[PATCH 0/9] Add support for the rest of Renesas RZ/G3S serial interfaces

[PATCH 0/9] Add support for the rest of Renesas RZ/G3S serial interfaces

[Claudiu]

From: Claudiu Beznea <claudiu.beznea.uj@bp.renesas.com>

Hi,

The Renesas RZ/G3S SoC has 6 serial interfaces. One of them is used
as debug console (and it is already enabled in the current code base).
Series adds support for the remaining ones.

Patches:
-    01/09 - adds clock, reset and power domain support for the serial
             interfaces
- 02-03/09 - serial driver fixes patches identified while adding RZ/G3S
             support
-    04/09 - extends suspend to RAM support on the serial driver for
             the RZ/G3S SoC
- 05-09/09 - add device tree support

Merge strategy, if any:
- patch 01/09 can go through Renesas tree
- patches 02-04/09 can go through serial tree
- patches 05-09/09 can go through Renesas tree

Thank you,
Claudiu Beznea


Claudiu Beznea (9):
  clk: renesas: r9a08g045: Add clock, reset and power domain for the
    remaining SCIFs
  serial: sh-sci: Check if TX data was written to device in .tx_empty()
  serial: sh-sci: Clean sci_ports[0] after at earlycon exit
  serial: sh-sci: Update the suspend/resume support
  arm64: dts: renesas: r9a08g045: Add the remaining SCIF interfaces
  arm64: dts: renesas: rzg3s-smarc: Fix the debug serial alias
  arm64: dts: renesas: rzg3s-smarc-switches: Add a header to describe
    different switches
  arm64: dts: renesas: rzg3s-smarc: Enable SCIF3
  arm64: dts: renesas: r9a08g045s33-smarc-pmod: Add overlay for SCIF1

 arch/arm64/boot/dts/renesas/Makefile          |  3 +
 arch/arm64/boot/dts/renesas/r9a08g045.dtsi    | 90 ++++++++++++++++++
 .../dts/renesas/r9a08g045s33-smarc-pmod.dtso  | 48 ++++++++++
 .../boot/dts/renesas/rzg3s-smarc-som.dtsi     | 25 +----
 .../boot/dts/renesas/rzg3s-smarc-switches.h   | 32 +++++++
 arch/arm64/boot/dts/renesas/rzg3s-smarc.dtsi  | 25 ++++-
 drivers/clk/renesas/r9a08g045-cpg.c           | 20 ++++
 drivers/tty/serial/sh-sci.c                   | 92 +++++++++++++++++--
 8 files changed, 301 insertions(+), 34 deletions(-)
 create mode 100644 arch/arm64/boot/dts/renesas/r9a08g045s33-smarc-pmod.dtso
 create mode 100644 arch/arm64/boot/dts/renesas/rzg3s-smarc-switches.h

-- 
2.39.2

[PATCH 1/9] clk: renesas: r9a08g045: Add clock, reset and power domain for the remaining SCIFs

[Claudiu]

From: Claudiu Beznea <claudiu.beznea.uj@bp.renesas.com>

The Renesas RZ/G3S SoC has 6 SCIF interfaces. SCIF0 is used as debug
console and is already enabled. Add the clock, reset and power domain
support for the remaining ones.

Signed-off-by: Claudiu Beznea <claudiu.beznea.uj@bp.renesas.com>
---
 drivers/clk/renesas/r9a08g045-cpg.c | 20 ++++++++++++++++++++
 1 file changed, 20 insertions(+)

diff --git a/drivers/clk/renesas/r9a08g045-cpg.c b/drivers/clk/renesas/r9a08g045-cpg.c
index b2ae8cdc4723..da6dfffa089a 100644
--- a/drivers/clk/renesas/r9a08g045-cpg.c
+++ b/drivers/clk/renesas/r9a08g045-cpg.c
@@ -224,6 +224,11 @@ static const struct rzg2l_mod_clk r9a08g045_mod_clks[] = {
 	DEF_MOD("i2c2_pclk",		R9A08G045_I2C2_PCLK, R9A08G045_CLK_P0, 0x580, 2),
 	DEF_MOD("i2c3_pclk",		R9A08G045_I2C3_PCLK, R9A08G045_CLK_P0, 0x580, 3),
 	DEF_MOD("scif0_clk_pck",	R9A08G045_SCIF0_CLK_PCK, R9A08G045_CLK_P0, 0x584, 0),
+	DEF_MOD("scif1_clk_pck",	R9A08G045_SCIF1_CLK_PCK, R9A08G045_CLK_P0, 0x584, 1),
+	DEF_MOD("scif2_clk_pck",	R9A08G045_SCIF2_CLK_PCK, R9A08G045_CLK_P0, 0x584, 2),
+	DEF_MOD("scif3_clk_pck",	R9A08G045_SCIF3_CLK_PCK, R9A08G045_CLK_P0, 0x584, 3),
+	DEF_MOD("scif4_clk_pck",	R9A08G045_SCIF4_CLK_PCK, R9A08G045_CLK_P0, 0x584, 4),
+	DEF_MOD("scif5_clk_pck",	R9A08G045_SCIF5_CLK_PCK, R9A08G045_CLK_P0, 0x584, 5),
 	DEF_MOD("gpio_hclk",		R9A08G045_GPIO_HCLK, R9A08G045_OSCCLK, 0x598, 0),
 	DEF_MOD("vbat_bclk",		R9A08G045_VBAT_BCLK, R9A08G045_OSCCLK, 0x614, 0),
 };
@@ -249,6 +254,11 @@ static const struct rzg2l_reset r9a08g045_resets[] = {
 	DEF_RST(R9A08G045_I2C2_MRST, 0x880, 2),
 	DEF_RST(R9A08G045_I2C3_MRST, 0x880, 3),
 	DEF_RST(R9A08G045_SCIF0_RST_SYSTEM_N, 0x884, 0),
+	DEF_RST(R9A08G045_SCIF1_RST_SYSTEM_N, 0x884, 1),
+	DEF_RST(R9A08G045_SCIF2_RST_SYSTEM_N, 0x884, 2),
+	DEF_RST(R9A08G045_SCIF3_RST_SYSTEM_N, 0x884, 3),
+	DEF_RST(R9A08G045_SCIF4_RST_SYSTEM_N, 0x884, 4),
+	DEF_RST(R9A08G045_SCIF5_RST_SYSTEM_N, 0x884, 5),
 	DEF_RST(R9A08G045_GPIO_RSTN, 0x898, 0),
 	DEF_RST(R9A08G045_GPIO_PORT_RESETN, 0x898, 1),
 	DEF_RST(R9A08G045_GPIO_SPARE_RESETN, 0x898, 2),
@@ -306,6 +316,16 @@ static const struct rzg2l_cpg_pm_domain_init_data r9a08g045_pm_domains[] = {
 				DEF_REG_CONF(CPG_BUS_MCPU2_MSTOP, BIT(13)), 0),
 	DEF_PD("scif0",		R9A08G045_PD_SCIF0,
 				DEF_REG_CONF(CPG_BUS_MCPU2_MSTOP, BIT(1)), 0),
+	DEF_PD("scif1",		R9A08G045_PD_SCIF1,
+				DEF_REG_CONF(CPG_BUS_MCPU2_MSTOP, BIT(2)), 0),
+	DEF_PD("scif2",		R9A08G045_PD_SCIF2,
+				DEF_REG_CONF(CPG_BUS_MCPU2_MSTOP, BIT(3)), 0),
+	DEF_PD("scif3",		R9A08G045_PD_SCIF3,
+				DEF_REG_CONF(CPG_BUS_MCPU2_MSTOP, BIT(4)), 0),
+	DEF_PD("scif4",		R9A08G045_PD_SCIF4,
+				DEF_REG_CONF(CPG_BUS_MCPU2_MSTOP, BIT(5)), 0),
+	DEF_PD("scif5",		R9A08G045_PD_SCIF5,
+				DEF_REG_CONF(CPG_BUS_MCPU3_MSTOP, BIT(4)), 0),
 	DEF_PD("vbat",		R9A08G045_PD_VBAT,
 				DEF_REG_CONF(CPG_BUS_MCPU3_MSTOP, BIT(8)),
 				GENPD_FLAG_ALWAYS_ON),
-- 
2.39.2

[PATCH 2/9] serial: sh-sci: Check if TX data was written to device in .tx_empty()

[Claudiu]

From: Claudiu Beznea <claudiu.beznea.uj@bp.renesas.com>

On the Renesas RZ/G3S, when doing suspend to RAM, the uart_suspend_port()
is called. The uart_suspend_port() calls 3 times the
struct uart_port::ops::tx_empty() before shutting down the port.

According to the documentation, the struct uart_port::ops::tx_empty()
API tests whether the transmitter FIFO and shifter for the port is
empty.

The Renesas RZ/G3S SCIFA IP reports the number of data units stored in the
transmit FIFO through the FDR (FIFO Data Count Register). The data units
in the FIFOs are written in the shift register and transmitted from there.
The TEND bit in the Serial Status Register reports if the data was
transmitted from the shift register.

In the previous code, in the tx_empty() API implemented by the sh-sci
driver, it is considered that the TX is empty if the hardware reports the
TEND bit set and the number of data units in the FIFO is zero.

According to the HW manual, the TEND bit has the following meaning:

0: Transmission is in the waiting state or in progress.
1: Transmission is completed.

It has been noticed that when opening the serial device w/o using it and
then switch to a power saving mode, the tx_empty() call in the
uart_port_suspend() function fails, leading to the "Unable to drain
transmitter" message being printed on the console. This is because the
TEND=0 if nothing has been transmitted and the FIFOs are empty. As the
TEND=0 has double meaning (waiting state, in progress) we can't
determined the scenario described above.

Add a software workaround for this. This sets a variable if any data has
been sent on the serial console (when using PIO) or if the DMA callback has
been called (meaning something has been transmitted).

Fixes: 73a19e4c0301 ("serial: sh-sci: Add DMA support.")
Cc: stable@vger.kernel.org
Signed-off-by: Claudiu Beznea <claudiu.beznea.uj@bp.renesas.com>
---
 drivers/tty/serial/sh-sci.c | 11 +++++++++++
 1 file changed, 11 insertions(+)

diff --git a/drivers/tty/serial/sh-sci.c b/drivers/tty/serial/sh-sci.c
index df523c744423..8e2d534401fa 100644
--- a/drivers/tty/serial/sh-sci.c
+++ b/drivers/tty/serial/sh-sci.c
@@ -153,6 +153,7 @@ struct sci_port {
 	int				rx_trigger;
 	struct timer_list		rx_fifo_timer;
 	int				rx_fifo_timeout;
+	atomic_t			first_time_tx;
 	u16				hscif_tot;
 
 	bool has_rtscts;
@@ -850,6 +851,7 @@ static void sci_transmit_chars(struct uart_port *port)
 {
 	struct tty_port *tport = &port->state->port;
 	unsigned int stopped = uart_tx_stopped(port);
+	struct sci_port *s = to_sci_port(port);
 	unsigned short status;
 	unsigned short ctrl;
 	int count;
@@ -885,6 +887,7 @@ static void sci_transmit_chars(struct uart_port *port)
 		}
 
 		sci_serial_out(port, SCxTDR, c);
+		atomic_set(&s->first_time_tx, 1);
 
 		port->icount.tx++;
 	} while (--count > 0);
@@ -1241,6 +1244,8 @@ static void sci_dma_tx_complete(void *arg)
 	if (kfifo_len(&tport->xmit_fifo) < WAKEUP_CHARS)
 		uart_write_wakeup(port);
 
+	atomic_set(&s->first_time_tx, 1);
+
 	if (!kfifo_is_empty(&tport->xmit_fifo)) {
 		s->cookie_tx = 0;
 		schedule_work(&s->work_tx);
@@ -2076,6 +2081,10 @@ static unsigned int sci_tx_empty(struct uart_port *port)
 {
 	unsigned short status = sci_serial_in(port, SCxSR);
 	unsigned short in_tx_fifo = sci_txfill(port);
+	struct sci_port *s = to_sci_port(port);
+
+	if (!atomic_read(&s->first_time_tx))
+		return TIOCSER_TEMT;
 
 	return (status & SCxSR_TEND(port)) && !in_tx_fifo ? TIOCSER_TEMT : 0;
 }
@@ -2247,6 +2256,7 @@ static int sci_startup(struct uart_port *port)
 
 	dev_dbg(port->dev, "%s(%d)\n", __func__, port->line);
 
+	atomic_set(&s->first_time_tx, 0);
 	sci_request_dma(port);
 
 	ret = sci_request_irq(s);
@@ -2267,6 +2277,7 @@ static void sci_shutdown(struct uart_port *port)
 	dev_dbg(port->dev, "%s(%d)\n", __func__, port->line);
 
 	s->autorts = false;
+	atomic_set(&s->first_time_tx, 0);
 	mctrl_gpio_disable_ms(to_sci_port(port)->gpios);
 
 	uart_port_lock_irqsave(port, &flags);
-- 
2.39.2

[PATCH 3/9] serial: sh-sci: Clean sci_ports[0] after at earlycon exit

[Claudiu]

From: Claudiu Beznea <claudiu.beznea.uj@bp.renesas.com>

The early_console_setup() function initializes the sci_ports[0].port with
an object of type struct uart_port obtained from the object of type
struct earlycon_device received as argument by the early_console_setup().

It may happen that later, when the rest of the serial ports are probed,
the serial port that was used as earlycon (e.g., port A) to be mapped to a
different position in sci_ports[] and the slot 0 to be used by a different
serial port (e.g., port B), as follows:

sci_ports[0] = port A
sci_ports[X] = port B

In this case, the new port mapped at index zero will have associated data
that was used for earlycon.

In case this happens, after Linux boot, any access to the serial port that
maps on sci_ports[0] (port A) will block the serial port that was used as
earlycon (port B).

To fix this, add early_console_exit() that clean the sci_ports[0] at
earlycon exit time.

Fixes: 0b0cced19ab1 ("serial: sh-sci: Add CONFIG_SERIAL_EARLYCON support")
Cc: stable@vger.kernel.org
Signed-off-by: Claudiu Beznea <claudiu.beznea.uj@bp.renesas.com>
---
 drivers/tty/serial/sh-sci.c | 28 ++++++++++++++++++++++++++++
 1 file changed, 28 insertions(+)

diff --git a/drivers/tty/serial/sh-sci.c b/drivers/tty/serial/sh-sci.c
index 8e2d534401fa..2f8188bdb251 100644
--- a/drivers/tty/serial/sh-sci.c
+++ b/drivers/tty/serial/sh-sci.c
@@ -3546,6 +3546,32 @@ sh_early_platform_init_buffer("earlyprintk", &sci_driver,
 #ifdef CONFIG_SERIAL_SH_SCI_EARLYCON
 static struct plat_sci_port port_cfg __initdata;
 
+static int early_console_exit(struct console *co)
+{
+	struct sci_port *sci_port = &sci_ports[0];
+	struct uart_port *port = &sci_port->port;
+	unsigned long flags;
+	int locked = 1;
+
+	if (port->sysrq)
+		locked = 0;
+	else if (oops_in_progress)
+		locked = uart_port_trylock_irqsave(port, &flags);
+	else
+		uart_port_lock_irqsave(port, &flags);
+
+	/*
+	 * Clean the slot used by earlycon. A new SCI device might
+	 * map to this slot.
+	 */
+	memset(sci_ports, 0, sizeof(*sci_port));
+
+	if (locked)
+		uart_port_unlock_irqrestore(port, flags);
+
+	return 0;
+}
+
 static int __init early_console_setup(struct earlycon_device *device,
 				      int type)
 {
@@ -3562,6 +3588,8 @@ static int __init early_console_setup(struct earlycon_device *device,
 		       SCSCR_RE | SCSCR_TE | port_cfg.scscr);
 
 	device->con->write = serial_console_write;
+	device->con->exit = early_console_exit;
+
 	return 0;
 }
 static int __init sci_early_console_setup(struct earlycon_device *device,
-- 
2.39.2

[PATCH 4/9] serial: sh-sci: Update the suspend/resume support

[Claudiu]

From: Claudiu Beznea <claudiu.beznea.uj@bp.renesas.com>

The Renesas RZ/G3S supports a power saving mode where power to most of the
SoC components is turned off. When returning from this power saving mode,
SoC components need to be re-configured.

The SCIFs on the Renesas RZ/G3S need to be re-configured as well when
returning from this power saving mode. The sh-sci code already configures
the SCIF clocks, power domain and registers by calling uart_resume_port()
in sci_resume(). On suspend path the SCIF UART ports are suspended
accordingly (by calling uart_suspend_port() in sci_suspend()). The only
missing setting is the reset signal. For this assert/de-assert the reset
signal on driver suspend/resume.

In case the no_console_suspend is specified by the user, the registers need
to be saved on suspend path and restore on resume path. To do this the
sci_console_setup() function was added. There is no need to cache/restore
the status or FIFO registers. Only the control registers. To differentiate
b/w these, the struct sci_port_params::regs was updated with a new member
that specifies if the register needs to be chached on suspend. Only the
RZ_SCIFA instances were updated with this new support as the hardware for
the rest of variants was missing for testing.

Signed-off-by: Claudiu Beznea <claudiu.beznea.uj@bp.renesas.com>
---
 drivers/tty/serial/sh-sci.c | 53 ++++++++++++++++++++++++++++++-------
 1 file changed, 44 insertions(+), 9 deletions(-)

diff --git a/drivers/tty/serial/sh-sci.c b/drivers/tty/serial/sh-sci.c
index 2f8188bdb251..9548b8e97b8f 100644
--- a/drivers/tty/serial/sh-sci.c
+++ b/drivers/tty/serial/sh-sci.c
@@ -101,7 +101,7 @@ enum SCI_CLKS {
 		if ((_port)->sampling_rate_mask & SCI_SR((_sr)))
 
 struct plat_sci_reg {
-	u8 offset, size;
+	u8 offset, size, suspend_cacheable;
 };
 
 struct sci_port_params {
@@ -134,6 +134,8 @@ struct sci_port {
 	struct dma_chan			*chan_tx;
 	struct dma_chan			*chan_rx;
 
+	struct reset_control		*rstc;
+
 #ifdef CONFIG_SERIAL_SH_SCI_DMA
 	struct dma_chan			*chan_tx_saved;
 	struct dma_chan			*chan_rx_saved;
@@ -154,6 +156,7 @@ struct sci_port {
 	struct timer_list		rx_fifo_timer;
 	int				rx_fifo_timeout;
 	atomic_t			first_time_tx;
+	unsigned int			console_cached_regs[SCIx_NR_REGS];
 	u16				hscif_tot;
 
 	bool has_rtscts;
@@ -298,17 +301,17 @@ static const struct sci_port_params sci_port_params[SCIx_NR_REGTYPES] = {
 	 */
 	[SCIx_RZ_SCIFA_REGTYPE] = {
 		.regs = {
-			[SCSMR]		= { 0x00, 16 },
-			[SCBRR]		= { 0x02,  8 },
-			[SCSCR]		= { 0x04, 16 },
+			[SCSMR]		= { 0x00, 16, 1 },
+			[SCBRR]		= { 0x02,  8, 1 },
+			[SCSCR]		= { 0x04, 16, 1 },
 			[SCxTDR]	= { 0x06,  8 },
 			[SCxSR]		= { 0x08, 16 },
 			[SCxRDR]	= { 0x0A,  8 },
-			[SCFCR]		= { 0x0C, 16 },
+			[SCFCR]		= { 0x0C, 16, 1 },
 			[SCFDR]		= { 0x0E, 16 },
-			[SCSPTR]	= { 0x10, 16 },
+			[SCSPTR]	= { 0x10, 16, 1 },
 			[SCLSR]		= { 0x12, 16 },
-			[SEMR]		= { 0x14, 8 },
+			[SEMR]		= { 0x14, 8, 1 },
 		},
 		.fifosize = 16,
 		.overrun_reg = SCLSR,
@@ -3365,6 +3368,7 @@ static struct plat_sci_port *sci_parse_dt(struct platform_device *pdev,
 	}
 
 	sp = &sci_ports[id];
+	sp->rstc = rstc;
 	*dev_id = id;
 
 	p->type = SCI_OF_TYPE(data);
@@ -3492,13 +3496,34 @@ static int sci_probe(struct platform_device *dev)
 	return 0;
 }
 
+static void sci_console_setup(struct sci_port *s, bool save)
+{
+	for (u16 i = 0; i < SCIx_NR_REGS; i++) {
+		struct uart_port *port = &s->port;
+
+		if (!s->params->regs[i].suspend_cacheable)
+			continue;
+
+		if (save)
+			s->console_cached_regs[i] = sci_serial_in(port, i);
+		else
+			sci_serial_out(port, i, s->console_cached_regs[i]);
+	}
+}
+
 static __maybe_unused int sci_suspend(struct device *dev)
 {
 	struct sci_port *sport = dev_get_drvdata(dev);
 
-	if (sport)
+	if (sport) {
 		uart_suspend_port(&sci_uart_driver, &sport->port);
 
+		if (!console_suspend_enabled && uart_console(&sport->port))
+			sci_console_setup(sport, true);
+		else
+			return reset_control_assert(sport->rstc);
+	}
+
 	return 0;
 }
 
@@ -3506,8 +3531,18 @@ static __maybe_unused int sci_resume(struct device *dev)
 {
 	struct sci_port *sport = dev_get_drvdata(dev);
 
-	if (sport)
+	if (sport) {
+		if (!console_suspend_enabled && uart_console(&sport->port)) {
+			sci_console_setup(sport, false);
+		} else {
+			int ret = reset_control_deassert(sport->rstc);
+
+			if (ret)
+				return ret;
+		}
+
 		uart_resume_port(&sci_uart_driver, &sport->port);
+	}
 
 	return 0;
 }
-- 
2.39.2

[PATCH 5/9] arm64: dts: renesas: r9a08g045: Add the remaining SCIF interfaces

[Claudiu]

From: Claudiu Beznea <claudiu.beznea.uj@bp.renesas.com>

The Renesas RZ/G3S SoC has 6 SCIF interfaces. SCIF0 is used as debug
console. Add the remaining ones.

Signed-off-by: Claudiu Beznea <claudiu.beznea.uj@bp.renesas.com>
---
 arch/arm64/boot/dts/renesas/r9a08g045.dtsi | 90 ++++++++++++++++++++++
 1 file changed, 90 insertions(+)

diff --git a/arch/arm64/boot/dts/renesas/r9a08g045.dtsi b/arch/arm64/boot/dts/renesas/r9a08g045.dtsi
index be8a0a768c65..5b15ff2482ab 100644
--- a/arch/arm64/boot/dts/renesas/r9a08g045.dtsi
+++ b/arch/arm64/boot/dts/renesas/r9a08g045.dtsi
@@ -73,6 +73,96 @@ scif0: serial@1004b800 {
 			status = "disabled";
 		};
 
+		scif1: serial@1004bc00 {
+			compatible = "renesas,scif-r9a08g045", "renesas,scif-r9a07g044";
+			reg = <0 0x1004bc00 0 0x400>;
+			interrupts = <GIC_SPI 325 IRQ_TYPE_LEVEL_HIGH>,
+				     <GIC_SPI 327 IRQ_TYPE_LEVEL_HIGH>,
+				     <GIC_SPI 328 IRQ_TYPE_LEVEL_HIGH>,
+				     <GIC_SPI 326 IRQ_TYPE_LEVEL_HIGH>,
+				     <GIC_SPI 329 IRQ_TYPE_LEVEL_HIGH>,
+				     <GIC_SPI 329 IRQ_TYPE_LEVEL_HIGH>;
+			interrupt-names = "eri", "rxi", "txi",
+					  "bri", "dri", "tei";
+			clocks = <&cpg CPG_MOD R9A08G045_SCIF1_CLK_PCK>;
+			clock-names = "fck";
+			power-domains = <&cpg>;
+			resets = <&cpg R9A08G045_SCIF1_RST_SYSTEM_N>;
+			status = "disabled";
+		};
+
+		scif2: serial@1004c000 {
+			compatible = "renesas,scif-r9a08g045", "renesas,scif-r9a07g044";
+			reg = <0 0x1004c000 0 0x400>;
+			interrupts = <GIC_SPI 330 IRQ_TYPE_LEVEL_HIGH>,
+				     <GIC_SPI 332 IRQ_TYPE_LEVEL_HIGH>,
+				     <GIC_SPI 333 IRQ_TYPE_LEVEL_HIGH>,
+				     <GIC_SPI 331 IRQ_TYPE_LEVEL_HIGH>,
+				     <GIC_SPI 334 IRQ_TYPE_LEVEL_HIGH>,
+				     <GIC_SPI 334 IRQ_TYPE_LEVEL_HIGH>;
+			interrupt-names = "eri", "rxi", "txi",
+					  "bri", "dri", "tei";
+			clocks = <&cpg CPG_MOD R9A08G045_SCIF2_CLK_PCK>;
+			clock-names = "fck";
+			power-domains = <&cpg>;
+			resets = <&cpg R9A08G045_SCIF2_RST_SYSTEM_N>;
+			status = "disabled";
+		};
+
+		scif3: serial@1004c400 {
+			compatible = "renesas,scif-r9a08g045", "renesas,scif-r9a07g044";
+			reg = <0 0x1004c400 0 0x400>;
+			interrupts = <GIC_SPI 335 IRQ_TYPE_LEVEL_HIGH>,
+				     <GIC_SPI 337 IRQ_TYPE_LEVEL_HIGH>,
+				     <GIC_SPI 338 IRQ_TYPE_LEVEL_HIGH>,
+				     <GIC_SPI 336 IRQ_TYPE_LEVEL_HIGH>,
+				     <GIC_SPI 339 IRQ_TYPE_LEVEL_HIGH>,
+				     <GIC_SPI 339 IRQ_TYPE_LEVEL_HIGH>;
+			interrupt-names = "eri", "rxi", "txi",
+					  "bri", "dri", "tei";
+			clocks = <&cpg CPG_MOD R9A08G045_SCIF3_CLK_PCK>;
+			clock-names = "fck";
+			power-domains = <&cpg>;
+			resets = <&cpg R9A08G045_SCIF3_RST_SYSTEM_N>;
+			status = "disabled";
+		};
+
+		scif4: serial@1004c800 {
+			compatible = "renesas,scif-r9a08g045", "renesas,scif-r9a07g044";
+			reg = <0 0x1004c800 0 0x400>;
+			interrupts = <GIC_SPI 340 IRQ_TYPE_LEVEL_HIGH>,
+				     <GIC_SPI 342 IRQ_TYPE_LEVEL_HIGH>,
+				     <GIC_SPI 343 IRQ_TYPE_LEVEL_HIGH>,
+				     <GIC_SPI 341 IRQ_TYPE_LEVEL_HIGH>,
+				     <GIC_SPI 344 IRQ_TYPE_LEVEL_HIGH>,
+				     <GIC_SPI 344 IRQ_TYPE_LEVEL_HIGH>;
+			interrupt-names = "eri", "rxi", "txi",
+					  "bri", "dri", "tei";
+			clocks = <&cpg CPG_MOD R9A08G045_SCIF4_CLK_PCK>;
+			clock-names = "fck";
+			power-domains = <&cpg>;
+			resets = <&cpg R9A08G045_SCIF4_RST_SYSTEM_N>;
+			status = "disabled";
+		};
+
+		scif5: serial@1004e000 {
+			compatible = "renesas,scif-r9a08g045", "renesas,scif-r9a07g044";
+			reg = <0 0x1004e000 0 0x400>;
+			interrupts = <GIC_SPI 345 IRQ_TYPE_LEVEL_HIGH>,
+				     <GIC_SPI 347 IRQ_TYPE_LEVEL_HIGH>,
+				     <GIC_SPI 348 IRQ_TYPE_LEVEL_HIGH>,
+				     <GIC_SPI 346 IRQ_TYPE_LEVEL_HIGH>,
+				     <GIC_SPI 349 IRQ_TYPE_LEVEL_HIGH>,
+				     <GIC_SPI 349 IRQ_TYPE_LEVEL_HIGH>;
+			interrupt-names = "eri", "rxi", "txi",
+					  "bri", "dri", "tei";
+			clocks = <&cpg CPG_MOD R9A08G045_SCIF5_CLK_PCK>;
+			clock-names = "fck";
+			power-domains = <&cpg>;
+			resets = <&cpg R9A08G045_SCIF5_RST_SYSTEM_N>;
+			status = "disabled";
+		};
+
 		rtc: rtc@1004ec00 {
 			compatible = "renesas,r9a08g045-rtca3", "renesas,rz-rtca3";
 			reg = <0 0x1004ec00 0 0x400>;
-- 
2.39.2

[PATCH 6/9] arm64: dts: renesas: rzg3s-smarc: Fix the debug serial alias

[Claudiu]

From: Claudiu Beznea <claudiu.beznea.uj@bp.renesas.com>

The debug serial of the RZ/G3S is SCIF0 which is routed on the Renesas
RZ SMARC Carrier II board on the SER3_UART. Use serial3 alias for it for
better hardware description. Along with it, the chosen properties were
moved to the device tree corresponding to the RZ SMARC Carrier II board.

Fixes: adb4f0c5699c ("arm64: dts: renesas: Add initial support for RZ/G3S SMARC SoM")
Fixes: d1ae4200bb26 ("arm64: dts: renesas: Add initial device tree for RZ SMARC Carrier-II Board")
Signed-off-by: Claudiu Beznea <claudiu.beznea.uj@bp.renesas.com>
---
 arch/arm64/boot/dts/renesas/rzg3s-smarc-som.dtsi | 5 -----
 arch/arm64/boot/dts/renesas/rzg3s-smarc.dtsi     | 7 ++++++-
 2 files changed, 6 insertions(+), 6 deletions(-)

diff --git a/arch/arm64/boot/dts/renesas/rzg3s-smarc-som.dtsi b/arch/arm64/boot/dts/renesas/rzg3s-smarc-som.dtsi
index 2ed01d391554..55c72c8a0735 100644
--- a/arch/arm64/boot/dts/renesas/rzg3s-smarc-som.dtsi
+++ b/arch/arm64/boot/dts/renesas/rzg3s-smarc-som.dtsi
@@ -43,11 +43,6 @@ aliases {
 #endif
 	};
 
-	chosen {
-		bootargs = "ignore_loglevel";
-		stdout-path = "serial0:115200n8";
-	};
-
 	memory@48000000 {
 		device_type = "memory";
 		/* First 128MB is reserved for secure area. */
diff --git a/arch/arm64/boot/dts/renesas/rzg3s-smarc.dtsi b/arch/arm64/boot/dts/renesas/rzg3s-smarc.dtsi
index 4509151344c4..33b9873b225a 100644
--- a/arch/arm64/boot/dts/renesas/rzg3s-smarc.dtsi
+++ b/arch/arm64/boot/dts/renesas/rzg3s-smarc.dtsi
@@ -12,10 +12,15 @@
 / {
 	aliases {
 		i2c0 = &i2c0;
-		serial0 = &scif0;
+		serial3 = &scif0;
 		mmc1 = &sdhi1;
 	};
 
+	chosen {
+		bootargs = "ignore_loglevel";
+		stdout-path = "serial3:115200n8";
+	};
+
 	keys {
 		compatible = "gpio-keys";
 
-- 
2.39.2

[PATCH 7/9] arm64: dts: renesas: rzg3s-smarc-switches: Add a header to describe different switches

[Claudiu]

From: Claudiu Beznea <claudiu.beznea.uj@bp.renesas.com>

There are different switches available on both the RZ/G3S SMARC Module and
RZ SMARC Carrier II boards. These switches are used to route different SoC
signals to different parts available on board.

These switches are described in device trees through macros. These macros
are set accordingly such that the resulted compiled dtb to describe the
on-board switches states.

Based on the SW_CONFIG3 switch state (populated on the module board), the
SCIF3 SoC interface is routed or not to an U(S)ART pin header available on
the carrier board. As the SCIF3 is accessible through the carrier board,
the device tree enables it in the carrier DTS. To be able to cope with
these type of configurations, add a header file where all the on-board
switches can be described and shared accordingly between module and carrier
board.

Commit prepares the code to enable SCIF3 on the RZ/G3S carrier device
tree.

Signed-off-by: Claudiu Beznea <claudiu.beznea.uj@bp.renesas.com>
---
 .../boot/dts/renesas/rzg3s-smarc-som.dtsi     | 20 +-----------
 .../boot/dts/renesas/rzg3s-smarc-switches.h   | 32 +++++++++++++++++++
 2 files changed, 33 insertions(+), 19 deletions(-)
 create mode 100644 arch/arm64/boot/dts/renesas/rzg3s-smarc-switches.h

diff --git a/arch/arm64/boot/dts/renesas/rzg3s-smarc-som.dtsi b/arch/arm64/boot/dts/renesas/rzg3s-smarc-som.dtsi
index 55c72c8a0735..5c88e130c89e 100644
--- a/arch/arm64/boot/dts/renesas/rzg3s-smarc-som.dtsi
+++ b/arch/arm64/boot/dts/renesas/rzg3s-smarc-som.dtsi
@@ -9,25 +9,7 @@
 #include <dt-bindings/gpio/gpio.h>
 #include <dt-bindings/pinctrl/rzg2l-pinctrl.h>
 
-/*
- * On-board switches' states:
- * @SW_OFF: switch's state is OFF
- * @SW_ON:  switch's state is ON
- */
-#define SW_OFF		0
-#define SW_ON		1
-
-/*
- * SW_CONFIG[x] switches' states:
- * @SW_CONFIG2:
- *	SW_OFF - SD0 is connected to eMMC
- *	SW_ON  - SD0 is connected to uSD0 card
- * @SW_CONFIG3:
- *	SW_OFF - SD2 is connected to SoC
- *	SW_ON  - SCIF1, SSI0, IRQ0, IRQ1 connected to SoC
- */
-#define SW_CONFIG2	SW_OFF
-#define SW_CONFIG3	SW_ON
+#include "rzg3s-smarc-switches.h"
 
 / {
 	compatible = "renesas,rzg3s-smarcm", "renesas,r9a08g045s33", "renesas,r9a08g045";
diff --git a/arch/arm64/boot/dts/renesas/rzg3s-smarc-switches.h b/arch/arm64/boot/dts/renesas/rzg3s-smarc-switches.h
new file mode 100644
index 000000000000..e2d9b953f627
--- /dev/null
+++ b/arch/arm64/boot/dts/renesas/rzg3s-smarc-switches.h
@@ -0,0 +1,32 @@
+/* SPDX-License-Identifier: GPL-2.0 */
+/*
+ * On-board switches for the Renesas RZ/G3S SMARC Module and RZ SMARC Carrier II
+ * boards.
+ *
+ * Copyright (C) 2024 Renesas Electronics Corp.
+ */
+
+#ifndef __RZG3S_SMARC_SWITCHES__
+#define __RZG3S_SMARC_SWITCHES__
+
+/*
+ * On-board switches' states:
+ * @SW_OFF: switch's state is OFF
+ * @SW_ON:  switch's state is ON
+ */
+#define SW_OFF		0
+#define SW_ON		1
+
+/*
+ * SW_CONFIG[x] switches' states:
+ * @SW_CONFIG2:
+ *	SW_OFF - SD0 is connected to eMMC
+ *	SW_ON  - SD0 is connected to uSD0 card
+ * @SW_CONFIG3:
+ *	SW_OFF - SD2 is connected to SoC
+ *	SW_ON  - SCIF3, SSI3, IRQ0, IRQ1 connected to SoC
+ */
+#define SW_CONFIG2	SW_OFF
+#define SW_CONFIG3	SW_ON
+
+#endif /* __RZG3S_SMARC_SWITCHES__ */
-- 
2.39.2

[PATCH 8/9] arm64: dts: renesas: rzg3s-smarc: Enable SCIF3

[Claudiu]

From: Claudiu Beznea <claudiu.beznea.uj@bp.renesas.com>

Enable SCIF3. It is routed on the RZ SMARC Carrier II board on SER1_UART
interface.

Signed-off-by: Claudiu Beznea <claudiu.beznea.uj@bp.renesas.com>
---
 arch/arm64/boot/dts/renesas/rzg3s-smarc.dtsi | 18 ++++++++++++++++++
 1 file changed, 18 insertions(+)

diff --git a/arch/arm64/boot/dts/renesas/rzg3s-smarc.dtsi b/arch/arm64/boot/dts/renesas/rzg3s-smarc.dtsi
index 33b9873b225a..1be21ece131e 100644
--- a/arch/arm64/boot/dts/renesas/rzg3s-smarc.dtsi
+++ b/arch/arm64/boot/dts/renesas/rzg3s-smarc.dtsi
@@ -9,9 +9,14 @@
 #include <dt-bindings/input/input.h>
 #include <dt-bindings/pinctrl/rzg2l-pinctrl.h>
 
+#include "rzg3s-smarc-switches.h"
+
 / {
 	aliases {
 		i2c0 = &i2c0;
+#if SW_CONFIG3 == SW_ON
+		serial1 = &scif3;
+#endif
 		serial3 = &scif0;
 		mmc1 = &sdhi1;
 	};
@@ -102,6 +107,11 @@ scif0_pins: scif0 {
 			 <RZG2L_PORT_PINMUX(6, 4, 1)>; /* TXD */
 	};
 
+	scif3_pins: scif3 {
+		pinmux = <RZG2L_PORT_PINMUX(17, 2, 7)>, /* RXD */
+			 <RZG2L_PORT_PINMUX(17, 3, 7)>; /* TXD */
+	};
+
 	sdhi1_pins: sd1 {
 		data {
 			pins = "SD1_DATA0", "SD1_DATA1", "SD1_DATA2", "SD1_DATA3";
@@ -141,6 +151,14 @@ &scif0 {
 	status = "okay";
 };
 
+#if SW_CONFIG3 == SW_ON
+&scif3 {
+	pinctrl-names = "default";
+	pinctrl-0 = <&scif3_pins>;
+	status = "okay";
+};
+#endif
+
 &sdhi1 {
 	pinctrl-0 = <&sdhi1_pins>;
 	pinctrl-1 = <&sdhi1_pins_uhs>;
-- 
2.39.2

[PATCH 9/9] arm64: dts: renesas: r9a08g045s33-smarc-pmod: Add overlay for SCIF1

[Claudiu]

From: Claudiu Beznea <claudiu.beznea.uj@bp.renesas.com>

Add DT overlay for SCIF1 (of the Renesas RZ/G3S SoC) routed through the
PMOD1_3A interface available on the Renesas RZ SMARC Carrier II board.

Signed-off-by: Claudiu Beznea <claudiu.beznea.uj@bp.renesas.com>
---
 arch/arm64/boot/dts/renesas/Makefile          |  3 ++
 .../dts/renesas/r9a08g045s33-smarc-pmod.dtso  | 48 +++++++++++++++++++
 2 files changed, 51 insertions(+)
 create mode 100644 arch/arm64/boot/dts/renesas/r9a08g045s33-smarc-pmod.dtso

diff --git a/arch/arm64/boot/dts/renesas/Makefile b/arch/arm64/boot/dts/renesas/Makefile
index 97228a3cb99c..7ad52630d350 100644
--- a/arch/arm64/boot/dts/renesas/Makefile
+++ b/arch/arm64/boot/dts/renesas/Makefile
@@ -137,6 +137,9 @@ r9a07g054l2-smarc-cru-csi-ov5645-dtbs := r9a07g054l2-smarc.dtb r9a07g054l2-smarc
 dtb-$(CONFIG_ARCH_R9A07G054) += r9a07g054l2-smarc-cru-csi-ov5645.dtb
 
 dtb-$(CONFIG_ARCH_R9A08G045) += r9a08g045s33-smarc.dtb
+dtb-$(CONFIG_ARCH_R9A07G043) += r9a08g045s33-smarc-pmod.dtbo
+r9a08g045s33-smarc-pmod-dtbs := r9a08g045s33-smarc.dtb r9a08g045s33-smarc-pmod.dtbo
+dtb-$(CONFIG_ARCH_R9A07G043) += r9a08g045s33-smarc-pmod.dtb
 
 dtb-$(CONFIG_ARCH_R9A09G011) += r9a09g011-v2mevk2.dtb
 
diff --git a/arch/arm64/boot/dts/renesas/r9a08g045s33-smarc-pmod.dtso b/arch/arm64/boot/dts/renesas/r9a08g045s33-smarc-pmod.dtso
new file mode 100644
index 000000000000..7d637ab110e1
--- /dev/null
+++ b/arch/arm64/boot/dts/renesas/r9a08g045s33-smarc-pmod.dtso
@@ -0,0 +1,48 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * Device Tree Source for the RZ/G3S SMARC Carrier II EVK PMOD parts
+ *
+ * Copyright (C) 2024 Renesas Electronics Corp.
+ *
+ *
+ * [Connection]
+ *
+ * SMARC Carrier II EVK
+ * +--------------------------------------------+
+ * |PMOD1_3A (PMOD1 PIN HEADER)			|
+ * |	SCIF1_CTS# (pin1)  (pin7)  PMOD1_GPIO10	|
+ * |	SCIF1_TXD  (pin2)  (pin8)  PMOD1_GPIO11	|
+ * |	SCIF1_RXD  (pin3)  (pin9)  PMOD1_GPIO12	|
+ * |	SCIF1_RTS# (pin4)  (pin10) PMOD1_GPIO13	|
+ * |	GND	   (pin5)  (pin11) GND		|
+ * |	PWR_PMOD1  (pin6)  (pin12) GND		|
+ * +--------------------------------------------+
+ *
+ */
+
+/dts-v1/;
+/plugin/;
+
+#include <dt-bindings/pinctrl/rzg2l-pinctrl.h>
+
+&{/} {
+	aliases {
+		serial0 = "/soc/serial@1004bc00";
+	};
+};
+
+&pinctrl {
+	scif1_pins: scif1-pins {
+		pinmux = <RZG2L_PORT_PINMUX(14, 0, 1)>, /* TXD */
+			 <RZG2L_PORT_PINMUX(14, 1, 1)>, /* RXD */
+			 <RZG2L_PORT_PINMUX(16, 0, 1)>, /* CTS */
+			 <RZG2L_PORT_PINMUX(16, 1, 1)>; /* RTS */
+	};
+};
+
+&scif1 {
+	pinctrl-names = "default";
+	pinctrl-0 = <&scif1_pins>;
+	uart-has-rtscts;
+	status = "okay";
+};
-- 
2.39.2

Re: [PATCH 2/9] serial: sh-sci: Check if TX data was written to device in .tx_empty()

[Greg KH]

On Wed, Nov 06, 2024 at 02:01:11PM +0200, Claudiu wrote:
> From: Claudiu Beznea <claudiu.beznea.uj@bp.renesas.com>
> 
> On the Renesas RZ/G3S, when doing suspend to RAM, the uart_suspend_port()
> is called. The uart_suspend_port() calls 3 times the
> struct uart_port::ops::tx_empty() before shutting down the port.
> 
> According to the documentation, the struct uart_port::ops::tx_empty()
> API tests whether the transmitter FIFO and shifter for the port is
> empty.
> 
> The Renesas RZ/G3S SCIFA IP reports the number of data units stored in the
> transmit FIFO through the FDR (FIFO Data Count Register). The data units
> in the FIFOs are written in the shift register and transmitted from there.
> The TEND bit in the Serial Status Register reports if the data was
> transmitted from the shift register.
> 
> In the previous code, in the tx_empty() API implemented by the sh-sci
> driver, it is considered that the TX is empty if the hardware reports the
> TEND bit set and the number of data units in the FIFO is zero.
> 
> According to the HW manual, the TEND bit has the following meaning:
> 
> 0: Transmission is in the waiting state or in progress.
> 1: Transmission is completed.
> 
> It has been noticed that when opening the serial device w/o using it and
> then switch to a power saving mode, the tx_empty() call in the
> uart_port_suspend() function fails, leading to the "Unable to drain
> transmitter" message being printed on the console. This is because the
> TEND=0 if nothing has been transmitted and the FIFOs are empty. As the
> TEND=0 has double meaning (waiting state, in progress) we can't
> determined the scenario described above.
> 
> Add a software workaround for this. This sets a variable if any data has
> been sent on the serial console (when using PIO) or if the DMA callback has
> been called (meaning something has been transmitted).
> 
> Fixes: 73a19e4c0301 ("serial: sh-sci: Add DMA support.")
> Cc: stable@vger.kernel.org
> Signed-off-by: Claudiu Beznea <claudiu.beznea.uj@bp.renesas.com>
> ---
>  drivers/tty/serial/sh-sci.c | 11 +++++++++++
>  1 file changed, 11 insertions(+)
> 
> diff --git a/drivers/tty/serial/sh-sci.c b/drivers/tty/serial/sh-sci.c
> index df523c744423..8e2d534401fa 100644
> --- a/drivers/tty/serial/sh-sci.c
> +++ b/drivers/tty/serial/sh-sci.c
> @@ -153,6 +153,7 @@ struct sci_port {
>  	int				rx_trigger;
>  	struct timer_list		rx_fifo_timer;
>  	int				rx_fifo_timeout;
> +	atomic_t			first_time_tx;

Don't use an atomic variable for an informational thing like this, it is
racy and doesn't work properly.  Either use a real lock (because you
care about the locking stuff here), or just use a boolean and live with
any potential races.



>  	u16				hscif_tot;
>  
>  	bool has_rtscts;
> @@ -850,6 +851,7 @@ static void sci_transmit_chars(struct uart_port *port)
>  {
>  	struct tty_port *tport = &port->state->port;
>  	unsigned int stopped = uart_tx_stopped(port);
> +	struct sci_port *s = to_sci_port(port);
>  	unsigned short status;
>  	unsigned short ctrl;
>  	int count;
> @@ -885,6 +887,7 @@ static void sci_transmit_chars(struct uart_port *port)
>  		}
>  
>  		sci_serial_out(port, SCxTDR, c);
> +		atomic_set(&s->first_time_tx, 1);
>  
>  		port->icount.tx++;
>  	} while (--count > 0);
> @@ -1241,6 +1244,8 @@ static void sci_dma_tx_complete(void *arg)
>  	if (kfifo_len(&tport->xmit_fifo) < WAKEUP_CHARS)
>  		uart_write_wakeup(port);
>  
> +	atomic_set(&s->first_time_tx, 1);
> +
>  	if (!kfifo_is_empty(&tport->xmit_fifo)) {
>  		s->cookie_tx = 0;
>  		schedule_work(&s->work_tx);
> @@ -2076,6 +2081,10 @@ static unsigned int sci_tx_empty(struct uart_port *port)
>  {
>  	unsigned short status = sci_serial_in(port, SCxSR);
>  	unsigned short in_tx_fifo = sci_txfill(port);
> +	struct sci_port *s = to_sci_port(port);
> +
> +	if (!atomic_read(&s->first_time_tx))
> +		return TIOCSER_TEMT;

See, what happens here if the value changes right after you check it?
Being an atomic doesn't mean anything :(

thanks,

greg k-h

Re: [PATCH 4/9] serial: sh-sci: Update the suspend/resume support

[Greg KH]

On Wed, Nov 06, 2024 at 02:01:13PM +0200, Claudiu wrote:
> From: Claudiu Beznea <claudiu.beznea.uj@bp.renesas.com>
> 
> The Renesas RZ/G3S supports a power saving mode where power to most of the
> SoC components is turned off. When returning from this power saving mode,
> SoC components need to be re-configured.
> 
> The SCIFs on the Renesas RZ/G3S need to be re-configured as well when
> returning from this power saving mode. The sh-sci code already configures
> the SCIF clocks, power domain and registers by calling uart_resume_port()
> in sci_resume(). On suspend path the SCIF UART ports are suspended
> accordingly (by calling uart_suspend_port() in sci_suspend()). The only
> missing setting is the reset signal. For this assert/de-assert the reset
> signal on driver suspend/resume.
> 
> In case the no_console_suspend is specified by the user, the registers need
> to be saved on suspend path and restore on resume path. To do this the
> sci_console_setup() function was added. There is no need to cache/restore
> the status or FIFO registers. Only the control registers. To differentiate
> b/w these, the struct sci_port_params::regs was updated with a new member
> that specifies if the register needs to be chached on suspend. Only the
> RZ_SCIFA instances were updated with this new support as the hardware for
> the rest of variants was missing for testing.
> 
> Signed-off-by: Claudiu Beznea <claudiu.beznea.uj@bp.renesas.com>
> ---
>  drivers/tty/serial/sh-sci.c | 53 ++++++++++++++++++++++++++++++-------
>  1 file changed, 44 insertions(+), 9 deletions(-)

Does not apply as I didn't take the atomic_t patch :(

thanks,

greg k-h

Re: [PATCH 2/9] serial: sh-sci: Check if TX data was written to device in .tx_empty()

[Claudiu Beznea]

Hi, Greg,

On 07.11.2024 10:47, Greg KH wrote:
> On Wed, Nov 06, 2024 at 02:01:11PM +0200, Claudiu wrote:
>> From: Claudiu Beznea <claudiu.beznea.uj@bp.renesas.com>
>>
>> On the Renesas RZ/G3S, when doing suspend to RAM, the uart_suspend_port()
>> is called. The uart_suspend_port() calls 3 times the
>> struct uart_port::ops::tx_empty() before shutting down the port.
>>
>> According to the documentation, the struct uart_port::ops::tx_empty()
>> API tests whether the transmitter FIFO and shifter for the port is
>> empty.
>>
>> The Renesas RZ/G3S SCIFA IP reports the number of data units stored in the
>> transmit FIFO through the FDR (FIFO Data Count Register). The data units
>> in the FIFOs are written in the shift register and transmitted from there.
>> The TEND bit in the Serial Status Register reports if the data was
>> transmitted from the shift register.
>>
>> In the previous code, in the tx_empty() API implemented by the sh-sci
>> driver, it is considered that the TX is empty if the hardware reports the
>> TEND bit set and the number of data units in the FIFO is zero.
>>
>> According to the HW manual, the TEND bit has the following meaning:
>>
>> 0: Transmission is in the waiting state or in progress.
>> 1: Transmission is completed.
>>
>> It has been noticed that when opening the serial device w/o using it and
>> then switch to a power saving mode, the tx_empty() call in the
>> uart_port_suspend() function fails, leading to the "Unable to drain
>> transmitter" message being printed on the console. This is because the
>> TEND=0 if nothing has been transmitted and the FIFOs are empty. As the
>> TEND=0 has double meaning (waiting state, in progress) we can't
>> determined the scenario described above.
>>
>> Add a software workaround for this. This sets a variable if any data has
>> been sent on the serial console (when using PIO) or if the DMA callback has
>> been called (meaning something has been transmitted).
>>
>> Fixes: 73a19e4c0301 ("serial: sh-sci: Add DMA support.")
>> Cc: stable@vger.kernel.org
>> Signed-off-by: Claudiu Beznea <claudiu.beznea.uj@bp.renesas.com>
>> ---
>>  drivers/tty/serial/sh-sci.c | 11 +++++++++++
>>  1 file changed, 11 insertions(+)
>>
>> diff --git a/drivers/tty/serial/sh-sci.c b/drivers/tty/serial/sh-sci.c
>> index df523c744423..8e2d534401fa 100644
>> --- a/drivers/tty/serial/sh-sci.c
>> +++ b/drivers/tty/serial/sh-sci.c
>> @@ -153,6 +153,7 @@ struct sci_port {
>>  	int				rx_trigger;
>>  	struct timer_list		rx_fifo_timer;
>>  	int				rx_fifo_timeout;
>> +	atomic_t			first_time_tx;
> 
> Don't use an atomic variable for an informational thing like this, it is
> racy and doesn't work properly.  Either use a real lock (because you
> care about the locking stuff here), or just use a boolean and live with
> any potential races.

OK, I'll drop it and use a boolean.

> 
> 
> 
>>  	u16				hscif_tot;
>>  
>>  	bool has_rtscts;
>> @@ -850,6 +851,7 @@ static void sci_transmit_chars(struct uart_port *port)
>>  {
>>  	struct tty_port *tport = &port->state->port;
>>  	unsigned int stopped = uart_tx_stopped(port);
>> +	struct sci_port *s = to_sci_port(port);
>>  	unsigned short status;
>>  	unsigned short ctrl;
>>  	int count;
>> @@ -885,6 +887,7 @@ static void sci_transmit_chars(struct uart_port *port)
>>  		}
>>  
>>  		sci_serial_out(port, SCxTDR, c);
>> +		atomic_set(&s->first_time_tx, 1);
>>  
>>  		port->icount.tx++;
>>  	} while (--count > 0);
>> @@ -1241,6 +1244,8 @@ static void sci_dma_tx_complete(void *arg)
>>  	if (kfifo_len(&tport->xmit_fifo) < WAKEUP_CHARS)
>>  		uart_write_wakeup(port);
>>  
>> +	atomic_set(&s->first_time_tx, 1);
>> +
>>  	if (!kfifo_is_empty(&tport->xmit_fifo)) {
>>  		s->cookie_tx = 0;
>>  		schedule_work(&s->work_tx);
>> @@ -2076,6 +2081,10 @@ static unsigned int sci_tx_empty(struct uart_port *port)
>>  {
>>  	unsigned short status = sci_serial_in(port, SCxSR);
>>  	unsigned short in_tx_fifo = sci_txfill(port);
>> +	struct sci_port *s = to_sci_port(port);
>> +
>> +	if (!atomic_read(&s->first_time_tx))
>> +		return TIOCSER_TEMT;
> 
> See, what happens here if the value changes right after you check it?

I agree. I am aware if it.

I chose this approach (w/o locking) as I noticed (as of my code checking)
that this function is called in kernel through uart_ioctl(),
uart_wait_until_sent(), uart_suspend_port().

The uart_wait_until_sent(), uart_suspend_port() are implementing a multiple
try approach when checking the ops::tx_timeout() return value.

I haven't checked any user space application but considered that it might
work in a similar way.

I will switch to a boolean in the next version.

Thank you,
Claudiu Beznea


> Being an atomic doesn't mean anything :(
> 
> thanks,
> 
> greg k-h

Re: [PATCH 3/9] serial: sh-sci: Clean sci_ports[0] after at earlycon exit

[Geert Uytterhoeven]

Hi Claudiu,

Thanks for your patch, which is now commit 3791ea69a4858b81 ("serial:
sh-sci: Clean sci_ports[0] after at earlycon exit") in tty/tty-next.

On Wed, Nov 6, 2024 at 1:02 PM Claudiu <claudiu.beznea@tuxon.dev> wrote:
> From: Claudiu Beznea <claudiu.beznea.uj@bp.renesas.com>
>
> The early_console_setup() function initializes the sci_ports[0].port with
> an object of type struct uart_port obtained from the object of type
> struct earlycon_device received as argument by the early_console_setup().
>
> It may happen that later, when the rest of the serial ports are probed,
> the serial port that was used as earlycon (e.g., port A) to be mapped to a
> different position in sci_ports[] and the slot 0 to be used by a different
> serial port (e.g., port B), as follows:
>
> sci_ports[0] = port A
> sci_ports[X] = port B

Haven't you mixed A and B?

> In this case, the new port mapped at index zero will have associated data
> that was used for earlycon.

Oops, do you have a simple reproducer for this?

> In case this happens, after Linux boot, any access to the serial port that
> maps on sci_ports[0] (port A) will block the serial port that was used as
> earlycon (port B).

Again, A <-> B?

> To fix this, add early_console_exit() that clean the sci_ports[0] at
> earlycon exit time.
>
> Fixes: 0b0cced19ab1 ("serial: sh-sci: Add CONFIG_SERIAL_EARLYCON support")
> Cc: stable@vger.kernel.org
> Signed-off-by: Claudiu Beznea <claudiu.beznea.uj@bp.renesas.com>

This causes a crash (lock-up without any output) when
CONFIG_DEBUG_SPINLOCK=y (e.g. CONFIG_PROVE_LOCKING=y).

> --- a/drivers/tty/serial/sh-sci.c
> +++ b/drivers/tty/serial/sh-sci.c
> @@ -3546,6 +3546,32 @@ sh_early_platform_init_buffer("earlyprintk", &sci_driver,
>  #ifdef CONFIG_SERIAL_SH_SCI_EARLYCON
>  static struct plat_sci_port port_cfg __initdata;
>
> +static int early_console_exit(struct console *co)
> +{
> +       struct sci_port *sci_port = &sci_ports[0];
> +       struct uart_port *port = &sci_port->port;
> +       unsigned long flags;
> +       int locked = 1;
> +
> +       if (port->sysrq)
> +               locked = 0;
> +       else if (oops_in_progress)
> +               locked = uart_port_trylock_irqsave(port, &flags);
> +       else
> +               uart_port_lock_irqsave(port, &flags);
> +
> +       /*
> +        * Clean the slot used by earlycon. A new SCI device might
> +        * map to this slot.
> +        */
> +       memset(sci_ports, 0, sizeof(*sci_port));

Nit: I'd rather use "*sci_port" instead of "sci_ports".

> +
> +       if (locked)
> +               uart_port_unlock_irqrestore(port, flags);

"BUG: spinlock bad magic", as you've just cleared the port, including
the spinlock.

I guess we can just remove all locking from this function to fix this?

However, could it happen that the new device taking slot 0 is probed
before the early console is terminated?  In that case, its active
sci_ports[] entry would be cleared when early_console_exit() is called.

Also, what happens if "earlycon keep_bootcon" is passed on the kernel
command line, and the new device takes slot 0?

Thanks!

> +
> +       return 0;
> +}
> +
>  static int __init early_console_setup(struct earlycon_device *device,
>                                       int type)
>  {
> @@ -3562,6 +3588,8 @@ static int __init early_console_setup(struct earlycon_device *device,
>                        SCSCR_RE | SCSCR_TE | port_cfg.scscr);
>
>         device->con->write = serial_console_write;
> +       device->con->exit = early_console_exit;
> +
>         return 0;
>  }
>  static int __init sci_early_console_setup(struct earlycon_device *device,

Gr{oetje,eeting}s,

                        Geert

-- 
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
                                -- Linus Torvalds

Re: [PATCH 3/9] serial: sh-sci: Clean sci_ports[0] after at earlycon exit

[Claudiu Beznea]

Hi, Geert,

On 27.11.2024 18:28, Geert Uytterhoeven wrote:
> Hi Claudiu,
> 
> Thanks for your patch, which is now commit 3791ea69a4858b81 ("serial:
> sh-sci: Clean sci_ports[0] after at earlycon exit") in tty/tty-next.
> 
> On Wed, Nov 6, 2024 at 1:02 PM Claudiu <claudiu.beznea@tuxon.dev> wrote:
>> From: Claudiu Beznea <claudiu.beznea.uj@bp.renesas.com>
>>
>> The early_console_setup() function initializes the sci_ports[0].port with
>> an object of type struct uart_port obtained from the object of type
>> struct earlycon_device received as argument by the early_console_setup().
>>
>> It may happen that later, when the rest of the serial ports are probed,
>> the serial port that was used as earlycon (e.g., port A) to be mapped to a
>> different position in sci_ports[] and the slot 0 to be used by a different
>> serial port (e.g., port B), as follows:
>>
>> sci_ports[0] = port A
>> sci_ports[X] = port B
> 
> Haven't you mixed A and B?
> 
>> In this case, the new port mapped at index zero will have associated data
>> that was used for earlycon.
> 
> Oops, do you have a simple reproducer for this?

It is reproducible with patches:
- [PATCH 6/9] arm64: dts: renesas: rzg3s-smarc: Fix the debug serial alias
- [PATCH 9/9] arm64: dts: renesas: r9a08g045s33-smarc-pmod: Add overlay for
SCIF1

After boot, cat /dev/ttySC0 will lead to the issue described.

> 
>> In case this happens, after Linux boot, any access to the serial port that
>> maps on sci_ports[0] (port A) will block the serial port that was used as
>> earlycon (port B).
> 
> Again, A <-> B?
> 
>> To fix this, add early_console_exit() that clean the sci_ports[0] at
>> earlycon exit time.
>>
>> Fixes: 0b0cced19ab1 ("serial: sh-sci: Add CONFIG_SERIAL_EARLYCON support")
>> Cc: stable@vger.kernel.org
>> Signed-off-by: Claudiu Beznea <claudiu.beznea.uj@bp.renesas.com>
> 
> This causes a crash (lock-up without any output) when
> CONFIG_DEBUG_SPINLOCK=y (e.g. CONFIG_PROVE_LOCKING=y).

I missed to check this. Thank you for testing it.

> 
>> --- a/drivers/tty/serial/sh-sci.c
>> +++ b/drivers/tty/serial/sh-sci.c
>> @@ -3546,6 +3546,32 @@ sh_early_platform_init_buffer("earlyprintk", &sci_driver,
>>  #ifdef CONFIG_SERIAL_SH_SCI_EARLYCON
>>  static struct plat_sci_port port_cfg __initdata;
>>
>> +static int early_console_exit(struct console *co)
>> +{
>> +       struct sci_port *sci_port = &sci_ports[0];
>> +       struct uart_port *port = &sci_port->port;
>> +       unsigned long flags;
>> +       int locked = 1;
>> +
>> +       if (port->sysrq)
>> +               locked = 0;
>> +       else if (oops_in_progress)
>> +               locked = uart_port_trylock_irqsave(port, &flags);
>> +       else
>> +               uart_port_lock_irqsave(port, &flags);
>> +
>> +       /*
>> +        * Clean the slot used by earlycon. A new SCI device might
>> +        * map to this slot.
>> +        */
>> +       memset(sci_ports, 0, sizeof(*sci_port));
> 
> Nit: I'd rather use "*sci_port" instead of "sci_ports".

That would be better, indeed.

> 
>> +
>> +       if (locked)
>> +               uart_port_unlock_irqrestore(port, flags);
> 
> "BUG: spinlock bad magic", as you've just cleared the port, including
> the spinlock.
> 
> I guess we can just remove all locking from this function to fix this?

I'll look to it.

> 
> However, could it happen that the new device taking slot 0 is probed
> before the early console is terminated?

I don't know to answer this. In my testing I haven't encountered it.

> In that case, its active
> sci_ports[] entry would be cleared when early_console_exit() is called.
> 
> Also, what happens if "earlycon keep_bootcon" is passed on the kernel
> command line, and the new device takes slot 0?

I checked it with earlycon and the serial device being on slot 0. In this
case it was OK.

> 
> Thanks!
> 
>> +
>> +       return 0;
>> +}
>> +
>>  static int __init early_console_setup(struct earlycon_device *device,
>>                                       int type)
>>  {
>> @@ -3562,6 +3588,8 @@ static int __init early_console_setup(struct earlycon_device *device,
>>                        SCSCR_RE | SCSCR_TE | port_cfg.scscr);
>>
>>         device->con->write = serial_console_write;
>> +       device->con->exit = early_console_exit;
>> +
>>         return 0;
>>  }
>>  static int __init sci_early_console_setup(struct earlycon_device *device,
> 
> Gr{oetje,eeting}s,
> 
>                         Geert
>