u-boot/drivers/spi/mxc_spi.c

694 lines
16 KiB
C
Raw Normal View History

// SPDX-License-Identifier: GPL-2.0+
/*
* Copyright (C) 2008, Guennadi Liakhovetski <lg@denx.de>
*/
#include <common.h>
#include <clk.h>
#include <dm.h>
#include <log.h>
SPI API improvements This patch gets rid of the spi_chipsel table and adds a handful of new functions that makes the SPI layer cleaner and more flexible. Instead of the spi_chipsel table, each board that wants to use SPI gets to implement three hooks: * spi_cs_activate(): Activates the chipselect for a given slave * spi_cs_deactivate(): Deactivates the chipselect for a given slave * spi_cs_is_valid(): Determines if the given bus/chipselect combination can be activated. Not all drivers may need those extra functions however. If that's the case, the board code may just leave them out (assuming they know what the driver needs) or rely on the linker to strip them out (assuming --gc-sections is being used.) To set up communication parameters for a given slave, the driver needs to call spi_setup_slave(). This returns a pointer to an opaque spi_slave struct which must be passed as a parameter to subsequent SPI calls. This struct can be freed by calling spi_free_slave(), but most driver probably don't want to do this. Before starting one or more SPI transfers, the driver must call spi_claim_bus() to gain exclusive access to the SPI bus and initialize the hardware. When all transfers are done, the driver must call spi_release_bus() to make the bus available to others, and possibly shut down the SPI controller hardware. spi_xfer() behaves mostly the same as before, but it now takes a spi_slave parameter instead of a spi_chipsel function pointer. It also got a new parameter, flags, which is used to specify chip select behaviour. This may be extended with other flags in the future. This patch has been build-tested on all powerpc and arm boards involved. I have not tested NIOS since I don't have a toolchain for it installed, so I expect some breakage there even though I've tried fixing up everything I could find by visual inspection. I have run-time tested this on AVR32 ATNGW100 using the atmel_spi and DataFlash drivers posted as a follow-up. I'd like some help testing other boards that use the existing SPI API. But most of all, I'd like some comments on the new API. Is this stuff usable for everyone? If not, why? Changed in v4: - Build fixes for various boards, drivers and commands - Provide common struct spi_slave definition that can be extended by drivers - Pass a struct spi_slave * to spi_cs_activate and spi_cs_deactivate - Make default bus and mode build-time configurable - Override default SPI bus ID and mode on mx32ads and imx31_litekit. Changed in v3: - Add opaque struct spi_slave for controller-specific data associated with a slave. - Add spi_claim_bus() and spi_release_bus() - Add spi_free_slave() - spi_setup() is now called spi_setup_slave() and returns a struct spi_slave - soft_spi now supports four SPI modes (CPOL|CPHA) - Add bus parameter to spi_setup_slave() - Convert the new i.MX32 SPI driver - Convert the new MC13783 RTC driver Changed in v2: - Convert the mpc8xxx_spi driver and the mpc8349emds board to the new API. Signed-off-by: Haavard Skinnemoen <hskinnemoen@atmel.com> Tested-by: Guennadi Liakhovetski <lg@denx.de>
2008-05-16 09:10:31 +00:00
#include <malloc.h>
#include <spi.h>
#include <asm/global_data.h>
#include <dm/device_compat.h>
#include <linux/bitops.h>
#include <linux/delay.h>
#include <linux/errno.h>
#include <asm/io.h>
#include <asm/gpio.h>
#include <asm/arch/imx-regs.h>
#include <asm/arch/clock.h>
imx: reorganize IMX code as other SOCs Change is consistent with other SOCs and it is in preparation for adding SOMs. SOC's related files are moved from cpu/ to mach-imx/<SOC>. This change is also coherent with the structure in kernel. Signed-off-by: Stefano Babic <sbabic@denx.de> CC: Fabio Estevam <fabio.estevam@nxp.com> CC: Akshay Bhat <akshaybhat@timesys.com> CC: Ken Lin <Ken.Lin@advantech.com.tw> CC: Marek Vasut <marek.vasut@gmail.com> CC: Heiko Schocher <hs@denx.de> CC: "Sébastien Szymanski" <sebastien.szymanski@armadeus.com> CC: Christian Gmeiner <christian.gmeiner@gmail.com> CC: Stefan Roese <sr@denx.de> CC: Patrick Bruenn <p.bruenn@beckhoff.com> CC: Troy Kisky <troy.kisky@boundarydevices.com> CC: Nikita Kiryanov <nikita@compulab.co.il> CC: Otavio Salvador <otavio@ossystems.com.br> CC: "Eric Bénard" <eric@eukrea.com> CC: Jagan Teki <jagan@amarulasolutions.com> CC: Ye Li <ye.li@nxp.com> CC: Peng Fan <peng.fan@nxp.com> CC: Adrian Alonso <adrian.alonso@nxp.com> CC: Alison Wang <b18965@freescale.com> CC: Tim Harvey <tharvey@gateworks.com> CC: Martin Donnelly <martin.donnelly@ge.com> CC: Marcin Niestroj <m.niestroj@grinn-global.com> CC: Lukasz Majewski <lukma@denx.de> CC: Adam Ford <aford173@gmail.com> CC: "Albert ARIBAUD (3ADEV)" <albert.aribaud@3adev.fr> CC: Boris Brezillon <boris.brezillon@free-electrons.com> CC: Soeren Moch <smoch@web.de> CC: Richard Hu <richard.hu@technexion.com> CC: Wig Cheng <wig.cheng@technexion.com> CC: Vanessa Maegima <vanessa.maegima@nxp.com> CC: Max Krummenacher <max.krummenacher@toradex.com> CC: Stefan Agner <stefan.agner@toradex.com> CC: Markus Niebel <Markus.Niebel@tq-group.com> CC: Breno Lima <breno.lima@nxp.com> CC: Francesco Montefoschi <francesco.montefoschi@udoo.org> CC: Jaehoon Chung <jh80.chung@samsung.com> CC: Scott Wood <oss@buserror.net> CC: Joe Hershberger <joe.hershberger@ni.com> CC: Anatolij Gustschin <agust@denx.de> CC: Simon Glass <sjg@chromium.org> CC: "Andrew F. Davis" <afd@ti.com> CC: "Łukasz Majewski" <l.majewski@samsung.com> CC: Patrice Chotard <patrice.chotard@st.com> CC: Nobuhiro Iwamatsu <iwamatsu@nigauri.org> CC: Hans de Goede <hdegoede@redhat.com> CC: Masahiro Yamada <yamada.masahiro@socionext.com> CC: Stephen Warren <swarren@nvidia.com> CC: Andre Przywara <andre.przywara@arm.com> CC: "Álvaro Fernández Rojas" <noltari@gmail.com> CC: York Sun <york.sun@nxp.com> CC: Xiaoliang Yang <xiaoliang.yang@nxp.com> CC: Chen-Yu Tsai <wens@csie.org> CC: George McCollister <george.mccollister@gmail.com> CC: Sven Ebenfeld <sven.ebenfeld@gmail.com> CC: Filip Brozovic <fbrozovic@gmail.com> CC: Petr Kulhavy <brain@jikos.cz> CC: Eric Nelson <eric@nelint.com> CC: Bai Ping <ping.bai@nxp.com> CC: Anson Huang <Anson.Huang@nxp.com> CC: Sanchayan Maity <maitysanchayan@gmail.com> CC: Lokesh Vutla <lokeshvutla@ti.com> CC: Patrick Delaunay <patrick.delaunay@st.com> CC: Gary Bisson <gary.bisson@boundarydevices.com> CC: Alexander Graf <agraf@suse.de> CC: u-boot@lists.denx.de Reviewed-by: Fabio Estevam <fabio.estevam@nxp.com> Reviewed-by: Christian Gmeiner <christian.gmeiner@gmail.com>
2017-06-29 08:16:06 +00:00
#include <asm/mach-imx/spi.h>
DECLARE_GLOBAL_DATA_PTR;
/* MX35 and older is CSPI */
#if defined(CONFIG_MX31)
#define MXC_CSPI
struct cspi_regs {
u32 rxdata;
u32 txdata;
u32 ctrl;
u32 intr;
u32 dma;
u32 stat;
u32 period;
u32 test;
};
#define MXC_CSPICTRL_EN BIT(0)
#define MXC_CSPICTRL_MODE BIT(1)
#define MXC_CSPICTRL_XCH BIT(2)
#define MXC_CSPICTRL_SMC BIT(3)
#define MXC_CSPICTRL_POL BIT(4)
#define MXC_CSPICTRL_PHA BIT(5)
#define MXC_CSPICTRL_SSCTL BIT(6)
#define MXC_CSPICTRL_SSPOL BIT(7)
#define MXC_CSPICTRL_DATARATE(x) (((x) & 0x7) << 16)
#define MXC_CSPICTRL_RXOVF BIT(6)
#define MXC_CSPIPERIOD_32KHZ BIT(15)
#define MAX_SPI_BYTES 4
#define MXC_CSPICTRL_CHIPSELECT(x) (((x) & 0x3) << 24)
#define MXC_CSPICTRL_BITCOUNT(x) (((x) & 0x1f) << 8)
#define MXC_CSPICTRL_TC BIT(8)
#define MXC_CSPICTRL_MAXBITS 0x1f
#else /* MX51 and newer is ECSPI */
#define MXC_ECSPI
struct cspi_regs {
u32 rxdata;
u32 txdata;
u32 ctrl;
u32 cfg;
u32 intr;
u32 dma;
u32 stat;
u32 period;
};
#define MXC_CSPICTRL_EN BIT(0)
#define MXC_CSPICTRL_MODE BIT(1)
#define MXC_CSPICTRL_XCH BIT(2)
#define MXC_CSPICTRL_MODE_MASK (0xf << 4)
#define MXC_CSPICTRL_CHIPSELECT(x) (((x) & 0x3) << 12)
#define MXC_CSPICTRL_BITCOUNT(x) (((x) & 0xfff) << 20)
#define MXC_CSPICTRL_PREDIV(x) (((x) & 0xF) << 12)
#define MXC_CSPICTRL_POSTDIV(x) (((x) & 0xF) << 8)
#define MXC_CSPICTRL_SELCHAN(x) (((x) & 0x3) << 18)
#define MXC_CSPICTRL_MAXBITS 0xfff
#define MXC_CSPICTRL_TC BIT(7)
#define MXC_CSPICTRL_RXOVF BIT(6)
#define MXC_CSPIPERIOD_32KHZ BIT(15)
#define MAX_SPI_BYTES 32
/* Bit position inside CTRL register to be associated with SS */
#define MXC_CSPICTRL_CHAN 18
/* Bit position inside CON register to be associated with SS */
#define MXC_CSPICON_PHA 0 /* SCLK phase control */
#define MXC_CSPICON_POL 4 /* SCLK polarity */
#define MXC_CSPICON_SSPOL 12 /* SS polarity */
#define MXC_CSPICON_CTL 20 /* inactive state of SCLK */
#endif
#ifdef CONFIG_MX27
/* i.MX27 has a completely wrong register layout and register definitions in the
* datasheet, the correct one is in the Freescale's Linux driver */
#error "i.MX27 CSPI not supported due to drastic differences in register definitions" \
"See linux mxc_spi driver from Freescale for details."
#endif
__weak int board_spi_cs_gpio(unsigned bus, unsigned cs)
{
return -1;
}
#define OUT MXC_GPIO_DIRECTION_OUT
#define reg_read readl
#define reg_write(a, v) writel(v, a)
#if !defined(CONFIG_SYS_SPI_MXC_WAIT)
#define CONFIG_SYS_SPI_MXC_WAIT (CONFIG_SYS_HZ/100) /* 10 ms */
#endif
#define MAX_CS_COUNT 4
SPI API improvements This patch gets rid of the spi_chipsel table and adds a handful of new functions that makes the SPI layer cleaner and more flexible. Instead of the spi_chipsel table, each board that wants to use SPI gets to implement three hooks: * spi_cs_activate(): Activates the chipselect for a given slave * spi_cs_deactivate(): Deactivates the chipselect for a given slave * spi_cs_is_valid(): Determines if the given bus/chipselect combination can be activated. Not all drivers may need those extra functions however. If that's the case, the board code may just leave them out (assuming they know what the driver needs) or rely on the linker to strip them out (assuming --gc-sections is being used.) To set up communication parameters for a given slave, the driver needs to call spi_setup_slave(). This returns a pointer to an opaque spi_slave struct which must be passed as a parameter to subsequent SPI calls. This struct can be freed by calling spi_free_slave(), but most driver probably don't want to do this. Before starting one or more SPI transfers, the driver must call spi_claim_bus() to gain exclusive access to the SPI bus and initialize the hardware. When all transfers are done, the driver must call spi_release_bus() to make the bus available to others, and possibly shut down the SPI controller hardware. spi_xfer() behaves mostly the same as before, but it now takes a spi_slave parameter instead of a spi_chipsel function pointer. It also got a new parameter, flags, which is used to specify chip select behaviour. This may be extended with other flags in the future. This patch has been build-tested on all powerpc and arm boards involved. I have not tested NIOS since I don't have a toolchain for it installed, so I expect some breakage there even though I've tried fixing up everything I could find by visual inspection. I have run-time tested this on AVR32 ATNGW100 using the atmel_spi and DataFlash drivers posted as a follow-up. I'd like some help testing other boards that use the existing SPI API. But most of all, I'd like some comments on the new API. Is this stuff usable for everyone? If not, why? Changed in v4: - Build fixes for various boards, drivers and commands - Provide common struct spi_slave definition that can be extended by drivers - Pass a struct spi_slave * to spi_cs_activate and spi_cs_deactivate - Make default bus and mode build-time configurable - Override default SPI bus ID and mode on mx32ads and imx31_litekit. Changed in v3: - Add opaque struct spi_slave for controller-specific data associated with a slave. - Add spi_claim_bus() and spi_release_bus() - Add spi_free_slave() - spi_setup() is now called spi_setup_slave() and returns a struct spi_slave - soft_spi now supports four SPI modes (CPOL|CPHA) - Add bus parameter to spi_setup_slave() - Convert the new i.MX32 SPI driver - Convert the new MC13783 RTC driver Changed in v2: - Convert the mpc8xxx_spi driver and the mpc8349emds board to the new API. Signed-off-by: Haavard Skinnemoen <hskinnemoen@atmel.com> Tested-by: Guennadi Liakhovetski <lg@denx.de>
2008-05-16 09:10:31 +00:00
struct mxc_spi_slave {
struct spi_slave slave;
unsigned long base;
u32 ctrl_reg;
#if defined(MXC_ECSPI)
u32 cfg_reg;
#endif
int gpio;
int ss_pol;
unsigned int max_hz;
unsigned int mode;
struct gpio_desc ss;
struct gpio_desc cs_gpios[MAX_CS_COUNT];
struct udevice *dev;
};
SPI API improvements This patch gets rid of the spi_chipsel table and adds a handful of new functions that makes the SPI layer cleaner and more flexible. Instead of the spi_chipsel table, each board that wants to use SPI gets to implement three hooks: * spi_cs_activate(): Activates the chipselect for a given slave * spi_cs_deactivate(): Deactivates the chipselect for a given slave * spi_cs_is_valid(): Determines if the given bus/chipselect combination can be activated. Not all drivers may need those extra functions however. If that's the case, the board code may just leave them out (assuming they know what the driver needs) or rely on the linker to strip them out (assuming --gc-sections is being used.) To set up communication parameters for a given slave, the driver needs to call spi_setup_slave(). This returns a pointer to an opaque spi_slave struct which must be passed as a parameter to subsequent SPI calls. This struct can be freed by calling spi_free_slave(), but most driver probably don't want to do this. Before starting one or more SPI transfers, the driver must call spi_claim_bus() to gain exclusive access to the SPI bus and initialize the hardware. When all transfers are done, the driver must call spi_release_bus() to make the bus available to others, and possibly shut down the SPI controller hardware. spi_xfer() behaves mostly the same as before, but it now takes a spi_slave parameter instead of a spi_chipsel function pointer. It also got a new parameter, flags, which is used to specify chip select behaviour. This may be extended with other flags in the future. This patch has been build-tested on all powerpc and arm boards involved. I have not tested NIOS since I don't have a toolchain for it installed, so I expect some breakage there even though I've tried fixing up everything I could find by visual inspection. I have run-time tested this on AVR32 ATNGW100 using the atmel_spi and DataFlash drivers posted as a follow-up. I'd like some help testing other boards that use the existing SPI API. But most of all, I'd like some comments on the new API. Is this stuff usable for everyone? If not, why? Changed in v4: - Build fixes for various boards, drivers and commands - Provide common struct spi_slave definition that can be extended by drivers - Pass a struct spi_slave * to spi_cs_activate and spi_cs_deactivate - Make default bus and mode build-time configurable - Override default SPI bus ID and mode on mx32ads and imx31_litekit. Changed in v3: - Add opaque struct spi_slave for controller-specific data associated with a slave. - Add spi_claim_bus() and spi_release_bus() - Add spi_free_slave() - spi_setup() is now called spi_setup_slave() and returns a struct spi_slave - soft_spi now supports four SPI modes (CPOL|CPHA) - Add bus parameter to spi_setup_slave() - Convert the new i.MX32 SPI driver - Convert the new MC13783 RTC driver Changed in v2: - Convert the mpc8xxx_spi driver and the mpc8349emds board to the new API. Signed-off-by: Haavard Skinnemoen <hskinnemoen@atmel.com> Tested-by: Guennadi Liakhovetski <lg@denx.de>
2008-05-16 09:10:31 +00:00
static inline struct mxc_spi_slave *to_mxc_spi_slave(struct spi_slave *slave)
{
return container_of(slave, struct mxc_spi_slave, slave);
}
static void mxc_spi_cs_activate(struct mxc_spi_slave *mxcs)
{
spi: Convert CONFIG_DM_SPI* to CONFIG_$(SPL_TPL_)DM_SPI* This change allows more fine tuning of driver model based SPI support in SPL and TPL. It is now possible to explicitly enable/disable the DM_SPI support in SPL and TPL via Kconfig option. Before this change it was necessary to use: /* SPI Flash Configs */ #if defined(CONFIG_SPL_BUILD) #undef CONFIG_DM_SPI #undef CONFIG_DM_SPI_FLASH #undef CONFIG_SPI_FLASH_MTD #endif in the ./include/configs/<board>.h, which is error prone and shall be avoided when we strive to switch to Kconfig. The goal of this patch: Provide distinction for DM_SPI support in both U-Boot proper and SPL (TPL). Valid use case is when U-Boot proper wants to use DM_SPI, but SPL must still support non DM driver. Another use case is the conversion of non DM/DTS SPI driver to support DM/DTS. When such driver needs to work in both SPL and U-Boot proper, the distinction is needed in Kconfig (also if SPL version of the driver supports OF_PLATDATA). In the end of the day one would have to support following use cases (in single driver file - e.g. mxs_spi.c): - U-Boot proper driver supporting DT/DTS - U-Boot proper driver without DT/DTS support (deprecated) - SPL driver without DT/DTS support - SPL (and TPL) driver with DT/DTS (when the SoC has enough resources to run full blown DT/DTS) - SPL driver with DT/DTS and SPL_OF_PLATDATA (when one have constrained environment with no fitImage and OF_LIBFDT support). Some boards do require SPI support (with DM) in SPL (TPL) and some only have DM_SPI{_FLASH} defined to allow compiling SPL. This patch converts #ifdef CONFIG_DM_SPI* to #if CONFIG_IS_ENABLED(DM_SPI) and provides corresponding defines in Kconfig. Signed-off-by: Lukasz Majewski <lukma@denx.de> Tested-by: Adam Ford <aford173@gmail.com> #da850-evm Signed-off-by: Hou Zhiqiang <Zhiqiang.Hou@nxp.com> [trini: Fixup a few platforms] Signed-off-by: Tom Rini <trini@konsulko.com>
2020-06-04 15:11:53 +00:00
#if CONFIG_IS_ENABLED(DM_SPI)
struct udevice *dev = mxcs->dev;
struct dm_spi_slave_plat *slave_plat = dev_get_parent_plat(dev);
u32 cs = slave_plat->cs;
if (!dm_gpio_is_valid(&mxcs->cs_gpios[cs]))
return;
dm_gpio_set_value(&mxcs->cs_gpios[cs], 1);
#else
if (mxcs->gpio > 0)
gpio_set_value(mxcs->gpio, mxcs->ss_pol);
#endif
}
static void mxc_spi_cs_deactivate(struct mxc_spi_slave *mxcs)
{
spi: Convert CONFIG_DM_SPI* to CONFIG_$(SPL_TPL_)DM_SPI* This change allows more fine tuning of driver model based SPI support in SPL and TPL. It is now possible to explicitly enable/disable the DM_SPI support in SPL and TPL via Kconfig option. Before this change it was necessary to use: /* SPI Flash Configs */ #if defined(CONFIG_SPL_BUILD) #undef CONFIG_DM_SPI #undef CONFIG_DM_SPI_FLASH #undef CONFIG_SPI_FLASH_MTD #endif in the ./include/configs/<board>.h, which is error prone and shall be avoided when we strive to switch to Kconfig. The goal of this patch: Provide distinction for DM_SPI support in both U-Boot proper and SPL (TPL). Valid use case is when U-Boot proper wants to use DM_SPI, but SPL must still support non DM driver. Another use case is the conversion of non DM/DTS SPI driver to support DM/DTS. When such driver needs to work in both SPL and U-Boot proper, the distinction is needed in Kconfig (also if SPL version of the driver supports OF_PLATDATA). In the end of the day one would have to support following use cases (in single driver file - e.g. mxs_spi.c): - U-Boot proper driver supporting DT/DTS - U-Boot proper driver without DT/DTS support (deprecated) - SPL driver without DT/DTS support - SPL (and TPL) driver with DT/DTS (when the SoC has enough resources to run full blown DT/DTS) - SPL driver with DT/DTS and SPL_OF_PLATDATA (when one have constrained environment with no fitImage and OF_LIBFDT support). Some boards do require SPI support (with DM) in SPL (TPL) and some only have DM_SPI{_FLASH} defined to allow compiling SPL. This patch converts #ifdef CONFIG_DM_SPI* to #if CONFIG_IS_ENABLED(DM_SPI) and provides corresponding defines in Kconfig. Signed-off-by: Lukasz Majewski <lukma@denx.de> Tested-by: Adam Ford <aford173@gmail.com> #da850-evm Signed-off-by: Hou Zhiqiang <Zhiqiang.Hou@nxp.com> [trini: Fixup a few platforms] Signed-off-by: Tom Rini <trini@konsulko.com>
2020-06-04 15:11:53 +00:00
#if CONFIG_IS_ENABLED(DM_SPI)
struct udevice *dev = mxcs->dev;
struct dm_spi_slave_plat *slave_plat = dev_get_parent_plat(dev);
u32 cs = slave_plat->cs;
if (!dm_gpio_is_valid(&mxcs->cs_gpios[cs]))
return;
dm_gpio_set_value(&mxcs->cs_gpios[cs], 0);
#else
if (mxcs->gpio > 0)
gpio_set_value(mxcs->gpio, !(mxcs->ss_pol));
#endif
}
u32 get_cspi_div(u32 div)
{
int i;
for (i = 0; i < 8; i++) {
if (div <= (4 << i))
return i;
}
return i;
}
#ifdef MXC_CSPI
static s32 spi_cfg_mxc(struct mxc_spi_slave *mxcs, unsigned int cs)
{
unsigned int ctrl_reg;
u32 clk_src;
u32 div;
unsigned int max_hz = mxcs->max_hz;
unsigned int mode = mxcs->mode;
clk_src = mxc_get_clock(MXC_CSPI_CLK);
div = DIV_ROUND_UP(clk_src, max_hz);
div = get_cspi_div(div);
debug("clk %d Hz, div %d, real clk %d Hz\n",
max_hz, div, clk_src / (4 << div));
ctrl_reg = MXC_CSPICTRL_CHIPSELECT(cs) |
MXC_CSPICTRL_BITCOUNT(MXC_CSPICTRL_MAXBITS) |
MXC_CSPICTRL_DATARATE(div) |
MXC_CSPICTRL_EN |
MXC_CSPICTRL_MODE;
if (mode & SPI_CPHA)
ctrl_reg |= MXC_CSPICTRL_PHA;
if (mode & SPI_CPOL)
ctrl_reg |= MXC_CSPICTRL_POL;
if (mode & SPI_CS_HIGH)
ctrl_reg |= MXC_CSPICTRL_SSPOL;
mxcs->ctrl_reg = ctrl_reg;
return 0;
}
#endif
#ifdef MXC_ECSPI
static s32 spi_cfg_mxc(struct mxc_spi_slave *mxcs, unsigned int cs)
{
u32 clk_src = mxc_get_clock(MXC_CSPI_CLK);
s32 reg_ctrl, reg_config;
u32 ss_pol = 0, sclkpol = 0, sclkpha = 0, sclkctl = 0;
u32 pre_div = 0, post_div = 0;
struct cspi_regs *regs = (struct cspi_regs *)mxcs->base;
unsigned int max_hz = mxcs->max_hz;
unsigned int mode = mxcs->mode;
/*
* Reset SPI and set all CSs to master mode, if toggling
* between slave and master mode we might see a glitch
* on the clock line
*/
reg_ctrl = MXC_CSPICTRL_MODE_MASK;
reg_write(&regs->ctrl, reg_ctrl);
reg_ctrl |= MXC_CSPICTRL_EN;
reg_write(&regs->ctrl, reg_ctrl);
if (clk_src > max_hz) {
pre_div = (clk_src - 1) / max_hz;
/* fls(1) = 1, fls(0x80000000) = 32, fls(16) = 5 */
post_div = fls(pre_div);
if (post_div > 4) {
post_div -= 4;
if (post_div >= 16) {
printf("Error: no divider for the freq: %d\n",
max_hz);
return -1;
}
pre_div >>= post_div;
} else {
post_div = 0;
}
}
debug("pre_div = %d, post_div=%d\n", pre_div, post_div);
reg_ctrl = (reg_ctrl & ~MXC_CSPICTRL_SELCHAN(3)) |
MXC_CSPICTRL_SELCHAN(cs);
reg_ctrl = (reg_ctrl & ~MXC_CSPICTRL_PREDIV(0x0F)) |
MXC_CSPICTRL_PREDIV(pre_div);
reg_ctrl = (reg_ctrl & ~MXC_CSPICTRL_POSTDIV(0x0F)) |
MXC_CSPICTRL_POSTDIV(post_div);
if (mode & SPI_CS_HIGH)
ss_pol = 1;
if (mode & SPI_CPOL) {
sclkpol = 1;
sclkctl = 1;
}
if (mode & SPI_CPHA)
sclkpha = 1;
reg_config = reg_read(&regs->cfg);
/*
* Configuration register setup
* The MX51 supports different setup for each SS
*/
reg_config = (reg_config & ~(1 << (cs + MXC_CSPICON_SSPOL))) |
(ss_pol << (cs + MXC_CSPICON_SSPOL));
reg_config = (reg_config & ~(1 << (cs + MXC_CSPICON_POL))) |
(sclkpol << (cs + MXC_CSPICON_POL));
reg_config = (reg_config & ~(1 << (cs + MXC_CSPICON_CTL))) |
(sclkctl << (cs + MXC_CSPICON_CTL));
reg_config = (reg_config & ~(1 << (cs + MXC_CSPICON_PHA))) |
(sclkpha << (cs + MXC_CSPICON_PHA));
debug("reg_ctrl = 0x%x\n", reg_ctrl);
reg_write(&regs->ctrl, reg_ctrl);
debug("reg_config = 0x%x\n", reg_config);
reg_write(&regs->cfg, reg_config);
/* save config register and control register */
mxcs->ctrl_reg = reg_ctrl;
mxcs->cfg_reg = reg_config;
/* clear interrupt reg */
reg_write(&regs->intr, 0);
reg_write(&regs->stat, MXC_CSPICTRL_TC | MXC_CSPICTRL_RXOVF);
return 0;
}
#endif
int spi_xchg_single(struct mxc_spi_slave *mxcs, unsigned int bitlen,
const u8 *dout, u8 *din, unsigned long flags)
{
int nbytes = DIV_ROUND_UP(bitlen, 8);
u32 data, cnt, i;
struct cspi_regs *regs = (struct cspi_regs *)mxcs->base;
u32 ts;
int status;
debug("%s: bitlen %d dout 0x%lx din 0x%lx\n",
__func__, bitlen, (ulong)dout, (ulong)din);
mxcs->ctrl_reg = (mxcs->ctrl_reg &
~MXC_CSPICTRL_BITCOUNT(MXC_CSPICTRL_MAXBITS)) |
MXC_CSPICTRL_BITCOUNT(bitlen - 1);
reg_write(&regs->ctrl, mxcs->ctrl_reg | MXC_CSPICTRL_EN);
#ifdef MXC_ECSPI
reg_write(&regs->cfg, mxcs->cfg_reg);
#endif
/* Clear interrupt register */
reg_write(&regs->stat, MXC_CSPICTRL_TC | MXC_CSPICTRL_RXOVF);
/*
* The SPI controller works only with words,
* check if less than a word is sent.
* Access to the FIFO is only 32 bit
*/
if (bitlen % 32) {
data = 0;
cnt = (bitlen % 32) / 8;
if (dout) {
for (i = 0; i < cnt; i++) {
data = (data << 8) | (*dout++ & 0xFF);
}
}
debug("Sending SPI 0x%x\n", data);
reg_write(&regs->txdata, data);
nbytes -= cnt;
}
data = 0;
while (nbytes > 0) {
data = 0;
if (dout) {
/* Buffer is not 32-bit aligned */
if ((unsigned long)dout & 0x03) {
data = 0;
for (i = 0; i < 4; i++)
data = (data << 8) | (*dout++ & 0xFF);
} else {
data = *(u32 *)dout;
data = cpu_to_be32(data);
dout += 4;
}
}
debug("Sending SPI 0x%x\n", data);
reg_write(&regs->txdata, data);
nbytes -= 4;
}
/* FIFO is written, now starts the transfer setting the XCH bit */
reg_write(&regs->ctrl, mxcs->ctrl_reg |
MXC_CSPICTRL_EN | MXC_CSPICTRL_XCH);
ts = get_timer(0);
status = reg_read(&regs->stat);
/* Wait until the TC (Transfer completed) bit is set */
while ((status & MXC_CSPICTRL_TC) == 0) {
if (get_timer(ts) > CONFIG_SYS_SPI_MXC_WAIT) {
printf("spi_xchg_single: Timeout!\n");
return -1;
}
status = reg_read(&regs->stat);
}
/* Transfer completed, clear any pending request */
reg_write(&regs->stat, MXC_CSPICTRL_TC | MXC_CSPICTRL_RXOVF);
nbytes = DIV_ROUND_UP(bitlen, 8);
cnt = nbytes % 32;
if (bitlen % 32) {
data = reg_read(&regs->rxdata);
cnt = (bitlen % 32) / 8;
data = cpu_to_be32(data) >> ((sizeof(data) - cnt) * 8);
debug("SPI Rx unaligned: 0x%x\n", data);
if (din) {
memcpy(din, &data, cnt);
din += cnt;
}
nbytes -= cnt;
}
while (nbytes > 0) {
u32 tmp;
tmp = reg_read(&regs->rxdata);
data = cpu_to_be32(tmp);
debug("SPI Rx: 0x%x 0x%x\n", tmp, data);
cnt = min_t(u32, nbytes, sizeof(data));
if (din) {
memcpy(din, &data, cnt);
din += cnt;
}
nbytes -= cnt;
}
return 0;
}
static int mxc_spi_xfer_internal(struct mxc_spi_slave *mxcs,
unsigned int bitlen, const void *dout,
void *din, unsigned long flags)
{
int n_bytes = DIV_ROUND_UP(bitlen, 8);
int n_bits;
int ret;
u32 blk_size;
u8 *p_outbuf = (u8 *)dout;
u8 *p_inbuf = (u8 *)din;
if (!mxcs)
return -EINVAL;
if (flags & SPI_XFER_BEGIN)
mxc_spi_cs_activate(mxcs);
while (n_bytes > 0) {
if (n_bytes < MAX_SPI_BYTES)
blk_size = n_bytes;
else
blk_size = MAX_SPI_BYTES;
n_bits = blk_size * 8;
ret = spi_xchg_single(mxcs, n_bits, p_outbuf, p_inbuf, 0);
if (ret)
return ret;
if (dout)
p_outbuf += blk_size;
if (din)
p_inbuf += blk_size;
n_bytes -= blk_size;
}
if (flags & SPI_XFER_END) {
mxc_spi_cs_deactivate(mxcs);
}
return 0;
}
static int mxc_spi_claim_bus_internal(struct mxc_spi_slave *mxcs, int cs)
{
struct cspi_regs *regs = (struct cspi_regs *)mxcs->base;
int ret;
reg_write(&regs->rxdata, 1);
udelay(1);
ret = spi_cfg_mxc(mxcs, cs);
if (ret) {
printf("mxc_spi: cannot setup SPI controller\n");
return ret;
}
reg_write(&regs->period, MXC_CSPIPERIOD_32KHZ);
reg_write(&regs->intr, 0);
return 0;
}
spi: Convert CONFIG_DM_SPI* to CONFIG_$(SPL_TPL_)DM_SPI* This change allows more fine tuning of driver model based SPI support in SPL and TPL. It is now possible to explicitly enable/disable the DM_SPI support in SPL and TPL via Kconfig option. Before this change it was necessary to use: /* SPI Flash Configs */ #if defined(CONFIG_SPL_BUILD) #undef CONFIG_DM_SPI #undef CONFIG_DM_SPI_FLASH #undef CONFIG_SPI_FLASH_MTD #endif in the ./include/configs/<board>.h, which is error prone and shall be avoided when we strive to switch to Kconfig. The goal of this patch: Provide distinction for DM_SPI support in both U-Boot proper and SPL (TPL). Valid use case is when U-Boot proper wants to use DM_SPI, but SPL must still support non DM driver. Another use case is the conversion of non DM/DTS SPI driver to support DM/DTS. When such driver needs to work in both SPL and U-Boot proper, the distinction is needed in Kconfig (also if SPL version of the driver supports OF_PLATDATA). In the end of the day one would have to support following use cases (in single driver file - e.g. mxs_spi.c): - U-Boot proper driver supporting DT/DTS - U-Boot proper driver without DT/DTS support (deprecated) - SPL driver without DT/DTS support - SPL (and TPL) driver with DT/DTS (when the SoC has enough resources to run full blown DT/DTS) - SPL driver with DT/DTS and SPL_OF_PLATDATA (when one have constrained environment with no fitImage and OF_LIBFDT support). Some boards do require SPI support (with DM) in SPL (TPL) and some only have DM_SPI{_FLASH} defined to allow compiling SPL. This patch converts #ifdef CONFIG_DM_SPI* to #if CONFIG_IS_ENABLED(DM_SPI) and provides corresponding defines in Kconfig. Signed-off-by: Lukasz Majewski <lukma@denx.de> Tested-by: Adam Ford <aford173@gmail.com> #da850-evm Signed-off-by: Hou Zhiqiang <Zhiqiang.Hou@nxp.com> [trini: Fixup a few platforms] Signed-off-by: Tom Rini <trini@konsulko.com>
2020-06-04 15:11:53 +00:00
#if !CONFIG_IS_ENABLED(DM_SPI)
int spi_xfer(struct spi_slave *slave, unsigned int bitlen, const void *dout,
void *din, unsigned long flags)
{
struct mxc_spi_slave *mxcs = to_mxc_spi_slave(slave);
return mxc_spi_xfer_internal(mxcs, bitlen, dout, din, flags);
}
/*
* Some SPI devices require active chip-select over multiple
* transactions, we achieve this using a GPIO. Still, the SPI
* controller has to be configured to use one of its own chipselects.
* To use this feature you have to implement board_spi_cs_gpio() to assign
* a gpio value for each cs (-1 if cs doesn't need to use gpio).
* You must use some unused on this SPI controller cs between 0 and 3.
*/
static int setup_cs_gpio(struct mxc_spi_slave *mxcs,
unsigned int bus, unsigned int cs)
{
int ret;
mxcs->gpio = board_spi_cs_gpio(bus, cs);
if (mxcs->gpio == -1)
return 0;
gpio_request(mxcs->gpio, "spi-cs");
ret = gpio_direction_output(mxcs->gpio, !(mxcs->ss_pol));
if (ret) {
printf("mxc_spi: cannot setup gpio %d\n", mxcs->gpio);
return -EINVAL;
}
return 0;
}
static unsigned long spi_bases[] = {
MXC_SPI_BASE_ADDRESSES
};
SPI API improvements This patch gets rid of the spi_chipsel table and adds a handful of new functions that makes the SPI layer cleaner and more flexible. Instead of the spi_chipsel table, each board that wants to use SPI gets to implement three hooks: * spi_cs_activate(): Activates the chipselect for a given slave * spi_cs_deactivate(): Deactivates the chipselect for a given slave * spi_cs_is_valid(): Determines if the given bus/chipselect combination can be activated. Not all drivers may need those extra functions however. If that's the case, the board code may just leave them out (assuming they know what the driver needs) or rely on the linker to strip them out (assuming --gc-sections is being used.) To set up communication parameters for a given slave, the driver needs to call spi_setup_slave(). This returns a pointer to an opaque spi_slave struct which must be passed as a parameter to subsequent SPI calls. This struct can be freed by calling spi_free_slave(), but most driver probably don't want to do this. Before starting one or more SPI transfers, the driver must call spi_claim_bus() to gain exclusive access to the SPI bus and initialize the hardware. When all transfers are done, the driver must call spi_release_bus() to make the bus available to others, and possibly shut down the SPI controller hardware. spi_xfer() behaves mostly the same as before, but it now takes a spi_slave parameter instead of a spi_chipsel function pointer. It also got a new parameter, flags, which is used to specify chip select behaviour. This may be extended with other flags in the future. This patch has been build-tested on all powerpc and arm boards involved. I have not tested NIOS since I don't have a toolchain for it installed, so I expect some breakage there even though I've tried fixing up everything I could find by visual inspection. I have run-time tested this on AVR32 ATNGW100 using the atmel_spi and DataFlash drivers posted as a follow-up. I'd like some help testing other boards that use the existing SPI API. But most of all, I'd like some comments on the new API. Is this stuff usable for everyone? If not, why? Changed in v4: - Build fixes for various boards, drivers and commands - Provide common struct spi_slave definition that can be extended by drivers - Pass a struct spi_slave * to spi_cs_activate and spi_cs_deactivate - Make default bus and mode build-time configurable - Override default SPI bus ID and mode on mx32ads and imx31_litekit. Changed in v3: - Add opaque struct spi_slave for controller-specific data associated with a slave. - Add spi_claim_bus() and spi_release_bus() - Add spi_free_slave() - spi_setup() is now called spi_setup_slave() and returns a struct spi_slave - soft_spi now supports four SPI modes (CPOL|CPHA) - Add bus parameter to spi_setup_slave() - Convert the new i.MX32 SPI driver - Convert the new MC13783 RTC driver Changed in v2: - Convert the mpc8xxx_spi driver and the mpc8349emds board to the new API. Signed-off-by: Haavard Skinnemoen <hskinnemoen@atmel.com> Tested-by: Guennadi Liakhovetski <lg@denx.de>
2008-05-16 09:10:31 +00:00
struct spi_slave *spi_setup_slave(unsigned int bus, unsigned int cs,
unsigned int max_hz, unsigned int mode)
{
SPI API improvements This patch gets rid of the spi_chipsel table and adds a handful of new functions that makes the SPI layer cleaner and more flexible. Instead of the spi_chipsel table, each board that wants to use SPI gets to implement three hooks: * spi_cs_activate(): Activates the chipselect for a given slave * spi_cs_deactivate(): Deactivates the chipselect for a given slave * spi_cs_is_valid(): Determines if the given bus/chipselect combination can be activated. Not all drivers may need those extra functions however. If that's the case, the board code may just leave them out (assuming they know what the driver needs) or rely on the linker to strip them out (assuming --gc-sections is being used.) To set up communication parameters for a given slave, the driver needs to call spi_setup_slave(). This returns a pointer to an opaque spi_slave struct which must be passed as a parameter to subsequent SPI calls. This struct can be freed by calling spi_free_slave(), but most driver probably don't want to do this. Before starting one or more SPI transfers, the driver must call spi_claim_bus() to gain exclusive access to the SPI bus and initialize the hardware. When all transfers are done, the driver must call spi_release_bus() to make the bus available to others, and possibly shut down the SPI controller hardware. spi_xfer() behaves mostly the same as before, but it now takes a spi_slave parameter instead of a spi_chipsel function pointer. It also got a new parameter, flags, which is used to specify chip select behaviour. This may be extended with other flags in the future. This patch has been build-tested on all powerpc and arm boards involved. I have not tested NIOS since I don't have a toolchain for it installed, so I expect some breakage there even though I've tried fixing up everything I could find by visual inspection. I have run-time tested this on AVR32 ATNGW100 using the atmel_spi and DataFlash drivers posted as a follow-up. I'd like some help testing other boards that use the existing SPI API. But most of all, I'd like some comments on the new API. Is this stuff usable for everyone? If not, why? Changed in v4: - Build fixes for various boards, drivers and commands - Provide common struct spi_slave definition that can be extended by drivers - Pass a struct spi_slave * to spi_cs_activate and spi_cs_deactivate - Make default bus and mode build-time configurable - Override default SPI bus ID and mode on mx32ads and imx31_litekit. Changed in v3: - Add opaque struct spi_slave for controller-specific data associated with a slave. - Add spi_claim_bus() and spi_release_bus() - Add spi_free_slave() - spi_setup() is now called spi_setup_slave() and returns a struct spi_slave - soft_spi now supports four SPI modes (CPOL|CPHA) - Add bus parameter to spi_setup_slave() - Convert the new i.MX32 SPI driver - Convert the new MC13783 RTC driver Changed in v2: - Convert the mpc8xxx_spi driver and the mpc8349emds board to the new API. Signed-off-by: Haavard Skinnemoen <hskinnemoen@atmel.com> Tested-by: Guennadi Liakhovetski <lg@denx.de>
2008-05-16 09:10:31 +00:00
struct mxc_spi_slave *mxcs;
int ret;
if (bus >= ARRAY_SIZE(spi_bases))
return NULL;
if (max_hz == 0) {
printf("Error: desired clock is 0\n");
return NULL;
}
mxcs = spi_alloc_slave(struct mxc_spi_slave, bus, cs);
if (!mxcs) {
puts("mxc_spi: SPI Slave not allocated !\n");
return NULL;
}
mxcs->ss_pol = (mode & SPI_CS_HIGH) ? 1 : 0;
ret = setup_cs_gpio(mxcs, bus, cs);
if (ret < 0) {
free(mxcs);
SPI API improvements This patch gets rid of the spi_chipsel table and adds a handful of new functions that makes the SPI layer cleaner and more flexible. Instead of the spi_chipsel table, each board that wants to use SPI gets to implement three hooks: * spi_cs_activate(): Activates the chipselect for a given slave * spi_cs_deactivate(): Deactivates the chipselect for a given slave * spi_cs_is_valid(): Determines if the given bus/chipselect combination can be activated. Not all drivers may need those extra functions however. If that's the case, the board code may just leave them out (assuming they know what the driver needs) or rely on the linker to strip them out (assuming --gc-sections is being used.) To set up communication parameters for a given slave, the driver needs to call spi_setup_slave(). This returns a pointer to an opaque spi_slave struct which must be passed as a parameter to subsequent SPI calls. This struct can be freed by calling spi_free_slave(), but most driver probably don't want to do this. Before starting one or more SPI transfers, the driver must call spi_claim_bus() to gain exclusive access to the SPI bus and initialize the hardware. When all transfers are done, the driver must call spi_release_bus() to make the bus available to others, and possibly shut down the SPI controller hardware. spi_xfer() behaves mostly the same as before, but it now takes a spi_slave parameter instead of a spi_chipsel function pointer. It also got a new parameter, flags, which is used to specify chip select behaviour. This may be extended with other flags in the future. This patch has been build-tested on all powerpc and arm boards involved. I have not tested NIOS since I don't have a toolchain for it installed, so I expect some breakage there even though I've tried fixing up everything I could find by visual inspection. I have run-time tested this on AVR32 ATNGW100 using the atmel_spi and DataFlash drivers posted as a follow-up. I'd like some help testing other boards that use the existing SPI API. But most of all, I'd like some comments on the new API. Is this stuff usable for everyone? If not, why? Changed in v4: - Build fixes for various boards, drivers and commands - Provide common struct spi_slave definition that can be extended by drivers - Pass a struct spi_slave * to spi_cs_activate and spi_cs_deactivate - Make default bus and mode build-time configurable - Override default SPI bus ID and mode on mx32ads and imx31_litekit. Changed in v3: - Add opaque struct spi_slave for controller-specific data associated with a slave. - Add spi_claim_bus() and spi_release_bus() - Add spi_free_slave() - spi_setup() is now called spi_setup_slave() and returns a struct spi_slave - soft_spi now supports four SPI modes (CPOL|CPHA) - Add bus parameter to spi_setup_slave() - Convert the new i.MX32 SPI driver - Convert the new MC13783 RTC driver Changed in v2: - Convert the mpc8xxx_spi driver and the mpc8349emds board to the new API. Signed-off-by: Haavard Skinnemoen <hskinnemoen@atmel.com> Tested-by: Guennadi Liakhovetski <lg@denx.de>
2008-05-16 09:10:31 +00:00
return NULL;
}
mxcs->base = spi_bases[bus];
mxcs->max_hz = max_hz;
mxcs->mode = mode;
SPI API improvements This patch gets rid of the spi_chipsel table and adds a handful of new functions that makes the SPI layer cleaner and more flexible. Instead of the spi_chipsel table, each board that wants to use SPI gets to implement three hooks: * spi_cs_activate(): Activates the chipselect for a given slave * spi_cs_deactivate(): Deactivates the chipselect for a given slave * spi_cs_is_valid(): Determines if the given bus/chipselect combination can be activated. Not all drivers may need those extra functions however. If that's the case, the board code may just leave them out (assuming they know what the driver needs) or rely on the linker to strip them out (assuming --gc-sections is being used.) To set up communication parameters for a given slave, the driver needs to call spi_setup_slave(). This returns a pointer to an opaque spi_slave struct which must be passed as a parameter to subsequent SPI calls. This struct can be freed by calling spi_free_slave(), but most driver probably don't want to do this. Before starting one or more SPI transfers, the driver must call spi_claim_bus() to gain exclusive access to the SPI bus and initialize the hardware. When all transfers are done, the driver must call spi_release_bus() to make the bus available to others, and possibly shut down the SPI controller hardware. spi_xfer() behaves mostly the same as before, but it now takes a spi_slave parameter instead of a spi_chipsel function pointer. It also got a new parameter, flags, which is used to specify chip select behaviour. This may be extended with other flags in the future. This patch has been build-tested on all powerpc and arm boards involved. I have not tested NIOS since I don't have a toolchain for it installed, so I expect some breakage there even though I've tried fixing up everything I could find by visual inspection. I have run-time tested this on AVR32 ATNGW100 using the atmel_spi and DataFlash drivers posted as a follow-up. I'd like some help testing other boards that use the existing SPI API. But most of all, I'd like some comments on the new API. Is this stuff usable for everyone? If not, why? Changed in v4: - Build fixes for various boards, drivers and commands - Provide common struct spi_slave definition that can be extended by drivers - Pass a struct spi_slave * to spi_cs_activate and spi_cs_deactivate - Make default bus and mode build-time configurable - Override default SPI bus ID and mode on mx32ads and imx31_litekit. Changed in v3: - Add opaque struct spi_slave for controller-specific data associated with a slave. - Add spi_claim_bus() and spi_release_bus() - Add spi_free_slave() - spi_setup() is now called spi_setup_slave() and returns a struct spi_slave - soft_spi now supports four SPI modes (CPOL|CPHA) - Add bus parameter to spi_setup_slave() - Convert the new i.MX32 SPI driver - Convert the new MC13783 RTC driver Changed in v2: - Convert the mpc8xxx_spi driver and the mpc8349emds board to the new API. Signed-off-by: Haavard Skinnemoen <hskinnemoen@atmel.com> Tested-by: Guennadi Liakhovetski <lg@denx.de>
2008-05-16 09:10:31 +00:00
return &mxcs->slave;
}
void spi_free_slave(struct spi_slave *slave)
{
struct mxc_spi_slave *mxcs = to_mxc_spi_slave(slave);
free(mxcs);
SPI API improvements This patch gets rid of the spi_chipsel table and adds a handful of new functions that makes the SPI layer cleaner and more flexible. Instead of the spi_chipsel table, each board that wants to use SPI gets to implement three hooks: * spi_cs_activate(): Activates the chipselect for a given slave * spi_cs_deactivate(): Deactivates the chipselect for a given slave * spi_cs_is_valid(): Determines if the given bus/chipselect combination can be activated. Not all drivers may need those extra functions however. If that's the case, the board code may just leave them out (assuming they know what the driver needs) or rely on the linker to strip them out (assuming --gc-sections is being used.) To set up communication parameters for a given slave, the driver needs to call spi_setup_slave(). This returns a pointer to an opaque spi_slave struct which must be passed as a parameter to subsequent SPI calls. This struct can be freed by calling spi_free_slave(), but most driver probably don't want to do this. Before starting one or more SPI transfers, the driver must call spi_claim_bus() to gain exclusive access to the SPI bus and initialize the hardware. When all transfers are done, the driver must call spi_release_bus() to make the bus available to others, and possibly shut down the SPI controller hardware. spi_xfer() behaves mostly the same as before, but it now takes a spi_slave parameter instead of a spi_chipsel function pointer. It also got a new parameter, flags, which is used to specify chip select behaviour. This may be extended with other flags in the future. This patch has been build-tested on all powerpc and arm boards involved. I have not tested NIOS since I don't have a toolchain for it installed, so I expect some breakage there even though I've tried fixing up everything I could find by visual inspection. I have run-time tested this on AVR32 ATNGW100 using the atmel_spi and DataFlash drivers posted as a follow-up. I'd like some help testing other boards that use the existing SPI API. But most of all, I'd like some comments on the new API. Is this stuff usable for everyone? If not, why? Changed in v4: - Build fixes for various boards, drivers and commands - Provide common struct spi_slave definition that can be extended by drivers - Pass a struct spi_slave * to spi_cs_activate and spi_cs_deactivate - Make default bus and mode build-time configurable - Override default SPI bus ID and mode on mx32ads and imx31_litekit. Changed in v3: - Add opaque struct spi_slave for controller-specific data associated with a slave. - Add spi_claim_bus() and spi_release_bus() - Add spi_free_slave() - spi_setup() is now called spi_setup_slave() and returns a struct spi_slave - soft_spi now supports four SPI modes (CPOL|CPHA) - Add bus parameter to spi_setup_slave() - Convert the new i.MX32 SPI driver - Convert the new MC13783 RTC driver Changed in v2: - Convert the mpc8xxx_spi driver and the mpc8349emds board to the new API. Signed-off-by: Haavard Skinnemoen <hskinnemoen@atmel.com> Tested-by: Guennadi Liakhovetski <lg@denx.de>
2008-05-16 09:10:31 +00:00
}
int spi_claim_bus(struct spi_slave *slave)
{
struct mxc_spi_slave *mxcs = to_mxc_spi_slave(slave);
return mxc_spi_claim_bus_internal(mxcs, slave->cs);
}
void spi_release_bus(struct spi_slave *slave)
{
/* TODO: Shut the controller down */
}
#else
static int mxc_spi_probe(struct udevice *bus)
{
struct mxc_spi_slave *mxcs = dev_get_plat(bus);
int ret;
int i;
ret = gpio_request_list_by_name(bus, "cs-gpios", mxcs->cs_gpios,
ARRAY_SIZE(mxcs->cs_gpios), 0);
if (ret < 0) {
pr_err("Can't get %s gpios! Error: %d", bus->name, ret);
return ret;
}
for (i = 0; i < ARRAY_SIZE(mxcs->cs_gpios); i++) {
if (!dm_gpio_is_valid(&mxcs->cs_gpios[i]))
continue;
ret = dm_gpio_set_dir_flags(&mxcs->cs_gpios[i],
GPIOD_IS_OUT | GPIOD_ACTIVE_LOW);
if (ret) {
dev_err(bus, "Setting cs %d error\n", i);
return ret;
}
}
mxcs->base = dev_read_addr(bus);
if (mxcs->base == FDT_ADDR_T_NONE)
return -ENODEV;
#if CONFIG_IS_ENABLED(CLK)
struct clk clk;
ret = clk_get_by_index(bus, 0, &clk);
if (ret)
return ret;
clk_enable(&clk);
mxcs->max_hz = clk_get_rate(&clk);
#else
int node = dev_of_offset(bus);
const void *blob = gd->fdt_blob;
mxcs->max_hz = fdtdec_get_int(blob, node, "spi-max-frequency",
20000000);
#endif
return 0;
}
SPI API improvements This patch gets rid of the spi_chipsel table and adds a handful of new functions that makes the SPI layer cleaner and more flexible. Instead of the spi_chipsel table, each board that wants to use SPI gets to implement three hooks: * spi_cs_activate(): Activates the chipselect for a given slave * spi_cs_deactivate(): Deactivates the chipselect for a given slave * spi_cs_is_valid(): Determines if the given bus/chipselect combination can be activated. Not all drivers may need those extra functions however. If that's the case, the board code may just leave them out (assuming they know what the driver needs) or rely on the linker to strip them out (assuming --gc-sections is being used.) To set up communication parameters for a given slave, the driver needs to call spi_setup_slave(). This returns a pointer to an opaque spi_slave struct which must be passed as a parameter to subsequent SPI calls. This struct can be freed by calling spi_free_slave(), but most driver probably don't want to do this. Before starting one or more SPI transfers, the driver must call spi_claim_bus() to gain exclusive access to the SPI bus and initialize the hardware. When all transfers are done, the driver must call spi_release_bus() to make the bus available to others, and possibly shut down the SPI controller hardware. spi_xfer() behaves mostly the same as before, but it now takes a spi_slave parameter instead of a spi_chipsel function pointer. It also got a new parameter, flags, which is used to specify chip select behaviour. This may be extended with other flags in the future. This patch has been build-tested on all powerpc and arm boards involved. I have not tested NIOS since I don't have a toolchain for it installed, so I expect some breakage there even though I've tried fixing up everything I could find by visual inspection. I have run-time tested this on AVR32 ATNGW100 using the atmel_spi and DataFlash drivers posted as a follow-up. I'd like some help testing other boards that use the existing SPI API. But most of all, I'd like some comments on the new API. Is this stuff usable for everyone? If not, why? Changed in v4: - Build fixes for various boards, drivers and commands - Provide common struct spi_slave definition that can be extended by drivers - Pass a struct spi_slave * to spi_cs_activate and spi_cs_deactivate - Make default bus and mode build-time configurable - Override default SPI bus ID and mode on mx32ads and imx31_litekit. Changed in v3: - Add opaque struct spi_slave for controller-specific data associated with a slave. - Add spi_claim_bus() and spi_release_bus() - Add spi_free_slave() - spi_setup() is now called spi_setup_slave() and returns a struct spi_slave - soft_spi now supports four SPI modes (CPOL|CPHA) - Add bus parameter to spi_setup_slave() - Convert the new i.MX32 SPI driver - Convert the new MC13783 RTC driver Changed in v2: - Convert the mpc8xxx_spi driver and the mpc8349emds board to the new API. Signed-off-by: Haavard Skinnemoen <hskinnemoen@atmel.com> Tested-by: Guennadi Liakhovetski <lg@denx.de>
2008-05-16 09:10:31 +00:00
static int mxc_spi_xfer(struct udevice *dev, unsigned int bitlen,
const void *dout, void *din, unsigned long flags)
SPI API improvements This patch gets rid of the spi_chipsel table and adds a handful of new functions that makes the SPI layer cleaner and more flexible. Instead of the spi_chipsel table, each board that wants to use SPI gets to implement three hooks: * spi_cs_activate(): Activates the chipselect for a given slave * spi_cs_deactivate(): Deactivates the chipselect for a given slave * spi_cs_is_valid(): Determines if the given bus/chipselect combination can be activated. Not all drivers may need those extra functions however. If that's the case, the board code may just leave them out (assuming they know what the driver needs) or rely on the linker to strip them out (assuming --gc-sections is being used.) To set up communication parameters for a given slave, the driver needs to call spi_setup_slave(). This returns a pointer to an opaque spi_slave struct which must be passed as a parameter to subsequent SPI calls. This struct can be freed by calling spi_free_slave(), but most driver probably don't want to do this. Before starting one or more SPI transfers, the driver must call spi_claim_bus() to gain exclusive access to the SPI bus and initialize the hardware. When all transfers are done, the driver must call spi_release_bus() to make the bus available to others, and possibly shut down the SPI controller hardware. spi_xfer() behaves mostly the same as before, but it now takes a spi_slave parameter instead of a spi_chipsel function pointer. It also got a new parameter, flags, which is used to specify chip select behaviour. This may be extended with other flags in the future. This patch has been build-tested on all powerpc and arm boards involved. I have not tested NIOS since I don't have a toolchain for it installed, so I expect some breakage there even though I've tried fixing up everything I could find by visual inspection. I have run-time tested this on AVR32 ATNGW100 using the atmel_spi and DataFlash drivers posted as a follow-up. I'd like some help testing other boards that use the existing SPI API. But most of all, I'd like some comments on the new API. Is this stuff usable for everyone? If not, why? Changed in v4: - Build fixes for various boards, drivers and commands - Provide common struct spi_slave definition that can be extended by drivers - Pass a struct spi_slave * to spi_cs_activate and spi_cs_deactivate - Make default bus and mode build-time configurable - Override default SPI bus ID and mode on mx32ads and imx31_litekit. Changed in v3: - Add opaque struct spi_slave for controller-specific data associated with a slave. - Add spi_claim_bus() and spi_release_bus() - Add spi_free_slave() - spi_setup() is now called spi_setup_slave() and returns a struct spi_slave - soft_spi now supports four SPI modes (CPOL|CPHA) - Add bus parameter to spi_setup_slave() - Convert the new i.MX32 SPI driver - Convert the new MC13783 RTC driver Changed in v2: - Convert the mpc8xxx_spi driver and the mpc8349emds board to the new API. Signed-off-by: Haavard Skinnemoen <hskinnemoen@atmel.com> Tested-by: Guennadi Liakhovetski <lg@denx.de>
2008-05-16 09:10:31 +00:00
{
struct mxc_spi_slave *mxcs = dev_get_plat(dev->parent);
return mxc_spi_xfer_internal(mxcs, bitlen, dout, din, flags);
}
static int mxc_spi_claim_bus(struct udevice *dev)
{
struct mxc_spi_slave *mxcs = dev_get_plat(dev->parent);
struct dm_spi_slave_plat *slave_plat = dev_get_parent_plat(dev);
mxcs->dev = dev;
return mxc_spi_claim_bus_internal(mxcs, slave_plat->cs);
SPI API improvements This patch gets rid of the spi_chipsel table and adds a handful of new functions that makes the SPI layer cleaner and more flexible. Instead of the spi_chipsel table, each board that wants to use SPI gets to implement three hooks: * spi_cs_activate(): Activates the chipselect for a given slave * spi_cs_deactivate(): Deactivates the chipselect for a given slave * spi_cs_is_valid(): Determines if the given bus/chipselect combination can be activated. Not all drivers may need those extra functions however. If that's the case, the board code may just leave them out (assuming they know what the driver needs) or rely on the linker to strip them out (assuming --gc-sections is being used.) To set up communication parameters for a given slave, the driver needs to call spi_setup_slave(). This returns a pointer to an opaque spi_slave struct which must be passed as a parameter to subsequent SPI calls. This struct can be freed by calling spi_free_slave(), but most driver probably don't want to do this. Before starting one or more SPI transfers, the driver must call spi_claim_bus() to gain exclusive access to the SPI bus and initialize the hardware. When all transfers are done, the driver must call spi_release_bus() to make the bus available to others, and possibly shut down the SPI controller hardware. spi_xfer() behaves mostly the same as before, but it now takes a spi_slave parameter instead of a spi_chipsel function pointer. It also got a new parameter, flags, which is used to specify chip select behaviour. This may be extended with other flags in the future. This patch has been build-tested on all powerpc and arm boards involved. I have not tested NIOS since I don't have a toolchain for it installed, so I expect some breakage there even though I've tried fixing up everything I could find by visual inspection. I have run-time tested this on AVR32 ATNGW100 using the atmel_spi and DataFlash drivers posted as a follow-up. I'd like some help testing other boards that use the existing SPI API. But most of all, I'd like some comments on the new API. Is this stuff usable for everyone? If not, why? Changed in v4: - Build fixes for various boards, drivers and commands - Provide common struct spi_slave definition that can be extended by drivers - Pass a struct spi_slave * to spi_cs_activate and spi_cs_deactivate - Make default bus and mode build-time configurable - Override default SPI bus ID and mode on mx32ads and imx31_litekit. Changed in v3: - Add opaque struct spi_slave for controller-specific data associated with a slave. - Add spi_claim_bus() and spi_release_bus() - Add spi_free_slave() - spi_setup() is now called spi_setup_slave() and returns a struct spi_slave - soft_spi now supports four SPI modes (CPOL|CPHA) - Add bus parameter to spi_setup_slave() - Convert the new i.MX32 SPI driver - Convert the new MC13783 RTC driver Changed in v2: - Convert the mpc8xxx_spi driver and the mpc8349emds board to the new API. Signed-off-by: Haavard Skinnemoen <hskinnemoen@atmel.com> Tested-by: Guennadi Liakhovetski <lg@denx.de>
2008-05-16 09:10:31 +00:00
}
static int mxc_spi_release_bus(struct udevice *dev)
{
return 0;
}
static int mxc_spi_set_speed(struct udevice *bus, uint speed)
{
struct mxc_spi_slave *mxcs = dev_get_plat(bus);
mxcs->max_hz = speed;
return 0;
}
static int mxc_spi_set_mode(struct udevice *bus, uint mode)
{
struct mxc_spi_slave *mxcs = dev_get_plat(bus);
mxcs->mode = mode;
mxcs->ss_pol = (mode & SPI_CS_HIGH) ? 1 : 0;
return 0;
}
static const struct dm_spi_ops mxc_spi_ops = {
.claim_bus = mxc_spi_claim_bus,
.release_bus = mxc_spi_release_bus,
.xfer = mxc_spi_xfer,
.set_speed = mxc_spi_set_speed,
.set_mode = mxc_spi_set_mode,
};
static const struct udevice_id mxc_spi_ids[] = {
{ .compatible = "fsl,imx51-ecspi" },
{ }
};
U_BOOT_DRIVER(mxc_spi) = {
.name = "mxc_spi",
.id = UCLASS_SPI,
.of_match = mxc_spi_ids,
.ops = &mxc_spi_ops,
.plat_auto = sizeof(struct mxc_spi_slave),
.probe = mxc_spi_probe,
};
#endif