linux/drivers/spi/spi-fsl-cpm.c
Christophe Leroy fc96ec826b
spi: fsl-cpm: Use 16 bit mode for large transfers with even size
On CPM, the RISC core is a lot more efficiant when doing transfers
in 16-bits chunks than in 8-bits chunks, but unfortunately the
words need to be byte swapped as seen in a previous commit.

So, for large tranfers with an even size, allocate a temporary tx
buffer and byte-swap data before and after transfer.

This change allows setting higher speed for transfer. For instance
on an MPC 8xx (CPM1 comms RISC processor), the documentation tells
that transfer in byte mode at 1 kbit/s uses 0.200% of CPM load
at 25 MHz while a word transfer at the same speed uses 0.032%
of CPM load. This means the speed can be 6 times higher in
word mode for the same CPM load.

For the time being, only do it on CPM1 as there must be a
trade-off between the CPM load reduction and the CPU load required
to byte swap the data.

Signed-off-by: Christophe Leroy <christophe.leroy@csgroup.eu>
Link: https://lore.kernel.org/r/f2e981f20f92dd28983c3949702a09248c23845c.1680371809.git.christophe.leroy@csgroup.eu
Signed-off-by: Mark Brown <broonie@kernel.org>
2023-04-05 12:33:34 +01:00

425 lines
10 KiB
C

// SPDX-License-Identifier: GPL-2.0-or-later
/*
* Freescale SPI controller driver cpm functions.
*
* Maintainer: Kumar Gala
*
* Copyright (C) 2006 Polycom, Inc.
* Copyright 2010 Freescale Semiconductor, Inc.
*
* CPM SPI and QE buffer descriptors mode support:
* Copyright (c) 2009 MontaVista Software, Inc.
* Author: Anton Vorontsov <avorontsov@ru.mvista.com>
*/
#include <asm/cpm.h>
#include <soc/fsl/qe/qe.h>
#include <linux/dma-mapping.h>
#include <linux/fsl_devices.h>
#include <linux/kernel.h>
#include <linux/module.h>
#include <linux/of_address.h>
#include <linux/spi/spi.h>
#include <linux/types.h>
#include <linux/platform_device.h>
#include <linux/byteorder/generic.h>
#include "spi-fsl-cpm.h"
#include "spi-fsl-lib.h"
#include "spi-fsl-spi.h"
/* CPM1 and CPM2 are mutually exclusive. */
#ifdef CONFIG_CPM1
#include <asm/cpm1.h>
#define CPM_SPI_CMD mk_cr_cmd(CPM_CR_CH_SPI, 0)
#else
#include <asm/cpm2.h>
#define CPM_SPI_CMD mk_cr_cmd(CPM_CR_SPI_PAGE, CPM_CR_SPI_SBLOCK, 0, 0)
#endif
#define SPIE_TXB 0x00000200 /* Last char is written to tx fifo */
#define SPIE_RXB 0x00000100 /* Last char is written to rx buf */
/* SPCOM register values */
#define SPCOM_STR (1 << 23) /* Start transmit */
#define SPI_PRAM_SIZE 0x100
#define SPI_MRBLR ((unsigned int)PAGE_SIZE)
static void *fsl_dummy_rx;
static DEFINE_MUTEX(fsl_dummy_rx_lock);
static int fsl_dummy_rx_refcnt;
void fsl_spi_cpm_reinit_txrx(struct mpc8xxx_spi *mspi)
{
if (mspi->flags & SPI_QE) {
qe_issue_cmd(QE_INIT_TX_RX, mspi->subblock,
QE_CR_PROTOCOL_UNSPECIFIED, 0);
} else {
if (mspi->flags & SPI_CPM1) {
out_be32(&mspi->pram->rstate, 0);
out_be16(&mspi->pram->rbptr,
in_be16(&mspi->pram->rbase));
out_be32(&mspi->pram->tstate, 0);
out_be16(&mspi->pram->tbptr,
in_be16(&mspi->pram->tbase));
} else {
cpm_command(CPM_SPI_CMD, CPM_CR_INIT_TRX);
}
}
}
EXPORT_SYMBOL_GPL(fsl_spi_cpm_reinit_txrx);
static void fsl_spi_cpm_bufs_start(struct mpc8xxx_spi *mspi)
{
struct cpm_buf_desc __iomem *tx_bd = mspi->tx_bd;
struct cpm_buf_desc __iomem *rx_bd = mspi->rx_bd;
unsigned int xfer_len = min(mspi->count, SPI_MRBLR);
unsigned int xfer_ofs;
struct fsl_spi_reg *reg_base = mspi->reg_base;
xfer_ofs = mspi->xfer_in_progress->len - mspi->count;
if (mspi->rx_dma == mspi->dma_dummy_rx)
out_be32(&rx_bd->cbd_bufaddr, mspi->rx_dma);
else
out_be32(&rx_bd->cbd_bufaddr, mspi->rx_dma + xfer_ofs);
out_be16(&rx_bd->cbd_datlen, 0);
out_be16(&rx_bd->cbd_sc, BD_SC_EMPTY | BD_SC_INTRPT | BD_SC_WRAP);
if (mspi->tx_dma == mspi->dma_dummy_tx)
out_be32(&tx_bd->cbd_bufaddr, mspi->tx_dma);
else
out_be32(&tx_bd->cbd_bufaddr, mspi->tx_dma + xfer_ofs);
out_be16(&tx_bd->cbd_datlen, xfer_len);
out_be16(&tx_bd->cbd_sc, BD_SC_READY | BD_SC_INTRPT | BD_SC_WRAP |
BD_SC_LAST);
/* start transfer */
mpc8xxx_spi_write_reg(&reg_base->command, SPCOM_STR);
}
int fsl_spi_cpm_bufs(struct mpc8xxx_spi *mspi,
struct spi_transfer *t, bool is_dma_mapped)
{
struct device *dev = mspi->dev;
struct fsl_spi_reg *reg_base = mspi->reg_base;
if (is_dma_mapped) {
mspi->map_tx_dma = 0;
mspi->map_rx_dma = 0;
} else {
mspi->map_tx_dma = 1;
mspi->map_rx_dma = 1;
}
if (!t->tx_buf) {
mspi->tx_dma = mspi->dma_dummy_tx;
mspi->map_tx_dma = 0;
}
if (!t->rx_buf) {
mspi->rx_dma = mspi->dma_dummy_rx;
mspi->map_rx_dma = 0;
}
if (t->bits_per_word == 16 && t->tx_buf) {
const u16 *src = t->tx_buf;
u16 *dst;
int i;
dst = kmalloc(t->len, GFP_KERNEL);
if (!dst)
return -ENOMEM;
for (i = 0; i < t->len >> 1; i++)
dst[i] = cpu_to_le16p(src + i);
mspi->tx = dst;
mspi->map_tx_dma = 1;
}
if (mspi->map_tx_dma) {
void *nonconst_tx = (void *)mspi->tx; /* shut up gcc */
mspi->tx_dma = dma_map_single(dev, nonconst_tx, t->len,
DMA_TO_DEVICE);
if (dma_mapping_error(dev, mspi->tx_dma)) {
dev_err(dev, "unable to map tx dma\n");
return -ENOMEM;
}
} else if (t->tx_buf) {
mspi->tx_dma = t->tx_dma;
}
if (mspi->map_rx_dma) {
mspi->rx_dma = dma_map_single(dev, mspi->rx, t->len,
DMA_FROM_DEVICE);
if (dma_mapping_error(dev, mspi->rx_dma)) {
dev_err(dev, "unable to map rx dma\n");
goto err_rx_dma;
}
} else if (t->rx_buf) {
mspi->rx_dma = t->rx_dma;
}
/* enable rx ints */
mpc8xxx_spi_write_reg(&reg_base->mask, SPIE_RXB);
mspi->xfer_in_progress = t;
mspi->count = t->len;
/* start CPM transfers */
fsl_spi_cpm_bufs_start(mspi);
return 0;
err_rx_dma:
if (mspi->map_tx_dma)
dma_unmap_single(dev, mspi->tx_dma, t->len, DMA_TO_DEVICE);
return -ENOMEM;
}
EXPORT_SYMBOL_GPL(fsl_spi_cpm_bufs);
void fsl_spi_cpm_bufs_complete(struct mpc8xxx_spi *mspi)
{
struct device *dev = mspi->dev;
struct spi_transfer *t = mspi->xfer_in_progress;
if (mspi->map_tx_dma)
dma_unmap_single(dev, mspi->tx_dma, t->len, DMA_TO_DEVICE);
if (mspi->map_rx_dma)
dma_unmap_single(dev, mspi->rx_dma, t->len, DMA_FROM_DEVICE);
mspi->xfer_in_progress = NULL;
if (t->bits_per_word == 16 && t->rx_buf) {
int i;
for (i = 0; i < t->len; i += 2)
le16_to_cpus(t->rx_buf + i);
}
}
EXPORT_SYMBOL_GPL(fsl_spi_cpm_bufs_complete);
void fsl_spi_cpm_irq(struct mpc8xxx_spi *mspi, u32 events)
{
u16 len;
struct fsl_spi_reg *reg_base = mspi->reg_base;
dev_dbg(mspi->dev, "%s: bd datlen %d, count %d\n", __func__,
in_be16(&mspi->rx_bd->cbd_datlen), mspi->count);
len = in_be16(&mspi->rx_bd->cbd_datlen);
if (len > mspi->count) {
WARN_ON(1);
len = mspi->count;
}
/* Clear the events */
mpc8xxx_spi_write_reg(&reg_base->event, events);
mspi->count -= len;
if (mspi->count)
fsl_spi_cpm_bufs_start(mspi);
else
complete(&mspi->done);
}
EXPORT_SYMBOL_GPL(fsl_spi_cpm_irq);
static void *fsl_spi_alloc_dummy_rx(void)
{
mutex_lock(&fsl_dummy_rx_lock);
if (!fsl_dummy_rx)
fsl_dummy_rx = kmalloc(SPI_MRBLR, GFP_KERNEL);
if (fsl_dummy_rx)
fsl_dummy_rx_refcnt++;
mutex_unlock(&fsl_dummy_rx_lock);
return fsl_dummy_rx;
}
static void fsl_spi_free_dummy_rx(void)
{
mutex_lock(&fsl_dummy_rx_lock);
switch (fsl_dummy_rx_refcnt) {
case 0:
WARN_ON(1);
break;
case 1:
kfree(fsl_dummy_rx);
fsl_dummy_rx = NULL;
fallthrough;
default:
fsl_dummy_rx_refcnt--;
break;
}
mutex_unlock(&fsl_dummy_rx_lock);
}
static unsigned long fsl_spi_cpm_get_pram(struct mpc8xxx_spi *mspi)
{
struct device *dev = mspi->dev;
struct device_node *np = dev->of_node;
const u32 *iprop;
int size;
void __iomem *spi_base;
unsigned long pram_ofs = -ENOMEM;
/* Can't use of_address_to_resource(), QE muram isn't at 0. */
iprop = of_get_property(np, "reg", &size);
/* QE with a fixed pram location? */
if (mspi->flags & SPI_QE && iprop && size == sizeof(*iprop) * 4)
return cpm_muram_alloc_fixed(iprop[2], SPI_PRAM_SIZE);
/* QE but with a dynamic pram location? */
if (mspi->flags & SPI_QE) {
pram_ofs = cpm_muram_alloc(SPI_PRAM_SIZE, 64);
qe_issue_cmd(QE_ASSIGN_PAGE_TO_DEVICE, mspi->subblock,
QE_CR_PROTOCOL_UNSPECIFIED, pram_ofs);
return pram_ofs;
}
spi_base = of_iomap(np, 1);
if (spi_base == NULL)
return -EINVAL;
if (mspi->flags & SPI_CPM2) {
pram_ofs = cpm_muram_alloc(SPI_PRAM_SIZE, 64);
out_be16(spi_base, pram_ofs);
}
iounmap(spi_base);
return pram_ofs;
}
int fsl_spi_cpm_init(struct mpc8xxx_spi *mspi)
{
struct device *dev = mspi->dev;
struct device_node *np = dev->of_node;
const u32 *iprop;
int size;
unsigned long bds_ofs;
if (!(mspi->flags & SPI_CPM_MODE))
return 0;
if (!fsl_spi_alloc_dummy_rx())
return -ENOMEM;
if (mspi->flags & SPI_QE) {
iprop = of_get_property(np, "cell-index", &size);
if (iprop && size == sizeof(*iprop))
mspi->subblock = *iprop;
switch (mspi->subblock) {
default:
dev_warn(dev, "cell-index unspecified, assuming SPI1\n");
fallthrough;
case 0:
mspi->subblock = QE_CR_SUBBLOCK_SPI1;
break;
case 1:
mspi->subblock = QE_CR_SUBBLOCK_SPI2;
break;
}
}
if (mspi->flags & SPI_CPM1) {
void *pram;
pram = devm_platform_ioremap_resource(to_platform_device(dev),
1);
if (IS_ERR(pram))
mspi->pram = NULL;
else
mspi->pram = pram;
} else {
unsigned long pram_ofs = fsl_spi_cpm_get_pram(mspi);
if (IS_ERR_VALUE(pram_ofs))
mspi->pram = NULL;
else
mspi->pram = cpm_muram_addr(pram_ofs);
}
if (mspi->pram == NULL) {
dev_err(dev, "can't allocate spi parameter ram\n");
goto err_pram;
}
bds_ofs = cpm_muram_alloc(sizeof(*mspi->tx_bd) +
sizeof(*mspi->rx_bd), 8);
if (IS_ERR_VALUE(bds_ofs)) {
dev_err(dev, "can't allocate bds\n");
goto err_bds;
}
mspi->dma_dummy_tx = dma_map_single(dev, ZERO_PAGE(0), PAGE_SIZE,
DMA_TO_DEVICE);
if (dma_mapping_error(dev, mspi->dma_dummy_tx)) {
dev_err(dev, "unable to map dummy tx buffer\n");
goto err_dummy_tx;
}
mspi->dma_dummy_rx = dma_map_single(dev, fsl_dummy_rx, SPI_MRBLR,
DMA_FROM_DEVICE);
if (dma_mapping_error(dev, mspi->dma_dummy_rx)) {
dev_err(dev, "unable to map dummy rx buffer\n");
goto err_dummy_rx;
}
mspi->tx_bd = cpm_muram_addr(bds_ofs);
mspi->rx_bd = cpm_muram_addr(bds_ofs + sizeof(*mspi->tx_bd));
/* Initialize parameter ram. */
out_be16(&mspi->pram->tbase, cpm_muram_offset(mspi->tx_bd));
out_be16(&mspi->pram->rbase, cpm_muram_offset(mspi->rx_bd));
out_8(&mspi->pram->tfcr, CPMFCR_EB | CPMFCR_GBL);
out_8(&mspi->pram->rfcr, CPMFCR_EB | CPMFCR_GBL);
out_be16(&mspi->pram->mrblr, SPI_MRBLR);
out_be32(&mspi->pram->rstate, 0);
out_be32(&mspi->pram->rdp, 0);
out_be16(&mspi->pram->rbptr, 0);
out_be16(&mspi->pram->rbc, 0);
out_be32(&mspi->pram->rxtmp, 0);
out_be32(&mspi->pram->tstate, 0);
out_be32(&mspi->pram->tdp, 0);
out_be16(&mspi->pram->tbptr, 0);
out_be16(&mspi->pram->tbc, 0);
out_be32(&mspi->pram->txtmp, 0);
return 0;
err_dummy_rx:
dma_unmap_single(dev, mspi->dma_dummy_tx, PAGE_SIZE, DMA_TO_DEVICE);
err_dummy_tx:
cpm_muram_free(bds_ofs);
err_bds:
if (!(mspi->flags & SPI_CPM1))
cpm_muram_free(cpm_muram_offset(mspi->pram));
err_pram:
fsl_spi_free_dummy_rx();
return -ENOMEM;
}
EXPORT_SYMBOL_GPL(fsl_spi_cpm_init);
void fsl_spi_cpm_free(struct mpc8xxx_spi *mspi)
{
struct device *dev = mspi->dev;
if (!(mspi->flags & SPI_CPM_MODE))
return;
dma_unmap_single(dev, mspi->dma_dummy_rx, SPI_MRBLR, DMA_FROM_DEVICE);
dma_unmap_single(dev, mspi->dma_dummy_tx, PAGE_SIZE, DMA_TO_DEVICE);
cpm_muram_free(cpm_muram_offset(mspi->tx_bd));
if (!(mspi->flags & SPI_CPM1))
cpm_muram_free(cpm_muram_offset(mspi->pram));
fsl_spi_free_dummy_rx();
}
EXPORT_SYMBOL_GPL(fsl_spi_cpm_free);
MODULE_LICENSE("GPL");