summaryrefslogtreecommitdiffstats
path: root/u-boot/drivers/spi/bfin_spi.c
diff options
context:
space:
mode:
Diffstat (limited to 'u-boot/drivers/spi/bfin_spi.c')
-rw-r--r--u-boot/drivers/spi/bfin_spi.c401
1 files changed, 401 insertions, 0 deletions
diff --git a/u-boot/drivers/spi/bfin_spi.c b/u-boot/drivers/spi/bfin_spi.c
new file mode 100644
index 0000000..d7e1474
--- /dev/null
+++ b/u-boot/drivers/spi/bfin_spi.c
@@ -0,0 +1,401 @@
+/*
+ * Driver for Blackfin On-Chip SPI device
+ *
+ * Copyright (c) 2005-2010 Analog Devices Inc.
+ *
+ * Licensed under the GPL-2 or later.
+ */
+
+/*#define DEBUG*/
+
+#include <common.h>
+#include <malloc.h>
+#include <spi.h>
+
+#include <asm/blackfin.h>
+#include <asm/dma.h>
+#include <asm/gpio.h>
+#include <asm/portmux.h>
+#include <asm/mach-common/bits/spi.h>
+
+struct bfin_spi_slave {
+ struct spi_slave slave;
+ void *mmr_base;
+ u16 ctl, baud, flg;
+};
+
+#define MAKE_SPI_FUNC(mmr, off) \
+static inline void write_##mmr(struct bfin_spi_slave *bss, u16 val) { bfin_write16(bss->mmr_base + off, val); } \
+static inline u16 read_##mmr(struct bfin_spi_slave *bss) { return bfin_read16(bss->mmr_base + off); }
+MAKE_SPI_FUNC(SPI_CTL, 0x00)
+MAKE_SPI_FUNC(SPI_FLG, 0x04)
+MAKE_SPI_FUNC(SPI_STAT, 0x08)
+MAKE_SPI_FUNC(SPI_TDBR, 0x0c)
+MAKE_SPI_FUNC(SPI_RDBR, 0x10)
+MAKE_SPI_FUNC(SPI_BAUD, 0x14)
+
+#define to_bfin_spi_slave(s) container_of(s, struct bfin_spi_slave, slave)
+
+#define gpio_cs(cs) ((cs) - MAX_CTRL_CS)
+#ifdef CONFIG_BFIN_SPI_GPIO_CS
+# define is_gpio_cs(cs) ((cs) > MAX_CTRL_CS)
+#else
+# define is_gpio_cs(cs) 0
+#endif
+
+int spi_cs_is_valid(unsigned int bus, unsigned int cs)
+{
+ if (is_gpio_cs(cs))
+ return gpio_is_valid(gpio_cs(cs));
+ else
+ return (cs >= 1 && cs <= MAX_CTRL_CS);
+}
+
+void spi_cs_activate(struct spi_slave *slave)
+{
+ struct bfin_spi_slave *bss = to_bfin_spi_slave(slave);
+
+ if (is_gpio_cs(slave->cs)) {
+ unsigned int cs = gpio_cs(slave->cs);
+ gpio_set_value(cs, bss->flg);
+ debug("%s: SPI_CS_GPIO:%x\n", __func__, gpio_get_value(cs));
+ } else {
+ write_SPI_FLG(bss,
+ (read_SPI_FLG(bss) &
+ ~((!bss->flg << 8) << slave->cs)) |
+ (1 << slave->cs));
+ debug("%s: SPI_FLG:%x\n", __func__, read_SPI_FLG(bss));
+ }
+
+ SSYNC();
+}
+
+void spi_cs_deactivate(struct spi_slave *slave)
+{
+ struct bfin_spi_slave *bss = to_bfin_spi_slave(slave);
+
+ if (is_gpio_cs(slave->cs)) {
+ unsigned int cs = gpio_cs(slave->cs);
+ gpio_set_value(cs, !bss->flg);
+ debug("%s: SPI_CS_GPIO:%x\n", __func__, gpio_get_value(cs));
+ } else {
+ u16 flg;
+
+ /* make sure we force the cs to deassert rather than let the
+ * pin float back up. otherwise, exact timings may not be
+ * met some of the time leading to random behavior (ugh).
+ */
+ flg = read_SPI_FLG(bss) | ((!bss->flg << 8) << slave->cs);
+ write_SPI_FLG(bss, flg);
+ SSYNC();
+ debug("%s: SPI_FLG:%x\n", __func__, read_SPI_FLG(bss));
+
+ flg &= ~(1 << slave->cs);
+ write_SPI_FLG(bss, flg);
+ debug("%s: SPI_FLG:%x\n", __func__, read_SPI_FLG(bss));
+ }
+
+ SSYNC();
+}
+
+void spi_init()
+{
+}
+
+#ifdef SPI_CTL
+# define SPI0_CTL SPI_CTL
+#endif
+
+#define SPI_PINS(n) \
+ [n] = { 0, P_SPI##n##_SCK, P_SPI##n##_MISO, P_SPI##n##_MOSI, 0 }
+static unsigned short pins[][5] = {
+#ifdef SPI0_CTL
+ SPI_PINS(0),
+#endif
+#ifdef SPI1_CTL
+ SPI_PINS(1),
+#endif
+#ifdef SPI2_CTL
+ SPI_PINS(2),
+#endif
+};
+
+#define SPI_CS_PINS(n) \
+ [n] = { \
+ P_SPI##n##_SSEL1, P_SPI##n##_SSEL2, P_SPI##n##_SSEL3, \
+ P_SPI##n##_SSEL4, P_SPI##n##_SSEL5, P_SPI##n##_SSEL6, \
+ P_SPI##n##_SSEL7, \
+ }
+static const unsigned short cs_pins[][7] = {
+#ifdef SPI0_CTL
+ SPI_CS_PINS(0),
+#endif
+#ifdef SPI1_CTL
+ SPI_CS_PINS(1),
+#endif
+#ifdef SPI2_CTL
+ SPI_CS_PINS(2),
+#endif
+};
+
+struct spi_slave *spi_setup_slave(unsigned int bus, unsigned int cs,
+ unsigned int max_hz, unsigned int mode)
+{
+ struct bfin_spi_slave *bss;
+ ulong sclk;
+ u32 mmr_base;
+ u32 baud;
+
+ if (!spi_cs_is_valid(bus, cs))
+ return NULL;
+
+ if (bus >= ARRAY_SIZE(pins) || pins[bus] == NULL) {
+ debug("%s: invalid bus %u\n", __func__, bus);
+ return NULL;
+ }
+ switch (bus) {
+#ifdef SPI0_CTL
+ case 0: mmr_base = SPI0_CTL; break;
+#endif
+#ifdef SPI1_CTL
+ case 1: mmr_base = SPI1_CTL; break;
+#endif
+#ifdef SPI2_CTL
+ case 2: mmr_base = SPI2_CTL; break;
+#endif
+ default: return NULL;
+ }
+
+ sclk = get_sclk();
+ baud = sclk / (2 * max_hz);
+ /* baud should be rounded up */
+ if (sclk % (2 * max_hz))
+ baud += 1;
+ if (baud < 2)
+ baud = 2;
+ else if (baud > (u16)-1)
+ baud = -1;
+
+ bss = malloc(sizeof(*bss));
+ if (!bss)
+ return NULL;
+
+ bss->slave.bus = bus;
+ bss->slave.cs = cs;
+ bss->mmr_base = (void *)mmr_base;
+ bss->ctl = SPE | MSTR | TDBR_CORE;
+ if (mode & SPI_CPHA) bss->ctl |= CPHA;
+ if (mode & SPI_CPOL) bss->ctl |= CPOL;
+ if (mode & SPI_LSB_FIRST) bss->ctl |= LSBF;
+ bss->baud = baud;
+ bss->flg = mode & SPI_CS_HIGH ? 1 : 0;
+
+ debug("%s: bus:%i cs:%i mmr:%x ctl:%x baud:%i flg:%i\n", __func__,
+ bus, cs, mmr_base, bss->ctl, baud, bss->flg);
+
+ return &bss->slave;
+}
+
+void spi_free_slave(struct spi_slave *slave)
+{
+ struct bfin_spi_slave *bss = to_bfin_spi_slave(slave);
+ free(bss);
+}
+
+int spi_claim_bus(struct spi_slave *slave)
+{
+ struct bfin_spi_slave *bss = to_bfin_spi_slave(slave);
+
+ debug("%s: bus:%i cs:%i\n", __func__, slave->bus, slave->cs);
+
+ if (is_gpio_cs(slave->cs)) {
+ unsigned int cs = gpio_cs(slave->cs);
+ gpio_request(cs, "bfin-spi");
+ gpio_direction_output(cs, !bss->flg);
+ pins[slave->bus][0] = P_DONTCARE;
+ } else
+ pins[slave->bus][0] = cs_pins[slave->bus][slave->cs - 1];
+ peripheral_request_list(pins[slave->bus], "bfin-spi");
+
+ write_SPI_CTL(bss, bss->ctl);
+ write_SPI_BAUD(bss, bss->baud);
+ SSYNC();
+
+ return 0;
+}
+
+void spi_release_bus(struct spi_slave *slave)
+{
+ struct bfin_spi_slave *bss = to_bfin_spi_slave(slave);
+
+ debug("%s: bus:%i cs:%i\n", __func__, slave->bus, slave->cs);
+
+ peripheral_free_list(pins[slave->bus]);
+ if (is_gpio_cs(slave->cs))
+ gpio_free(gpio_cs(slave->cs));
+
+ write_SPI_CTL(bss, 0);
+ SSYNC();
+}
+
+#ifdef __ADSPBF54x__
+# define SPI_DMA_BASE DMA4_NEXT_DESC_PTR
+#elif defined(__ADSPBF533__) || defined(__ADSPBF532__) || defined(__ADSPBF531__) || \
+ defined(__ADSPBF538__) || defined(__ADSPBF539__)
+# define SPI_DMA_BASE DMA5_NEXT_DESC_PTR
+#elif defined(__ADSPBF561__)
+# define SPI_DMA_BASE DMA2_4_NEXT_DESC_PTR
+#elif defined(__ADSPBF537__) || defined(__ADSPBF536__) || defined(__ADSPBF534__) || \
+ defined(__ADSPBF52x__) || defined(__ADSPBF51x__)
+# define SPI_DMA_BASE DMA7_NEXT_DESC_PTR
+#else
+# error "Please provide SPI DMA channel defines"
+#endif
+static volatile struct dma_register *dma = (void *)SPI_DMA_BASE;
+
+#ifndef CONFIG_BFIN_SPI_IDLE_VAL
+# define CONFIG_BFIN_SPI_IDLE_VAL 0xff
+#endif
+
+#ifdef CONFIG_BFIN_SPI_NO_DMA
+# define SPI_DMA 0
+#else
+# define SPI_DMA 1
+#endif
+
+static int spi_dma_xfer(struct bfin_spi_slave *bss, const u8 *tx, u8 *rx,
+ uint bytes)
+{
+ int ret = -1;
+ u16 ndsize, spi_config, dma_config;
+ struct dmasg dmasg[2];
+ const u8 *buf;
+
+ if (tx) {
+ debug("%s: doing half duplex TX\n", __func__);
+ buf = tx;
+ spi_config = TDBR_DMA;
+ dma_config = 0;
+ } else {
+ debug("%s: doing half duplex RX\n", __func__);
+ buf = rx;
+ spi_config = RDBR_DMA;
+ dma_config = WNR;
+ }
+
+ dmasg[0].start_addr = (unsigned long)buf;
+ dmasg[0].x_modify = 1;
+ dma_config |= WDSIZE_8 | DMAEN;
+ if (bytes <= 65536) {
+ blackfin_dcache_flush_invalidate_range(buf, buf + bytes);
+ ndsize = NDSIZE_5;
+ dmasg[0].cfg = NDSIZE_0 | dma_config | FLOW_STOP | DI_EN;
+ dmasg[0].x_count = bytes;
+ } else {
+ blackfin_dcache_flush_invalidate_range(buf, buf + 65536 - 1);
+ ndsize = NDSIZE_7;
+ dmasg[0].cfg = NDSIZE_5 | dma_config | FLOW_ARRAY | DMA2D;
+ dmasg[0].x_count = 0; /* 2^16 */
+ dmasg[0].y_count = bytes >> 16; /* count / 2^16 */
+ dmasg[0].y_modify = 1;
+ dmasg[1].start_addr = (unsigned long)(buf + (bytes & ~0xFFFF));
+ dmasg[1].cfg = NDSIZE_0 | dma_config | FLOW_STOP | DI_EN;
+ dmasg[1].x_count = bytes & 0xFFFF; /* count % 2^16 */
+ dmasg[1].x_modify = 1;
+ }
+
+ dma->cfg = 0;
+ dma->irq_status = DMA_DONE | DMA_ERR;
+ dma->curr_desc_ptr = dmasg;
+ write_SPI_CTL(bss, (bss->ctl & ~TDBR_CORE));
+ write_SPI_STAT(bss, -1);
+ SSYNC();
+
+ write_SPI_TDBR(bss, CONFIG_BFIN_SPI_IDLE_VAL);
+ dma->cfg = ndsize | FLOW_ARRAY | DMAEN;
+ write_SPI_CTL(bss, (bss->ctl & ~TDBR_CORE) | spi_config);
+ SSYNC();
+
+ /*
+ * We already invalidated the first 64k,
+ * now while we just wait invalidate the remaining part.
+ * Its not likely that the DMA is going to overtake
+ */
+ if (bytes > 65536)
+ blackfin_dcache_flush_invalidate_range(buf + 65536, buf + bytes);
+
+ while (!(dma->irq_status & DMA_DONE))
+ if (ctrlc())
+ goto done;
+
+ dma->cfg = 0;
+
+ ret = 0;
+ done:
+ write_SPI_CTL(bss, bss->ctl);
+ return ret;
+}
+
+static int spi_pio_xfer(struct bfin_spi_slave *bss, const u8 *tx, u8 *rx,
+ uint bytes)
+{
+ /* todo: take advantage of hardware fifos */
+ while (bytes--) {
+ u8 value = (tx ? *tx++ : CONFIG_BFIN_SPI_IDLE_VAL);
+ debug("%s: tx:%x ", __func__, value);
+ write_SPI_TDBR(bss, value);
+ SSYNC();
+ while ((read_SPI_STAT(bss) & TXS))
+ if (ctrlc())
+ return -1;
+ while (!(read_SPI_STAT(bss) & SPIF))
+ if (ctrlc())
+ return -1;
+ while (!(read_SPI_STAT(bss) & RXS))
+ if (ctrlc())
+ return -1;
+ value = read_SPI_RDBR(bss);
+ if (rx)
+ *rx++ = value;
+ debug("rx:%x\n", value);
+ }
+
+ return 0;
+}
+
+int spi_xfer(struct spi_slave *slave, unsigned int bitlen, const void *dout,
+ void *din, unsigned long flags)
+{
+ struct bfin_spi_slave *bss = to_bfin_spi_slave(slave);
+ const u8 *tx = dout;
+ u8 *rx = din;
+ uint bytes = bitlen / 8;
+ int ret = 0;
+
+ debug("%s: bus:%i cs:%i bitlen:%i bytes:%i flags:%lx\n", __func__,
+ slave->bus, slave->cs, bitlen, bytes, flags);
+
+ if (bitlen == 0)
+ goto done;
+
+ /* we can only do 8 bit transfers */
+ if (bitlen % 8) {
+ flags |= SPI_XFER_END;
+ goto done;
+ }
+
+ if (flags & SPI_XFER_BEGIN)
+ spi_cs_activate(slave);
+
+ /* TX DMA doesn't work quite right */
+ if (SPI_DMA && bytes > 6 && (!tx /*|| !rx*/))
+ ret = spi_dma_xfer(bss, tx, rx, bytes);
+ else
+ ret = spi_pio_xfer(bss, tx, rx, bytes);
+
+ done:
+ if (flags & SPI_XFER_END)
+ spi_cs_deactivate(slave);
+
+ return ret;
+}