[PATCH v2 0/2] Chunk splitting of spi transfers

[PATCH v2 0/2] Chunk splitting of spi transfers

[Meghana Madhyastha]

I've added bcm2835_spi_transfer_one_message in spi-bcm2835. This calls
spi_split_transfers_maxsize to split large chunks for spi dma transfers. 
I then removed chunk splitting in the tinydrm spi helper (as now the core is
handling the chunk splitting). However, although the SPI HW should be able to
accomodate upto 65535 bytes for dma transfers, the splitting of chunks to 65535
bytes results in a dma transfer time out error. However, when the chunks are
split to <64  bytes it seems to work fine.

Changes in v2:
-Patch 2 did not exist in v1.

Meghana Madhyastha (2):
  drm/tinydrm: Remove chunk splitting in tinydrm_spi_transfer
  spi/spi-bcm2835: Add bcm2835_spi_transfer_one_message in spi-bcm2835.c

 drivers/gpu/drm/tinydrm/core/tinydrm-helpers.c | 48 ++++----------------------
 drivers/gpu/drm/tinydrm/mipi-dbi.c             | 10 ++----
 drivers/spi/spi-bcm2835.c                      | 29 ++++++++--------
 drivers/spi/spi.c                              |  5 ++-
 include/linux/spi/spi.h                        |  2 ++
 5 files changed, 27 insertions(+), 67 deletions(-)

-- 
2.11.0

_______________________________________________
dri-devel mailing list
dri-devel@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/dri-devel

[PATCH v2 1/2] drm/tinydrm: Remove chunk splitting in tinydrm_spi_transfer

[Meghana Madhyastha]

-Remove chunk splitting in tinydrm_spi_transfer in tinydrm-helpers as The spi core will split a buffer into max_dma_len chunks for the
 spi controller driver to handle.
-Remove automatic byte swapping in tinydrm_spi_transfer as it doesn't have users.
-Remove the upper bound check on dma transfer length in bcm2835_spi_can_dma().

Signed-off-by: Meghana Madhyastha <meghana.madhyastha@gmail.com>
---
 drivers/gpu/drm/tinydrm/core/tinydrm-helpers.c | 48 ++++----------------------
 drivers/gpu/drm/tinydrm/mipi-dbi.c             | 10 ++----
 drivers/spi/spi-bcm2835.c                      | 15 +-------
 3 files changed, 9 insertions(+), 64 deletions(-)

diff --git a/drivers/gpu/drm/tinydrm/core/tinydrm-helpers.c b/drivers/gpu/drm/tinydrm/core/tinydrm-helpers.c
index bf96072d1b97..6064099e8e63 100644
--- a/drivers/gpu/drm/tinydrm/core/tinydrm-helpers.c
+++ b/drivers/gpu/drm/tinydrm/core/tinydrm-helpers.c
@@ -452,62 +452,26 @@ int tinydrm_spi_transfer(struct spi_device *spi, u32 speed_hz,
 	struct spi_transfer tr = {
 		.bits_per_word = bpw,
 		.speed_hz = speed_hz,
+		.tx_buf = buf,
+		.len = len
 	};
 	struct spi_message m;
-	u16 *swap_buf = NULL;
 	size_t max_chunk;
-	size_t chunk;
-	int ret = 0;
 
-	if (WARN_ON_ONCE(bpw != 8 && bpw != 16))
-		return -EINVAL;
-
-	max_chunk = tinydrm_spi_max_transfer_size(spi, 0);
+	max_chunk = SIZE_MAX;
 
 	if (drm_debug & DRM_UT_DRIVER)
 		pr_debug("[drm:%s] bpw=%u, max_chunk=%zu, transfers:\n",
-			 __func__, bpw, max_chunk);
-
-	if (bpw == 16 && !tinydrm_spi_bpw_supported(spi, 16)) {
-		tr.bits_per_word = 8;
-		if (tinydrm_machine_little_endian()) {
-			swap_buf = kmalloc(min(len, max_chunk), GFP_KERNEL);
-			if (!swap_buf)
-				return -ENOMEM;
-		}
-	}
+			__func__, bpw, max_chunk);
 
 	spi_message_init(&m);
 	if (header)
 		spi_message_add_tail(header, &m);
 	spi_message_add_tail(&tr, &m);
 
-	while (len) {
-		chunk = min(len, max_chunk);
-
-		tr.tx_buf = buf;
-		tr.len = chunk;
-
-		if (swap_buf) {
-			const u16 *buf16 = buf;
-			unsigned int i;
-
-			for (i = 0; i < chunk / 2; i++)
-				swap_buf[i] = swab16(buf16[i]);
-
-			tr.tx_buf = swap_buf;
-		}
-
-		buf += chunk;
-		len -= chunk;
-
-		tinydrm_dbg_spi_message(spi, &m);
-		ret = spi_sync(spi, &m);
-		if (ret)
-			return ret;
-	}
+	tinydrm_dbg_spi_message(spi, &m);
 
-	return 0;
+	return spi_sync(spi, &m);
 }
 EXPORT_SYMBOL(tinydrm_spi_transfer);
 
diff --git a/drivers/gpu/drm/tinydrm/mipi-dbi.c b/drivers/gpu/drm/tinydrm/mipi-dbi.c
index 75dd65c57e74..c8af2d65c2ad 100644
--- a/drivers/gpu/drm/tinydrm/mipi-dbi.c
+++ b/drivers/gpu/drm/tinydrm/mipi-dbi.c
@@ -886,15 +886,9 @@ static int mipi_dbi_typec3_command(struct mipi_dbi *mipi, u8 cmd,
 int mipi_dbi_spi_init(struct spi_device *spi, struct mipi_dbi *mipi,
 		      struct gpio_desc *dc)
 {
-	size_t tx_size = tinydrm_spi_max_transfer_size(spi, 0);
 	struct device *dev = &spi->dev;
 	int ret;
 
-	if (tx_size < 16) {
-		DRM_ERROR("SPI transmit buffer too small: %zu\n", tx_size);
-		return -EINVAL;
-	}
-
 	/*
 	 * Even though it's not the SPI device that does DMA (the master does),
 	 * the dma mask is necessary for the dma_alloc_wc() in
@@ -924,8 +918,8 @@ int mipi_dbi_spi_init(struct spi_device *spi, struct mipi_dbi *mipi,
 			mipi->swap_bytes = true;
 	} else {
 		mipi->command = mipi_dbi_typec1_command;
-		mipi->tx_buf9_len = tx_size;
-		mipi->tx_buf9 = devm_kmalloc(dev, tx_size, GFP_KERNEL);
+		mipi->tx_buf9_len = SZ_16K;
+		mipi->tx_buf9 = devm_kmalloc(dev, mipi->tx_buf9_len, GFP_KERNEL);
 		if (!mipi->tx_buf9)
 			return -ENOMEM;
 	}
diff --git a/drivers/spi/spi-bcm2835.c b/drivers/spi/spi-bcm2835.c
index f35cc10772f6..2fd650891c07 100644
--- a/drivers/spi/spi-bcm2835.c
+++ b/drivers/spi/spi-bcm2835.c
@@ -365,19 +365,6 @@ static bool bcm2835_spi_can_dma(struct spi_master *master,
 	if (tfr->len < BCM2835_SPI_DMA_MIN_LENGTH)
 		return false;
 
-	/* BCM2835_SPI_DLEN has defined a max transfer size as
-	 * 16 bit, so max is 65535
-	 * we can revisit this by using an alternative transfer
-	 * method - ideally this would get done without any more
-	 * interaction...
-	 */
-	if (tfr->len > 65535) {
-		dev_warn_once(&spi->dev,
-			      "transfer size of %d too big for dma-transfer\n",
-			      tfr->len);
-		return false;
-	}
-
 	/* if we run rx/tx_buf with word aligned addresses then we are OK */
 	if ((((size_t)tfr->rx_buf & 3) == 0) &&
 	    (((size_t)tfr->tx_buf & 3) == 0))
@@ -461,7 +448,7 @@ static void bcm2835_dma_init(struct spi_master *master, struct device *dev)
 
 	/* all went well, so set can_dma */
 	master->can_dma = bcm2835_spi_can_dma;
-	master->max_dma_len = 65535; /* limitation by BCM2835_SPI_DLEN */
+	master->max_dma_len = 65528; /* limitation by BCM2835_SPI_DLEN */
 	/* need to do TX AND RX DMA, so we need dummy buffers */
 	master->flags = SPI_MASTER_MUST_RX | SPI_MASTER_MUST_TX;
 
-- 
2.11.0

_______________________________________________
dri-devel mailing list
dri-devel@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/dri-devel

Re: [PATCH v2 1/2] drm/tinydrm: Remove chunk splitting in tinydrm_spi_transfer

[Lukas Wunner]

On Sat, Feb 24, 2018 at 06:16:46PM +0000, Meghana Madhyastha wrote:
> -Remove chunk splitting in tinydrm_spi_transfer in tinydrm-helpers as The spi core will split a buffer into max_dma_len chunks for the
>  spi controller driver to handle.
> -Remove automatic byte swapping in tinydrm_spi_transfer as it doesn't have users.
> -Remove the upper bound check on dma transfer length in bcm2835_spi_can_dma().

AFAICS you need to reverse the order of the two patches, so that the
splitting is added to spi-bcm2835 before you remove it from the DRM
drivers.  Otherwise there's no splitting at all in-between the patches,
which may cause issues for someone hitting this patch when bisecting.

It would be good if you could explain the rationale of the patch
(the "why") in more detail.  You've provided the rationale in the
cover letter but the cover letter isn't recorded in the git history.
Right now the commit message only summarizes the "what".

Also, please wrap the commit message at 72 chars.


> --- a/drivers/spi/spi-bcm2835.c
> +++ b/drivers/spi/spi-bcm2835.c
> @@ -365,19 +365,6 @@ static bool bcm2835_spi_can_dma(struct spi_master *master,
>  	if (tfr->len < BCM2835_SPI_DMA_MIN_LENGTH)
>  		return false;
>  
> -	/* BCM2835_SPI_DLEN has defined a max transfer size as
> -	 * 16 bit, so max is 65535
> -	 * we can revisit this by using an alternative transfer
> -	 * method - ideally this would get done without any more
> -	 * interaction...
> -	 */
> -	if (tfr->len > 65535) {
> -		dev_warn_once(&spi->dev,
> -			      "transfer size of %d too big for dma-transfer\n",
> -			      tfr->len);
> -		return false;
> -	}
> -
>  	/* if we run rx/tx_buf with word aligned addresses then we are OK */
>  	if ((((size_t)tfr->rx_buf & 3) == 0) &&
>  	    (((size_t)tfr->tx_buf & 3) == 0))

Move this hunk to the other patch so that it's together with the
changes that remove the 65535 limitation.  It would probably good
to retain a portion of the comment to explain why the splitting is
necessary.


> @@ -461,7 +448,7 @@ static void bcm2835_dma_init(struct spi_master *master, struct device *dev)
>  
>  	/* all went well, so set can_dma */
>  	master->can_dma = bcm2835_spi_can_dma;
> -	master->max_dma_len = 65535; /* limitation by BCM2835_SPI_DLEN */
> +	master->max_dma_len = 65528; /* limitation by BCM2835_SPI_DLEN */

Spurious unexplained change.

Thanks,

Lukas
_______________________________________________
dri-devel mailing list
dri-devel@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/dri-devel

Re: [PATCH v2 1/2] drm/tinydrm: Remove chunk splitting in tinydrm_spi_transfer

[Noralf Trønnes]

Den 25.02.2018 18.03, skrev Lukas Wunner:
> On Sat, Feb 24, 2018 at 06:16:46PM +0000, Meghana Madhyastha wrote:
>> -Remove chunk splitting in tinydrm_spi_transfer in tinydrm-helpers as The spi core will split a buffer into max_dma_len chunks for the
>>   spi controller driver to handle.
>> -Remove automatic byte swapping in tinydrm_spi_transfer as it doesn't have users.
>> -Remove the upper bound check on dma transfer length in bcm2835_spi_can_dma().
> AFAICS you need to reverse the order of the two patches, so that the
> splitting is added to spi-bcm2835 before you remove it from the DRM
> drivers.  Otherwise there's no splitting at all in-between the patches,
> which may cause issues for someone hitting this patch when bisecting.
>
> It would be good if you could explain the rationale of the patch
> (the "why") in more detail.  You've provided the rationale in the
> cover letter but the cover letter isn't recorded in the git history.
> Right now the commit message only summarizes the "what".

Yeah, the why is important.

Currently clients using spi-bcm2835 need to know about the max_dma_len
limit to ensure DMA transfers. Trying to transfer larger buffers will
result in a warning and a slow PIO transfer.
tinydrm works around this by splitting the framebuffers before transfer
(a legacy from fbtft which tinydrm grew out of).

So the change is about removing the upper bound on DMA SPI transfers for
this particular spi controller driver.

(When I checked this a year ago, there was one more controller driver with
an upper bound on DMA transfers, but I don't remember which)

Noralf.


> Also, please wrap the commit message at 72 chars.
>
>
>> --- a/drivers/spi/spi-bcm2835.c
>> +++ b/drivers/spi/spi-bcm2835.c
>> @@ -365,19 +365,6 @@ static bool bcm2835_spi_can_dma(struct spi_master *master,
>>   	if (tfr->len < BCM2835_SPI_DMA_MIN_LENGTH)
>>   		return false;
>>   
>> -	/* BCM2835_SPI_DLEN has defined a max transfer size as
>> -	 * 16 bit, so max is 65535
>> -	 * we can revisit this by using an alternative transfer
>> -	 * method - ideally this would get done without any more
>> -	 * interaction...
>> -	 */
>> -	if (tfr->len > 65535) {
>> -		dev_warn_once(&spi->dev,
>> -			      "transfer size of %d too big for dma-transfer\n",
>> -			      tfr->len);
>> -		return false;
>> -	}
>> -
>>   	/* if we run rx/tx_buf with word aligned addresses then we are OK */
>>   	if ((((size_t)tfr->rx_buf & 3) == 0) &&
>>   	    (((size_t)tfr->tx_buf & 3) == 0))
> Move this hunk to the other patch so that it's together with the
> changes that remove the 65535 limitation.  It would probably good
> to retain a portion of the comment to explain why the splitting is
> necessary.
>
>
>> @@ -461,7 +448,7 @@ static void bcm2835_dma_init(struct spi_master *master, struct device *dev)
>>   
>>   	/* all went well, so set can_dma */
>>   	master->can_dma = bcm2835_spi_can_dma;
>> -	master->max_dma_len = 65535; /* limitation by BCM2835_SPI_DLEN */
>> +	master->max_dma_len = 65528; /* limitation by BCM2835_SPI_DLEN */
> Spurious unexplained change.
>
> Thanks,
>
> Lukas

_______________________________________________
dri-devel mailing list
dri-devel@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/dri-devel

[PATCH v2 2/2] spi/spi-bcm2835: Add bcm2835_spi_transfer_one_message in spi-bcm2835.c

[Meghana Madhyastha]

Add bcm2835_spi_transfer_one_message in spi-bcm2835.c. This function calls
the helper spi_split_transfers_maxsize before calling spi_transfer_one_message
to split the message into smaller chunks to be able to use dma.

Signed-off-by: Meghana Madhyastha <meghana.madhyastha@gmail.com>
---
 drivers/spi/spi-bcm2835.c | 14 ++++++++++++++
 drivers/spi/spi.c         |  5 ++---
 include/linux/spi/spi.h   |  2 ++
 3 files changed, 18 insertions(+), 3 deletions(-)

diff --git a/drivers/spi/spi-bcm2835.c b/drivers/spi/spi-bcm2835.c
index 2fd650891c07..68d35407e7a3 100644
--- a/drivers/spi/spi-bcm2835.c
+++ b/drivers/spi/spi-bcm2835.c
@@ -579,6 +579,19 @@ static int bcm2835_spi_transfer_one(struct spi_master *master,
 	return bcm2835_spi_transfer_one_irq(master, spi, tfr, cs);
 }
 
+static int bcm2835_spi_transfer_one_message(struct spi_controller *ctlr,
+					struct spi_message *msg)
+{
+	int ret;
+	gfp_t gfp_flags = GFP_KERNEL | GFP_DMA;
+	size_t max_transfer_size = 64;
+	ret = spi_split_transfers_maxsize(ctlr, msg, max_transfer_size, gfp_flags);
+	if (ret)
+		return ret;
+
+	return spi_transfer_one_message(ctlr, msg);
+}
+
 static int bcm2835_spi_prepare_message(struct spi_master *master,
 				       struct spi_message *msg)
 {
@@ -739,6 +752,7 @@ static int bcm2835_spi_probe(struct platform_device *pdev)
 	master->setup = bcm2835_spi_setup;
 	master->set_cs = bcm2835_spi_set_cs;
 	master->transfer_one = bcm2835_spi_transfer_one;
+	master->transfer_one_message = bcm2835_spi_transfer_one_message;
 	master->handle_err = bcm2835_spi_handle_err;
 	master->prepare_message = bcm2835_spi_prepare_message;
 	master->dev.of_node = pdev->dev.of_node;
diff --git a/drivers/spi/spi.c b/drivers/spi/spi.c
index b33a727a0158..3a352a0a6cdf 100644
--- a/drivers/spi/spi.c
+++ b/drivers/spi/spi.c
@@ -1003,7 +1003,7 @@ static int spi_map_msg(struct spi_controller *ctlr, struct spi_message *msg)
  * drivers which implement a transfer_one() operation.  It provides
  * standard handling of delays and chip select management.
  */
-static int spi_transfer_one_message(struct spi_controller *ctlr,
+int spi_transfer_one_message(struct spi_controller *ctlr,
 				    struct spi_message *msg)
 {
 	struct spi_transfer *xfer;
@@ -1026,7 +1026,6 @@ static int spi_transfer_one_message(struct spi_controller *ctlr,
 
 		if (xfer->tx_buf || xfer->rx_buf) {
 			reinit_completion(&ctlr->xfer_completion);
-
 			ret = ctlr->transfer_one(ctlr, msg->spi, xfer);
 			if (ret < 0) {
 				SPI_STATISTICS_INCREMENT_FIELD(statm,
@@ -1111,6 +1110,7 @@ static int spi_transfer_one_message(struct spi_controller *ctlr,
 
 	return ret;
 }
+EXPORT_SYMBOL_GPL(spi_transfer_one_message);
 
 /**
  * spi_finalize_current_transfer - report completion of a transfer
@@ -2583,7 +2583,6 @@ static int __spi_split_transfer_maxsize(struct spi_controller *ctlr,
 
 	/* calculate how many we have to replace */
 	count = DIV_ROUND_UP(xfer->len, maxsize);
-
 	/* create replacement */
 	srt = spi_replace_transfers(msg, xfer, 1, count, NULL, 0, gfp);
 	if (IS_ERR(srt))
diff --git a/include/linux/spi/spi.h b/include/linux/spi/spi.h
index bc6bb325d1bf..49b6dbce3e6d 100644
--- a/include/linux/spi/spi.h
+++ b/include/linux/spi/spi.h
@@ -622,6 +622,8 @@ extern void spi_finalize_current_transfer(struct spi_controller *ctlr);
 /* the spi driver core manages memory for the spi_controller classdev */
 extern struct spi_controller *__spi_alloc_controller(struct device *host,
 						unsigned int size, bool slave);
+extern int spi_transfer_one_message(struct spi_controller *ctlr,
+						struct spi_message *msg);
 
 static inline struct spi_controller *spi_alloc_master(struct device *host,
 						      unsigned int size)
-- 
2.11.0

_______________________________________________
dri-devel mailing list
dri-devel@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/dri-devel

Re: [PATCH v2 2/2] spi/spi-bcm2835: Add bcm2835_spi_transfer_one_message in spi-bcm2835.c

[Lukas Wunner]

On Sat, Feb 24, 2018 at 06:17:12PM +0000, Meghana Madhyastha wrote:
> --- a/drivers/spi/spi.c
> +++ b/drivers/spi/spi.c
> @@ -1003,7 +1003,7 @@ static int spi_map_msg(struct spi_controller *ctlr, struct spi_message *msg)
>   * drivers which implement a transfer_one() operation.  It provides
>   * standard handling of delays and chip select management.
>   */
> -static int spi_transfer_one_message(struct spi_controller *ctlr,
> +int spi_transfer_one_message(struct spi_controller *ctlr,

An alternative to exporting this function would be to have it check
if any of the transfers in the message exceed ctlr->max_dma_len,
and automatically split those.  That's for Mark to decide.


>  		if (xfer->tx_buf || xfer->rx_buf) {
>  			reinit_completion(&ctlr->xfer_completion);
> -

Spurious deleted newline.


>  	/* calculate how many we have to replace */
>  	count = DIV_ROUND_UP(xfer->len, maxsize);
> -

Another one.

Thanks,

Lukas
_______________________________________________
dri-devel mailing list
dri-devel@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/dri-devel

Re: [PATCH v2 0/2] Chunk splitting of spi transfers

[Lukas Wunner]

[cc += linux-rpi-kernel@lists.infradead.org]

On Sat, Feb 24, 2018 at 06:15:59PM +0000, Meghana Madhyastha wrote:
> I've added bcm2835_spi_transfer_one_message in spi-bcm2835. This calls
> spi_split_transfers_maxsize to split large chunks for spi dma transfers. 
> I then removed chunk splitting in the tinydrm spi helper (as now the core
> is handling the chunk splitting). However, although the SPI HW should be
> able to accomodate up to 65535 bytes for dma transfers, the splitting of
> chunks to 65535 bytes results in a dma transfer time out error. However,
> when the chunks are split to < 64 bytes it seems to work fine.

Hm, that is really odd, how did you test this exactly, what did you
use as SPI slave?  It contradicts our own experience, we're using
Micrel KSZ8851 Ethernet chips as SPI slave on spi0 of a BCM2837
and can send/receive messages via DMA to the tune of several hundred
bytes without any issues.  In fact, for messages < 96 bytes, DMA is
not used at all, so you've probably been using interrupt mode,
see the BCM2835_SPI_DMA_MIN_LENGTH macro in spi-bcm2835.c.

Thanks,

Lukas
_______________________________________________
dri-devel mailing list
dri-devel@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/dri-devel

Re: [PATCH v2 0/2] Chunk splitting of spi transfers

[Noralf Trønnes]

Den 25.02.2018 14.19, skrev Lukas Wunner:
> [cc += linux-rpi-kernel@lists.infradead.org]
>
> On Sat, Feb 24, 2018 at 06:15:59PM +0000, Meghana Madhyastha wrote:
>> I've added bcm2835_spi_transfer_one_message in spi-bcm2835. This calls
>> spi_split_transfers_maxsize to split large chunks for spi dma transfers.
>> I then removed chunk splitting in the tinydrm spi helper (as now the core
>> is handling the chunk splitting). However, although the SPI HW should be
>> able to accomodate up to 65535 bytes for dma transfers, the splitting of
>> chunks to 65535 bytes results in a dma transfer time out error. However,
>> when the chunks are split to < 64 bytes it seems to work fine.
> Hm, that is really odd, how did you test this exactly, what did you
> use as SPI slave?  It contradicts our own experience, we're using
> Micrel KSZ8851 Ethernet chips as SPI slave on spi0 of a BCM2837
> and can send/receive messages via DMA to the tune of several hundred
> bytes without any issues.  In fact, for messages < 96 bytes, DMA is
> not used at all, so you've probably been using interrupt mode,
> see the BCM2835_SPI_DMA_MIN_LENGTH macro in spi-bcm2835.c.

Meghana,
I suggest using modetest with page flipping to check that you don't
regress performance wise:

$ ./libdrm/tests/modetest/modetest -M mi0283qt -s 28:320x240@RG16 -v
setting mode 320x240-0Hz@RG16 on connectors 28, crtc 30
freq: 29.88Hz
freq: 29.96Hz
freq: 29.99Hz
^C

Noralf.

_______________________________________________
dri-devel mailing list
dri-devel@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/dri-devel

Re: [PATCH v2 0/2] Chunk splitting of spi transfers

[Meghana Madhyastha]

On Sun, Feb 25, 2018 at 02:19:10PM +0100, Lukas Wunner wrote:
> [cc += linux-rpi-kernel@lists.infradead.org]
> 
> On Sat, Feb 24, 2018 at 06:15:59PM +0000, Meghana Madhyastha wrote:
> > I've added bcm2835_spi_transfer_one_message in spi-bcm2835. This calls
> > spi_split_transfers_maxsize to split large chunks for spi dma transfers. 
> > I then removed chunk splitting in the tinydrm spi helper (as now the core
> > is handling the chunk splitting). However, although the SPI HW should be
> > able to accomodate up to 65535 bytes for dma transfers, the splitting of
> > chunks to 65535 bytes results in a dma transfer time out error. However,
> > when the chunks are split to < 64 bytes it seems to work fine.
> 
> Hm, that is really odd, how did you test this exactly, what did you
> use as SPI slave?  It contradicts our own experience, we're using
> Micrel KSZ8851 Ethernet chips as SPI slave on spi0 of a BCM2837
> and can send/receive messages via DMA to the tune of several hundred
> bytes without any issues.  In fact, for messages < 96 bytes, DMA is
> not used at all, so you've probably been using interrupt mode,
> see the BCM2835_SPI_DMA_MIN_LENGTH macro in spi-bcm2835.c.

Hi Lukas,

I think you are right. I checked it and its not using the DMA mode which
is why its working with 64 bytes.
Noralf, that leaves us back to the
initial time out problem. I've tried doing the message splitting in
spi_sync as well as spi_pump_messages. Martin had explained that DMA
will wait for
the SPI HW to set the send_more_data line, but the SPI-HW itself will
stop triggering it when SPI_LEN is 0 causing DMA to wait forever. I
thought if we split it before itself, the SPI_LEN will not go to zero
thus preventing this problem, however it didn't work and started
hanging. So I'm a little uncertain as to how to proceed and debug what
exactly has caused the time out due to the asynchronous methods.

Thanks and regards,
Meghana

> Thanks,
> 
> Lukas
_______________________________________________
dri-devel mailing list
dri-devel@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/dri-devel

Re: [PATCH v2 0/2] Chunk splitting of spi transfers

[Noralf Trønnes]

Den 02.03.2018 12.11, skrev Meghana Madhyastha:
> On Sun, Feb 25, 2018 at 02:19:10PM +0100, Lukas Wunner wrote:
>> [cc += linux-rpi-kernel@lists.infradead.org]
>>
>> On Sat, Feb 24, 2018 at 06:15:59PM +0000, Meghana Madhyastha wrote:
>>> I've added bcm2835_spi_transfer_one_message in spi-bcm2835. This calls
>>> spi_split_transfers_maxsize to split large chunks for spi dma transfers.
>>> I then removed chunk splitting in the tinydrm spi helper (as now the core
>>> is handling the chunk splitting). However, although the SPI HW should be
>>> able to accomodate up to 65535 bytes for dma transfers, the splitting of
>>> chunks to 65535 bytes results in a dma transfer time out error. However,
>>> when the chunks are split to < 64 bytes it seems to work fine.
>> Hm, that is really odd, how did you test this exactly, what did you
>> use as SPI slave?  It contradicts our own experience, we're using
>> Micrel KSZ8851 Ethernet chips as SPI slave on spi0 of a BCM2837
>> and can send/receive messages via DMA to the tune of several hundred
>> bytes without any issues.  In fact, for messages < 96 bytes, DMA is
>> not used at all, so you've probably been using interrupt mode,
>> see the BCM2835_SPI_DMA_MIN_LENGTH macro in spi-bcm2835.c.
> Hi Lukas,
>
> I think you are right. I checked it and its not using the DMA mode which
> is why its working with 64 bytes.
> Noralf, that leaves us back to the
> initial time out problem. I've tried doing the message splitting in
> spi_sync as well as spi_pump_messages. Martin had explained that DMA
> will wait for
> the SPI HW to set the send_more_data line, but the SPI-HW itself will
> stop triggering it when SPI_LEN is 0 causing DMA to wait forever. I
> thought if we split it before itself, the SPI_LEN will not go to zero
> thus preventing this problem, however it didn't work and started
> hanging. So I'm a little uncertain as to how to proceed and debug what
> exactly has caused the time out due to the asynchronous methods.

I did a quick test and at least this is working:

int tinydrm_spi_transfer(struct spi_device *spi, u32 speed_hz,
              struct spi_transfer *header, u8 bpw, const void *buf,
              size_t len)
{
     struct spi_transfer tr = {
         .bits_per_word = bpw,
         .speed_hz = speed_hz,
         .tx_buf = buf,
         .len = len,
     };
     struct spi_message m;
     size_t maxsize;
     int ret;

     maxsize = tinydrm_spi_max_transfer_size(spi, 0);

     if (drm_debug & DRM_UT_DRIVER)
         pr_debug("[drm:%s] bpw=%u, maxsize=%zu, transfers:\n",
              __func__, bpw, maxsize);

     spi_message_init(&m);
     m.spi = spi;
     if (header)
         spi_message_add_tail(header, &m);
     spi_message_add_tail(&tr, &m);

     ret = spi_split_transfers_maxsize(spi->controller, &m, maxsize, 
GFP_KERNEL);
     if (ret)
         return ret;

     tinydrm_dbg_spi_message(spi, &m);

     return spi_sync(spi, &m);
}
EXPORT_SYMBOL(tinydrm_spi_transfer);


Log:
[   39.015644] [drm:mipi_dbi_fb_dirty [mipi_dbi]] Flushing [FB:36] x1=0, 
x2=320, y1=0, y2=240

[   39.018079] [drm:mipi_dbi_typec3_command [mipi_dbi]] cmd=2a, par=00 
00 01 3f
[   39.018129] [drm:tinydrm_spi_transfer] bpw=8, maxsize=65532, transfers:
[   39.018152]     tr(0): speed=10MHz, bpw=8, len=1, tx_buf=[2a]
[   39.018231] [drm:tinydrm_spi_transfer] bpw=8, maxsize=65532, transfers:
[   39.018248]     tr(0): speed=10MHz, bpw=8, len=4, tx_buf=[00 00 01 3f]

[   39.018330] [drm:mipi_dbi_typec3_command [mipi_dbi]] cmd=2b, par=00 
00 00 ef
[   39.018347] [drm:tinydrm_spi_transfer] bpw=8, maxsize=65532, transfers:
[   39.018362]     tr(0): speed=10MHz, bpw=8, len=1, tx_buf=[2b]
[   39.018396] [drm:tinydrm_spi_transfer] bpw=8, maxsize=65532, transfers:
[   39.018428]     tr(0): speed=10MHz, bpw=8, len=4, tx_buf=[00 00 00 ef]

[   39.018487] [drm:mipi_dbi_typec3_command [mipi_dbi]] cmd=2c, len=153600
[   39.018502] [drm:tinydrm_spi_transfer] bpw=8, maxsize=65532, transfers:
[   39.018517]     tr(0): speed=10MHz, bpw=8, len=1, tx_buf=[2c]
[   39.018565] [drm:tinydrm_spi_transfer] bpw=8, maxsize=65532, transfers:
[   39.018594]     tr(0): speed=48MHz, bpw=8, len=65532, tx_buf=[c6 18 
c6 18 c6 18 c6 18 c6 18 c6 18 c6 18 c6 18 ...]
[   39.018608]     tr(1): speed=48MHz, bpw=8, len=65532, tx_buf=[06 18 
06 18 06 18 06 18 06 18 06 18 06 18 06 18 ...]
[   39.018621]     tr(2): speed=48MHz, bpw=8, len=22536, tx_buf=[10 82 
10 82 10 82 10 82 10 82 10 82 18 e3 18 e3 ...]


Noralf.


_______________________________________________
dri-devel mailing list
dri-devel@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/dri-devel