linux-spi.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
* [PATCH AUTOSEL 5.15 2/3] spi: sun6i: reduce DMA RX transfer width to single byte
       [not found] <20230914015536.51984-1-sashal@kernel.org>
@ 2023-09-14  1:55 ` Sasha Levin
  2023-09-14  1:55 ` [PATCH AUTOSEL 5.15 3/3] spi: sun6i: fix race between DMA RX transfer completion and RX FIFO drain Sasha Levin
  1 sibling, 0 replies; 2+ messages in thread
From: Sasha Levin @ 2023-09-14  1:55 UTC (permalink / raw)
  To: linux-kernel, stable
  Cc: Tobias Schramm, Mark Brown, Sasha Levin, wens, jernej.skrabec,
	samuel, linux-spi, linux-arm-kernel, linux-sunxi

From: Tobias Schramm <t.schramm@manjaro.org>

[ Upstream commit 171f8a49f212e87a8b04087568e1b3d132e36a18 ]

Through empirical testing it has been determined that sometimes RX SPI
transfers with DMA enabled return corrupted data. This is down to single
or even multiple bytes lost during DMA transfer from SPI peripheral to
memory. It seems the RX FIFO within the SPI peripheral can become
confused when performing bus read accesses wider than a single byte to it
during an active SPI transfer.

This patch reduces the width of individual DMA read accesses to the
RX FIFO to a single byte to mitigate that issue.

Signed-off-by: Tobias Schramm <t.schramm@manjaro.org>
Link: https://lore.kernel.org/r/20230827152558.5368-2-t.schramm@manjaro.org
Signed-off-by: Mark Brown <broonie@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 drivers/spi/spi-sun6i.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/spi/spi-sun6i.c b/drivers/spi/spi-sun6i.c
index 23ad052528dbe..2bfe87873edb3 100644
--- a/drivers/spi/spi-sun6i.c
+++ b/drivers/spi/spi-sun6i.c
@@ -200,7 +200,7 @@ static int sun6i_spi_prepare_dma(struct sun6i_spi *sspi,
 		struct dma_slave_config rxconf = {
 			.direction = DMA_DEV_TO_MEM,
 			.src_addr = sspi->dma_addr_rx,
-			.src_addr_width = DMA_SLAVE_BUSWIDTH_4_BYTES,
+			.src_addr_width = DMA_SLAVE_BUSWIDTH_1_BYTE,
 			.src_maxburst = 8,
 		};
 
-- 
2.40.1


^ permalink raw reply related	[flat|nested] 2+ messages in thread

* [PATCH AUTOSEL 5.15 3/3] spi: sun6i: fix race between DMA RX transfer completion and RX FIFO drain
       [not found] <20230914015536.51984-1-sashal@kernel.org>
  2023-09-14  1:55 ` [PATCH AUTOSEL 5.15 2/3] spi: sun6i: reduce DMA RX transfer width to single byte Sasha Levin
@ 2023-09-14  1:55 ` Sasha Levin
  1 sibling, 0 replies; 2+ messages in thread
From: Sasha Levin @ 2023-09-14  1:55 UTC (permalink / raw)
  To: linux-kernel, stable
  Cc: Tobias Schramm, Mark Brown, Sasha Levin, wens, jernej.skrabec,
	samuel, linux-spi, linux-arm-kernel, linux-sunxi

From: Tobias Schramm <t.schramm@manjaro.org>

[ Upstream commit 1f11f4202caf5710204d334fe63392052783876d ]

Previously the transfer complete IRQ immediately drained to RX FIFO to
read any data remaining in FIFO to the RX buffer. This behaviour is
correct when dealing with SPI in interrupt mode. However in DMA mode the
transfer complete interrupt still fires as soon as all bytes to be
transferred have been stored in the FIFO. At that point data in the FIFO
still needs to be picked up by the DMA engine. Thus the drain procedure
and DMA engine end up racing to read from RX FIFO, corrupting any data
read. Additionally the RX buffer pointer is never adjusted according to
DMA progress in DMA mode, thus calling the RX FIFO drain procedure in DMA
mode is a bug.
Fix corruptions in DMA RX mode by draining RX FIFO only in interrupt mode.
Also wait for completion of RX DMA when in DMA mode before returning to
ensure all data has been copied to the supplied memory buffer.

Signed-off-by: Tobias Schramm <t.schramm@manjaro.org>
Link: https://lore.kernel.org/r/20230827152558.5368-3-t.schramm@manjaro.org
Signed-off-by: Mark Brown <broonie@kernel.org>
Signed-off-by: Sasha Levin <sashal@kernel.org>
---
 drivers/spi/spi-sun6i.c | 29 ++++++++++++++++++++++++++++-
 1 file changed, 28 insertions(+), 1 deletion(-)

diff --git a/drivers/spi/spi-sun6i.c b/drivers/spi/spi-sun6i.c
index 2bfe87873edb3..d79853ba7792a 100644
--- a/drivers/spi/spi-sun6i.c
+++ b/drivers/spi/spi-sun6i.c
@@ -95,6 +95,7 @@ struct sun6i_spi {
 	struct reset_control	*rstc;
 
 	struct completion	done;
+	struct completion	dma_rx_done;
 
 	const u8		*tx_buf;
 	u8			*rx_buf;
@@ -189,6 +190,13 @@ static size_t sun6i_spi_max_transfer_size(struct spi_device *spi)
 	return SUN6I_MAX_XFER_SIZE - 1;
 }
 
+static void sun6i_spi_dma_rx_cb(void *param)
+{
+	struct sun6i_spi *sspi = param;
+
+	complete(&sspi->dma_rx_done);
+}
+
 static int sun6i_spi_prepare_dma(struct sun6i_spi *sspi,
 				 struct spi_transfer *tfr)
 {
@@ -213,6 +221,8 @@ static int sun6i_spi_prepare_dma(struct sun6i_spi *sspi,
 						 DMA_PREP_INTERRUPT);
 		if (!rxdesc)
 			return -EINVAL;
+		rxdesc->callback_param = sspi;
+		rxdesc->callback = sun6i_spi_dma_rx_cb;
 	}
 
 	txdesc = NULL;
@@ -268,6 +278,7 @@ static int sun6i_spi_transfer_one(struct spi_master *master,
 		return -EINVAL;
 
 	reinit_completion(&sspi->done);
+	reinit_completion(&sspi->dma_rx_done);
 	sspi->tx_buf = tfr->tx_buf;
 	sspi->rx_buf = tfr->rx_buf;
 	sspi->len = tfr->len;
@@ -426,6 +437,22 @@ static int sun6i_spi_transfer_one(struct spi_master *master,
 	start = jiffies;
 	timeout = wait_for_completion_timeout(&sspi->done,
 					      msecs_to_jiffies(tx_time));
+
+	if (!use_dma) {
+		sun6i_spi_drain_fifo(sspi);
+	} else {
+		if (timeout && rx_len) {
+			/*
+			 * Even though RX on the peripheral side has finished
+			 * RX DMA might still be in flight
+			 */
+			timeout = wait_for_completion_timeout(&sspi->dma_rx_done,
+							      timeout);
+			if (!timeout)
+				dev_warn(&master->dev, "RX DMA timeout\n");
+		}
+	}
+
 	end = jiffies;
 	if (!timeout) {
 		dev_warn(&master->dev,
@@ -453,7 +480,6 @@ static irqreturn_t sun6i_spi_handler(int irq, void *dev_id)
 	/* Transfer complete */
 	if (status & SUN6I_INT_CTL_TC) {
 		sun6i_spi_write(sspi, SUN6I_INT_STA_REG, SUN6I_INT_CTL_TC);
-		sun6i_spi_drain_fifo(sspi);
 		complete(&sspi->done);
 		return IRQ_HANDLED;
 	}
@@ -611,6 +637,7 @@ static int sun6i_spi_probe(struct platform_device *pdev)
 	}
 
 	init_completion(&sspi->done);
+	init_completion(&sspi->dma_rx_done);
 
 	sspi->rstc = devm_reset_control_get_exclusive(&pdev->dev, NULL);
 	if (IS_ERR(sspi->rstc)) {
-- 
2.40.1


^ permalink raw reply related	[flat|nested] 2+ messages in thread

end of thread, other threads:[~2023-09-14  2:11 UTC | newest]

Thread overview: 2+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
     [not found] <20230914015536.51984-1-sashal@kernel.org>
2023-09-14  1:55 ` [PATCH AUTOSEL 5.15 2/3] spi: sun6i: reduce DMA RX transfer width to single byte Sasha Levin
2023-09-14  1:55 ` [PATCH AUTOSEL 5.15 3/3] spi: sun6i: fix race between DMA RX transfer completion and RX FIFO drain Sasha Levin

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).