		dev_err(up->dev, "Error, %d BAUD rate is too fast.\n", baud);
+		return;
+	}
+	rate = priv->real_rates[best_index];
+	rc = clk_set_rate(priv->baud_mux_clk, rate);
+	if (rc)
+		dev_err(up->dev, "Error selecting BAUD MUX clock\n");
+
+	/* Error over 3 percent will cause data errors */
+	if (best_percent > 300)

drivers/tty/serial/8250/8250_bcm7271.c new file +5 -0

+	i += (i / 2);
+	priv->char_wait = ns_to_ktime(i);
+
+	up->uartclk = rate;
+}

drivers/tty/serial/8250/8250_bcm7271.c new file +12 -0

+static int __maybe_unused brcmuart_resume(struct device *dev)
+{
+	struct brcmuart_priv *priv = dev_get_drvdata(dev);
+	int ret;
+
+	ret = clk_prepare_enable(priv->baud_mux_clk);
+	if (ret)
+		dev_err(dev, "Error enabling BAUD MUX clock\n");
+
+	/*
+	 * The hardware goes back to it's default after suspend
+	 * so get the "clk" back in sync.

drivers/tty/serial/8250/8250_bcm7271.c new file +2 -0

+	return 0;
+}

Quoted 9 lines · Al Cooper

Add a UART driver for the new Broadcom 8250 based STB UART. The new UART is backward compatible with the standard 8250, but has some additional features. The new features include a high accuracy baud rate clock system and DMA support.

The driver will use the new optional BAUD MUX clock to select the best one of the four master clocks (81MHz, 108MHz, 64MHz and 48MHz) to feed the baud rate selection logic for any requested baud rate. This allows for more accurate BAUD rates when high speed baud rates are selected.

...

Quoted 2 lines

drivers/tty/serial/8250/8250_bcm7271.c new file +0 -0

...

drivers/tty/serial/8250/8250_bcm7271.c line 367

Quoted 7 lines

drivers/tty/serial/8250/8250_bcm7271.c new file +7 -0

+static void brcmuart_rx_isr(struct uart_port *up, u32 rx_isr)
+{
+     struct brcmuart_priv *priv = up->private_data;
+     struct device *dev = up->dev;
+     u32 rx_done_isr;
+     u32 check_isr;
+     char seq_err[] = "RX buffer ready out of sequence, restarting RX DMA\n";

What's the purpose of this on-stack variable?

Quoted 20 lines

drivers/tty/serial/8250/8250_bcm7271.c new file +18 -0

+static void init_real_clk_rates(struct device *dev, struct brcmuart_priv *priv)
+{
+     int x;
+     int rc;
+
+     priv->default_mux_rate = clk_get_rate(priv->baud_mux_clk);
+     dev_dbg(dev, "Default BAUD MUX Clock rate is %lu\n",
+             priv->default_mux_rate);
+
+     for (x = 0; x < ARRAY_SIZE(priv->real_rates); x++) {
+             if (priv->rate_table[x] == 0) {
+                     priv->real_rates[x] = 0;
+                     continue;
+             }
+             rc = clk_set_rate(priv->baud_mux_clk, priv->rate_table[x]);
+             if (rc) {
+                     dev_err(dev, "Error selecting BAUD MUX clock for %u\n",
+                             priv->rate_table[x]);

drivers/tty/serial/8250/8250_bcm7271.c line 551

Quoted 6 lines

drivers/tty/serial/8250/8250_bcm7271.c new file +6 -0

+                     priv->real_rates[x] = priv->rate_table[x];
+             } else {
+                     priv->real_rates[x] = clk_get_rate(priv->baud_mux_clk);
+             }
+     }
+      clk_set_rate(priv->baud_mux_clk, priv->default_mux_rate);

This is only weirdly indented.

Quoted 71 lines

drivers/tty/serial/8250/8250_bcm7271.c new file +69 -0

+}
+
+static void set_clock_mux(struct uart_port *up, struct brcmuart_priv *priv,
+                     u32 baud)
+{
+     u32 percent;
+     u32 best_percent = UINT_MAX;
+     u32 quot;
+     u32 best_quot = 1;
+     u32 rate;
+     int best_index = -1;
+     u64 hires_rate;
+     u64 hires_baud;
+     u64 hires_err;
+     int rc;
+     int i;
+     int real_baud;
+
+     /* If the Baud Mux Clock was not specified, just return */
+     if (priv->baud_mux_clk == NULL)
+             return;
+
+     /* Find the closest match for specified baud */
+     for (i = 0; i < ARRAY_SIZE(priv->real_rates); i++) {
+             if (priv->real_rates[i] == 0)
+                     continue;
+             rate = priv->real_rates[i] / 16;
+             quot = DIV_ROUND_CLOSEST(rate, baud);
+             if (!quot)
+                     continue;
+
+             /* increase resolution to get xx.xx percent */
+             hires_rate = (u64)rate * 10000;
+             hires_baud = (u64)baud * 10000;
+
+             hires_err = div_u64(hires_rate, (u64)quot);
+
+             /* get the delta */
+             if (hires_err > hires_baud)
+                     hires_err = (hires_err - hires_baud);
+             else
+                     hires_err = (hires_baud - hires_err);
+
+             percent = (unsigned long)DIV_ROUND_CLOSEST_ULL(hires_err, baud);
+             dev_dbg(up->dev,
+                     "Baud rate: %u, MUX Clk: %u, Error: %u.%u%%\n",
+                     baud, priv->real_rates[i], percent / 100,
+                     percent % 100);
+             if (percent < best_percent) {
+                     best_percent = percent;
+                     best_index = i;
+                     best_quot = quot;
+             }
+     }
+     if (best_index == -1) {
+             dev_err(up->dev, "Error, %d BAUD rate is too fast.\n", baud);
+             return;
+     }
+     rate = priv->real_rates[best_index];
+     rc = clk_set_rate(priv->baud_mux_clk, rate);
+     if (rc)
+             dev_err(up->dev, "Error selecting BAUD MUX clock\n");
+
+     /* Error over 3 percent will cause data errors */
+     if (best_percent > 300)
+             dev_err(up->dev, "Error, baud: %d has %u.%u%% error\n",
+                     baud, percent / 100, percent % 100);
+
+     real_baud = rate / 16 / best_quot;

drivers/tty/serial/8250/8250_bcm7271.c line 626

Quoted 6 lines

drivers/tty/serial/8250/8250_bcm7271.c new file +6 -0

+     dev_dbg(up->dev, "Selecting BAUD MUX rate: %u\n", rate);
+     dev_dbg(up->dev, "Requested baud: %u, Actual baud: %u\n",
+             baud, real_baud);
+
+     /* calc nanoseconds for 1.5 characters time at the given baud rate */
+     i = 1000000000 / real_baud / 10;

NSEC_PER_SEC here?

Quoted 7 lines

drivers/tty/serial/8250/8250_bcm7271.c new file +5 -0

+     i += (i / 2);
+     priv->char_wait = ns_to_ktime(i);
+
+     up->uartclk = rate;
+}

...

Quoted 17 lines

drivers/tty/serial/8250/8250_bcm7271.c new file +17 -0

+static int __maybe_unused brcmuart_resume(struct device *dev)
+{
+     struct brcmuart_priv *priv = dev_get_drvdata(dev);
+     int ret;
+
+     ret = clk_prepare_enable(priv->baud_mux_clk);
+     if (ret)
+             dev_err(dev, "Error enabling BAUD MUX clock\n");
+
+     /*
+      * The hardware goes back to it's default after suspend
+      * so get the "clk" back in sync.
+      */
+     ret = clk_set_rate(priv->baud_mux_clk, priv->default_mux_rate);
+     if (ret)
+             dev_err(dev, "Error restoring default BAUD MUX clock\n");
+     if (priv->dma_enabled) {

drivers/tty/serial/8250/8250_bcm7271.c line 1078

Quoted 5 lines

drivers/tty/serial/8250/8250_bcm7271.c new file +5 -0

+             brcmuart_arbitration(priv, 1);
+             brcmuart_init_dma_hardware(priv);
+             start_rx_dma(serial8250_get_port(priv->line));
+     }
+     serial8250_resume_port(priv->line);

All these cannot fail? Or the above can, so does proceeding further without an error make sense?

Quoted 3 lines

drivers/tty/serial/8250/8250_bcm7271.c new file +2 -0

+     return 0;
+}

....

Quoted 6 lines

drivers/tty/serial/8250/Kconfig +0 -0

@@ -501,6 +501,7 @@ config SERIAL_8250_PXA
        applicable to both devicetree and legacy boards, and early console is
        part of its support.
 

drivers/tty/serial/8250/Kconfig config SERIAL_8250_PXA · line 504

Quoted 1 line

drivers/tty/serial/8250/Kconfig +1 -0

+

Why adding a newline here?

drivers/tty/serial/8250/Kconfig config SERIAL_8250_PXA · line 507

Quoted 5 lines

drivers/tty/serial/8250/Kconfig +0 -0

  config SERIAL_8250_TEGRA
      tristate "8250 support for Tegra serial ports"
      default SERIAL_8250

regards,
--
js
suse labs