[U-Boot] [PATCH v2] spi: Add support SH Queued SPI driver

Jagan Teki jagannadh.teki at gmail.com
Fri Aug 9 10:45:05 CEST 2013


Hi,

On 08-08-2013 15:02, Nobuhiro Iwamatsu wrote:
> This patch adds a driver for Renesas SoC's Queued SPI bus.
> This supports with 8 bits per transfer to use with SPI flash.
>
> Signed-off-by: Kouei Abe <kouei.abe.cp at renesas.com>
> Signed-off-by: Nobuhiro Iwamatsu <nobuhiro.iwamatsu.yj at renesas.com>
> ---
>   v2: Change "SH QSPI" to "SH QSPI (Queued SPI)".
>       Remove magic number.
>
>   drivers/spi/Makefile  |   1 +
>   drivers/spi/sh_qspi.c | 232 ++++++++++++++++++++++++++++++++++++++++++++++++++
>   drivers/spi/sh_qspi.h |  52 +++++++++++

Please move the structure decelerations on to sh_qspi.c
FYI: I am trying to prepare the spi driver code to more readable

<< header file inclusion >>
<< Register bit masks >>

<< MISC macro definitions >>

<< controller reg structure >>

<< controller private slave structure >>

<< inline func defination >>

<< spi_xfer_sub() >>

<< spi_setup_slave_sub >>

<< spi_cs_is_valid >>

<< spi_cs_activate >>

<< spi_cs_deactivate >>

<< spi_init >>

<< spi_setup_slave >>
{
	spi_setup_slave_sub()
}

<< spi_free_slave >>

<< spi_claim_bus >>

<< spi_release_bus >>

<< spi_xfer code >>
{
   spi_xfer_sub()
}

I am just trying to do the above format atleast from the drivers which 
are pushing now onwards.

Please see the reff driver
http://git.denx.de/?p=u-boot/u-boot-spi.git;a=commitdiff;h=1465d055f9d7a81edacf30c9d20a1b51dfcbfa8d

Let me know your views.

>   3 files changed, 285 insertions(+)
>   create mode 100644 drivers/spi/sh_qspi.c
>   create mode 100644 drivers/spi/sh_qspi.h
>
> diff --git a/drivers/spi/Makefile b/drivers/spi/Makefile
> index 019132e..f71c089 100644
> --- a/drivers/spi/Makefile
> +++ b/drivers/spi/Makefile
> @@ -33,6 +33,7 @@ COBJS-$(CONFIG_OC_TINY_SPI) += oc_tiny_spi.o
>   COBJS-$(CONFIG_OMAP3_SPI) += omap3_spi.o
>   COBJS-$(CONFIG_SOFT_SPI) += soft_spi.o
>   COBJS-$(CONFIG_SH_SPI) += sh_spi.o
> +COBJS-$(CONFIG_SH_QSPI) += sh_qspi.o
>   COBJS-$(CONFIG_FSL_ESPI) += fsl_espi.o
>   COBJS-$(CONFIG_FDT_SPI) += fdt_spi.o
>   COBJS-$(CONFIG_TEGRA20_SFLASH) += tegra20_sflash.o
> diff --git a/drivers/spi/sh_qspi.c b/drivers/spi/sh_qspi.c
> new file mode 100644
> index 0000000..25ce60a
> --- /dev/null
> +++ b/drivers/spi/sh_qspi.c
> @@ -0,0 +1,232 @@
> +/*
> + * drivers/spi/sh_qspi.c
Not required, i guess.

> + * SH QSPI (Queued SPI) driver
> + *
> + * Copyright (C) 2013 Renesas Electronics Corporation
> + * Copyright (C) 2013 Nobuhiro Iwamatsu <nobuhiro.iwamatsu.yj at renesas.com>
> + *
> + * SPDX-License-Identifier:	GPL-2.0
> + */
> +
> +#include <common.h>
> +#include <malloc.h>
> +#include <spi.h>
> +#include <asm/io.h>
> +#include "sh_qspi.h"
> +
> +static void sh_qspi_writeb(unsigned char data, unsigned char *reg)
> +{
> +	writeb(data, reg);
> +}
> +
> +static void sh_qspi_writew(unsigned short data, unsigned short *reg)
> +{
> +	writew(data, reg);
> +}
> +
> +static void sh_qspi_writel(unsigned long data, unsigned long *reg)
> +{
> +	writel(data, reg);
> +}
> +
> +static unsigned char sh_qspi_readb(unsigned char *reg)
> +{
> +	return readb(reg);
> +}
> +
> +static unsigned short sh_qspi_readw(unsigned short *reg)
> +{
> +	return readw(reg);
> +}
> +
> +void spi_init(void)
> +{
> +}

Can you use read* and write* directly instead of sh_qspi_write* 
sh_qspi_read*

Do you have any specific reason for using these?

> +
> +/* SPCR */
> +#define SPCR_MSTR 0x08
> +#define SPCR_SPE 0x40
> +/* SPSR */
> +#define SPSR_SPRFF 0x80
> +#define SPSR_SPTEF 0x20
> +/* SPPCR */
> +#define SPPCR_IO3FV 0x04
> +#define SPPCR_IO2FV 0x02
> +#define SPPCR_IO1FV 0x01
> +/* SPBDCR */
> +#define SPBDCR_RXBC0 (1 << 0)
> +/* SPCMD */
> +#define SPCMD_SCKDEN (1 << 15)
> +#define SPCMD_SLNDEN (1 << 14)
> +#define SPCMD_SPNDEN (1 << 13)
> +#define SPCMD_SSLKP (1 << 7)
> +#define SPCMD_BRDV0 (1 << 2)
> +#define SPCMD_INIT1 \
> +	(SPCMD_SCKDEN|SPCMD_SLNDEN|SPCMD_SPNDEN|SPCMD_SSLKP|SPCMD_BRDV0)
> +#define SPCMD_INIT2 (SPCMD_SPNDEN|SPCMD_SSLKP|SPCMD_BRDV0)
> +/* SPBFCR */
> +#define SPBFCR_TXRST (1 << 7)
> +#define SPBFCR_RXRST (1 << 6)
> +
> +struct spi_slave *spi_setup_slave(unsigned int bus, unsigned int cs,
> +		unsigned int max_hz, unsigned int mode)
> +{
> +	struct sh_qspi *ss;
> +
> +	if (!spi_cs_is_valid(bus, cs))
> +		return NULL;
> +
> +	ss = spi_alloc_slave(struct sh_qspi, bus, cs);
> +	if (!ss)
> +		return NULL;
> +
> +	ss->regs = (struct sh_qspi_regs *)CONFIG_SH_QSPI_BASE;
> +
> +	/* QSPI initialize */
> +	sh_qspi_writeb(SPCR_MSTR, &ss->regs->spcr);
> +	sh_qspi_writeb(0x00, &ss->regs->sslp);
> +	sh_qspi_writeb(SPPCR_IO3FV|SPPCR_IO2FV, &ss->regs->sppcr);
> +
> +	/* Set bit rate. See 58.3.8 Quad Serial Peripheral Interface */
> +	sh_qspi_writeb(0x01, &ss->regs->spbr);
> +
> +	sh_qspi_writeb(0x00, &ss->regs->spdcr);
> +	sh_qspi_writeb(0x00, &ss->regs->spckd);
> +	sh_qspi_writeb(0x00, &ss->regs->sslnd);
> +	sh_qspi_writeb(0x00, &ss->regs->spnd);
> +	sh_qspi_writew(SPCMD_INIT1, &ss->regs->spcmd0);
> +	sh_qspi_writew(SPCMD_INIT2, &ss->regs->spcmd0);
> +	sh_qspi_writeb(SPBFCR_TXRST|SPBFCR_RXRST, &ss->regs->spbfcr);
> +	sh_qspi_writeb(0x00, &ss->regs->spbfcr);
> +	sh_qspi_writeb(0x00, &ss->regs->spscr);
> +	sh_qspi_writeb(SPCR_SPE|SPCR_MSTR, &ss->regs->spcr);
> +
> +	return &ss->slave;
> +}
> +
> +void spi_free_slave(struct spi_slave *slave)
> +{
> +	struct sh_qspi *spi = to_sh_qspi(slave);
> +
> +	free(spi);
> +}
> +
> +int spi_claim_bus(struct spi_slave *slave)
> +{
> +	return 0;
> +}
> +
> +void spi_release_bus(struct spi_slave *slave)
> +{
> +}
> +
> +static int sh_qspi_xfer(struct sh_qspi *ss, unsigned char *tdata,
> +			unsigned char *rdata, unsigned long flags)
> +{
> +	while (!(sh_qspi_readb(&ss->regs->spsr) & SPSR_SPTEF)) {
> +		if (ctrlc())
> +			return 1;
> +		udelay(10);
> +	}
> +
> +	sh_qspi_writeb(*tdata, (unsigned char *)(&ss->regs->spdr));
> +
> +	while ((sh_qspi_readw(&ss->regs->spbdcr) != SPBDCR_RXBC0)) {
> +		int i = 100;
> +
> +		if (ctrlc())
> +			return 1;
> +		while (i--)
> +			;
> +	}
> +
> +	while (!(sh_qspi_readb(&ss->regs->spsr) & SPSR_SPRFF)) {
> +		if (ctrlc())
> +			return 1;
> +		udelay(10);
> +	}
> +
> +	*rdata = sh_qspi_readb((unsigned char *)(&ss->regs->spdr));
> +
> +	return 0;
> +}
> +
> +int spi_xfer(struct spi_slave *slave, unsigned int bitlen, const void *dout,
> +	     void *din, unsigned long flags)
> +{
> +	struct sh_qspi *ss = to_sh_qspi(slave);
> +	unsigned int nbyte;
> +	int ret = 0;
> +	unsigned char *tdata, *rdata, dtdata = 0, drdata;
> +
> +	if (dout == NULL && din == NULL) {
> +		if (flags & SPI_XFER_END)
> +			sh_qspi_writeb(SPCR_MSTR, &ss->regs->spcr);
> +		return 0;
> +	}
> +
> +	if (bitlen % 8) {
> +		printf("%s: bitlen is not 8bit alined %d", __func__, bitlen);
> +		return 1;
> +	}
> +
> +	nbyte = bitlen / 8;
> +
> +	if (flags & SPI_XFER_BEGIN) {
> +		sh_qspi_writeb(SPCR_MSTR, &ss->regs->spcr);
> +
> +		sh_qspi_writew(SPCMD_INIT1, &ss->regs->spcmd0);
> +
> +		if (flags & SPI_XFER_END)
> +			sh_qspi_writel(nbyte, &ss->regs->spbmul0);
> +		else
> +			/* Set 1048576 byte */
> +			sh_qspi_writel(0x100000, &ss->regs->spbmul0);
> +		sh_qspi_writeb(SPBFCR_TXRST|SPBFCR_RXRST, &ss->regs->spbfcr);
> +		sh_qspi_writeb(0x00, &ss->regs->spbfcr);
> +		sh_qspi_writeb(0x00, &ss->regs->spscr);
> +		sh_qspi_writeb(SPCR_SPE|SPCR_MSTR, &ss->regs->spcr);
> +	}
> +
> +	if (dout != NULL)
> +		tdata = (unsigned char *)dout;
> +	else
> +		tdata = &dtdata;
> +
> +	if (din != NULL)
> +		rdata = din;
> +	else
> +		rdata = &drdata;
> +
> +	while (nbyte > 0) {
> +		ret = sh_qspi_xfer(ss, tdata, rdata, flags);
> +		if (ret)
> +			break;
> +
> +		if (dout != NULL)
> +			tdata++;
> +
> +		if (din != NULL)
> +			rdata++;
> +
> +		nbyte--;
> +	}
> +
> +	if (flags & SPI_XFER_END)
> +		sh_qspi_writeb(SPCR_MSTR, &ss->regs->spcr);
> +
> +	return ret;
> +}
> +
> +int  spi_cs_is_valid(unsigned int bus, unsigned int cs)
> +{
> +	return 1;
> +}
> +
> +void spi_cs_activate(struct spi_slave *slave)
> +{
> +}
> +
> +void spi_cs_deactivate(struct spi_slave *slave)
> +{
> +}

I will give few more comments, please fix these if you agree with the 
driver format i suggested.
if not please let me know your views.

--
Thanks,
Jagan



More information about the U-Boot mailing list