[PATCH] spi: stm32: enable controller before asserting CS

* [PATCH] spi: stm32: enable controller before asserting CS
@ 2023-12-01 21:40 Ben Wolsieffer
  2023-12-01 21:50 ` Mark Brown
  0 siblings, 1 reply; 5+ messages in thread
From: Ben Wolsieffer @ 2023-12-01 21:40 UTC (permalink / raw)
  To: linux-spi, linux-stm32, linux-arm-kernel, linux-kernel
  Cc: Mark Brown, Maxime Coquelin, Alexandre Torgue, Alain Volmat,
	Ben Wolsieffer

On the STM32F4/7, the SPI pins float while the controller is disabled.
Currently, the controller is enabled in the transfer_one() callback,
which runs after CS is asserted. Therefore, there is a period where the
SPI pins are floating while CS is asserted, making it possible for stray
signals to disrupt communications. An analogous problem occurs at the
end of the transfer when the controller is disabled before CS is
released.

This problem can be reliably observed by enabling the pull-up (if
CPOL=0) or pull-down (if CPOL=1) on the clock pin. This will cause two
extra unintended clock edges per transfer, when the controller is
enabled and disabled.

This patch fixes the bug by enabling the controller in prepare_message()
and disabling it in unprepare_message(), which are called while CS is
not asserted.

Note that bug is likely not present on the STM32H7, because it supports
the AFCNTR bit (and this driver sets it), which keeps the SPI pins
driven even while the controller is disabled.

This patch has been tested on an STM32F746 with a MAX14830 UART
expander.

Signed-off-by: Ben Wolsieffer <ben.wolsieffer@hefring.com>
---
 drivers/spi/spi-stm32.c | 14 ++------------
 1 file changed, 2 insertions(+), 12 deletions(-)

diff --git a/drivers/spi/spi-stm32.c b/drivers/spi/spi-stm32.c
index 94df3836834c..885f53a51441 100644
--- a/drivers/spi/spi-stm32.c
+++ b/drivers/spi/spi-stm32.c
@@ -948,10 +948,8 @@ static irqreturn_t stm32fx_spi_irq_event(int irq, void *dev_id)
 static irqreturn_t stm32fx_spi_irq_thread(int irq, void *dev_id)
 {
 	struct spi_controller *ctrl = dev_id;
-	struct stm32_spi *spi = spi_controller_get_devdata(ctrl);
 
 	spi_finalize_current_transfer(ctrl);
-	stm32fx_spi_disable(spi);
 
 	return IRQ_HANDLED;
 }
@@ -1118,6 +1116,8 @@ static int stm32_spi_prepare_msg(struct spi_controller *ctrl,
 			 ~clrb) | setb,
 			spi->base + spi->cfg->regs->cpol.reg);
 
+	stm32_spi_enable(spi);
+
 	spin_unlock_irqrestore(&spi->lock, flags);
 
 	return 0;
@@ -1135,7 +1135,6 @@ static void stm32fx_spi_dma_tx_cb(void *data)
 
 	if (spi->cur_comm == SPI_SIMPLEX_TX || spi->cur_comm == SPI_3WIRE_TX) {
 		spi_finalize_current_transfer(spi->ctrl);
-		stm32fx_spi_disable(spi);
 	}
 }
 
@@ -1150,7 +1149,6 @@ static void stm32_spi_dma_rx_cb(void *data)
 	struct stm32_spi *spi = data;
 
 	spi_finalize_current_transfer(spi->ctrl);
-	spi->cfg->disable(spi);
 }
 
 /**
@@ -1235,8 +1233,6 @@ static int stm32fx_spi_transfer_one_irq(struct stm32_spi *spi)
 
 	stm32_spi_set_bits(spi, STM32FX_SPI_CR2, cr2);
 
-	stm32_spi_enable(spi);
-
 	/* starting data transfer when buffer is loaded */
 	if (spi->tx_buf)
 		spi->cfg->write_tx(spi);
@@ -1273,8 +1269,6 @@ static int stm32h7_spi_transfer_one_irq(struct stm32_spi *spi)
 
 	spin_lock_irqsave(&spi->lock, flags);
 
-	stm32_spi_enable(spi);
-
 	/* Be sure to have data in fifo before starting data transfer */
 	if (spi->tx_buf)
 		stm32h7_spi_write_txfifo(spi);
@@ -1306,8 +1300,6 @@ static void stm32fx_spi_transfer_one_dma_start(struct stm32_spi *spi)
 		 */
 		stm32_spi_set_bits(spi, STM32FX_SPI_CR2, STM32FX_SPI_CR2_ERRIE);
 	}
-
-	stm32_spi_enable(spi);
 }
 
 /**
@@ -1341,8 +1333,6 @@ static void stm32h7_spi_transfer_one_dma_start(struct stm32_spi *spi)
 
 	stm32_spi_set_bits(spi, STM32H7_SPI_IER, ier);
 
-	stm32_spi_enable(spi);
-
 	if (STM32_SPI_MASTER_MODE(spi))
 		stm32_spi_set_bits(spi, STM32H7_SPI_CR1, STM32H7_SPI_CR1_CSTART);
 }
-- 
2.42.1


_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

^ permalink raw reply related	[flat|nested] 5+ messages in thread