|
|
|
/*
|
|
|
|
* Freescale i.MX28 SPI driver
|
|
|
|
*
|
|
|
|
* Copyright (C) 2011 Marek Vasut <marek.vasut@gmail.com>
|
|
|
|
* on behalf of DENX Software Engineering GmbH
|
|
|
|
*
|
|
|
|
* SPDX-License-Identifier: GPL-2.0+
|
|
|
|
*
|
|
|
|
* NOTE: This driver only supports the SPI-controller chipselects,
|
|
|
|
* GPIO driven chipselects are not supported.
|
|
|
|
*/
|
|
|
|
|
|
|
|
#include <common.h>
|
|
|
|
#include <malloc.h>
|
|
|
|
#include <spi.h>
|
|
|
|
#include <asm/errno.h>
|
|
|
|
#include <asm/io.h>
|
|
|
|
#include <asm/arch/clock.h>
|
|
|
|
#include <asm/arch/imx-regs.h>
|
|
|
|
#include <asm/arch/sys_proto.h>
|
|
|
|
#include <asm/imx-common/dma.h>
|
|
|
|
|
|
|
|
#define MXS_SPI_MAX_TIMEOUT 1000000
|
|
|
|
#define MXS_SPI_PORT_OFFSET 0x2000
|
|
|
|
#define MXS_SSP_CHIPSELECT_MASK 0x00300000
|
|
|
|
#define MXS_SSP_CHIPSELECT_SHIFT 20
|
|
|
|
|
|
|
|
#define MXSSSP_SMALL_TRANSFER 512
|
|
|
|
|
|
|
|
struct mxs_spi_slave {
|
|
|
|
struct spi_slave slave;
|
|
|
|
uint32_t max_khz;
|
|
|
|
uint32_t mode;
|
|
|
|
struct mxs_ssp_regs *regs;
|
|
|
|
};
|
|
|
|
|
|
|
|
static inline struct mxs_spi_slave *to_mxs_slave(struct spi_slave *slave)
|
|
|
|
{
|
|
|
|
return container_of(slave, struct mxs_spi_slave, slave);
|
|
|
|
}
|
|
|
|
|
|
|
|
void spi_init(void)
|
|
|
|
{
|
|
|
|
}
|
|
|
|
|
|
|
|
int spi_cs_is_valid(unsigned int bus, unsigned int cs)
|
|
|
|
{
|
|
|
|
/* MXS SPI: 4 ports and 3 chip selects maximum */
|
mxs: mmc: spi: dma: Better wrap the MXS differences
This patch streamlines the differences between the MX23 and MX28 by
implementing a few helper functions to handle different DMA channel
mapping, different clock domain for SSP block and fixes a few minor
bugs.
First of all, the DMA channel mapping is now fixed in dma.h by defining
the actual channel map for both MX23 and MX28. Thus, MX23 now does no
longer use MX28 channel map which was wrong. Also, there is a fix for
MX28 DMA channel map, where the last four channels were incorrect.
Next, because correct DMA channel map is in place, the mxs_dma_init_channel()
call now bases the channel ID starting from SSP port #0. This removes the
need for DMA channel offset being added and cleans up the code. For the
same reason, the SSP0 offset can now be used in mxs_dma_desc_append(), thus
no need to adjust dma channel number in the driver either.
Lastly, the SSP clock ID is now retrieved by calling mxs_ssp_clock_by_bus()
which handles the fact that MX23 has shared SSP clock for both ports, while
MX28 has per-port SSP clock.
Finally, the mxs_ssp_bus_id_valid() pulls out two implementations of the
same functionality from MMC and SPI driver into common code.
Signed-off-by: Marek Vasut <marex@denx.de>
Cc: Fabio Estevam <fabio.estevam@freescale.com>
Cc: Otavio Salvador <otavio@ossystems.com.br>
Cc: Stefano Babic <sbabic@denx.de>
12 years ago
|
|
|
if (!mxs_ssp_bus_id_valid(bus) || cs > 2)
|
|
|
|
return 0;
|
|
|
|
else
|
|
|
|
return 1;
|
|
|
|
}
|
|
|
|
|
|
|
|
struct spi_slave *spi_setup_slave(unsigned int bus, unsigned int cs,
|
|
|
|
unsigned int max_hz, unsigned int mode)
|
|
|
|
{
|
|
|
|
struct mxs_spi_slave *mxs_slave;
|
|
|
|
|
|
|
|
if (!spi_cs_is_valid(bus, cs)) {
|
|
|
|
printf("mxs_spi: invalid bus %d / chip select %d\n", bus, cs);
|
|
|
|
return NULL;
|
|
|
|
}
|
|
|
|
|
|
|
|
mxs_slave = spi_alloc_slave(struct mxs_spi_slave, bus, cs);
|
|
|
|
if (!mxs_slave)
|
|
|
|
return NULL;
|
|
|
|
|
mxs: mmc: spi: dma: Better wrap the MXS differences
This patch streamlines the differences between the MX23 and MX28 by
implementing a few helper functions to handle different DMA channel
mapping, different clock domain for SSP block and fixes a few minor
bugs.
First of all, the DMA channel mapping is now fixed in dma.h by defining
the actual channel map for both MX23 and MX28. Thus, MX23 now does no
longer use MX28 channel map which was wrong. Also, there is a fix for
MX28 DMA channel map, where the last four channels were incorrect.
Next, because correct DMA channel map is in place, the mxs_dma_init_channel()
call now bases the channel ID starting from SSP port #0. This removes the
need for DMA channel offset being added and cleans up the code. For the
same reason, the SSP0 offset can now be used in mxs_dma_desc_append(), thus
no need to adjust dma channel number in the driver either.
Lastly, the SSP clock ID is now retrieved by calling mxs_ssp_clock_by_bus()
which handles the fact that MX23 has shared SSP clock for both ports, while
MX28 has per-port SSP clock.
Finally, the mxs_ssp_bus_id_valid() pulls out two implementations of the
same functionality from MMC and SPI driver into common code.
Signed-off-by: Marek Vasut <marex@denx.de>
Cc: Fabio Estevam <fabio.estevam@freescale.com>
Cc: Otavio Salvador <otavio@ossystems.com.br>
Cc: Stefano Babic <sbabic@denx.de>
12 years ago
|
|
|
if (mxs_dma_init_channel(MXS_DMA_CHANNEL_AHB_APBH_SSP0 + bus))
|
|
|
|
goto err_init;
|
|
|
|
|
|
|
|
mxs_slave->max_khz = max_hz / 1000;
|
|
|
|
mxs_slave->mode = mode;
|
|
|
|
mxs_slave->regs = mxs_ssp_regs_by_bus(bus);
|
|
|
|
|
|
|
|
return &mxs_slave->slave;
|
|
|
|
|
|
|
|
err_init:
|
|
|
|
free(mxs_slave);
|
|
|
|
return NULL;
|
|
|
|
}
|
|
|
|
|
|
|
|
void spi_free_slave(struct spi_slave *slave)
|
|
|
|
{
|
|
|
|
struct mxs_spi_slave *mxs_slave = to_mxs_slave(slave);
|
|
|
|
free(mxs_slave);
|
|
|
|
}
|
|
|
|
|
|
|
|
int spi_claim_bus(struct spi_slave *slave)
|
|
|
|
{
|
|
|
|
struct mxs_spi_slave *mxs_slave = to_mxs_slave(slave);
|
|
|
|
struct mxs_ssp_regs *ssp_regs = mxs_slave->regs;
|
|
|
|
uint32_t reg = 0;
|
|
|
|
|
|
|
|
mxs_reset_block(&ssp_regs->hw_ssp_ctrl0_reg);
|
|
|
|
|
|
|
|
writel((slave->cs << MXS_SSP_CHIPSELECT_SHIFT) |
|
|
|
|
SSP_CTRL0_BUS_WIDTH_ONE_BIT,
|
|
|
|
&ssp_regs->hw_ssp_ctrl0);
|
|
|
|
|
|
|
|
reg = SSP_CTRL1_SSP_MODE_SPI | SSP_CTRL1_WORD_LENGTH_EIGHT_BITS;
|
|
|
|
reg |= (mxs_slave->mode & SPI_CPOL) ? SSP_CTRL1_POLARITY : 0;
|
|
|
|
reg |= (mxs_slave->mode & SPI_CPHA) ? SSP_CTRL1_PHASE : 0;
|
|
|
|
writel(reg, &ssp_regs->hw_ssp_ctrl1);
|
|
|
|
|
|
|
|
writel(0, &ssp_regs->hw_ssp_cmd0);
|
|
|
|
|
|
|
|
mxs_set_ssp_busclock(slave->bus, mxs_slave->max_khz);
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
void spi_release_bus(struct spi_slave *slave)
|
|
|
|
{
|
|
|
|
}
|
|
|
|
|
|
|
|
static void mxs_spi_start_xfer(struct mxs_ssp_regs *ssp_regs)
|
|
|
|
{
|
|
|
|
writel(SSP_CTRL0_LOCK_CS, &ssp_regs->hw_ssp_ctrl0_set);
|
|
|
|
writel(SSP_CTRL0_IGNORE_CRC, &ssp_regs->hw_ssp_ctrl0_clr);
|
|
|
|
}
|
|
|
|
|
|
|
|
static void mxs_spi_end_xfer(struct mxs_ssp_regs *ssp_regs)
|
|
|
|
{
|
|
|
|
writel(SSP_CTRL0_LOCK_CS, &ssp_regs->hw_ssp_ctrl0_clr);
|
|
|
|
writel(SSP_CTRL0_IGNORE_CRC, &ssp_regs->hw_ssp_ctrl0_set);
|
|
|
|
}
|
|
|
|
|
|
|
|
static int mxs_spi_xfer_pio(struct mxs_spi_slave *slave,
|
|
|
|
char *data, int length, int write, unsigned long flags)
|
|
|
|
{
|
|
|
|
struct mxs_ssp_regs *ssp_regs = slave->regs;
|
|
|
|
|
|
|
|
if (flags & SPI_XFER_BEGIN)
|
|
|
|
mxs_spi_start_xfer(ssp_regs);
|
|
|
|
|
|
|
|
while (length--) {
|
|
|
|
/* We transfer 1 byte */
|
|
|
|
#if defined(CONFIG_MX23)
|
|
|
|
writel(SSP_CTRL0_XFER_COUNT_MASK, &ssp_regs->hw_ssp_ctrl0_clr);
|
|
|
|
writel(1, &ssp_regs->hw_ssp_ctrl0_set);
|
|
|
|
#elif defined(CONFIG_MX28)
|
|
|
|
writel(1, &ssp_regs->hw_ssp_xfer_size);
|
|
|
|
#endif
|
|
|
|
|
|
|
|
if ((flags & SPI_XFER_END) && !length)
|
|
|
|
mxs_spi_end_xfer(ssp_regs);
|
|
|
|
|
|
|
|
if (write)
|
|
|
|
writel(SSP_CTRL0_READ, &ssp_regs->hw_ssp_ctrl0_clr);
|
|
|
|
else
|
|
|
|
writel(SSP_CTRL0_READ, &ssp_regs->hw_ssp_ctrl0_set);
|
|
|
|
|
|
|
|
writel(SSP_CTRL0_RUN, &ssp_regs->hw_ssp_ctrl0_set);
|
|
|
|
|
|
|
|
if (mxs_wait_mask_set(&ssp_regs->hw_ssp_ctrl0_reg,
|
|
|
|
SSP_CTRL0_RUN, MXS_SPI_MAX_TIMEOUT)) {
|
|
|
|
printf("MXS SPI: Timeout waiting for start\n");
|
|
|
|
return -ETIMEDOUT;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (write)
|
|
|
|
writel(*data++, &ssp_regs->hw_ssp_data);
|
|
|
|
|
|
|
|
writel(SSP_CTRL0_DATA_XFER, &ssp_regs->hw_ssp_ctrl0_set);
|
|
|
|
|
|
|
|
if (!write) {
|
|
|
|
if (mxs_wait_mask_clr(&ssp_regs->hw_ssp_status_reg,
|
|
|
|
SSP_STATUS_FIFO_EMPTY, MXS_SPI_MAX_TIMEOUT)) {
|
|
|
|
printf("MXS SPI: Timeout waiting for data\n");
|
|
|
|
return -ETIMEDOUT;
|
|
|
|
}
|
|
|
|
|
|
|
|
*data = readl(&ssp_regs->hw_ssp_data);
|
|
|
|
data++;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (mxs_wait_mask_clr(&ssp_regs->hw_ssp_ctrl0_reg,
|
|
|
|
SSP_CTRL0_RUN, MXS_SPI_MAX_TIMEOUT)) {
|
|
|
|
printf("MXS SPI: Timeout waiting for finish\n");
|
|
|
|
return -ETIMEDOUT;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
static int mxs_spi_xfer_dma(struct mxs_spi_slave *slave,
|
|
|
|
char *data, int length, int write, unsigned long flags)
|
|
|
|
{
|
|
|
|
const int xfer_max_sz = 0xff00;
|
|
|
|
const int desc_count = DIV_ROUND_UP(length, xfer_max_sz) + 1;
|
|
|
|
struct mxs_ssp_regs *ssp_regs = slave->regs;
|
|
|
|
struct mxs_dma_desc *dp;
|
|
|
|
uint32_t ctrl0;
|
|
|
|
uint32_t cache_data_count;
|
|
|
|
const uint32_t dstart = (uint32_t)data;
|
|
|
|
int dmach;
|
|
|
|
int tl;
|
MX28: SPI: Fix the DMA chaining
It turns out that in order for the SPI DMA to properly support
continuous transfers longer than 65280 bytes, there are some very
important parts that were left out from the documentation.
Firstly, the XFER_SIZE register is not written with the whole length
of a transfer, but is written by each and every chained descriptor
with the length of the descriptors data buffer.
Next, unlike the demo code supplied by FSL, which only writes one PIO
word per descriptor, this does not apply if the descriptors are chained,
since the XFER_SIZE register must be written. Therefore, it is essential
to use four PIO words, CTRL0, CMD0, CMD1, XFER_SIZE. CMD0 and CMD1 are
written with zero, since they don't apply. The DMA programs the PIO words
in an incrementing order, so four PIO words.
Finally, unlike the demo code supplied by FSL, the SSP_CTRL0_IGNORE_CRC
must not be set during the whole transfer, but it must be set only on the
last descriptor in the chain.
Signed-off-by: Marek Vasut <marex@denx.de>
Cc: Fabio Estevam <festevam@gmail.com>
Cc: Otavio Salvador <otavio@ossystems.com.br>
Cc: Stefano Babic <sbabic@denx.de>
13 years ago
|
|
|
int ret = 0;
|
|
|
|
|
|
|
|
#if defined(CONFIG_MX23)
|
|
|
|
const int mxs_spi_pio_words = 1;
|
|
|
|
#elif defined(CONFIG_MX28)
|
|
|
|
const int mxs_spi_pio_words = 4;
|
|
|
|
#endif
|
|
|
|
|
|
|
|
ALLOC_CACHE_ALIGN_BUFFER(struct mxs_dma_desc, desc, desc_count);
|
|
|
|
|
|
|
|
memset(desc, 0, sizeof(struct mxs_dma_desc) * desc_count);
|
|
|
|
|
|
|
|
ctrl0 = readl(&ssp_regs->hw_ssp_ctrl0);
|
|
|
|
ctrl0 |= SSP_CTRL0_DATA_XFER;
|
|
|
|
|
|
|
|
if (flags & SPI_XFER_BEGIN)
|
|
|
|
ctrl0 |= SSP_CTRL0_LOCK_CS;
|
|
|
|
if (!write)
|
|
|
|
ctrl0 |= SSP_CTRL0_READ;
|
|
|
|
|
|
|
|
if (length % ARCH_DMA_MINALIGN)
|
|
|
|
cache_data_count = roundup(length, ARCH_DMA_MINALIGN);
|
|
|
|
else
|
|
|
|
cache_data_count = length;
|
|
|
|
|
|
|
|
/* Flush data to DRAM so DMA can pick them up */
|
|
|
|
if (write)
|
|
|
|
flush_dcache_range(dstart, dstart + cache_data_count);
|
|
|
|
|
|
|
|
/* Invalidate the area, so no writeback into the RAM races with DMA */
|
|
|
|
invalidate_dcache_range(dstart, dstart + cache_data_count);
|
|
|
|
|
|
|
|
dmach = MXS_DMA_CHANNEL_AHB_APBH_SSP0 + slave->slave.bus;
|
|
|
|
|
|
|
|
dp = desc;
|
|
|
|
while (length) {
|
|
|
|
dp->address = (dma_addr_t)dp;
|
|
|
|
dp->cmd.address = (dma_addr_t)data;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* This is correct, even though it does indeed look insane.
|
|
|
|
* I hereby have to, wholeheartedly, thank Freescale Inc.,
|
|
|
|
* for always inventing insane hardware and keeping me busy
|
|
|
|
* and employed ;-)
|
|
|
|
*/
|
|
|
|
if (write)
|
|
|
|
dp->cmd.data = MXS_DMA_DESC_COMMAND_DMA_READ;
|
|
|
|
else
|
|
|
|
dp->cmd.data = MXS_DMA_DESC_COMMAND_DMA_WRITE;
|
|
|
|
|
|
|
|
/*
|
|
|
|
* The DMA controller can transfer large chunks (64kB) at
|
|
|
|
* time by setting the transfer length to 0. Setting tl to
|
|
|
|
* 0x10000 will overflow below and make .data contain 0.
|
|
|
|
* Otherwise, 0xff00 is the transfer maximum.
|
|
|
|
*/
|
|
|
|
if (length >= 0x10000)
|
|
|
|
tl = 0x10000;
|
|
|
|
else
|
|
|
|
tl = min(length, xfer_max_sz);
|
|
|
|
|
|
|
|
dp->cmd.data |=
|
MX28: SPI: Fix the DMA chaining
It turns out that in order for the SPI DMA to properly support
continuous transfers longer than 65280 bytes, there are some very
important parts that were left out from the documentation.
Firstly, the XFER_SIZE register is not written with the whole length
of a transfer, but is written by each and every chained descriptor
with the length of the descriptors data buffer.
Next, unlike the demo code supplied by FSL, which only writes one PIO
word per descriptor, this does not apply if the descriptors are chained,
since the XFER_SIZE register must be written. Therefore, it is essential
to use four PIO words, CTRL0, CMD0, CMD1, XFER_SIZE. CMD0 and CMD1 are
written with zero, since they don't apply. The DMA programs the PIO words
in an incrementing order, so four PIO words.
Finally, unlike the demo code supplied by FSL, the SSP_CTRL0_IGNORE_CRC
must not be set during the whole transfer, but it must be set only on the
last descriptor in the chain.
Signed-off-by: Marek Vasut <marex@denx.de>
Cc: Fabio Estevam <festevam@gmail.com>
Cc: Otavio Salvador <otavio@ossystems.com.br>
Cc: Stefano Babic <sbabic@denx.de>
13 years ago
|
|
|
((tl & 0xffff) << MXS_DMA_DESC_BYTES_OFFSET) |
|
|
|
|
(mxs_spi_pio_words << MXS_DMA_DESC_PIO_WORDS_OFFSET) |
|
|
|
|
MXS_DMA_DESC_HALT_ON_TERMINATE |
|
|
|
|
MXS_DMA_DESC_TERMINATE_FLUSH;
|
|
|
|
|
|
|
|
data += tl;
|
|
|
|
length -= tl;
|
|
|
|
|
MX28: SPI: Fix the DMA chaining
It turns out that in order for the SPI DMA to properly support
continuous transfers longer than 65280 bytes, there are some very
important parts that were left out from the documentation.
Firstly, the XFER_SIZE register is not written with the whole length
of a transfer, but is written by each and every chained descriptor
with the length of the descriptors data buffer.
Next, unlike the demo code supplied by FSL, which only writes one PIO
word per descriptor, this does not apply if the descriptors are chained,
since the XFER_SIZE register must be written. Therefore, it is essential
to use four PIO words, CTRL0, CMD0, CMD1, XFER_SIZE. CMD0 and CMD1 are
written with zero, since they don't apply. The DMA programs the PIO words
in an incrementing order, so four PIO words.
Finally, unlike the demo code supplied by FSL, the SSP_CTRL0_IGNORE_CRC
must not be set during the whole transfer, but it must be set only on the
last descriptor in the chain.
Signed-off-by: Marek Vasut <marex@denx.de>
Cc: Fabio Estevam <festevam@gmail.com>
Cc: Otavio Salvador <otavio@ossystems.com.br>
Cc: Stefano Babic <sbabic@denx.de>
13 years ago
|
|
|
if (!length) {
|
|
|
|
dp->cmd.data |= MXS_DMA_DESC_IRQ | MXS_DMA_DESC_DEC_SEM;
|
|
|
|
|
|
|
|
if (flags & SPI_XFER_END) {
|
|
|
|
ctrl0 &= ~SSP_CTRL0_LOCK_CS;
|
|
|
|
ctrl0 |= SSP_CTRL0_IGNORE_CRC;
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Write CTRL0, CMD0, CMD1 and XFER_SIZE registers in
|
|
|
|
* case of MX28, write only CTRL0 in case of MX23 due
|
|
|
|
* to the difference in register layout. It is utterly
|
MX28: SPI: Fix the DMA chaining
It turns out that in order for the SPI DMA to properly support
continuous transfers longer than 65280 bytes, there are some very
important parts that were left out from the documentation.
Firstly, the XFER_SIZE register is not written with the whole length
of a transfer, but is written by each and every chained descriptor
with the length of the descriptors data buffer.
Next, unlike the demo code supplied by FSL, which only writes one PIO
word per descriptor, this does not apply if the descriptors are chained,
since the XFER_SIZE register must be written. Therefore, it is essential
to use four PIO words, CTRL0, CMD0, CMD1, XFER_SIZE. CMD0 and CMD1 are
written with zero, since they don't apply. The DMA programs the PIO words
in an incrementing order, so four PIO words.
Finally, unlike the demo code supplied by FSL, the SSP_CTRL0_IGNORE_CRC
must not be set during the whole transfer, but it must be set only on the
last descriptor in the chain.
Signed-off-by: Marek Vasut <marex@denx.de>
Cc: Fabio Estevam <festevam@gmail.com>
Cc: Otavio Salvador <otavio@ossystems.com.br>
Cc: Stefano Babic <sbabic@denx.de>
13 years ago
|
|
|
* essential that the XFER_SIZE register is written on
|
|
|
|
* a per-descriptor basis with the same size as is the
|
|
|
|
* descriptor!
|
|
|
|
*/
|
|
|
|
dp->cmd.pio_words[0] = ctrl0;
|
|
|
|
#ifdef CONFIG_MX28
|
MX28: SPI: Fix the DMA chaining
It turns out that in order for the SPI DMA to properly support
continuous transfers longer than 65280 bytes, there are some very
important parts that were left out from the documentation.
Firstly, the XFER_SIZE register is not written with the whole length
of a transfer, but is written by each and every chained descriptor
with the length of the descriptors data buffer.
Next, unlike the demo code supplied by FSL, which only writes one PIO
word per descriptor, this does not apply if the descriptors are chained,
since the XFER_SIZE register must be written. Therefore, it is essential
to use four PIO words, CTRL0, CMD0, CMD1, XFER_SIZE. CMD0 and CMD1 are
written with zero, since they don't apply. The DMA programs the PIO words
in an incrementing order, so four PIO words.
Finally, unlike the demo code supplied by FSL, the SSP_CTRL0_IGNORE_CRC
must not be set during the whole transfer, but it must be set only on the
last descriptor in the chain.
Signed-off-by: Marek Vasut <marex@denx.de>
Cc: Fabio Estevam <festevam@gmail.com>
Cc: Otavio Salvador <otavio@ossystems.com.br>
Cc: Stefano Babic <sbabic@denx.de>
13 years ago
|
|
|
dp->cmd.pio_words[1] = 0;
|
|
|
|
dp->cmd.pio_words[2] = 0;
|
|
|
|
dp->cmd.pio_words[3] = tl;
|
|
|
|
#endif
|
MX28: SPI: Fix the DMA chaining
It turns out that in order for the SPI DMA to properly support
continuous transfers longer than 65280 bytes, there are some very
important parts that were left out from the documentation.
Firstly, the XFER_SIZE register is not written with the whole length
of a transfer, but is written by each and every chained descriptor
with the length of the descriptors data buffer.
Next, unlike the demo code supplied by FSL, which only writes one PIO
word per descriptor, this does not apply if the descriptors are chained,
since the XFER_SIZE register must be written. Therefore, it is essential
to use four PIO words, CTRL0, CMD0, CMD1, XFER_SIZE. CMD0 and CMD1 are
written with zero, since they don't apply. The DMA programs the PIO words
in an incrementing order, so four PIO words.
Finally, unlike the demo code supplied by FSL, the SSP_CTRL0_IGNORE_CRC
must not be set during the whole transfer, but it must be set only on the
last descriptor in the chain.
Signed-off-by: Marek Vasut <marex@denx.de>
Cc: Fabio Estevam <festevam@gmail.com>
Cc: Otavio Salvador <otavio@ossystems.com.br>
Cc: Stefano Babic <sbabic@denx.de>
13 years ago
|
|
|
|
|
|
|
mxs_dma_desc_append(dmach, dp);
|
|
|
|
|
|
|
|
dp++;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (mxs_dma_go(dmach))
|
MX28: SPI: Fix the DMA chaining
It turns out that in order for the SPI DMA to properly support
continuous transfers longer than 65280 bytes, there are some very
important parts that were left out from the documentation.
Firstly, the XFER_SIZE register is not written with the whole length
of a transfer, but is written by each and every chained descriptor
with the length of the descriptors data buffer.
Next, unlike the demo code supplied by FSL, which only writes one PIO
word per descriptor, this does not apply if the descriptors are chained,
since the XFER_SIZE register must be written. Therefore, it is essential
to use four PIO words, CTRL0, CMD0, CMD1, XFER_SIZE. CMD0 and CMD1 are
written with zero, since they don't apply. The DMA programs the PIO words
in an incrementing order, so four PIO words.
Finally, unlike the demo code supplied by FSL, the SSP_CTRL0_IGNORE_CRC
must not be set during the whole transfer, but it must be set only on the
last descriptor in the chain.
Signed-off-by: Marek Vasut <marex@denx.de>
Cc: Fabio Estevam <festevam@gmail.com>
Cc: Otavio Salvador <otavio@ossystems.com.br>
Cc: Stefano Babic <sbabic@denx.de>
13 years ago
|
|
|
ret = -EINVAL;
|
|
|
|
|
|
|
|
/* The data arrived into DRAM, invalidate cache over them */
|
|
|
|
if (!write)
|
|
|
|
invalidate_dcache_range(dstart, dstart + cache_data_count);
|
|
|
|
|
MX28: SPI: Fix the DMA chaining
It turns out that in order for the SPI DMA to properly support
continuous transfers longer than 65280 bytes, there are some very
important parts that were left out from the documentation.
Firstly, the XFER_SIZE register is not written with the whole length
of a transfer, but is written by each and every chained descriptor
with the length of the descriptors data buffer.
Next, unlike the demo code supplied by FSL, which only writes one PIO
word per descriptor, this does not apply if the descriptors are chained,
since the XFER_SIZE register must be written. Therefore, it is essential
to use four PIO words, CTRL0, CMD0, CMD1, XFER_SIZE. CMD0 and CMD1 are
written with zero, since they don't apply. The DMA programs the PIO words
in an incrementing order, so four PIO words.
Finally, unlike the demo code supplied by FSL, the SSP_CTRL0_IGNORE_CRC
must not be set during the whole transfer, but it must be set only on the
last descriptor in the chain.
Signed-off-by: Marek Vasut <marex@denx.de>
Cc: Fabio Estevam <festevam@gmail.com>
Cc: Otavio Salvador <otavio@ossystems.com.br>
Cc: Stefano Babic <sbabic@denx.de>
13 years ago
|
|
|
return ret;
|
|
|
|
}
|
|
|
|
|
|
|
|
int spi_xfer(struct spi_slave *slave, unsigned int bitlen,
|
|
|
|
const void *dout, void *din, unsigned long flags)
|
|
|
|
{
|
|
|
|
struct mxs_spi_slave *mxs_slave = to_mxs_slave(slave);
|
|
|
|
struct mxs_ssp_regs *ssp_regs = mxs_slave->regs;
|
|
|
|
int len = bitlen / 8;
|
|
|
|
char dummy;
|
|
|
|
int write = 0;
|
|
|
|
char *data = NULL;
|
|
|
|
int dma = 1;
|
|
|
|
|
|
|
|
if (bitlen == 0) {
|
|
|
|
if (flags & SPI_XFER_END) {
|
|
|
|
din = (void *)&dummy;
|
|
|
|
len = 1;
|
|
|
|
} else
|
|
|
|
return 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
/* Half-duplex only */
|
|
|
|
if (din && dout)
|
|
|
|
return -EINVAL;
|
|
|
|
/* No data */
|
|
|
|
if (!din && !dout)
|
|
|
|
return 0;
|
|
|
|
|
|
|
|
if (dout) {
|
|
|
|
data = (char *)dout;
|
|
|
|
write = 1;
|
|
|
|
} else if (din) {
|
|
|
|
data = (char *)din;
|
|
|
|
write = 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
/*
|
|
|
|
* Check for alignment, if the buffer is aligned, do DMA transfer,
|
|
|
|
* PIO otherwise. This is a temporary workaround until proper bounce
|
|
|
|
* buffer is in place.
|
|
|
|
*/
|
|
|
|
if (dma) {
|
|
|
|
if (((uint32_t)data) & (ARCH_DMA_MINALIGN - 1))
|
|
|
|
dma = 0;
|
|
|
|
if (((uint32_t)len) & (ARCH_DMA_MINALIGN - 1))
|
|
|
|
dma = 0;
|
|
|
|
}
|
|
|
|
|
|
|
|
if (!dma || (len < MXSSSP_SMALL_TRANSFER)) {
|
|
|
|
writel(SSP_CTRL1_DMA_ENABLE, &ssp_regs->hw_ssp_ctrl1_clr);
|
|
|
|
return mxs_spi_xfer_pio(mxs_slave, data, len, write, flags);
|
|
|
|
} else {
|
|
|
|
writel(SSP_CTRL1_DMA_ENABLE, &ssp_regs->hw_ssp_ctrl1_set);
|
|
|
|
return mxs_spi_xfer_dma(mxs_slave, data, len, write, flags);
|
|
|
|
}
|
|
|
|
}
|