[PATCH 00/13] Add DMA support for RZ/T2H RSPI

[PATCH 00/13] Add DMA support for RZ/T2H RSPI

[Cosmin Tanislav]

The DMA controller can be used to transfer data to and from the SPI
controller without involving the CPU for each word of a SPI transfer.

Add support for DMA mode, and do some other cleanups while touching the
same code.

The dts changes in this series depend on the DMA series [1].

[1]: https://lore.kernel.org/lkml/20251201124911.572395-1-cosmin-gabriel.tanislav.xa@renesas.com/

Cosmin Tanislav (13):
  spi: rzv2h-rspi: fix rzv2h_rspi_transfer_one() indentation
  spi: rzv2h-rspi: remove call to spi_finalize_current_transfer()
  spi: rzv2h-rspi: do not set SPI_TRANS_FAIL_IO
  spi: rzv2h-rspi: use device-managed APIs
  spi: rzv2h-rspi: store RX interrupt in state
  spi: rzv2h-rspi: set MUST_RX/MUST_TX
  spi: rzv2h-rspi: set TX FIFO threshold to 0
  spi: rzv2h-rspi: enable TX buffer empty interrupt
  spi: rzv2h-rspi: split out PIO transfer
  dt-bindings: spi: renesas,rzv2h-rspi: document optional support for
    DMA
  spi: rzv2h-rspi: add support for DMA mode
  arm64: dts: renesas: r9a09g077: wire up DMA support for SPI
  arm64: dts: renesas: r9a09g087: wire up DMA support for SPI

 .../bindings/spi/renesas,rzv2h-rspi.yaml      |   8 +
 arch/arm64/boot/dts/renesas/r9a09g077.dtsi    |   8 +
 arch/arm64/boot/dts/renesas/r9a09g087.dtsi    |   8 +
 drivers/spi/spi-rzv2h-rspi.c                  | 281 ++++++++++++++----
 4 files changed, 242 insertions(+), 63 deletions(-)

-- 
2.52.0

[PATCH 01/13] spi: rzv2h-rspi: fix rzv2h_rspi_transfer_one() indentation

[Cosmin Tanislav]

Add the missing space to align to open pararenthesis.

Signed-off-by: Cosmin Tanislav <cosmin-gabriel.tanislav.xa@renesas.com>
---
 drivers/spi/spi-rzv2h-rspi.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/drivers/spi/spi-rzv2h-rspi.c b/drivers/spi/spi-rzv2h-rspi.c
index 1db7e4e5d64e..8cffc9cb6b6b 100644
--- a/drivers/spi/spi-rzv2h-rspi.c
+++ b/drivers/spi/spi-rzv2h-rspi.c
@@ -225,8 +225,8 @@ static int rzv2h_rspi_receive(struct rzv2h_rspi_priv *rspi, void *rxbuf,
 }
 
 static int rzv2h_rspi_transfer_one(struct spi_controller *controller,
-				  struct spi_device *spi,
-				  struct spi_transfer *transfer)
+				   struct spi_device *spi,
+				   struct spi_transfer *transfer)
 {
 	struct rzv2h_rspi_priv *rspi = spi_controller_get_devdata(controller);
 	unsigned int words_to_transfer, i;
-- 
2.52.0

[PATCH 02/13] spi: rzv2h-rspi: remove call to spi_finalize_current_transfer()

[Cosmin Tanislav]

A call to spi_finalize_current_transfer() is only needed when the SPI
transfer is completed outside of the current context, when the
.transfer_one() implementation returns > 0.

Since the SPI transfer is completed in the current context, and we
return 0 from .transfer_one(), the SPI core assumes that the transfer
has completed and it does not wait for the completion variable that
would be set by a call to spi_finalize_current_transfer().

Remove the call to spi_finalize_current_transfer().

Signed-off-by: Cosmin Tanislav <cosmin-gabriel.tanislav.xa@renesas.com>
---
 drivers/spi/spi-rzv2h-rspi.c | 2 --
 1 file changed, 2 deletions(-)

diff --git a/drivers/spi/spi-rzv2h-rspi.c b/drivers/spi/spi-rzv2h-rspi.c
index 8cffc9cb6b6b..beea4fb83d10 100644
--- a/drivers/spi/spi-rzv2h-rspi.c
+++ b/drivers/spi/spi-rzv2h-rspi.c
@@ -250,8 +250,6 @@ static int rzv2h_rspi_transfer_one(struct spi_controller *controller,
 	if (ret)
 		transfer->error = SPI_TRANS_FAIL_IO;
 
-	spi_finalize_current_transfer(controller);
-
 	return ret;
 }
 
-- 
2.52.0

[PATCH 03/13] spi: rzv2h-rspi: do not set SPI_TRANS_FAIL_IO

[Cosmin Tanislav]

Setting SPI_TRANS_FAIL_IO has no effect if the transfer completes in the
current context, as it is only handled when .transfer_one() returns > 0,
when the SPI core must wait for the SPI transfer to complete.

Do not set SPI_TRANS_FAIL_IO as we either return an error or 0, since we
do our own waiting.

Signed-off-by: Cosmin Tanislav <cosmin-gabriel.tanislav.xa@renesas.com>
---
 drivers/spi/spi-rzv2h-rspi.c | 3 ---
 1 file changed, 3 deletions(-)

diff --git a/drivers/spi/spi-rzv2h-rspi.c b/drivers/spi/spi-rzv2h-rspi.c
index beea4fb83d10..02424d4e722a 100644
--- a/drivers/spi/spi-rzv2h-rspi.c
+++ b/drivers/spi/spi-rzv2h-rspi.c
@@ -247,9 +247,6 @@ static int rzv2h_rspi_transfer_one(struct spi_controller *controller,
 
 	rzv2h_rspi_clear_all_irqs(rspi);
 
-	if (ret)
-		transfer->error = SPI_TRANS_FAIL_IO;
-
 	return ret;
 }
 
-- 
2.52.0

[PATCH 04/13] spi: rzv2h-rspi: use device-managed APIs

[Cosmin Tanislav]

Non-device-managed APIs were initially used here to avoid the buggy
interaction between PM domains and device-managed actions.

Commit f99508074e78 ("PM: domains: Detach on device_unbind_cleanup()")
fixed the interaction between PM domains and device-managed actions.

Simplify the code by using device-managed actions to unregister the SPI
controller and to assert and release the resets.

Signed-off-by: Cosmin Tanislav <cosmin-gabriel.tanislav.xa@renesas.com>
---
 drivers/spi/spi-rzv2h-rspi.c | 58 ++++++++++++------------------------
 1 file changed, 19 insertions(+), 39 deletions(-)

diff --git a/drivers/spi/spi-rzv2h-rspi.c b/drivers/spi/spi-rzv2h-rspi.c
index 02424d4e722a..6163ada3ccbb 100644
--- a/drivers/spi/spi-rzv2h-rspi.c
+++ b/drivers/spi/spi-rzv2h-rspi.c
@@ -93,7 +93,6 @@ struct rzv2h_rspi_info {
 };
 
 struct rzv2h_rspi_priv {
-	struct reset_control_bulk_data resets[RSPI_RESET_NUM];
 	struct spi_controller *controller;
 	const struct rzv2h_rspi_info *info;
 	void __iomem *base;
@@ -533,6 +532,7 @@ static int rzv2h_rspi_probe(struct platform_device *pdev)
 	struct spi_controller *controller;
 	struct device *dev = &pdev->dev;
 	struct rzv2h_rspi_priv *rspi;
+	struct reset_control *reset;
 	struct clk_bulk_data *clks;
 	int irq_rx, ret, i;
 	long tclk_rate;
@@ -568,28 +568,29 @@ static int rzv2h_rspi_probe(struct platform_device *pdev)
 	if (!rspi->tclk)
 		return dev_err_probe(dev, -EINVAL, "Failed to get tclk\n");
 
-	rspi->resets[0].id = "presetn";
-	rspi->resets[1].id = "tresetn";
-	ret = devm_reset_control_bulk_get_optional_exclusive(dev, RSPI_RESET_NUM,
-							     rspi->resets);
-	if (ret)
-		return dev_err_probe(dev, ret, "cannot get resets\n");
+	reset = devm_reset_control_get_optional_exclusive_deasserted(&pdev->dev,
+								     "presetn");
+	if (IS_ERR(reset))
+		return dev_err_probe(&pdev->dev, PTR_ERR(reset),
+				     "cannot get presetn reset\n");
+
+	reset = devm_reset_control_get_optional_exclusive_deasserted(&pdev->dev,
+								     "tresetn");
+	if (IS_ERR(reset))
+		return dev_err_probe(&pdev->dev, PTR_ERR(reset),
+				     "cannot get tresetn reset\n");
 
 	irq_rx = platform_get_irq_byname(pdev, "rx");
 	if (irq_rx < 0)
 		return dev_err_probe(dev, irq_rx, "cannot get IRQ 'rx'\n");
 
-	ret = reset_control_bulk_deassert(RSPI_RESET_NUM, rspi->resets);
-	if (ret)
-		return dev_err_probe(dev, ret, "failed to deassert resets\n");
-
 	init_waitqueue_head(&rspi->wait);
 
 	ret = devm_request_irq(dev, irq_rx, rzv2h_rx_irq_handler, 0,
 			       dev_name(dev), rspi);
 	if (ret) {
 		dev_err(dev, "cannot request `rx` IRQ\n");
-		goto quit_resets;
+		return ret;
 	}
 
 	controller->mode_bits = SPI_CPHA | SPI_CPOL | SPI_CS_HIGH |
@@ -601,20 +602,16 @@ static int rzv2h_rspi_probe(struct platform_device *pdev)
 	controller->transfer_one = rzv2h_rspi_transfer_one;
 
 	tclk_rate = clk_round_rate(rspi->tclk, 0);
-	if (tclk_rate < 0) {
-		ret = tclk_rate;
-		goto quit_resets;
-	}
+	if (tclk_rate < 0)
+		return tclk_rate;
 
 	controller->min_speed_hz = rzv2h_rspi_calc_bitrate(tclk_rate,
 							   RSPI_SPBR_SPR_MAX,
 							   RSPI_SPCMD_BRDV_MAX);
 
 	tclk_rate = clk_round_rate(rspi->tclk, ULONG_MAX);
-	if (tclk_rate < 0) {
-		ret = tclk_rate;
-		goto quit_resets;
-	}
+	if (tclk_rate < 0)
+		return tclk_rate;
 
 	controller->max_speed_hz = rzv2h_rspi_calc_bitrate(tclk_rate,
 							   RSPI_SPBR_SPR_MIN,
@@ -622,29 +619,13 @@ static int rzv2h_rspi_probe(struct platform_device *pdev)
 
 	device_set_node(&controller->dev, dev_fwnode(dev));
 
-	ret = spi_register_controller(controller);
-	if (ret) {
+	ret = devm_spi_register_controller(dev, controller);
+	if (ret)
 		dev_err(dev, "register controller failed\n");
-		goto quit_resets;
-	}
-
-	return 0;
-
-quit_resets:
-	reset_control_bulk_assert(RSPI_RESET_NUM, rspi->resets);
 
 	return ret;
 }
 
-static void rzv2h_rspi_remove(struct platform_device *pdev)
-{
-	struct rzv2h_rspi_priv *rspi = platform_get_drvdata(pdev);
-
-	spi_unregister_controller(rspi->controller);
-
-	reset_control_bulk_assert(RSPI_RESET_NUM, rspi->resets);
-}
-
 static const struct rzv2h_rspi_info rzv2h_info = {
 	.find_tclk_rate = rzv2h_rspi_find_rate_fixed,
 	.tclk_name = "tclk",
@@ -669,7 +650,6 @@ MODULE_DEVICE_TABLE(of, rzv2h_rspi_match);
 
 static struct platform_driver rzv2h_rspi_drv = {
 	.probe = rzv2h_rspi_probe,
-	.remove = rzv2h_rspi_remove,
 	.driver = {
 		.name = "rzv2h_rspi",
 		.of_match_table = rzv2h_rspi_match,
-- 
2.52.0

[PATCH 05/13] spi: rzv2h-rspi: store RX interrupt in state

[Cosmin Tanislav]

In preparation for implementing DMA support, store the RX interrupt
number in the private state, to allow disabling it during DMA.

Signed-off-by: Cosmin Tanislav <cosmin-gabriel.tanislav.xa@renesas.com>
---
 drivers/spi/spi-rzv2h-rspi.c | 11 ++++++-----
 1 file changed, 6 insertions(+), 5 deletions(-)

diff --git a/drivers/spi/spi-rzv2h-rspi.c b/drivers/spi/spi-rzv2h-rspi.c
index 6163ada3ccbb..50fd7ddef58d 100644
--- a/drivers/spi/spi-rzv2h-rspi.c
+++ b/drivers/spi/spi-rzv2h-rspi.c
@@ -100,6 +100,7 @@ struct rzv2h_rspi_priv {
 	struct clk *pclk;
 	wait_queue_head_t wait;
 	unsigned int bytes_per_word;
+	int irq_rx;
 	u32 last_speed_hz;
 	u32 freq;
 	u16 status;
@@ -534,8 +535,8 @@ static int rzv2h_rspi_probe(struct platform_device *pdev)
 	struct rzv2h_rspi_priv *rspi;
 	struct reset_control *reset;
 	struct clk_bulk_data *clks;
-	int irq_rx, ret, i;
 	long tclk_rate;
+	int ret, i;
 
 	controller = devm_spi_alloc_host(dev, sizeof(*rspi));
 	if (!controller)
@@ -580,13 +581,13 @@ static int rzv2h_rspi_probe(struct platform_device *pdev)
 		return dev_err_probe(&pdev->dev, PTR_ERR(reset),
 				     "cannot get tresetn reset\n");
 
-	irq_rx = platform_get_irq_byname(pdev, "rx");
-	if (irq_rx < 0)
-		return dev_err_probe(dev, irq_rx, "cannot get IRQ 'rx'\n");
+	rspi->irq_rx = platform_get_irq_byname(pdev, "rx");
+	if (rspi->irq_rx < 0)
+		return dev_err_probe(dev, rspi->irq_rx, "cannot get IRQ 'rx'\n");
 
 	init_waitqueue_head(&rspi->wait);
 
-	ret = devm_request_irq(dev, irq_rx, rzv2h_rx_irq_handler, 0,
+	ret = devm_request_irq(dev, rspi->irq_rx, rzv2h_rx_irq_handler, 0,
 			       dev_name(dev), rspi);
 	if (ret) {
 		dev_err(dev, "cannot request `rx` IRQ\n");
-- 
2.52.0

[PATCH 06/13] spi: rzv2h-rspi: set MUST_RX/MUST_TX

[Cosmin Tanislav]

In preparation for implementing DMA support, set MUST_RX and MUST_TX
flags on the controller so that we always receive non-NULL buffers.

The PIO mode already handles this manually by checking if rx_buf/tx_buf
are set on the transfer, and doing a dummy read/write if not.

DMA will not be able to implement this special handling, and although
the SPI controller advertises support for transmit-only or receive-only
transfers via SPCR's register TXMD bitfield, it does not seem to work.

Remove the special handling for PIO and let the SPI controller core
handle it.

Signed-off-by: Cosmin Tanislav <cosmin-gabriel.tanislav.xa@renesas.com>
---
 drivers/spi/spi-rzv2h-rspi.c | 11 +++--------
 1 file changed, 3 insertions(+), 8 deletions(-)

diff --git a/drivers/spi/spi-rzv2h-rspi.c b/drivers/spi/spi-rzv2h-rspi.c
index 50fd7ddef58d..f0bbbd21c763 100644
--- a/drivers/spi/spi-rzv2h-rspi.c
+++ b/drivers/spi/spi-rzv2h-rspi.c
@@ -113,11 +113,7 @@ struct rzv2h_rspi_priv {
 static inline void rzv2h_rspi_tx_##type(struct rzv2h_rspi_priv *rspi,	\
 					const void *txbuf,		\
 					unsigned int index) {		\
-	type buf = 0;							\
-									\
-	if (txbuf)							\
-		buf = ((type *)txbuf)[index];				\
-									\
+	type buf = ((type *)txbuf)[index];				\
 	func(buf, rspi->base + RSPI_SPDR);				\
 }
 
@@ -126,9 +122,7 @@ static inline void rzv2h_rspi_rx_##type(struct rzv2h_rspi_priv *rspi,	\
 					void *rxbuf,			\
 					unsigned int index) {		\
 	type buf = func(rspi->base + RSPI_SPDR);			\
-									\
-	if (rxbuf)							\
-		((type *)rxbuf)[index] = buf;				\
+	((type *)rxbuf)[index] = buf;					\
 }
 
 RZV2H_RSPI_TX(writel, u32)
@@ -596,6 +590,7 @@ static int rzv2h_rspi_probe(struct platform_device *pdev)
 
 	controller->mode_bits = SPI_CPHA | SPI_CPOL | SPI_CS_HIGH |
 				SPI_LSB_FIRST | SPI_LOOP;
+	controller->flags = SPI_CONTROLLER_MUST_RX | SPI_CONTROLLER_MUST_TX;
 	controller->bits_per_word_mask = SPI_BPW_RANGE_MASK(4, 32);
 	controller->prepare_message = rzv2h_rspi_prepare_message;
 	controller->unprepare_message = rzv2h_rspi_unprepare_message;
-- 
2.52.0

[PATCH 07/13] spi: rzv2h-rspi: set TX FIFO threshold to 0

[Cosmin Tanislav]

In PIO mode we send data word-by-word, and wait for the received data
to be available after each sent word, making no use of the TX interrupt.

In DMA mode, we need to set the RX and TX FIFO thresholds to 0, as
described in the User Manual.

In preparation for implementing DMA support, set TX FIFO threshold to 0,
as RX FIFO threshold is already 0.

Signed-off-by: Cosmin Tanislav <cosmin-gabriel.tanislav.xa@renesas.com>
---
 drivers/spi/spi-rzv2h-rspi.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/spi/spi-rzv2h-rspi.c b/drivers/spi/spi-rzv2h-rspi.c
index f0bbbd21c763..83bb0b7400b2 100644
--- a/drivers/spi/spi-rzv2h-rspi.c
+++ b/drivers/spi/spi-rzv2h-rspi.c
@@ -501,7 +501,7 @@ static int rzv2h_rspi_prepare_message(struct spi_controller *ctlr,
 		writeb(0, rspi->base + RSPI_SSLP);
 
 	/* Setup FIFO thresholds */
-	conf16 = FIELD_PREP(RSPI_SPDCR2_TTRG, rspi->info->fifo_size - 1);
+	conf16 = FIELD_PREP(RSPI_SPDCR2_TTRG, 0);
 	conf16 |= FIELD_PREP(RSPI_SPDCR2_RTRG, 0);
 	writew(conf16, rspi->base + RSPI_SPDCR2);
 
-- 
2.52.0

[PATCH 08/13] spi: rzv2h-rspi: enable TX buffer empty interrupt

[Cosmin Tanislav]

In preparation for implementing DMA support, enable the transmit buffer
empty interrupt, which is necessary for DMA to write more data to the
FIFO.

This does not affect the PIO mode as we do not even request the TX
interrupt line.

Signed-off-by: Cosmin Tanislav <cosmin-gabriel.tanislav.xa@renesas.com>
---
 drivers/spi/spi-rzv2h-rspi.c | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/drivers/spi/spi-rzv2h-rspi.c b/drivers/spi/spi-rzv2h-rspi.c
index 83bb0b7400b2..b31ef2f31f1b 100644
--- a/drivers/spi/spi-rzv2h-rspi.c
+++ b/drivers/spi/spi-rzv2h-rspi.c
@@ -37,6 +37,7 @@
 /* Register SPCR */
 #define RSPI_SPCR_BPEN		BIT(31)
 #define RSPI_SPCR_MSTR		BIT(30)
+#define RSPI_SPCR_SPTIE		BIT(20)
 #define RSPI_SPCR_SPRIE		BIT(17)
 #define RSPI_SPCR_SCKASE	BIT(12)
 #define RSPI_SPCR_SPE		BIT(0)
@@ -474,6 +475,9 @@ static int rzv2h_rspi_prepare_message(struct spi_controller *ctlr,
 	/* SPI receive buffer full interrupt enable */
 	conf32 |= RSPI_SPCR_SPRIE;
 
+	/* SPI transmit buffer empty interrupt enable */
+	conf32 |= RSPI_SPCR_SPTIE;
+
 	/* Bypass synchronization circuit */
 	conf32 |= FIELD_PREP(RSPI_SPCR_BPEN, rspi->use_pclk);
 
-- 
2.52.0

[PATCH 09/13] spi: rzv2h-rspi: split out PIO transfer

[Cosmin Tanislav]

In preparation for implementing DMA support, split out the PIO transfer
code into its own function.

Signed-off-by: Cosmin Tanislav <cosmin-gabriel.tanislav.xa@renesas.com>
---
 drivers/spi/spi-rzv2h-rspi.c | 27 ++++++++++++++++++++-------
 1 file changed, 20 insertions(+), 7 deletions(-)

diff --git a/drivers/spi/spi-rzv2h-rspi.c b/drivers/spi/spi-rzv2h-rspi.c
index b31ef2f31f1b..9f5bc047b485 100644
--- a/drivers/spi/spi-rzv2h-rspi.c
+++ b/drivers/spi/spi-rzv2h-rspi.c
@@ -219,17 +219,14 @@ static int rzv2h_rspi_receive(struct rzv2h_rspi_priv *rspi, void *rxbuf,
 	return 0;
 }
 
-static int rzv2h_rspi_transfer_one(struct spi_controller *controller,
+static int rzv2h_rspi_transfer_pio(struct rzv2h_rspi_priv *rspi,
 				   struct spi_device *spi,
-				   struct spi_transfer *transfer)
+				   struct spi_transfer *transfer,
+				   unsigned int words_to_transfer)
 {
-	struct rzv2h_rspi_priv *rspi = spi_controller_get_devdata(controller);
-	unsigned int words_to_transfer, i;
+	unsigned int i;
 	int ret = 0;
 
-	transfer->effective_speed_hz = rspi->freq;
-	words_to_transfer = transfer->len / rspi->bytes_per_word;
-
 	for (i = 0; i < words_to_transfer; i++) {
 		rzv2h_rspi_clear_all_irqs(rspi);
 
@@ -240,6 +237,22 @@ static int rzv2h_rspi_transfer_one(struct spi_controller *controller,
 			break;
 	}
 
+	return ret;
+}
+
+static int rzv2h_rspi_transfer_one(struct spi_controller *controller,
+				   struct spi_device *spi,
+				   struct spi_transfer *transfer)
+{
+	struct rzv2h_rspi_priv *rspi = spi_controller_get_devdata(controller);
+	unsigned int words_to_transfer;
+	int ret;
+
+	transfer->effective_speed_hz = rspi->freq;
+	words_to_transfer = transfer->len / rspi->bytes_per_word;
+
+	ret = rzv2h_rspi_transfer_pio(rspi, spi, transfer, words_to_transfer);
+
 	rzv2h_rspi_clear_all_irqs(rspi);
 
 	return ret;
-- 
2.52.0

[PATCH 10/13] dt-bindings: spi: renesas,rzv2h-rspi: document optional support for DMA

[Cosmin Tanislav]

The IP supports using DMA for reading and writing data from the FIFO,
document it.

Signed-off-by: Cosmin Tanislav <cosmin-gabriel.tanislav.xa@renesas.com>
---
 .../devicetree/bindings/spi/renesas,rzv2h-rspi.yaml       | 8 ++++++++
 1 file changed, 8 insertions(+)

diff --git a/Documentation/devicetree/bindings/spi/renesas,rzv2h-rspi.yaml b/Documentation/devicetree/bindings/spi/renesas,rzv2h-rspi.yaml
index 069557a587b5..a588b112e11e 100644
--- a/Documentation/devicetree/bindings/spi/renesas,rzv2h-rspi.yaml
+++ b/Documentation/devicetree/bindings/spi/renesas,rzv2h-rspi.yaml
@@ -57,6 +57,14 @@ properties:
       - const: presetn
       - const: tresetn
 
+  dmas:
+    maxItems: 2
+
+  dma-names:
+    items:
+      - const: rx
+      - const: tx
+
   power-domains:
     maxItems: 1
 
-- 
2.52.0

[PATCH 11/13] spi: rzv2h-rspi: add support for DMA mode

[Cosmin Tanislav]

The DMA controller can be used to transfer data to and from the SPI
controller without involving the CPU for each word of a SPI transfer.

Add support for DMA mode.

Signed-off-by: Cosmin Tanislav <cosmin-gabriel.tanislav.xa@renesas.com>
---
 drivers/spi/spi-rzv2h-rspi.c | 169 ++++++++++++++++++++++++++++++++++-
 1 file changed, 168 insertions(+), 1 deletion(-)

diff --git a/drivers/spi/spi-rzv2h-rspi.c b/drivers/spi/spi-rzv2h-rspi.c
index 9f5bc047b485..aae916882915 100644
--- a/drivers/spi/spi-rzv2h-rspi.c
+++ b/drivers/spi/spi-rzv2h-rspi.c
@@ -9,6 +9,7 @@
 #include <linux/bitops.h>
 #include <linux/bits.h>
 #include <linux/clk.h>
+#include <linux/dmaengine.h>
 #include <linux/interrupt.h>
 #include <linux/io.h>
 #include <linux/limits.h>
@@ -21,6 +22,8 @@
 #include <linux/spi/spi.h>
 #include <linux/wait.h>
 
+#include "internals.h"
+
 /* Registers */
 #define RSPI_SPDR		0x00
 #define RSPI_SPCR		0x08
@@ -96,6 +99,7 @@ struct rzv2h_rspi_info {
 struct rzv2h_rspi_priv {
 	struct spi_controller *controller;
 	const struct rzv2h_rspi_info *info;
+	struct platform_device *pdev;
 	void __iomem *base;
 	struct clk *tclk;
 	struct clk *pclk;
@@ -108,6 +112,7 @@ struct rzv2h_rspi_priv {
 	u8 spr;
 	u8 brdv;
 	bool use_pclk;
+	bool dma_callbacked;
 };
 
 #define RZV2H_RSPI_TX(func, type)					\
@@ -219,6 +224,20 @@ static int rzv2h_rspi_receive(struct rzv2h_rspi_priv *rspi, void *rxbuf,
 	return 0;
 }
 
+static bool rzv2h_rspi_can_dma(struct spi_controller *ctlr, struct spi_device *spi,
+			       struct spi_transfer *xfer)
+{
+	struct rzv2h_rspi_priv *rspi = spi_controller_get_devdata(ctlr);
+
+	if (ctlr->fallback)
+		return false;
+
+	if (!ctlr->dma_tx || !ctlr->dma_rx)
+		return false;
+
+	return xfer->len > rspi->info->fifo_size;
+}
+
 static int rzv2h_rspi_transfer_pio(struct rzv2h_rspi_priv *rspi,
 				   struct spi_device *spi,
 				   struct spi_transfer *transfer,
@@ -240,21 +259,149 @@ static int rzv2h_rspi_transfer_pio(struct rzv2h_rspi_priv *rspi,
 	return ret;
 }
 
+static void rzv2h_rspi_dma_complete(void *arg)
+{
+	struct rzv2h_rspi_priv *rspi = arg;
+
+	rspi->dma_callbacked = 1;
+	wake_up_interruptible(&rspi->wait);
+}
+
+static struct dma_async_tx_descriptor *
+rzv2h_rspi_setup_dma_channel(struct rzv2h_rspi_priv *rspi,
+			     struct dma_chan *chan, struct sg_table *sg,
+			     enum dma_slave_buswidth width,
+			     enum dma_transfer_direction direction)
+{
+	struct dma_slave_config config = {
+		.dst_addr = rspi->pdev->resource->start + RSPI_SPDR,
+		.src_addr = rspi->pdev->resource->start + RSPI_SPDR,
+		.dst_addr_width = width,
+		.src_addr_width = width,
+		.direction = direction,
+	};
+	struct dma_async_tx_descriptor *desc;
+	int ret;
+
+	ret = dmaengine_slave_config(chan, &config);
+	if (ret)
+		return ERR_PTR(ret);
+
+	desc = dmaengine_prep_slave_sg(chan, sg->sgl, sg->nents, direction,
+				       DMA_PREP_INTERRUPT | DMA_CTRL_ACK);
+	if (!desc)
+		return ERR_PTR(-EAGAIN);
+
+	if (direction == DMA_DEV_TO_MEM) {
+		desc->callback = rzv2h_rspi_dma_complete;
+		desc->callback_param = rspi;
+	}
+
+	return desc;
+}
+
+static enum dma_slave_buswidth
+rzv2h_rspi_dma_width(struct rzv2h_rspi_priv *rspi)
+{
+	switch (rspi->bytes_per_word) {
+	case 4:
+		return DMA_SLAVE_BUSWIDTH_4_BYTES;
+	case 2:
+		return DMA_SLAVE_BUSWIDTH_2_BYTES;
+	case 1:
+		return DMA_SLAVE_BUSWIDTH_1_BYTE;
+	default:
+		return DMA_SLAVE_BUSWIDTH_UNDEFINED;
+	}
+}
+
+static int rzv2h_rspi_transfer_dma(struct rzv2h_rspi_priv *rspi,
+				   struct spi_device *spi,
+				   struct spi_transfer *transfer,
+				   unsigned int words_to_transfer)
+{
+	struct dma_async_tx_descriptor *tx_desc = NULL, *rx_desc = NULL;
+	enum dma_slave_buswidth width;
+	dma_cookie_t cookie;
+	int ret;
+
+	width = rzv2h_rspi_dma_width(rspi);
+	if (width == DMA_SLAVE_BUSWIDTH_UNDEFINED)
+		return -EINVAL;
+
+	rx_desc = rzv2h_rspi_setup_dma_channel(rspi, rspi->controller->dma_rx,
+					       &transfer->rx_sg, width,
+					       DMA_DEV_TO_MEM);
+	if (IS_ERR(rx_desc))
+		return PTR_ERR(rx_desc);
+
+	tx_desc = rzv2h_rspi_setup_dma_channel(rspi, rspi->controller->dma_tx,
+					       &transfer->tx_sg, width,
+					       DMA_MEM_TO_DEV);
+	if (IS_ERR(tx_desc))
+		return PTR_ERR(tx_desc);
+
+	cookie = dmaengine_submit(rx_desc);
+	if (dma_submit_error(cookie))
+		return cookie;
+
+	cookie = dmaengine_submit(tx_desc);
+	if (dma_submit_error(cookie)) {
+		dmaengine_terminate_sync(rspi->controller->dma_rx);
+		return cookie;
+	}
+
+	/*
+	 * DMA transfer does not need IRQs to be enabled.
+	 * For PIO, we only use RX IRQ, so disable that.
+	 */
+	disable_irq(rspi->irq_rx);
+
+	rspi->dma_callbacked = 0;
+
+	dma_async_issue_pending(rspi->controller->dma_rx);
+	dma_async_issue_pending(rspi->controller->dma_tx);
+	rzv2h_rspi_clear_all_irqs(rspi);
+
+	ret = wait_event_interruptible_timeout(rspi->wait, rspi->dma_callbacked, HZ);
+	if (ret) {
+		dmaengine_synchronize(rspi->controller->dma_tx);
+		dmaengine_synchronize(rspi->controller->dma_rx);
+		ret = 0;
+	} else {
+		dmaengine_terminate_sync(rspi->controller->dma_tx);
+		dmaengine_terminate_sync(rspi->controller->dma_rx);
+		ret = -ETIMEDOUT;
+	}
+
+	enable_irq(rspi->irq_rx);
+
+	return ret;
+}
+
 static int rzv2h_rspi_transfer_one(struct spi_controller *controller,
 				   struct spi_device *spi,
 				   struct spi_transfer *transfer)
 {
 	struct rzv2h_rspi_priv *rspi = spi_controller_get_devdata(controller);
+	bool is_dma = spi_xfer_is_dma_mapped(controller, spi, transfer);
 	unsigned int words_to_transfer;
 	int ret;
 
 	transfer->effective_speed_hz = rspi->freq;
 	words_to_transfer = transfer->len / rspi->bytes_per_word;
 
-	ret = rzv2h_rspi_transfer_pio(rspi, spi, transfer, words_to_transfer);
+	if (is_dma)
+		ret = rzv2h_rspi_transfer_dma(rspi, spi, transfer, words_to_transfer);
+	else
+		ret = rzv2h_rspi_transfer_pio(rspi, spi, transfer, words_to_transfer);
 
 	rzv2h_rspi_clear_all_irqs(rspi);
 
+	if (is_dma && ret == -EAGAIN)
+		/* Retry with PIO */
+		transfer->error = SPI_TRANS_FAIL_NO_START;
+
 	return ret;
 }
 
@@ -557,6 +704,7 @@ static int rzv2h_rspi_probe(struct platform_device *pdev)
 	platform_set_drvdata(pdev, rspi);
 
 	rspi->controller = controller;
+	rspi->pdev = pdev;
 
 	rspi->info = device_get_match_data(dev);
 
@@ -613,6 +761,7 @@ static int rzv2h_rspi_probe(struct platform_device *pdev)
 	controller->unprepare_message = rzv2h_rspi_unprepare_message;
 	controller->num_chipselect = 4;
 	controller->transfer_one = rzv2h_rspi_transfer_one;
+	controller->can_dma = rzv2h_rspi_can_dma;
 
 	tclk_rate = clk_round_rate(rspi->tclk, 0);
 	if (tclk_rate < 0)
@@ -630,6 +779,24 @@ static int rzv2h_rspi_probe(struct platform_device *pdev)
 							   RSPI_SPBR_SPR_MIN,
 							   RSPI_SPCMD_BRDV_MIN);
 
+	controller->dma_tx = devm_dma_request_chan(dev, "tx");
+	if (IS_ERR(controller->dma_tx)) {
+		ret = dev_warn_probe(dev, PTR_ERR(controller->dma_tx),
+				     "failed to request TX DMA channel\n");
+		if (ret == -EPROBE_DEFER)
+			return ret;
+		controller->dma_tx = NULL;
+	}
+
+	controller->dma_rx = devm_dma_request_chan(dev, "rx");
+	if (IS_ERR(controller->dma_rx)) {
+		ret = dev_warn_probe(dev, PTR_ERR(controller->dma_rx),
+				     "failed to request RX DMA channel\n");
+		if (ret == -EPROBE_DEFER)
+			return ret;
+		controller->dma_rx = NULL;
+	}
+
 	device_set_node(&controller->dev, dev_fwnode(dev));
 
 	ret = devm_spi_register_controller(dev, controller);
-- 
2.52.0

[PATCH 12/13] arm64: dts: renesas: r9a09g077: wire up DMA support for SPI

[Cosmin Tanislav]

RZ/T2H (R9A09G077) has three DMA controllers that can be used by
peripherals like SPI to offload data transfers from the CPU.

Wire up the DMA channels for the SPI peripherals.

Signed-off-by: Cosmin Tanislav <cosmin-gabriel.tanislav.xa@renesas.com>
---
 arch/arm64/boot/dts/renesas/r9a09g077.dtsi | 8 ++++++++
 1 file changed, 8 insertions(+)

diff --git a/arch/arm64/boot/dts/renesas/r9a09g077.dtsi b/arch/arm64/boot/dts/renesas/r9a09g077.dtsi
index ee11efb68638..34f5a4d26f29 100644
--- a/arch/arm64/boot/dts/renesas/r9a09g077.dtsi
+++ b/arch/arm64/boot/dts/renesas/r9a09g077.dtsi
@@ -200,6 +200,8 @@ rspi0: spi@80007000 {
 			clocks = <&cpg CPG_CORE R9A09G077_CLK_PCLKM>,
 				 <&cpg CPG_MOD 104>;
 			clock-names = "pclk", "pclkspi";
+			dmas = <&dmac0 0x267a>, <&dmac0 0x267b>;
+			dma-names = "rx", "tx";
 			power-domains = <&cpg>;
 			#address-cells = <1>;
 			#size-cells = <0>;
@@ -218,6 +220,8 @@ rspi1: spi@80007400 {
 			clocks = <&cpg CPG_CORE R9A09G077_CLK_PCLKM>,
 				 <&cpg CPG_MOD 105>;
 			clock-names = "pclk", "pclkspi";
+			dmas = <&dmac0 0x267f>, <&dmac0 0x2680>;
+			dma-names = "rx", "tx";
 			power-domains = <&cpg>;
 			#address-cells = <1>;
 			#size-cells = <0>;
@@ -236,6 +240,8 @@ rspi2: spi@80007800 {
 			clocks = <&cpg CPG_CORE R9A09G077_CLK_PCLKM>,
 				 <&cpg CPG_MOD 106>;
 			clock-names = "pclk", "pclkspi";
+			dmas = <&dmac0 0x2684>, <&dmac0 0x2685>;
+			dma-names = "rx", "tx";
 			power-domains = <&cpg>;
 			#address-cells = <1>;
 			#size-cells = <0>;
@@ -254,6 +260,8 @@ rspi3: spi@81007000 {
 			clocks = <&cpg CPG_CORE R9A09G077_CLK_PCLKM>,
 				 <&cpg CPG_MOD 602>;
 			clock-names = "pclk", "pclkspi";
+			dmas = <&dmac0 0x2689>, <&dmac0 0x268a>;
+			dma-names = "rx", "tx";
 			power-domains = <&cpg>;
 			#address-cells = <1>;
 			#size-cells = <0>;
-- 
2.52.0

[PATCH 13/13] arm64: dts: renesas: r9a09g087: wire up DMA support for SPI

[Cosmin Tanislav]

RZ/N2H (R9A09G087) has three DMA controllers that can be used by
peripherals like SPI to offload data transfers from the CPU.

Wire up the DMA channels for the SPI peripherals.

Signed-off-by: Cosmin Tanislav <cosmin-gabriel.tanislav.xa@renesas.com>
---
 arch/arm64/boot/dts/renesas/r9a09g087.dtsi | 8 ++++++++
 1 file changed, 8 insertions(+)

diff --git a/arch/arm64/boot/dts/renesas/r9a09g087.dtsi b/arch/arm64/boot/dts/renesas/r9a09g087.dtsi
index 7b1f2c1c9e85..61c1b3713b43 100644
--- a/arch/arm64/boot/dts/renesas/r9a09g087.dtsi
+++ b/arch/arm64/boot/dts/renesas/r9a09g087.dtsi
@@ -200,6 +200,8 @@ rspi0: spi@80007000 {
 			clocks = <&cpg CPG_CORE R9A09G087_CLK_PCLKM>,
 				 <&cpg CPG_MOD 104>;
 			clock-names = "pclk", "pclkspi";
+			dmas = <&dmac0 0x267a>, <&dmac0 0x267b>;
+			dma-names = "rx", "tx";
 			power-domains = <&cpg>;
 			#address-cells = <1>;
 			#size-cells = <0>;
@@ -218,6 +220,8 @@ rspi1: spi@80007400 {
 			clocks = <&cpg CPG_CORE R9A09G087_CLK_PCLKM>,
 				 <&cpg CPG_MOD 105>;
 			clock-names = "pclk", "pclkspi";
+			dmas = <&dmac0 0x267f>, <&dmac0 0x2680>;
+			dma-names = "rx", "tx";
 			power-domains = <&cpg>;
 			#address-cells = <1>;
 			#size-cells = <0>;
@@ -236,6 +240,8 @@ rspi2: spi@80007800 {
 			clocks = <&cpg CPG_CORE R9A09G087_CLK_PCLKM>,
 				 <&cpg CPG_MOD 106>;
 			clock-names = "pclk", "pclkspi";
+			dmas = <&dmac0 0x2684>, <&dmac0 0x2685>;
+			dma-names = "rx", "tx";
 			power-domains = <&cpg>;
 			#address-cells = <1>;
 			#size-cells = <0>;
@@ -254,6 +260,8 @@ rspi3: spi@81007000 {
 			clocks = <&cpg CPG_CORE R9A09G087_CLK_PCLKM>,
 				 <&cpg CPG_MOD 602>;
 			clock-names = "pclk", "pclkspi";
+			dmas = <&dmac0 0x2689>, <&dmac0 0x268a>;
+			dma-names = "rx", "tx";
 			power-domains = <&cpg>;
 			#address-cells = <1>;
 			#size-cells = <0>;
-- 
2.52.0

Re: [PATCH 10/13] dt-bindings: spi: renesas,rzv2h-rspi: document optional support for DMA

[Rob Herring (Arm)]

On Mon, 01 Dec 2025 15:42:26 +0200, Cosmin Tanislav wrote:
> The IP supports using DMA for reading and writing data from the FIFO,
> document it.
> 
> Signed-off-by: Cosmin Tanislav <cosmin-gabriel.tanislav.xa@renesas.com>
> ---
>  .../devicetree/bindings/spi/renesas,rzv2h-rspi.yaml       | 8 ++++++++
>  1 file changed, 8 insertions(+)
> 

Acked-by: Rob Herring (Arm) <robh@kernel.org>

Re: (subset) [PATCH 00/13] Add DMA support for RZ/T2H RSPI

[Mark Brown]

On Mon, 01 Dec 2025 15:42:16 +0200, Cosmin Tanislav wrote:
> The DMA controller can be used to transfer data to and from the SPI
> controller without involving the CPU for each word of a SPI transfer.
> 
> Add support for DMA mode, and do some other cleanups while touching the
> same code.
> 
> The dts changes in this series depend on the DMA series [1].
> 
> [...]

Applied to

   https://git.kernel.org/pub/scm/linux/kernel/git/broonie/spi.git for-next

Thanks!

[01/13] spi: rzv2h-rspi: fix rzv2h_rspi_transfer_one() indentation
        commit: fb0140774aff45b8dc326987cc1da89484ecb081
[02/13] spi: rzv2h-rspi: remove call to spi_finalize_current_transfer()
        commit: 9e4830b35dc0d522f45e1ec3ee5b1ff1648afe1b
[03/13] spi: rzv2h-rspi: do not set SPI_TRANS_FAIL_IO
        commit: 218917659df165cff72439480929e68a6e127b55
[04/13] spi: rzv2h-rspi: use device-managed APIs
        commit: b73ac782828f27c2217a17bd26aa8710769f032d
[05/13] spi: rzv2h-rspi: store RX interrupt in state
        commit: 28b590bd4c6a051ec61cf286a46a8b14846e6fcf
[06/13] spi: rzv2h-rspi: set MUST_RX/MUST_TX
        commit: 6f9026b5a18acdf190d1622831b100aacfca0eb3
[07/13] spi: rzv2h-rspi: set TX FIFO threshold to 0
        commit: a886baaaa6e12a9b7d5a9687d11d3b895f1b87c9
[08/13] spi: rzv2h-rspi: enable TX buffer empty interrupt
        commit: d49eea07de5851e1b8941ad6b6179be7ec36a986
[09/13] spi: rzv2h-rspi: split out PIO transfer
        commit: 1e5e10df8b9be71ca64435cbe7c96b189e5ee293
[10/13] dt-bindings: spi: renesas,rzv2h-rspi: document optional support for DMA
        commit: 163345e356722e98ba57cd120787d6e991da7b1d
[11/13] spi: rzv2h-rspi: add support for DMA mode
        commit: fa08b566860bca8ebf9300090b85174c34de7ca5

All being well this means that it will be integrated into the linux-next
tree (usually sometime in the next 24 hours) and sent to Linus during
the next merge window (or sooner if it is a bug fix), however if
problems are discovered then the patch may be dropped or reverted.

You may get further e-mails resulting from automated or manual testing
and review of the tree, please engage with people reporting problems and
send followup patches addressing any issues that are reported if needed.

If any updates are required or you are submitting further changes they
should be sent as incremental updates against current git, existing
patches will not be replaced.

Please add any relevant lists and maintainers to the CCs when replying
to this mail.

Thanks,
Mark

Re: [PATCH 12/13] arm64: dts: renesas: r9a09g077: wire up DMA support for SPI

[Geert Uytterhoeven]

Hi Cosmin,

On Mon, 1 Dec 2025 at 14:44, Cosmin Tanislav
<cosmin-gabriel.tanislav.xa@renesas.com> wrote:
> RZ/T2H (R9A09G077) has three DMA controllers that can be used by
> peripherals like SPI to offload data transfers from the CPU.
>
> Wire up the DMA channels for the SPI peripherals.
>
> Signed-off-by: Cosmin Tanislav <cosmin-gabriel.tanislav.xa@renesas.com>

Thanks for your patch!

> --- a/arch/arm64/boot/dts/renesas/r9a09g077.dtsi
> +++ b/arch/arm64/boot/dts/renesas/r9a09g077.dtsi
> @@ -200,6 +200,8 @@ rspi0: spi@80007000 {
>                         clocks = <&cpg CPG_CORE R9A09G077_CLK_PCLKM>,
>                                  <&cpg CPG_MOD 104>;
>                         clock-names = "pclk", "pclkspi";
> +                       dmas = <&dmac0 0x267a>, <&dmac0 0x267b>;
> +                       dma-names = "rx", "tx";

RZ/T2H does not seem to have restrictions about which DMA controllers
can be used by which SPI instance.  Hence shouldn't these point to
all three DMA controllers?

    dmas = <&dmac0 0x267a>, <&dmac0 0x267b>,
           <&dmac1 0x267a>, <&dmac1 0x267b>,
           <&dmac2 0x267a>, <&dmac2 0x267b>;
    dma-names = "rx", "tx", "rx", "tx", "rx", "tx";

Note that this requires updating the DT bindings, as they currently
restrict dma to two entries.

The rest LGTM, so
Reviewed-by: Geert Uytterhoeven <geert+renesas@glider.be>

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 13/13] arm64: dts: renesas: r9a09g087: wire up DMA support for SPI

[Geert Uytterhoeven]

Hi Cosmin,

On Mon, 1 Dec 2025 at 14:44, Cosmin Tanislav
<cosmin-gabriel.tanislav.xa@renesas.com> wrote:
> RZ/N2H (R9A09G087) has three DMA controllers that can be used by
> peripherals like SPI to offload data transfers from the CPU.
>
> Wire up the DMA channels for the SPI peripherals.
>
> Signed-off-by: Cosmin Tanislav <cosmin-gabriel.tanislav.xa@renesas.com>

Thanks for your patch!

> --- a/arch/arm64/boot/dts/renesas/r9a09g087.dtsi
> +++ b/arch/arm64/boot/dts/renesas/r9a09g087.dtsi
> @@ -200,6 +200,8 @@ rspi0: spi@80007000 {
>                         clocks = <&cpg CPG_CORE R9A09G087_CLK_PCLKM>,
>                                  <&cpg CPG_MOD 104>;
>                         clock-names = "pclk", "pclkspi";
> +                       dmas = <&dmac0 0x267a>, <&dmac0 0x267b>;
> +                       dma-names = "rx", "tx";

Same comment as for RZ/T2H:
RZ/N2H does not seem to have restrictions about which DMA controllers
can be used by which SPI instance.  Hence shouldn't these point to
all three DMA controllers?

    dmas = <&dmac0 0x267a>, <&dmac0 0x267b>,
           <&dmac1 0x267a>, <&dmac1 0x267b>,
           <&dmac2 0x267a>, <&dmac2 0x267b>;
    dma-names = "rx", "tx", "rx", "tx", "rx", "tx";

Note that this requires updating the DT bindings, as they currently
restrict dma to two entries.

The rest LGTM, so
Reviewed-by: Geert Uytterhoeven <geert+renesas@glider.be>

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 12/13] arm64: dts: renesas: r9a09g077: wire up DMA support for SPI

[Cosmin-Gabriel Tanislav]

> From: Geert Uytterhoeven <geert@linux-m68k.org>
> Sent: Tuesday, December 23, 2025 4:25 PM
> 
> Hi Cosmin,
> 
> On Mon, 1 Dec 2025 at 14:44, Cosmin Tanislav
> <cosmin-gabriel.tanislav.xa@renesas.com> wrote:
> > RZ/T2H (R9A09G077) has three DMA controllers that can be used by
> > peripherals like SPI to offload data transfers from the CPU.
> >
> > Wire up the DMA channels for the SPI peripherals.
> >
> > Signed-off-by: Cosmin Tanislav <cosmin-gabriel.tanislav.xa@renesas.com>
> 
> Thanks for your patch!
> 
> > --- a/arch/arm64/boot/dts/renesas/r9a09g077.dtsi
> > +++ b/arch/arm64/boot/dts/renesas/r9a09g077.dtsi
> > @@ -200,6 +200,8 @@ rspi0: spi@80007000 {
> >                         clocks = <&cpg CPG_CORE R9A09G077_CLK_PCLKM>,
> >                                  <&cpg CPG_MOD 104>;
> >                         clock-names = "pclk", "pclkspi";
> > +                       dmas = <&dmac0 0x267a>, <&dmac0 0x267b>;
> > +                       dma-names = "rx", "tx";
> 
> RZ/T2H does not seem to have restrictions about which DMA controllers
> can be used by which SPI instance.  Hence shouldn't these point to
> all three DMA controllers?

It does seem like there's no restriction about which DMA controller to
use.

> 
>     dmas = <&dmac0 0x267a>, <&dmac0 0x267b>,
>            <&dmac1 0x267a>, <&dmac1 0x267b>,
>            <&dmac2 0x267a>, <&dmac2 0x267b>;
>     dma-names = "rx", "tx", "rx", "tx", "rx", "tx";
> 

I was not aware that the DMA core supports this. I will add the other DMA
controllers to the list.

> Note that this requires updating the DT bindings, as they currently
> restrict dma to two entries.
> 

dma-names:
  items:
    enum:
      - rx
      - tx

This should work fine, right?

> The rest LGTM, so
> Reviewed-by: Geert Uytterhoeven <geert+renesas@glider.be>
> 
> 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 12/13] arm64: dts: renesas: r9a09g077: wire up DMA support for SPI

[Geert Uytterhoeven]

Hi Cosmin,

On Tue, 23 Dec 2025 at 15:42, Cosmin-Gabriel Tanislav
<cosmin-gabriel.tanislav.xa@renesas.com> wrote:
> > From: Geert Uytterhoeven <geert@linux-m68k.org>
> > On Mon, 1 Dec 2025 at 14:44, Cosmin Tanislav
> > <cosmin-gabriel.tanislav.xa@renesas.com> wrote:
> > > RZ/T2H (R9A09G077) has three DMA controllers that can be used by
> > > peripherals like SPI to offload data transfers from the CPU.
> > >
> > > Wire up the DMA channels for the SPI peripherals.
> > >
> > > Signed-off-by: Cosmin Tanislav <cosmin-gabriel.tanislav.xa@renesas.com>
> >
> > Thanks for your patch!
> >
> > > --- a/arch/arm64/boot/dts/renesas/r9a09g077.dtsi
> > > +++ b/arch/arm64/boot/dts/renesas/r9a09g077.dtsi
> > > @@ -200,6 +200,8 @@ rspi0: spi@80007000 {
> > >                         clocks = <&cpg CPG_CORE R9A09G077_CLK_PCLKM>,
> > >                                  <&cpg CPG_MOD 104>;
> > >                         clock-names = "pclk", "pclkspi";
> > > +                       dmas = <&dmac0 0x267a>, <&dmac0 0x267b>;
> > > +                       dma-names = "rx", "tx";
> >
> > RZ/T2H does not seem to have restrictions about which DMA controllers
> > can be used by which SPI instance.  Hence shouldn't these point to
> > all three DMA controllers?
>
> It does seem like there's no restriction about which DMA controller to
> use.
>
> >     dmas = <&dmac0 0x267a>, <&dmac0 0x267b>,
> >            <&dmac1 0x267a>, <&dmac1 0x267b>,
> >            <&dmac2 0x267a>, <&dmac2 0x267b>;
> >     dma-names = "rx", "tx", "rx", "tx", "rx", "tx";
> >
>
> I was not aware that the DMA core supports this. I will add the other DMA
> controllers to the list.
>
> > Note that this requires updating the DT bindings, as they currently
> > restrict dma to two entries.
>
> dma-names:
>   items:
>     enum:
>       - rx
>       - tx
>
> This should work fine, right?

Yes, dma-names is OK.  But currently dmas has "maxItems: 2".

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