* [PATCH] spi: ti-qspi: use 128 bit transfer mode for writing to flash
@ 2015-08-20 10:30 Vignesh R
2015-08-20 17:55 ` Mark Brown
0 siblings, 1 reply; 3+ messages in thread
From: Vignesh R @ 2015-08-20 10:30 UTC (permalink / raw)
To: Mark Brown; +Cc: linux-spi, linux-kernel, linux-omap, Vignesh R
TI QSPI has four 32 bit data regsiters which can be used to transfer 16
bytes of data at once. The register group QSPI_SPI_DATA_REG_3,
QSPI_SPI_DATA_REG_2, QSPI_SPI_DATA_REG_1 and QSPI_SPI_DATA_REG is
treated as a single 128-bit word for shifting data in and out. The bit
at QSPI_SPI_DATA_REG_3[31] position is the first bit to be shifted out
in case of 128 bit transfer mode. Therefore the first byte to be written
to flash should be at QSPI_SPI_DATA_REG_3[31-25] position.
Instead of writing 1 byte at a time when interacting with spi-nor flash,
make use of all the four registers so that 16 bytes can be transferred
in one go. This reduces number of register writes and Word Complete
interrupts for a given transfer message size, thereby increasing the
write performance.
Without this patch the raw flash write speed is ~100KB/s, with this
patch the write speed increases to ~400 kB/s on DRA74 EVM.
Signed-off-by: Vignesh R <vigneshr@ti.com>
---
drivers/spi/spi-ti-qspi.c | 34 ++++++++++++++++++++++++++++++----
1 file changed, 30 insertions(+), 4 deletions(-)
diff --git a/drivers/spi/spi-ti-qspi.c b/drivers/spi/spi-ti-qspi.c
index 45844a227c5e..f4cea6834fad 100644
--- a/drivers/spi/spi-ti-qspi.c
+++ b/drivers/spi/spi-ti-qspi.c
@@ -96,6 +96,8 @@ struct ti_qspi {
#define QSPI_INVAL (4 << 16)
#define QSPI_WC_CMD_INT_EN (1 << 14)
#define QSPI_FLEN(n) ((n - 1) << 0)
+#define QSPI_WLEN_MAX_BITS 128
+#define QSPI_WLEN_MAX_BYTES 16
/* STATUS REGISTER */
#define BUSY 0x01
@@ -224,14 +226,16 @@ static inline u32 qspi_is_busy(struct ti_qspi *qspi)
static int qspi_write_msg(struct ti_qspi *qspi, struct spi_transfer *t)
{
- int wlen, count;
+ int wlen, count, xfer_len;
unsigned int cmd;
const u8 *txbuf;
+ u32 data;
txbuf = t->tx_buf;
cmd = qspi->cmd | QSPI_WR_SNGL;
count = t->len;
wlen = t->bits_per_word >> 3; /* in bytes */
+ xfer_len = wlen;
while (count) {
if (qspi_is_busy(qspi))
@@ -241,7 +245,29 @@ static int qspi_write_msg(struct ti_qspi *qspi, struct spi_transfer *t)
case 1:
dev_dbg(qspi->dev, "tx cmd %08x dc %08x data %02x\n",
cmd, qspi->dc, *txbuf);
- writeb(*txbuf, qspi->base + QSPI_SPI_DATA_REG);
+ if (count >= QSPI_WLEN_MAX_BYTES) {
+ u32 *txp = (u32 *)txbuf;
+
+ data = cpu_to_be32(*txp++);
+ writel(data, qspi->base +
+ QSPI_SPI_DATA_REG_3);
+ data = cpu_to_be32(*txp++);
+ writel(data, qspi->base +
+ QSPI_SPI_DATA_REG_2);
+ data = cpu_to_be32(*txp++);
+ writel(data, qspi->base +
+ QSPI_SPI_DATA_REG_1);
+ data = cpu_to_be32(*txp++);
+ writel(data, qspi->base +
+ QSPI_SPI_DATA_REG);
+ xfer_len = QSPI_WLEN_MAX_BYTES;
+ cmd |= QSPI_WLEN(QSPI_WLEN_MAX_BITS);
+ } else {
+ writeb(*txbuf, qspi->base + QSPI_SPI_DATA_REG);
+ cmd = qspi->cmd | QSPI_WR_SNGL;
+ xfer_len = wlen;
+ cmd |= QSPI_WLEN(wlen);
+ }
break;
case 2:
dev_dbg(qspi->dev, "tx cmd %08x dc %08x data %04x\n",
@@ -261,8 +287,8 @@ static int qspi_write_msg(struct ti_qspi *qspi, struct spi_transfer *t)
dev_err(qspi->dev, "write timed out\n");
return -ETIMEDOUT;
}
- txbuf += wlen;
- count -= wlen;
+ txbuf += xfer_len;
+ count -= xfer_len;
}
return 0;
--
2.5.0
^ permalink raw reply related [flat|nested] 3+ messages in thread
* Re: [PATCH] spi: ti-qspi: use 128 bit transfer mode for writing to flash
2015-08-20 10:30 [PATCH] spi: ti-qspi: use 128 bit transfer mode for writing to flash Vignesh R
@ 2015-08-20 17:55 ` Mark Brown
2015-08-25 6:54 ` Vignesh R
0 siblings, 1 reply; 3+ messages in thread
From: Mark Brown @ 2015-08-20 17:55 UTC (permalink / raw)
To: Vignesh R; +Cc: linux-spi, linux-kernel, linux-omap
[-- Attachment #1: Type: text/plain, Size: 1029 bytes --]
On Thu, Aug 20, 2015 at 04:00:59PM +0530, Vignesh R wrote:
> - writeb(*txbuf, qspi->base + QSPI_SPI_DATA_REG);
> + if (count >= QSPI_WLEN_MAX_BYTES) {
> + u32 *txp = (u32 *)txbuf;
> +
> + data = cpu_to_be32(*txp++);
> + writel(data, qspi->base +
> + QSPI_SPI_DATA_REG_3);
> + data = cpu_to_be32(*txp++);
> + writel(data, qspi->base +
> + QSPI_SPI_DATA_REG_2);
> + data = cpu_to_be32(*txp++);
> + writel(data, qspi->base +
> + QSPI_SPI_DATA_REG_1);
> + data = cpu_to_be32(*txp++);
> + writel(data, qspi->base +
> + QSPI_SPI_DATA_REG);
> + xfer_len = QSPI_WLEN_MAX_BYTES;
> + cmd |= QSPI_WLEN(QSPI_WLEN_MAX_BITS);
> + } else {
> + writeb(*txbuf, qspi->base + QSPI_SPI_DATA_REG);
> + cmd = qspi->cmd | QSPI_WR_SNGL;
> + xfer_len = wlen;
> + cmd |= QSPI_WLEN(wlen);
> + }
It's a bit sad that this isn't able to do a Duff's device type thing and
only kicks in for the full 128 bit FIFO size, it looks like it could do
any number of words.
[-- Attachment #2: Digital signature --]
[-- Type: application/pgp-signature, Size: 473 bytes --]
^ permalink raw reply [flat|nested] 3+ messages in thread
* Re: [PATCH] spi: ti-qspi: use 128 bit transfer mode for writing to flash
2015-08-20 17:55 ` Mark Brown
@ 2015-08-25 6:54 ` Vignesh R
0 siblings, 0 replies; 3+ messages in thread
From: Vignesh R @ 2015-08-25 6:54 UTC (permalink / raw)
To: Mark Brown
Cc: linux-spi@vger.kernel.org, linux-kernel@vger.kernel.org,
linux-omap@vger.kernel.org
On 08/20/2015 11:25 PM, Mark Brown wrote:
> On Thu, Aug 20, 2015 at 04:00:59PM +0530, Vignesh R wrote:
>
>> - writeb(*txbuf, qspi->base + QSPI_SPI_DATA_REG);
>> + if (count >= QSPI_WLEN_MAX_BYTES) {
>> + u32 *txp = (u32 *)txbuf;
>> +
>> + data = cpu_to_be32(*txp++);
>> + writel(data, qspi->base +
>> + QSPI_SPI_DATA_REG_3);
>> + data = cpu_to_be32(*txp++);
>> + writel(data, qspi->base +
>> + QSPI_SPI_DATA_REG_2);
>> + data = cpu_to_be32(*txp++);
>> + writel(data, qspi->base +
>> + QSPI_SPI_DATA_REG_1);
>> + data = cpu_to_be32(*txp++);
>> + writel(data, qspi->base +
>> + QSPI_SPI_DATA_REG);
>> + xfer_len = QSPI_WLEN_MAX_BYTES;
>> + cmd |= QSPI_WLEN(QSPI_WLEN_MAX_BITS);
>> + } else {
>> + writeb(*txbuf, qspi->base + QSPI_SPI_DATA_REG);
>> + cmd = qspi->cmd | QSPI_WR_SNGL;
>> + xfer_len = wlen;
>> + cmd |= QSPI_WLEN(wlen);
>> + }
>
> It's a bit sad that this isn't able to do a Duff's device type thing and
> only kicks in for the full 128 bit FIFO size, it looks like it could do
> any number of words.
>
Yes, any number of bytes can be transfered (max 16 bytes).
I will try to work on your suggestion. Thanks!
--
Regards
Vignesh
^ permalink raw reply [flat|nested] 3+ messages in thread
end of thread, other threads:[~2015-08-25 6:55 UTC | newest]
Thread overview: 3+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2015-08-20 10:30 [PATCH] spi: ti-qspi: use 128 bit transfer mode for writing to flash Vignesh R
2015-08-20 17:55 ` Mark Brown
2015-08-25 6:54 ` Vignesh R
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox