v4.19.13 snapshot.
diff --git a/drivers/dma/ti/Kconfig b/drivers/dma/ti/Kconfig
new file mode 100644
index 0000000..e5e74e1
--- /dev/null
+++ b/drivers/dma/ti/Kconfig
@@ -0,0 +1,37 @@
+#
+# Texas Instruments DMA drivers
+#
+
+config TI_CPPI41
+ tristate "Texas Instruments CPPI 4.1 DMA support"
+ depends on (ARCH_OMAP || ARCH_DAVINCI_DA8XX)
+ select DMA_ENGINE
+ help
+ The Communications Port Programming Interface (CPPI) 4.1 DMA engine
+ is currently used by the USB driver on AM335x and DA8xx platforms.
+
+config TI_EDMA
+ tristate "Texas Instruments EDMA support"
+ depends on ARCH_DAVINCI || ARCH_OMAP || ARCH_KEYSTONE || COMPILE_TEST
+ select DMA_ENGINE
+ select DMA_VIRTUAL_CHANNELS
+ select TI_DMA_CROSSBAR if (ARCH_OMAP || COMPILE_TEST)
+ default y
+ help
+ Enable support for the TI EDMA (Enhanced DMA) controller. This DMA
+ engine is found on TI DaVinci, AM33xx, AM43xx, DRA7xx and Keystone 2
+ parts.
+
+config DMA_OMAP
+ tristate "Texas Instruments sDMA (omap-dma) support"
+ depends on ARCH_OMAP || COMPILE_TEST
+ select DMA_ENGINE
+ select DMA_VIRTUAL_CHANNELS
+ select TI_DMA_CROSSBAR if (SOC_DRA7XX || COMPILE_TEST)
+ default y
+ help
+ Enable support for the TI sDMA (System DMA or DMA4) controller. This
+ DMA engine is found on OMAP and DRA7xx parts.
+
+config TI_DMA_CROSSBAR
+ bool
diff --git a/drivers/dma/ti/Makefile b/drivers/dma/ti/Makefile
new file mode 100644
index 0000000..113e59e
--- /dev/null
+++ b/drivers/dma/ti/Makefile
@@ -0,0 +1,5 @@
+# SPDX-License-Identifier: GPL-2.0
+obj-$(CONFIG_TI_CPPI41) += cppi41.o
+obj-$(CONFIG_TI_EDMA) += edma.o
+obj-$(CONFIG_DMA_OMAP) += omap-dma.o
+obj-$(CONFIG_TI_DMA_CROSSBAR) += dma-crossbar.o
diff --git a/drivers/dma/ti/cppi41.c b/drivers/dma/ti/cppi41.c
new file mode 100644
index 0000000..e507ec3
--- /dev/null
+++ b/drivers/dma/ti/cppi41.c
@@ -0,0 +1,1237 @@
+#include <linux/delay.h>
+#include <linux/dmaengine.h>
+#include <linux/dma-mapping.h>
+#include <linux/platform_device.h>
+#include <linux/module.h>
+#include <linux/of.h>
+#include <linux/slab.h>
+#include <linux/of_dma.h>
+#include <linux/of_irq.h>
+#include <linux/dmapool.h>
+#include <linux/interrupt.h>
+#include <linux/of_address.h>
+#include <linux/pm_runtime.h>
+#include "../dmaengine.h"
+
+#define DESC_TYPE 27
+#define DESC_TYPE_HOST 0x10
+#define DESC_TYPE_TEARD 0x13
+
+#define TD_DESC_IS_RX (1 << 16)
+#define TD_DESC_DMA_NUM 10
+
+#define DESC_LENGTH_BITS_NUM 21
+
+#define DESC_TYPE_USB (5 << 26)
+#define DESC_PD_COMPLETE (1 << 31)
+
+/* DMA engine */
+#define DMA_TDFDQ 4
+#define DMA_TXGCR(x) (0x800 + (x) * 0x20)
+#define DMA_RXGCR(x) (0x808 + (x) * 0x20)
+#define RXHPCRA0 4
+
+#define GCR_CHAN_ENABLE (1 << 31)
+#define GCR_TEARDOWN (1 << 30)
+#define GCR_STARV_RETRY (1 << 24)
+#define GCR_DESC_TYPE_HOST (1 << 14)
+
+/* DMA scheduler */
+#define DMA_SCHED_CTRL 0
+#define DMA_SCHED_CTRL_EN (1 << 31)
+#define DMA_SCHED_WORD(x) ((x) * 4 + 0x800)
+
+#define SCHED_ENTRY0_CHAN(x) ((x) << 0)
+#define SCHED_ENTRY0_IS_RX (1 << 7)
+
+#define SCHED_ENTRY1_CHAN(x) ((x) << 8)
+#define SCHED_ENTRY1_IS_RX (1 << 15)
+
+#define SCHED_ENTRY2_CHAN(x) ((x) << 16)
+#define SCHED_ENTRY2_IS_RX (1 << 23)
+
+#define SCHED_ENTRY3_CHAN(x) ((x) << 24)
+#define SCHED_ENTRY3_IS_RX (1 << 31)
+
+/* Queue manager */
+/* 4 KiB of memory for descriptors, 2 for each endpoint */
+#define ALLOC_DECS_NUM 128
+#define DESCS_AREAS 1
+#define TOTAL_DESCS_NUM (ALLOC_DECS_NUM * DESCS_AREAS)
+#define QMGR_SCRATCH_SIZE (TOTAL_DESCS_NUM * 4)
+
+#define QMGR_LRAM0_BASE 0x80
+#define QMGR_LRAM_SIZE 0x84
+#define QMGR_LRAM1_BASE 0x88
+#define QMGR_MEMBASE(x) (0x1000 + (x) * 0x10)
+#define QMGR_MEMCTRL(x) (0x1004 + (x) * 0x10)
+#define QMGR_MEMCTRL_IDX_SH 16
+#define QMGR_MEMCTRL_DESC_SH 8
+
+#define QMGR_PEND(x) (0x90 + (x) * 4)
+
+#define QMGR_PENDING_SLOT_Q(x) (x / 32)
+#define QMGR_PENDING_BIT_Q(x) (x % 32)
+
+#define QMGR_QUEUE_A(n) (0x2000 + (n) * 0x10)
+#define QMGR_QUEUE_B(n) (0x2004 + (n) * 0x10)
+#define QMGR_QUEUE_C(n) (0x2008 + (n) * 0x10)
+#define QMGR_QUEUE_D(n) (0x200c + (n) * 0x10)
+
+/* Packet Descriptor */
+#define PD2_ZERO_LENGTH (1 << 19)
+
+struct cppi41_channel {
+ struct dma_chan chan;
+ struct dma_async_tx_descriptor txd;
+ struct cppi41_dd *cdd;
+ struct cppi41_desc *desc;
+ dma_addr_t desc_phys;
+ void __iomem *gcr_reg;
+ int is_tx;
+ u32 residue;
+
+ unsigned int q_num;
+ unsigned int q_comp_num;
+ unsigned int port_num;
+
+ unsigned td_retry;
+ unsigned td_queued:1;
+ unsigned td_seen:1;
+ unsigned td_desc_seen:1;
+
+ struct list_head node; /* Node for pending list */
+};
+
+struct cppi41_desc {
+ u32 pd0;
+ u32 pd1;
+ u32 pd2;
+ u32 pd3;
+ u32 pd4;
+ u32 pd5;
+ u32 pd6;
+ u32 pd7;
+} __aligned(32);
+
+struct chan_queues {
+ u16 submit;
+ u16 complete;
+};
+
+struct cppi41_dd {
+ struct dma_device ddev;
+
+ void *qmgr_scratch;
+ dma_addr_t scratch_phys;
+
+ struct cppi41_desc *cd;
+ dma_addr_t descs_phys;
+ u32 first_td_desc;
+ struct cppi41_channel *chan_busy[ALLOC_DECS_NUM];
+
+ void __iomem *ctrl_mem;
+ void __iomem *sched_mem;
+ void __iomem *qmgr_mem;
+ unsigned int irq;
+ const struct chan_queues *queues_rx;
+ const struct chan_queues *queues_tx;
+ struct chan_queues td_queue;
+ u16 first_completion_queue;
+ u16 qmgr_num_pend;
+ u32 n_chans;
+ u8 platform;
+
+ struct list_head pending; /* Pending queued transfers */
+ spinlock_t lock; /* Lock for pending list */
+
+ /* context for suspend/resume */
+ unsigned int dma_tdfdq;
+
+ bool is_suspended;
+};
+
+static struct chan_queues am335x_usb_queues_tx[] = {
+ /* USB0 ENDP 1 */
+ [ 0] = { .submit = 32, .complete = 93},
+ [ 1] = { .submit = 34, .complete = 94},
+ [ 2] = { .submit = 36, .complete = 95},
+ [ 3] = { .submit = 38, .complete = 96},
+ [ 4] = { .submit = 40, .complete = 97},
+ [ 5] = { .submit = 42, .complete = 98},
+ [ 6] = { .submit = 44, .complete = 99},
+ [ 7] = { .submit = 46, .complete = 100},
+ [ 8] = { .submit = 48, .complete = 101},
+ [ 9] = { .submit = 50, .complete = 102},
+ [10] = { .submit = 52, .complete = 103},
+ [11] = { .submit = 54, .complete = 104},
+ [12] = { .submit = 56, .complete = 105},
+ [13] = { .submit = 58, .complete = 106},
+ [14] = { .submit = 60, .complete = 107},
+
+ /* USB1 ENDP1 */
+ [15] = { .submit = 62, .complete = 125},
+ [16] = { .submit = 64, .complete = 126},
+ [17] = { .submit = 66, .complete = 127},
+ [18] = { .submit = 68, .complete = 128},
+ [19] = { .submit = 70, .complete = 129},
+ [20] = { .submit = 72, .complete = 130},
+ [21] = { .submit = 74, .complete = 131},
+ [22] = { .submit = 76, .complete = 132},
+ [23] = { .submit = 78, .complete = 133},
+ [24] = { .submit = 80, .complete = 134},
+ [25] = { .submit = 82, .complete = 135},
+ [26] = { .submit = 84, .complete = 136},
+ [27] = { .submit = 86, .complete = 137},
+ [28] = { .submit = 88, .complete = 138},
+ [29] = { .submit = 90, .complete = 139},
+};
+
+static const struct chan_queues am335x_usb_queues_rx[] = {
+ /* USB0 ENDP 1 */
+ [ 0] = { .submit = 1, .complete = 109},
+ [ 1] = { .submit = 2, .complete = 110},
+ [ 2] = { .submit = 3, .complete = 111},
+ [ 3] = { .submit = 4, .complete = 112},
+ [ 4] = { .submit = 5, .complete = 113},
+ [ 5] = { .submit = 6, .complete = 114},
+ [ 6] = { .submit = 7, .complete = 115},
+ [ 7] = { .submit = 8, .complete = 116},
+ [ 8] = { .submit = 9, .complete = 117},
+ [ 9] = { .submit = 10, .complete = 118},
+ [10] = { .submit = 11, .complete = 119},
+ [11] = { .submit = 12, .complete = 120},
+ [12] = { .submit = 13, .complete = 121},
+ [13] = { .submit = 14, .complete = 122},
+ [14] = { .submit = 15, .complete = 123},
+
+ /* USB1 ENDP 1 */
+ [15] = { .submit = 16, .complete = 141},
+ [16] = { .submit = 17, .complete = 142},
+ [17] = { .submit = 18, .complete = 143},
+ [18] = { .submit = 19, .complete = 144},
+ [19] = { .submit = 20, .complete = 145},
+ [20] = { .submit = 21, .complete = 146},
+ [21] = { .submit = 22, .complete = 147},
+ [22] = { .submit = 23, .complete = 148},
+ [23] = { .submit = 24, .complete = 149},
+ [24] = { .submit = 25, .complete = 150},
+ [25] = { .submit = 26, .complete = 151},
+ [26] = { .submit = 27, .complete = 152},
+ [27] = { .submit = 28, .complete = 153},
+ [28] = { .submit = 29, .complete = 154},
+ [29] = { .submit = 30, .complete = 155},
+};
+
+static const struct chan_queues da8xx_usb_queues_tx[] = {
+ [0] = { .submit = 16, .complete = 24},
+ [1] = { .submit = 18, .complete = 24},
+ [2] = { .submit = 20, .complete = 24},
+ [3] = { .submit = 22, .complete = 24},
+};
+
+static const struct chan_queues da8xx_usb_queues_rx[] = {
+ [0] = { .submit = 1, .complete = 26},
+ [1] = { .submit = 3, .complete = 26},
+ [2] = { .submit = 5, .complete = 26},
+ [3] = { .submit = 7, .complete = 26},
+};
+
+struct cppi_glue_infos {
+ const struct chan_queues *queues_rx;
+ const struct chan_queues *queues_tx;
+ struct chan_queues td_queue;
+ u16 first_completion_queue;
+ u16 qmgr_num_pend;
+};
+
+static struct cppi41_channel *to_cpp41_chan(struct dma_chan *c)
+{
+ return container_of(c, struct cppi41_channel, chan);
+}
+
+static struct cppi41_channel *desc_to_chan(struct cppi41_dd *cdd, u32 desc)
+{
+ struct cppi41_channel *c;
+ u32 descs_size;
+ u32 desc_num;
+
+ descs_size = sizeof(struct cppi41_desc) * ALLOC_DECS_NUM;
+
+ if (!((desc >= cdd->descs_phys) &&
+ (desc < (cdd->descs_phys + descs_size)))) {
+ return NULL;
+ }
+
+ desc_num = (desc - cdd->descs_phys) / sizeof(struct cppi41_desc);
+ BUG_ON(desc_num >= ALLOC_DECS_NUM);
+ c = cdd->chan_busy[desc_num];
+ cdd->chan_busy[desc_num] = NULL;
+
+ /* Usecount for chan_busy[], paired with push_desc_queue() */
+ pm_runtime_put(cdd->ddev.dev);
+
+ return c;
+}
+
+static void cppi_writel(u32 val, void *__iomem *mem)
+{
+ __raw_writel(val, mem);
+}
+
+static u32 cppi_readl(void *__iomem *mem)
+{
+ return __raw_readl(mem);
+}
+
+static u32 pd_trans_len(u32 val)
+{
+ return val & ((1 << (DESC_LENGTH_BITS_NUM + 1)) - 1);
+}
+
+static u32 cppi41_pop_desc(struct cppi41_dd *cdd, unsigned queue_num)
+{
+ u32 desc;
+
+ desc = cppi_readl(cdd->qmgr_mem + QMGR_QUEUE_D(queue_num));
+ desc &= ~0x1f;
+ return desc;
+}
+
+static irqreturn_t cppi41_irq(int irq, void *data)
+{
+ struct cppi41_dd *cdd = data;
+ u16 first_completion_queue = cdd->first_completion_queue;
+ u16 qmgr_num_pend = cdd->qmgr_num_pend;
+ struct cppi41_channel *c;
+ int i;
+
+ for (i = QMGR_PENDING_SLOT_Q(first_completion_queue); i < qmgr_num_pend;
+ i++) {
+ u32 val;
+ u32 q_num;
+
+ val = cppi_readl(cdd->qmgr_mem + QMGR_PEND(i));
+ if (i == QMGR_PENDING_SLOT_Q(first_completion_queue) && val) {
+ u32 mask;
+ /* set corresponding bit for completetion Q 93 */
+ mask = 1 << QMGR_PENDING_BIT_Q(first_completion_queue);
+ /* not set all bits for queues less than Q 93 */
+ mask--;
+ /* now invert and keep only Q 93+ set */
+ val &= ~mask;
+ }
+
+ if (val)
+ __iormb();
+
+ while (val) {
+ u32 desc, len;
+
+ /*
+ * This should never trigger, see the comments in
+ * push_desc_queue()
+ */
+ WARN_ON(cdd->is_suspended);
+
+ q_num = __fls(val);
+ val &= ~(1 << q_num);
+ q_num += 32 * i;
+ desc = cppi41_pop_desc(cdd, q_num);
+ c = desc_to_chan(cdd, desc);
+ if (WARN_ON(!c)) {
+ pr_err("%s() q %d desc %08x\n", __func__,
+ q_num, desc);
+ continue;
+ }
+
+ if (c->desc->pd2 & PD2_ZERO_LENGTH)
+ len = 0;
+ else
+ len = pd_trans_len(c->desc->pd0);
+
+ c->residue = pd_trans_len(c->desc->pd6) - len;
+ dma_cookie_complete(&c->txd);
+ dmaengine_desc_get_callback_invoke(&c->txd, NULL);
+ }
+ }
+ return IRQ_HANDLED;
+}
+
+static dma_cookie_t cppi41_tx_submit(struct dma_async_tx_descriptor *tx)
+{
+ dma_cookie_t cookie;
+
+ cookie = dma_cookie_assign(tx);
+
+ return cookie;
+}
+
+static int cppi41_dma_alloc_chan_resources(struct dma_chan *chan)
+{
+ struct cppi41_channel *c = to_cpp41_chan(chan);
+ struct cppi41_dd *cdd = c->cdd;
+ int error;
+
+ error = pm_runtime_get_sync(cdd->ddev.dev);
+ if (error < 0) {
+ dev_err(cdd->ddev.dev, "%s pm runtime get: %i\n",
+ __func__, error);
+ pm_runtime_put_noidle(cdd->ddev.dev);
+
+ return error;
+ }
+
+ dma_cookie_init(chan);
+ dma_async_tx_descriptor_init(&c->txd, chan);
+ c->txd.tx_submit = cppi41_tx_submit;
+
+ if (!c->is_tx)
+ cppi_writel(c->q_num, c->gcr_reg + RXHPCRA0);
+
+ pm_runtime_mark_last_busy(cdd->ddev.dev);
+ pm_runtime_put_autosuspend(cdd->ddev.dev);
+
+ return 0;
+}
+
+static void cppi41_dma_free_chan_resources(struct dma_chan *chan)
+{
+ struct cppi41_channel *c = to_cpp41_chan(chan);
+ struct cppi41_dd *cdd = c->cdd;
+ int error;
+
+ error = pm_runtime_get_sync(cdd->ddev.dev);
+ if (error < 0) {
+ pm_runtime_put_noidle(cdd->ddev.dev);
+
+ return;
+ }
+
+ WARN_ON(!list_empty(&cdd->pending));
+
+ pm_runtime_mark_last_busy(cdd->ddev.dev);
+ pm_runtime_put_autosuspend(cdd->ddev.dev);
+}
+
+static enum dma_status cppi41_dma_tx_status(struct dma_chan *chan,
+ dma_cookie_t cookie, struct dma_tx_state *txstate)
+{
+ struct cppi41_channel *c = to_cpp41_chan(chan);
+ enum dma_status ret;
+
+ ret = dma_cookie_status(chan, cookie, txstate);
+
+ dma_set_residue(txstate, c->residue);
+
+ return ret;
+}
+
+static void push_desc_queue(struct cppi41_channel *c)
+{
+ struct cppi41_dd *cdd = c->cdd;
+ u32 desc_num;
+ u32 desc_phys;
+ u32 reg;
+
+ c->residue = 0;
+
+ reg = GCR_CHAN_ENABLE;
+ if (!c->is_tx) {
+ reg |= GCR_STARV_RETRY;
+ reg |= GCR_DESC_TYPE_HOST;
+ reg |= c->q_comp_num;
+ }
+
+ cppi_writel(reg, c->gcr_reg);
+
+ /*
+ * We don't use writel() but __raw_writel() so we have to make sure
+ * that the DMA descriptor in coherent memory made to the main memory
+ * before starting the dma engine.
+ */
+ __iowmb();
+
+ /*
+ * DMA transfers can take at least 200ms to complete with USB mass
+ * storage connected. To prevent autosuspend timeouts, we must use
+ * pm_runtime_get/put() when chan_busy[] is modified. This will get
+ * cleared in desc_to_chan() or cppi41_stop_chan() depending on the
+ * outcome of the transfer.
+ */
+ pm_runtime_get(cdd->ddev.dev);
+
+ desc_phys = lower_32_bits(c->desc_phys);
+ desc_num = (desc_phys - cdd->descs_phys) / sizeof(struct cppi41_desc);
+ WARN_ON(cdd->chan_busy[desc_num]);
+ cdd->chan_busy[desc_num] = c;
+
+ reg = (sizeof(struct cppi41_desc) - 24) / 4;
+ reg |= desc_phys;
+ cppi_writel(reg, cdd->qmgr_mem + QMGR_QUEUE_D(c->q_num));
+}
+
+/*
+ * Caller must hold cdd->lock to prevent push_desc_queue()
+ * getting called out of order. We have both cppi41_dma_issue_pending()
+ * and cppi41_runtime_resume() call this function.
+ */
+static void cppi41_run_queue(struct cppi41_dd *cdd)
+{
+ struct cppi41_channel *c, *_c;
+
+ list_for_each_entry_safe(c, _c, &cdd->pending, node) {
+ push_desc_queue(c);
+ list_del(&c->node);
+ }
+}
+
+static void cppi41_dma_issue_pending(struct dma_chan *chan)
+{
+ struct cppi41_channel *c = to_cpp41_chan(chan);
+ struct cppi41_dd *cdd = c->cdd;
+ unsigned long flags;
+ int error;
+
+ error = pm_runtime_get(cdd->ddev.dev);
+ if ((error != -EINPROGRESS) && error < 0) {
+ pm_runtime_put_noidle(cdd->ddev.dev);
+ dev_err(cdd->ddev.dev, "Failed to pm_runtime_get: %i\n",
+ error);
+
+ return;
+ }
+
+ spin_lock_irqsave(&cdd->lock, flags);
+ list_add_tail(&c->node, &cdd->pending);
+ if (!cdd->is_suspended)
+ cppi41_run_queue(cdd);
+ spin_unlock_irqrestore(&cdd->lock, flags);
+
+ pm_runtime_mark_last_busy(cdd->ddev.dev);
+ pm_runtime_put_autosuspend(cdd->ddev.dev);
+}
+
+static u32 get_host_pd0(u32 length)
+{
+ u32 reg;
+
+ reg = DESC_TYPE_HOST << DESC_TYPE;
+ reg |= length;
+
+ return reg;
+}
+
+static u32 get_host_pd1(struct cppi41_channel *c)
+{
+ u32 reg;
+
+ reg = 0;
+
+ return reg;
+}
+
+static u32 get_host_pd2(struct cppi41_channel *c)
+{
+ u32 reg;
+
+ reg = DESC_TYPE_USB;
+ reg |= c->q_comp_num;
+
+ return reg;
+}
+
+static u32 get_host_pd3(u32 length)
+{
+ u32 reg;
+
+ /* PD3 = packet size */
+ reg = length;
+
+ return reg;
+}
+
+static u32 get_host_pd6(u32 length)
+{
+ u32 reg;
+
+ /* PD6 buffer size */
+ reg = DESC_PD_COMPLETE;
+ reg |= length;
+
+ return reg;
+}
+
+static u32 get_host_pd4_or_7(u32 addr)
+{
+ u32 reg;
+
+ reg = addr;
+
+ return reg;
+}
+
+static u32 get_host_pd5(void)
+{
+ u32 reg;
+
+ reg = 0;
+
+ return reg;
+}
+
+static struct dma_async_tx_descriptor *cppi41_dma_prep_slave_sg(
+ struct dma_chan *chan, struct scatterlist *sgl, unsigned sg_len,
+ enum dma_transfer_direction dir, unsigned long tx_flags, void *context)
+{
+ struct cppi41_channel *c = to_cpp41_chan(chan);
+ struct cppi41_desc *d;
+ struct scatterlist *sg;
+ unsigned int i;
+
+ d = c->desc;
+ for_each_sg(sgl, sg, sg_len, i) {
+ u32 addr;
+ u32 len;
+
+ /* We need to use more than one desc once musb supports sg */
+ addr = lower_32_bits(sg_dma_address(sg));
+ len = sg_dma_len(sg);
+
+ d->pd0 = get_host_pd0(len);
+ d->pd1 = get_host_pd1(c);
+ d->pd2 = get_host_pd2(c);
+ d->pd3 = get_host_pd3(len);
+ d->pd4 = get_host_pd4_or_7(addr);
+ d->pd5 = get_host_pd5();
+ d->pd6 = get_host_pd6(len);
+ d->pd7 = get_host_pd4_or_7(addr);
+
+ d++;
+ }
+
+ return &c->txd;
+}
+
+static void cppi41_compute_td_desc(struct cppi41_desc *d)
+{
+ d->pd0 = DESC_TYPE_TEARD << DESC_TYPE;
+}
+
+static int cppi41_tear_down_chan(struct cppi41_channel *c)
+{
+ struct dmaengine_result abort_result;
+ struct cppi41_dd *cdd = c->cdd;
+ struct cppi41_desc *td;
+ u32 reg;
+ u32 desc_phys;
+ u32 td_desc_phys;
+
+ td = cdd->cd;
+ td += cdd->first_td_desc;
+
+ td_desc_phys = cdd->descs_phys;
+ td_desc_phys += cdd->first_td_desc * sizeof(struct cppi41_desc);
+
+ if (!c->td_queued) {
+ cppi41_compute_td_desc(td);
+ __iowmb();
+
+ reg = (sizeof(struct cppi41_desc) - 24) / 4;
+ reg |= td_desc_phys;
+ cppi_writel(reg, cdd->qmgr_mem +
+ QMGR_QUEUE_D(cdd->td_queue.submit));
+
+ reg = GCR_CHAN_ENABLE;
+ if (!c->is_tx) {
+ reg |= GCR_STARV_RETRY;
+ reg |= GCR_DESC_TYPE_HOST;
+ reg |= cdd->td_queue.complete;
+ }
+ reg |= GCR_TEARDOWN;
+ cppi_writel(reg, c->gcr_reg);
+ c->td_queued = 1;
+ c->td_retry = 500;
+ }
+
+ if (!c->td_seen || !c->td_desc_seen) {
+
+ desc_phys = cppi41_pop_desc(cdd, cdd->td_queue.complete);
+ if (!desc_phys && c->is_tx)
+ desc_phys = cppi41_pop_desc(cdd, c->q_comp_num);
+
+ if (desc_phys == c->desc_phys) {
+ c->td_desc_seen = 1;
+
+ } else if (desc_phys == td_desc_phys) {
+ u32 pd0;
+
+ __iormb();
+ pd0 = td->pd0;
+ WARN_ON((pd0 >> DESC_TYPE) != DESC_TYPE_TEARD);
+ WARN_ON(!c->is_tx && !(pd0 & TD_DESC_IS_RX));
+ WARN_ON((pd0 & 0x1f) != c->port_num);
+ c->td_seen = 1;
+ } else if (desc_phys) {
+ WARN_ON_ONCE(1);
+ }
+ }
+ c->td_retry--;
+ /*
+ * If the TX descriptor / channel is in use, the caller needs to poke
+ * his TD bit multiple times. After that he hardware releases the
+ * transfer descriptor followed by TD descriptor. Waiting seems not to
+ * cause any difference.
+ * RX seems to be thrown out right away. However once the TearDown
+ * descriptor gets through we are done. If we have seens the transfer
+ * descriptor before the TD we fetch it from enqueue, it has to be
+ * there waiting for us.
+ */
+ if (!c->td_seen && c->td_retry) {
+ udelay(1);
+ return -EAGAIN;
+ }
+ WARN_ON(!c->td_retry);
+
+ if (!c->td_desc_seen) {
+ desc_phys = cppi41_pop_desc(cdd, c->q_num);
+ if (!desc_phys)
+ desc_phys = cppi41_pop_desc(cdd, c->q_comp_num);
+ WARN_ON(!desc_phys);
+ }
+
+ c->td_queued = 0;
+ c->td_seen = 0;
+ c->td_desc_seen = 0;
+ cppi_writel(0, c->gcr_reg);
+
+ /* Invoke the callback to do the necessary clean-up */
+ abort_result.result = DMA_TRANS_ABORTED;
+ dma_cookie_complete(&c->txd);
+ dmaengine_desc_get_callback_invoke(&c->txd, &abort_result);
+
+ return 0;
+}
+
+static int cppi41_stop_chan(struct dma_chan *chan)
+{
+ struct cppi41_channel *c = to_cpp41_chan(chan);
+ struct cppi41_dd *cdd = c->cdd;
+ u32 desc_num;
+ u32 desc_phys;
+ int ret;
+
+ desc_phys = lower_32_bits(c->desc_phys);
+ desc_num = (desc_phys - cdd->descs_phys) / sizeof(struct cppi41_desc);
+ if (!cdd->chan_busy[desc_num]) {
+ struct cppi41_channel *cc, *_ct;
+
+ /*
+ * channels might still be in the pendling list if
+ * cppi41_dma_issue_pending() is called after
+ * cppi41_runtime_suspend() is called
+ */
+ list_for_each_entry_safe(cc, _ct, &cdd->pending, node) {
+ if (cc != c)
+ continue;
+ list_del(&cc->node);
+ break;
+ }
+ return 0;
+ }
+
+ ret = cppi41_tear_down_chan(c);
+ if (ret)
+ return ret;
+
+ WARN_ON(!cdd->chan_busy[desc_num]);
+ cdd->chan_busy[desc_num] = NULL;
+
+ /* Usecount for chan_busy[], paired with push_desc_queue() */
+ pm_runtime_put(cdd->ddev.dev);
+
+ return 0;
+}
+
+static int cppi41_add_chans(struct device *dev, struct cppi41_dd *cdd)
+{
+ struct cppi41_channel *cchan, *chans;
+ int i;
+ u32 n_chans = cdd->n_chans;
+
+ /*
+ * The channels can only be used as TX or as RX. So we add twice
+ * that much dma channels because USB can only do RX or TX.
+ */
+ n_chans *= 2;
+
+ chans = devm_kcalloc(dev, n_chans, sizeof(*chans), GFP_KERNEL);
+ if (!chans)
+ return -ENOMEM;
+
+ for (i = 0; i < n_chans; i++) {
+ cchan = &chans[i];
+
+ cchan->cdd = cdd;
+ if (i & 1) {
+ cchan->gcr_reg = cdd->ctrl_mem + DMA_TXGCR(i >> 1);
+ cchan->is_tx = 1;
+ } else {
+ cchan->gcr_reg = cdd->ctrl_mem + DMA_RXGCR(i >> 1);
+ cchan->is_tx = 0;
+ }
+ cchan->port_num = i >> 1;
+ cchan->desc = &cdd->cd[i];
+ cchan->desc_phys = cdd->descs_phys;
+ cchan->desc_phys += i * sizeof(struct cppi41_desc);
+ cchan->chan.device = &cdd->ddev;
+ list_add_tail(&cchan->chan.device_node, &cdd->ddev.channels);
+ }
+ cdd->first_td_desc = n_chans;
+
+ return 0;
+}
+
+static void purge_descs(struct device *dev, struct cppi41_dd *cdd)
+{
+ unsigned int mem_decs;
+ int i;
+
+ mem_decs = ALLOC_DECS_NUM * sizeof(struct cppi41_desc);
+
+ for (i = 0; i < DESCS_AREAS; i++) {
+
+ cppi_writel(0, cdd->qmgr_mem + QMGR_MEMBASE(i));
+ cppi_writel(0, cdd->qmgr_mem + QMGR_MEMCTRL(i));
+
+ dma_free_coherent(dev, mem_decs, cdd->cd,
+ cdd->descs_phys);
+ }
+}
+
+static void disable_sched(struct cppi41_dd *cdd)
+{
+ cppi_writel(0, cdd->sched_mem + DMA_SCHED_CTRL);
+}
+
+static void deinit_cppi41(struct device *dev, struct cppi41_dd *cdd)
+{
+ disable_sched(cdd);
+
+ purge_descs(dev, cdd);
+
+ cppi_writel(0, cdd->qmgr_mem + QMGR_LRAM0_BASE);
+ cppi_writel(0, cdd->qmgr_mem + QMGR_LRAM0_BASE);
+ dma_free_coherent(dev, QMGR_SCRATCH_SIZE, cdd->qmgr_scratch,
+ cdd->scratch_phys);
+}
+
+static int init_descs(struct device *dev, struct cppi41_dd *cdd)
+{
+ unsigned int desc_size;
+ unsigned int mem_decs;
+ int i;
+ u32 reg;
+ u32 idx;
+
+ BUILD_BUG_ON(sizeof(struct cppi41_desc) &
+ (sizeof(struct cppi41_desc) - 1));
+ BUILD_BUG_ON(sizeof(struct cppi41_desc) < 32);
+ BUILD_BUG_ON(ALLOC_DECS_NUM < 32);
+
+ desc_size = sizeof(struct cppi41_desc);
+ mem_decs = ALLOC_DECS_NUM * desc_size;
+
+ idx = 0;
+ for (i = 0; i < DESCS_AREAS; i++) {
+
+ reg = idx << QMGR_MEMCTRL_IDX_SH;
+ reg |= (ilog2(desc_size) - 5) << QMGR_MEMCTRL_DESC_SH;
+ reg |= ilog2(ALLOC_DECS_NUM) - 5;
+
+ BUILD_BUG_ON(DESCS_AREAS != 1);
+ cdd->cd = dma_alloc_coherent(dev, mem_decs,
+ &cdd->descs_phys, GFP_KERNEL);
+ if (!cdd->cd)
+ return -ENOMEM;
+
+ cppi_writel(cdd->descs_phys, cdd->qmgr_mem + QMGR_MEMBASE(i));
+ cppi_writel(reg, cdd->qmgr_mem + QMGR_MEMCTRL(i));
+
+ idx += ALLOC_DECS_NUM;
+ }
+ return 0;
+}
+
+static void init_sched(struct cppi41_dd *cdd)
+{
+ unsigned ch;
+ unsigned word;
+ u32 reg;
+
+ word = 0;
+ cppi_writel(0, cdd->sched_mem + DMA_SCHED_CTRL);
+ for (ch = 0; ch < cdd->n_chans; ch += 2) {
+
+ reg = SCHED_ENTRY0_CHAN(ch);
+ reg |= SCHED_ENTRY1_CHAN(ch) | SCHED_ENTRY1_IS_RX;
+
+ reg |= SCHED_ENTRY2_CHAN(ch + 1);
+ reg |= SCHED_ENTRY3_CHAN(ch + 1) | SCHED_ENTRY3_IS_RX;
+ cppi_writel(reg, cdd->sched_mem + DMA_SCHED_WORD(word));
+ word++;
+ }
+ reg = cdd->n_chans * 2 - 1;
+ reg |= DMA_SCHED_CTRL_EN;
+ cppi_writel(reg, cdd->sched_mem + DMA_SCHED_CTRL);
+}
+
+static int init_cppi41(struct device *dev, struct cppi41_dd *cdd)
+{
+ int ret;
+
+ BUILD_BUG_ON(QMGR_SCRATCH_SIZE > ((1 << 14) - 1));
+ cdd->qmgr_scratch = dma_alloc_coherent(dev, QMGR_SCRATCH_SIZE,
+ &cdd->scratch_phys, GFP_KERNEL);
+ if (!cdd->qmgr_scratch)
+ return -ENOMEM;
+
+ cppi_writel(cdd->scratch_phys, cdd->qmgr_mem + QMGR_LRAM0_BASE);
+ cppi_writel(TOTAL_DESCS_NUM, cdd->qmgr_mem + QMGR_LRAM_SIZE);
+ cppi_writel(0, cdd->qmgr_mem + QMGR_LRAM1_BASE);
+
+ ret = init_descs(dev, cdd);
+ if (ret)
+ goto err_td;
+
+ cppi_writel(cdd->td_queue.submit, cdd->ctrl_mem + DMA_TDFDQ);
+ init_sched(cdd);
+
+ return 0;
+err_td:
+ deinit_cppi41(dev, cdd);
+ return ret;
+}
+
+static struct platform_driver cpp41_dma_driver;
+/*
+ * The param format is:
+ * X Y
+ * X: Port
+ * Y: 0 = RX else TX
+ */
+#define INFO_PORT 0
+#define INFO_IS_TX 1
+
+static bool cpp41_dma_filter_fn(struct dma_chan *chan, void *param)
+{
+ struct cppi41_channel *cchan;
+ struct cppi41_dd *cdd;
+ const struct chan_queues *queues;
+ u32 *num = param;
+
+ if (chan->device->dev->driver != &cpp41_dma_driver.driver)
+ return false;
+
+ cchan = to_cpp41_chan(chan);
+
+ if (cchan->port_num != num[INFO_PORT])
+ return false;
+
+ if (cchan->is_tx && !num[INFO_IS_TX])
+ return false;
+ cdd = cchan->cdd;
+ if (cchan->is_tx)
+ queues = cdd->queues_tx;
+ else
+ queues = cdd->queues_rx;
+
+ BUILD_BUG_ON(ARRAY_SIZE(am335x_usb_queues_rx) !=
+ ARRAY_SIZE(am335x_usb_queues_tx));
+ if (WARN_ON(cchan->port_num >= ARRAY_SIZE(am335x_usb_queues_rx)))
+ return false;
+
+ cchan->q_num = queues[cchan->port_num].submit;
+ cchan->q_comp_num = queues[cchan->port_num].complete;
+ return true;
+}
+
+static struct of_dma_filter_info cpp41_dma_info = {
+ .filter_fn = cpp41_dma_filter_fn,
+};
+
+static struct dma_chan *cppi41_dma_xlate(struct of_phandle_args *dma_spec,
+ struct of_dma *ofdma)
+{
+ int count = dma_spec->args_count;
+ struct of_dma_filter_info *info = ofdma->of_dma_data;
+
+ if (!info || !info->filter_fn)
+ return NULL;
+
+ if (count != 2)
+ return NULL;
+
+ return dma_request_channel(info->dma_cap, info->filter_fn,
+ &dma_spec->args[0]);
+}
+
+static const struct cppi_glue_infos am335x_usb_infos = {
+ .queues_rx = am335x_usb_queues_rx,
+ .queues_tx = am335x_usb_queues_tx,
+ .td_queue = { .submit = 31, .complete = 0 },
+ .first_completion_queue = 93,
+ .qmgr_num_pend = 5,
+};
+
+static const struct cppi_glue_infos da8xx_usb_infos = {
+ .queues_rx = da8xx_usb_queues_rx,
+ .queues_tx = da8xx_usb_queues_tx,
+ .td_queue = { .submit = 31, .complete = 0 },
+ .first_completion_queue = 24,
+ .qmgr_num_pend = 2,
+};
+
+static const struct of_device_id cppi41_dma_ids[] = {
+ { .compatible = "ti,am3359-cppi41", .data = &am335x_usb_infos},
+ { .compatible = "ti,da830-cppi41", .data = &da8xx_usb_infos},
+ {},
+};
+MODULE_DEVICE_TABLE(of, cppi41_dma_ids);
+
+static const struct cppi_glue_infos *get_glue_info(struct device *dev)
+{
+ const struct of_device_id *of_id;
+
+ of_id = of_match_node(cppi41_dma_ids, dev->of_node);
+ if (!of_id)
+ return NULL;
+ return of_id->data;
+}
+
+#define CPPI41_DMA_BUSWIDTHS (BIT(DMA_SLAVE_BUSWIDTH_1_BYTE) | \
+ BIT(DMA_SLAVE_BUSWIDTH_2_BYTES) | \
+ BIT(DMA_SLAVE_BUSWIDTH_3_BYTES) | \
+ BIT(DMA_SLAVE_BUSWIDTH_4_BYTES))
+
+static int cppi41_dma_probe(struct platform_device *pdev)
+{
+ struct cppi41_dd *cdd;
+ struct device *dev = &pdev->dev;
+ const struct cppi_glue_infos *glue_info;
+ struct resource *mem;
+ int index;
+ int irq;
+ int ret;
+
+ glue_info = get_glue_info(dev);
+ if (!glue_info)
+ return -EINVAL;
+
+ cdd = devm_kzalloc(&pdev->dev, sizeof(*cdd), GFP_KERNEL);
+ if (!cdd)
+ return -ENOMEM;
+
+ dma_cap_set(DMA_SLAVE, cdd->ddev.cap_mask);
+ cdd->ddev.device_alloc_chan_resources = cppi41_dma_alloc_chan_resources;
+ cdd->ddev.device_free_chan_resources = cppi41_dma_free_chan_resources;
+ cdd->ddev.device_tx_status = cppi41_dma_tx_status;
+ cdd->ddev.device_issue_pending = cppi41_dma_issue_pending;
+ cdd->ddev.device_prep_slave_sg = cppi41_dma_prep_slave_sg;
+ cdd->ddev.device_terminate_all = cppi41_stop_chan;
+ cdd->ddev.directions = BIT(DMA_DEV_TO_MEM) | BIT(DMA_MEM_TO_DEV);
+ cdd->ddev.src_addr_widths = CPPI41_DMA_BUSWIDTHS;
+ cdd->ddev.dst_addr_widths = CPPI41_DMA_BUSWIDTHS;
+ cdd->ddev.residue_granularity = DMA_RESIDUE_GRANULARITY_BURST;
+ cdd->ddev.dev = dev;
+ INIT_LIST_HEAD(&cdd->ddev.channels);
+ cpp41_dma_info.dma_cap = cdd->ddev.cap_mask;
+
+ index = of_property_match_string(dev->of_node,
+ "reg-names", "controller");
+ if (index < 0)
+ return index;
+
+ mem = platform_get_resource(pdev, IORESOURCE_MEM, index);
+ cdd->ctrl_mem = devm_ioremap_resource(dev, mem);
+ if (IS_ERR(cdd->ctrl_mem))
+ return PTR_ERR(cdd->ctrl_mem);
+
+ mem = platform_get_resource(pdev, IORESOURCE_MEM, index + 1);
+ cdd->sched_mem = devm_ioremap_resource(dev, mem);
+ if (IS_ERR(cdd->sched_mem))
+ return PTR_ERR(cdd->sched_mem);
+
+ mem = platform_get_resource(pdev, IORESOURCE_MEM, index + 2);
+ cdd->qmgr_mem = devm_ioremap_resource(dev, mem);
+ if (IS_ERR(cdd->qmgr_mem))
+ return PTR_ERR(cdd->qmgr_mem);
+
+ spin_lock_init(&cdd->lock);
+ INIT_LIST_HEAD(&cdd->pending);
+
+ platform_set_drvdata(pdev, cdd);
+
+ pm_runtime_enable(dev);
+ pm_runtime_set_autosuspend_delay(dev, 100);
+ pm_runtime_use_autosuspend(dev);
+ ret = pm_runtime_get_sync(dev);
+ if (ret < 0)
+ goto err_get_sync;
+
+ cdd->queues_rx = glue_info->queues_rx;
+ cdd->queues_tx = glue_info->queues_tx;
+ cdd->td_queue = glue_info->td_queue;
+ cdd->qmgr_num_pend = glue_info->qmgr_num_pend;
+ cdd->first_completion_queue = glue_info->first_completion_queue;
+
+ ret = of_property_read_u32(dev->of_node,
+ "#dma-channels", &cdd->n_chans);
+ if (ret)
+ goto err_get_n_chans;
+
+ ret = init_cppi41(dev, cdd);
+ if (ret)
+ goto err_init_cppi;
+
+ ret = cppi41_add_chans(dev, cdd);
+ if (ret)
+ goto err_chans;
+
+ irq = irq_of_parse_and_map(dev->of_node, 0);
+ if (!irq) {
+ ret = -EINVAL;
+ goto err_chans;
+ }
+
+ ret = devm_request_irq(&pdev->dev, irq, cppi41_irq, IRQF_SHARED,
+ dev_name(dev), cdd);
+ if (ret)
+ goto err_chans;
+ cdd->irq = irq;
+
+ ret = dma_async_device_register(&cdd->ddev);
+ if (ret)
+ goto err_chans;
+
+ ret = of_dma_controller_register(dev->of_node,
+ cppi41_dma_xlate, &cpp41_dma_info);
+ if (ret)
+ goto err_of;
+
+ pm_runtime_mark_last_busy(dev);
+ pm_runtime_put_autosuspend(dev);
+
+ return 0;
+err_of:
+ dma_async_device_unregister(&cdd->ddev);
+err_chans:
+ deinit_cppi41(dev, cdd);
+err_init_cppi:
+ pm_runtime_dont_use_autosuspend(dev);
+err_get_n_chans:
+err_get_sync:
+ pm_runtime_put_sync(dev);
+ pm_runtime_disable(dev);
+ return ret;
+}
+
+static int cppi41_dma_remove(struct platform_device *pdev)
+{
+ struct cppi41_dd *cdd = platform_get_drvdata(pdev);
+ int error;
+
+ error = pm_runtime_get_sync(&pdev->dev);
+ if (error < 0)
+ dev_err(&pdev->dev, "%s could not pm_runtime_get: %i\n",
+ __func__, error);
+ of_dma_controller_free(pdev->dev.of_node);
+ dma_async_device_unregister(&cdd->ddev);
+
+ devm_free_irq(&pdev->dev, cdd->irq, cdd);
+ deinit_cppi41(&pdev->dev, cdd);
+ pm_runtime_dont_use_autosuspend(&pdev->dev);
+ pm_runtime_put_sync(&pdev->dev);
+ pm_runtime_disable(&pdev->dev);
+ return 0;
+}
+
+static int __maybe_unused cppi41_suspend(struct device *dev)
+{
+ struct cppi41_dd *cdd = dev_get_drvdata(dev);
+
+ cdd->dma_tdfdq = cppi_readl(cdd->ctrl_mem + DMA_TDFDQ);
+ disable_sched(cdd);
+
+ return 0;
+}
+
+static int __maybe_unused cppi41_resume(struct device *dev)
+{
+ struct cppi41_dd *cdd = dev_get_drvdata(dev);
+ struct cppi41_channel *c;
+ int i;
+
+ for (i = 0; i < DESCS_AREAS; i++)
+ cppi_writel(cdd->descs_phys, cdd->qmgr_mem + QMGR_MEMBASE(i));
+
+ list_for_each_entry(c, &cdd->ddev.channels, chan.device_node)
+ if (!c->is_tx)
+ cppi_writel(c->q_num, c->gcr_reg + RXHPCRA0);
+
+ init_sched(cdd);
+
+ cppi_writel(cdd->dma_tdfdq, cdd->ctrl_mem + DMA_TDFDQ);
+ cppi_writel(cdd->scratch_phys, cdd->qmgr_mem + QMGR_LRAM0_BASE);
+ cppi_writel(QMGR_SCRATCH_SIZE, cdd->qmgr_mem + QMGR_LRAM_SIZE);
+ cppi_writel(0, cdd->qmgr_mem + QMGR_LRAM1_BASE);
+
+ return 0;
+}
+
+static int __maybe_unused cppi41_runtime_suspend(struct device *dev)
+{
+ struct cppi41_dd *cdd = dev_get_drvdata(dev);
+ unsigned long flags;
+
+ spin_lock_irqsave(&cdd->lock, flags);
+ cdd->is_suspended = true;
+ WARN_ON(!list_empty(&cdd->pending));
+ spin_unlock_irqrestore(&cdd->lock, flags);
+
+ return 0;
+}
+
+static int __maybe_unused cppi41_runtime_resume(struct device *dev)
+{
+ struct cppi41_dd *cdd = dev_get_drvdata(dev);
+ unsigned long flags;
+
+ spin_lock_irqsave(&cdd->lock, flags);
+ cdd->is_suspended = false;
+ cppi41_run_queue(cdd);
+ spin_unlock_irqrestore(&cdd->lock, flags);
+
+ return 0;
+}
+
+static const struct dev_pm_ops cppi41_pm_ops = {
+ SET_LATE_SYSTEM_SLEEP_PM_OPS(cppi41_suspend, cppi41_resume)
+ SET_RUNTIME_PM_OPS(cppi41_runtime_suspend,
+ cppi41_runtime_resume,
+ NULL)
+};
+
+static struct platform_driver cpp41_dma_driver = {
+ .probe = cppi41_dma_probe,
+ .remove = cppi41_dma_remove,
+ .driver = {
+ .name = "cppi41-dma-engine",
+ .pm = &cppi41_pm_ops,
+ .of_match_table = of_match_ptr(cppi41_dma_ids),
+ },
+};
+
+module_platform_driver(cpp41_dma_driver);
+MODULE_LICENSE("GPL");
+MODULE_AUTHOR("Sebastian Andrzej Siewior <bigeasy@linutronix.de>");
diff --git a/drivers/dma/ti/dma-crossbar.c b/drivers/dma/ti/dma-crossbar.c
new file mode 100644
index 0000000..9272b17
--- /dev/null
+++ b/drivers/dma/ti/dma-crossbar.c
@@ -0,0 +1,478 @@
+/*
+ * Copyright (C) 2015 Texas Instruments Incorporated - http://www.ti.com
+ * Author: Peter Ujfalusi <peter.ujfalusi@ti.com>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ *
+ */
+#include <linux/slab.h>
+#include <linux/err.h>
+#include <linux/init.h>
+#include <linux/list.h>
+#include <linux/io.h>
+#include <linux/of_address.h>
+#include <linux/of_device.h>
+#include <linux/of_dma.h>
+
+#define TI_XBAR_DRA7 0
+#define TI_XBAR_AM335X 1
+static const u32 ti_xbar_type[] = {
+ [TI_XBAR_DRA7] = TI_XBAR_DRA7,
+ [TI_XBAR_AM335X] = TI_XBAR_AM335X,
+};
+
+static const struct of_device_id ti_dma_xbar_match[] = {
+ {
+ .compatible = "ti,dra7-dma-crossbar",
+ .data = &ti_xbar_type[TI_XBAR_DRA7],
+ },
+ {
+ .compatible = "ti,am335x-edma-crossbar",
+ .data = &ti_xbar_type[TI_XBAR_AM335X],
+ },
+ {},
+};
+
+/* Crossbar on AM335x/AM437x family */
+#define TI_AM335X_XBAR_LINES 64
+
+struct ti_am335x_xbar_data {
+ void __iomem *iomem;
+
+ struct dma_router dmarouter;
+
+ u32 xbar_events; /* maximum number of events to select in xbar */
+ u32 dma_requests; /* number of DMA requests on eDMA */
+};
+
+struct ti_am335x_xbar_map {
+ u16 dma_line;
+ u8 mux_val;
+};
+
+static inline void ti_am335x_xbar_write(void __iomem *iomem, int event, u8 val)
+{
+ /*
+ * TPCC_EVT_MUX_60_63 register layout is different than the
+ * rest, in the sense, that event 63 is mapped to lowest byte
+ * and event 60 is mapped to highest, handle it separately.
+ */
+ if (event >= 60 && event <= 63)
+ writeb_relaxed(val, iomem + (63 - event % 4));
+ else
+ writeb_relaxed(val, iomem + event);
+}
+
+static void ti_am335x_xbar_free(struct device *dev, void *route_data)
+{
+ struct ti_am335x_xbar_data *xbar = dev_get_drvdata(dev);
+ struct ti_am335x_xbar_map *map = route_data;
+
+ dev_dbg(dev, "Unmapping XBAR event %u on channel %u\n",
+ map->mux_val, map->dma_line);
+
+ ti_am335x_xbar_write(xbar->iomem, map->dma_line, 0);
+ kfree(map);
+}
+
+static void *ti_am335x_xbar_route_allocate(struct of_phandle_args *dma_spec,
+ struct of_dma *ofdma)
+{
+ struct platform_device *pdev = of_find_device_by_node(ofdma->of_node);
+ struct ti_am335x_xbar_data *xbar = platform_get_drvdata(pdev);
+ struct ti_am335x_xbar_map *map;
+
+ if (dma_spec->args_count != 3)
+ return ERR_PTR(-EINVAL);
+
+ if (dma_spec->args[2] >= xbar->xbar_events) {
+ dev_err(&pdev->dev, "Invalid XBAR event number: %d\n",
+ dma_spec->args[2]);
+ return ERR_PTR(-EINVAL);
+ }
+
+ if (dma_spec->args[0] >= xbar->dma_requests) {
+ dev_err(&pdev->dev, "Invalid DMA request line number: %d\n",
+ dma_spec->args[0]);
+ return ERR_PTR(-EINVAL);
+ }
+
+ /* The of_node_put() will be done in the core for the node */
+ dma_spec->np = of_parse_phandle(ofdma->of_node, "dma-masters", 0);
+ if (!dma_spec->np) {
+ dev_err(&pdev->dev, "Can't get DMA master\n");
+ return ERR_PTR(-EINVAL);
+ }
+
+ map = kzalloc(sizeof(*map), GFP_KERNEL);
+ if (!map) {
+ of_node_put(dma_spec->np);
+ return ERR_PTR(-ENOMEM);
+ }
+
+ map->dma_line = (u16)dma_spec->args[0];
+ map->mux_val = (u8)dma_spec->args[2];
+
+ dma_spec->args[2] = 0;
+ dma_spec->args_count = 2;
+
+ dev_dbg(&pdev->dev, "Mapping XBAR event%u to DMA%u\n",
+ map->mux_val, map->dma_line);
+
+ ti_am335x_xbar_write(xbar->iomem, map->dma_line, map->mux_val);
+
+ return map;
+}
+
+static const struct of_device_id ti_am335x_master_match[] = {
+ { .compatible = "ti,edma3-tpcc", },
+ {},
+};
+
+static int ti_am335x_xbar_probe(struct platform_device *pdev)
+{
+ struct device_node *node = pdev->dev.of_node;
+ const struct of_device_id *match;
+ struct device_node *dma_node;
+ struct ti_am335x_xbar_data *xbar;
+ struct resource *res;
+ void __iomem *iomem;
+ int i, ret;
+
+ if (!node)
+ return -ENODEV;
+
+ xbar = devm_kzalloc(&pdev->dev, sizeof(*xbar), GFP_KERNEL);
+ if (!xbar)
+ return -ENOMEM;
+
+ dma_node = of_parse_phandle(node, "dma-masters", 0);
+ if (!dma_node) {
+ dev_err(&pdev->dev, "Can't get DMA master node\n");
+ return -ENODEV;
+ }
+
+ match = of_match_node(ti_am335x_master_match, dma_node);
+ if (!match) {
+ dev_err(&pdev->dev, "DMA master is not supported\n");
+ of_node_put(dma_node);
+ return -EINVAL;
+ }
+
+ if (of_property_read_u32(dma_node, "dma-requests",
+ &xbar->dma_requests)) {
+ dev_info(&pdev->dev,
+ "Missing XBAR output information, using %u.\n",
+ TI_AM335X_XBAR_LINES);
+ xbar->dma_requests = TI_AM335X_XBAR_LINES;
+ }
+ of_node_put(dma_node);
+
+ if (of_property_read_u32(node, "dma-requests", &xbar->xbar_events)) {
+ dev_info(&pdev->dev,
+ "Missing XBAR input information, using %u.\n",
+ TI_AM335X_XBAR_LINES);
+ xbar->xbar_events = TI_AM335X_XBAR_LINES;
+ }
+
+ res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
+ iomem = devm_ioremap_resource(&pdev->dev, res);
+ if (IS_ERR(iomem))
+ return PTR_ERR(iomem);
+
+ xbar->iomem = iomem;
+
+ xbar->dmarouter.dev = &pdev->dev;
+ xbar->dmarouter.route_free = ti_am335x_xbar_free;
+
+ platform_set_drvdata(pdev, xbar);
+
+ /* Reset the crossbar */
+ for (i = 0; i < xbar->dma_requests; i++)
+ ti_am335x_xbar_write(xbar->iomem, i, 0);
+
+ ret = of_dma_router_register(node, ti_am335x_xbar_route_allocate,
+ &xbar->dmarouter);
+
+ return ret;
+}
+
+/* Crossbar on DRA7xx family */
+#define TI_DRA7_XBAR_OUTPUTS 127
+#define TI_DRA7_XBAR_INPUTS 256
+
+struct ti_dra7_xbar_data {
+ void __iomem *iomem;
+
+ struct dma_router dmarouter;
+ struct mutex mutex;
+ unsigned long *dma_inuse;
+
+ u16 safe_val; /* Value to rest the crossbar lines */
+ u32 xbar_requests; /* number of DMA requests connected to XBAR */
+ u32 dma_requests; /* number of DMA requests forwarded to DMA */
+ u32 dma_offset;
+};
+
+struct ti_dra7_xbar_map {
+ u16 xbar_in;
+ int xbar_out;
+};
+
+static inline void ti_dra7_xbar_write(void __iomem *iomem, int xbar, u16 val)
+{
+ writew_relaxed(val, iomem + (xbar * 2));
+}
+
+static void ti_dra7_xbar_free(struct device *dev, void *route_data)
+{
+ struct ti_dra7_xbar_data *xbar = dev_get_drvdata(dev);
+ struct ti_dra7_xbar_map *map = route_data;
+
+ dev_dbg(dev, "Unmapping XBAR%u (was routed to %d)\n",
+ map->xbar_in, map->xbar_out);
+
+ ti_dra7_xbar_write(xbar->iomem, map->xbar_out, xbar->safe_val);
+ mutex_lock(&xbar->mutex);
+ clear_bit(map->xbar_out, xbar->dma_inuse);
+ mutex_unlock(&xbar->mutex);
+ kfree(map);
+}
+
+static void *ti_dra7_xbar_route_allocate(struct of_phandle_args *dma_spec,
+ struct of_dma *ofdma)
+{
+ struct platform_device *pdev = of_find_device_by_node(ofdma->of_node);
+ struct ti_dra7_xbar_data *xbar = platform_get_drvdata(pdev);
+ struct ti_dra7_xbar_map *map;
+
+ if (dma_spec->args[0] >= xbar->xbar_requests) {
+ dev_err(&pdev->dev, "Invalid XBAR request number: %d\n",
+ dma_spec->args[0]);
+ return ERR_PTR(-EINVAL);
+ }
+
+ /* The of_node_put() will be done in the core for the node */
+ dma_spec->np = of_parse_phandle(ofdma->of_node, "dma-masters", 0);
+ if (!dma_spec->np) {
+ dev_err(&pdev->dev, "Can't get DMA master\n");
+ return ERR_PTR(-EINVAL);
+ }
+
+ map = kzalloc(sizeof(*map), GFP_KERNEL);
+ if (!map) {
+ of_node_put(dma_spec->np);
+ return ERR_PTR(-ENOMEM);
+ }
+
+ mutex_lock(&xbar->mutex);
+ map->xbar_out = find_first_zero_bit(xbar->dma_inuse,
+ xbar->dma_requests);
+ if (map->xbar_out == xbar->dma_requests) {
+ mutex_unlock(&xbar->mutex);
+ dev_err(&pdev->dev, "Run out of free DMA requests\n");
+ kfree(map);
+ return ERR_PTR(-ENOMEM);
+ }
+ set_bit(map->xbar_out, xbar->dma_inuse);
+ mutex_unlock(&xbar->mutex);
+
+ map->xbar_in = (u16)dma_spec->args[0];
+
+ dma_spec->args[0] = map->xbar_out + xbar->dma_offset;
+
+ dev_dbg(&pdev->dev, "Mapping XBAR%u to DMA%d\n",
+ map->xbar_in, map->xbar_out);
+
+ ti_dra7_xbar_write(xbar->iomem, map->xbar_out, map->xbar_in);
+
+ return map;
+}
+
+#define TI_XBAR_EDMA_OFFSET 0
+#define TI_XBAR_SDMA_OFFSET 1
+static const u32 ti_dma_offset[] = {
+ [TI_XBAR_EDMA_OFFSET] = 0,
+ [TI_XBAR_SDMA_OFFSET] = 1,
+};
+
+static const struct of_device_id ti_dra7_master_match[] = {
+ {
+ .compatible = "ti,omap4430-sdma",
+ .data = &ti_dma_offset[TI_XBAR_SDMA_OFFSET],
+ },
+ {
+ .compatible = "ti,edma3",
+ .data = &ti_dma_offset[TI_XBAR_EDMA_OFFSET],
+ },
+ {
+ .compatible = "ti,edma3-tpcc",
+ .data = &ti_dma_offset[TI_XBAR_EDMA_OFFSET],
+ },
+ {},
+};
+
+static inline void ti_dra7_xbar_reserve(int offset, int len, unsigned long *p)
+{
+ for (; len > 0; len--)
+ set_bit(offset + (len - 1), p);
+}
+
+static int ti_dra7_xbar_probe(struct platform_device *pdev)
+{
+ struct device_node *node = pdev->dev.of_node;
+ const struct of_device_id *match;
+ struct device_node *dma_node;
+ struct ti_dra7_xbar_data *xbar;
+ struct property *prop;
+ struct resource *res;
+ u32 safe_val;
+ int sz;
+ void __iomem *iomem;
+ int i, ret;
+
+ if (!node)
+ return -ENODEV;
+
+ xbar = devm_kzalloc(&pdev->dev, sizeof(*xbar), GFP_KERNEL);
+ if (!xbar)
+ return -ENOMEM;
+
+ dma_node = of_parse_phandle(node, "dma-masters", 0);
+ if (!dma_node) {
+ dev_err(&pdev->dev, "Can't get DMA master node\n");
+ return -ENODEV;
+ }
+
+ match = of_match_node(ti_dra7_master_match, dma_node);
+ if (!match) {
+ dev_err(&pdev->dev, "DMA master is not supported\n");
+ of_node_put(dma_node);
+ return -EINVAL;
+ }
+
+ if (of_property_read_u32(dma_node, "dma-requests",
+ &xbar->dma_requests)) {
+ dev_info(&pdev->dev,
+ "Missing XBAR output information, using %u.\n",
+ TI_DRA7_XBAR_OUTPUTS);
+ xbar->dma_requests = TI_DRA7_XBAR_OUTPUTS;
+ }
+ of_node_put(dma_node);
+
+ xbar->dma_inuse = devm_kcalloc(&pdev->dev,
+ BITS_TO_LONGS(xbar->dma_requests),
+ sizeof(unsigned long), GFP_KERNEL);
+ if (!xbar->dma_inuse)
+ return -ENOMEM;
+
+ if (of_property_read_u32(node, "dma-requests", &xbar->xbar_requests)) {
+ dev_info(&pdev->dev,
+ "Missing XBAR input information, using %u.\n",
+ TI_DRA7_XBAR_INPUTS);
+ xbar->xbar_requests = TI_DRA7_XBAR_INPUTS;
+ }
+
+ if (!of_property_read_u32(node, "ti,dma-safe-map", &safe_val))
+ xbar->safe_val = (u16)safe_val;
+
+
+ prop = of_find_property(node, "ti,reserved-dma-request-ranges", &sz);
+ if (prop) {
+ const char pname[] = "ti,reserved-dma-request-ranges";
+ u32 (*rsv_events)[2];
+ size_t nelm = sz / sizeof(*rsv_events);
+ int i;
+
+ if (!nelm)
+ return -EINVAL;
+
+ rsv_events = kcalloc(nelm, sizeof(*rsv_events), GFP_KERNEL);
+ if (!rsv_events)
+ return -ENOMEM;
+
+ ret = of_property_read_u32_array(node, pname, (u32 *)rsv_events,
+ nelm * 2);
+ if (ret)
+ return ret;
+
+ for (i = 0; i < nelm; i++) {
+ ti_dra7_xbar_reserve(rsv_events[i][0], rsv_events[i][1],
+ xbar->dma_inuse);
+ }
+ kfree(rsv_events);
+ }
+
+ res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
+ iomem = devm_ioremap_resource(&pdev->dev, res);
+ if (IS_ERR(iomem))
+ return PTR_ERR(iomem);
+
+ xbar->iomem = iomem;
+
+ xbar->dmarouter.dev = &pdev->dev;
+ xbar->dmarouter.route_free = ti_dra7_xbar_free;
+ xbar->dma_offset = *(u32 *)match->data;
+
+ mutex_init(&xbar->mutex);
+ platform_set_drvdata(pdev, xbar);
+
+ /* Reset the crossbar */
+ for (i = 0; i < xbar->dma_requests; i++) {
+ if (!test_bit(i, xbar->dma_inuse))
+ ti_dra7_xbar_write(xbar->iomem, i, xbar->safe_val);
+ }
+
+ ret = of_dma_router_register(node, ti_dra7_xbar_route_allocate,
+ &xbar->dmarouter);
+ if (ret) {
+ /* Restore the defaults for the crossbar */
+ for (i = 0; i < xbar->dma_requests; i++) {
+ if (!test_bit(i, xbar->dma_inuse))
+ ti_dra7_xbar_write(xbar->iomem, i, i);
+ }
+ }
+
+ return ret;
+}
+
+static int ti_dma_xbar_probe(struct platform_device *pdev)
+{
+ const struct of_device_id *match;
+ int ret;
+
+ match = of_match_node(ti_dma_xbar_match, pdev->dev.of_node);
+ if (unlikely(!match))
+ return -EINVAL;
+
+ switch (*(u32 *)match->data) {
+ case TI_XBAR_DRA7:
+ ret = ti_dra7_xbar_probe(pdev);
+ break;
+ case TI_XBAR_AM335X:
+ ret = ti_am335x_xbar_probe(pdev);
+ break;
+ default:
+ dev_err(&pdev->dev, "Unsupported crossbar\n");
+ ret = -ENODEV;
+ break;
+ }
+
+ return ret;
+}
+
+static struct platform_driver ti_dma_xbar_driver = {
+ .driver = {
+ .name = "ti-dma-crossbar",
+ .of_match_table = of_match_ptr(ti_dma_xbar_match),
+ },
+ .probe = ti_dma_xbar_probe,
+};
+
+static int omap_dmaxbar_init(void)
+{
+ return platform_driver_register(&ti_dma_xbar_driver);
+}
+arch_initcall(omap_dmaxbar_init);
diff --git a/drivers/dma/ti/edma.c b/drivers/dma/ti/edma.c
new file mode 100644
index 0000000..ceabdea
--- /dev/null
+++ b/drivers/dma/ti/edma.c
@@ -0,0 +1,2565 @@
+/*
+ * TI EDMA DMA engine driver
+ *
+ * Copyright 2012 Texas Instruments
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License as
+ * published by the Free Software Foundation version 2.
+ *
+ * This program is distributed "as is" WITHOUT ANY WARRANTY of any
+ * kind, whether express or implied; without even the implied warranty
+ * of MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
+ * GNU General Public License for more details.
+ */
+
+#include <linux/dmaengine.h>
+#include <linux/dma-mapping.h>
+#include <linux/edma.h>
+#include <linux/err.h>
+#include <linux/init.h>
+#include <linux/interrupt.h>
+#include <linux/list.h>
+#include <linux/module.h>
+#include <linux/platform_device.h>
+#include <linux/slab.h>
+#include <linux/spinlock.h>
+#include <linux/of.h>
+#include <linux/of_dma.h>
+#include <linux/of_irq.h>
+#include <linux/of_address.h>
+#include <linux/of_device.h>
+#include <linux/pm_runtime.h>
+
+#include <linux/platform_data/edma.h>
+
+#include "../dmaengine.h"
+#include "../virt-dma.h"
+
+/* Offsets matching "struct edmacc_param" */
+#define PARM_OPT 0x00
+#define PARM_SRC 0x04
+#define PARM_A_B_CNT 0x08
+#define PARM_DST 0x0c
+#define PARM_SRC_DST_BIDX 0x10
+#define PARM_LINK_BCNTRLD 0x14
+#define PARM_SRC_DST_CIDX 0x18
+#define PARM_CCNT 0x1c
+
+#define PARM_SIZE 0x20
+
+/* Offsets for EDMA CC global channel registers and their shadows */
+#define SH_ER 0x00 /* 64 bits */
+#define SH_ECR 0x08 /* 64 bits */
+#define SH_ESR 0x10 /* 64 bits */
+#define SH_CER 0x18 /* 64 bits */
+#define SH_EER 0x20 /* 64 bits */
+#define SH_EECR 0x28 /* 64 bits */
+#define SH_EESR 0x30 /* 64 bits */
+#define SH_SER 0x38 /* 64 bits */
+#define SH_SECR 0x40 /* 64 bits */
+#define SH_IER 0x50 /* 64 bits */
+#define SH_IECR 0x58 /* 64 bits */
+#define SH_IESR 0x60 /* 64 bits */
+#define SH_IPR 0x68 /* 64 bits */
+#define SH_ICR 0x70 /* 64 bits */
+#define SH_IEVAL 0x78
+#define SH_QER 0x80
+#define SH_QEER 0x84
+#define SH_QEECR 0x88
+#define SH_QEESR 0x8c
+#define SH_QSER 0x90
+#define SH_QSECR 0x94
+#define SH_SIZE 0x200
+
+/* Offsets for EDMA CC global registers */
+#define EDMA_REV 0x0000
+#define EDMA_CCCFG 0x0004
+#define EDMA_QCHMAP 0x0200 /* 8 registers */
+#define EDMA_DMAQNUM 0x0240 /* 8 registers (4 on OMAP-L1xx) */
+#define EDMA_QDMAQNUM 0x0260
+#define EDMA_QUETCMAP 0x0280
+#define EDMA_QUEPRI 0x0284
+#define EDMA_EMR 0x0300 /* 64 bits */
+#define EDMA_EMCR 0x0308 /* 64 bits */
+#define EDMA_QEMR 0x0310
+#define EDMA_QEMCR 0x0314
+#define EDMA_CCERR 0x0318
+#define EDMA_CCERRCLR 0x031c
+#define EDMA_EEVAL 0x0320
+#define EDMA_DRAE 0x0340 /* 4 x 64 bits*/
+#define EDMA_QRAE 0x0380 /* 4 registers */
+#define EDMA_QUEEVTENTRY 0x0400 /* 2 x 16 registers */
+#define EDMA_QSTAT 0x0600 /* 2 registers */
+#define EDMA_QWMTHRA 0x0620
+#define EDMA_QWMTHRB 0x0624
+#define EDMA_CCSTAT 0x0640
+
+#define EDMA_M 0x1000 /* global channel registers */
+#define EDMA_ECR 0x1008
+#define EDMA_ECRH 0x100C
+#define EDMA_SHADOW0 0x2000 /* 4 shadow regions */
+#define EDMA_PARM 0x4000 /* PaRAM entries */
+
+#define PARM_OFFSET(param_no) (EDMA_PARM + ((param_no) << 5))
+
+#define EDMA_DCHMAP 0x0100 /* 64 registers */
+
+/* CCCFG register */
+#define GET_NUM_DMACH(x) (x & 0x7) /* bits 0-2 */
+#define GET_NUM_QDMACH(x) ((x & 0x70) >> 4) /* bits 4-6 */
+#define GET_NUM_PAENTRY(x) ((x & 0x7000) >> 12) /* bits 12-14 */
+#define GET_NUM_EVQUE(x) ((x & 0x70000) >> 16) /* bits 16-18 */
+#define GET_NUM_REGN(x) ((x & 0x300000) >> 20) /* bits 20-21 */
+#define CHMAP_EXIST BIT(24)
+
+/* CCSTAT register */
+#define EDMA_CCSTAT_ACTV BIT(4)
+
+/*
+ * Max of 20 segments per channel to conserve PaRAM slots
+ * Also note that MAX_NR_SG should be atleast the no.of periods
+ * that are required for ASoC, otherwise DMA prep calls will
+ * fail. Today davinci-pcm is the only user of this driver and
+ * requires atleast 17 slots, so we setup the default to 20.
+ */
+#define MAX_NR_SG 20
+#define EDMA_MAX_SLOTS MAX_NR_SG
+#define EDMA_DESCRIPTORS 16
+
+#define EDMA_CHANNEL_ANY -1 /* for edma_alloc_channel() */
+#define EDMA_SLOT_ANY -1 /* for edma_alloc_slot() */
+#define EDMA_CONT_PARAMS_ANY 1001
+#define EDMA_CONT_PARAMS_FIXED_EXACT 1002
+#define EDMA_CONT_PARAMS_FIXED_NOT_EXACT 1003
+
+/* PaRAM slots are laid out like this */
+struct edmacc_param {
+ u32 opt;
+ u32 src;
+ u32 a_b_cnt;
+ u32 dst;
+ u32 src_dst_bidx;
+ u32 link_bcntrld;
+ u32 src_dst_cidx;
+ u32 ccnt;
+} __packed;
+
+/* fields in edmacc_param.opt */
+#define SAM BIT(0)
+#define DAM BIT(1)
+#define SYNCDIM BIT(2)
+#define STATIC BIT(3)
+#define EDMA_FWID (0x07 << 8)
+#define TCCMODE BIT(11)
+#define EDMA_TCC(t) ((t) << 12)
+#define TCINTEN BIT(20)
+#define ITCINTEN BIT(21)
+#define TCCHEN BIT(22)
+#define ITCCHEN BIT(23)
+
+struct edma_pset {
+ u32 len;
+ dma_addr_t addr;
+ struct edmacc_param param;
+};
+
+struct edma_desc {
+ struct virt_dma_desc vdesc;
+ struct list_head node;
+ enum dma_transfer_direction direction;
+ int cyclic;
+ int absync;
+ int pset_nr;
+ struct edma_chan *echan;
+ int processed;
+
+ /*
+ * The following 4 elements are used for residue accounting.
+ *
+ * - processed_stat: the number of SG elements we have traversed
+ * so far to cover accounting. This is updated directly to processed
+ * during edma_callback and is always <= processed, because processed
+ * refers to the number of pending transfer (programmed to EDMA
+ * controller), where as processed_stat tracks number of transfers
+ * accounted for so far.
+ *
+ * - residue: The amount of bytes we have left to transfer for this desc
+ *
+ * - residue_stat: The residue in bytes of data we have covered
+ * so far for accounting. This is updated directly to residue
+ * during callbacks to keep it current.
+ *
+ * - sg_len: Tracks the length of the current intermediate transfer,
+ * this is required to update the residue during intermediate transfer
+ * completion callback.
+ */
+ int processed_stat;
+ u32 sg_len;
+ u32 residue;
+ u32 residue_stat;
+
+ struct edma_pset pset[0];
+};
+
+struct edma_cc;
+
+struct edma_tc {
+ struct device_node *node;
+ u16 id;
+};
+
+struct edma_chan {
+ struct virt_dma_chan vchan;
+ struct list_head node;
+ struct edma_desc *edesc;
+ struct edma_cc *ecc;
+ struct edma_tc *tc;
+ int ch_num;
+ bool alloced;
+ bool hw_triggered;
+ int slot[EDMA_MAX_SLOTS];
+ int missed;
+ struct dma_slave_config cfg;
+};
+
+struct edma_cc {
+ struct device *dev;
+ struct edma_soc_info *info;
+ void __iomem *base;
+ int id;
+ bool legacy_mode;
+
+ /* eDMA3 resource information */
+ unsigned num_channels;
+ unsigned num_qchannels;
+ unsigned num_region;
+ unsigned num_slots;
+ unsigned num_tc;
+ bool chmap_exist;
+ enum dma_event_q default_queue;
+
+ unsigned int ccint;
+ unsigned int ccerrint;
+
+ /*
+ * The slot_inuse bit for each PaRAM slot is clear unless the slot is
+ * in use by Linux or if it is allocated to be used by DSP.
+ */
+ unsigned long *slot_inuse;
+
+ struct dma_device dma_slave;
+ struct dma_device *dma_memcpy;
+ struct edma_chan *slave_chans;
+ struct edma_tc *tc_list;
+ int dummy_slot;
+};
+
+/* dummy param set used to (re)initialize parameter RAM slots */
+static const struct edmacc_param dummy_paramset = {
+ .link_bcntrld = 0xffff,
+ .ccnt = 1,
+};
+
+#define EDMA_BINDING_LEGACY 0
+#define EDMA_BINDING_TPCC 1
+static const u32 edma_binding_type[] = {
+ [EDMA_BINDING_LEGACY] = EDMA_BINDING_LEGACY,
+ [EDMA_BINDING_TPCC] = EDMA_BINDING_TPCC,
+};
+
+static const struct of_device_id edma_of_ids[] = {
+ {
+ .compatible = "ti,edma3",
+ .data = &edma_binding_type[EDMA_BINDING_LEGACY],
+ },
+ {
+ .compatible = "ti,edma3-tpcc",
+ .data = &edma_binding_type[EDMA_BINDING_TPCC],
+ },
+ {}
+};
+MODULE_DEVICE_TABLE(of, edma_of_ids);
+
+static const struct of_device_id edma_tptc_of_ids[] = {
+ { .compatible = "ti,edma3-tptc", },
+ {}
+};
+MODULE_DEVICE_TABLE(of, edma_tptc_of_ids);
+
+static inline unsigned int edma_read(struct edma_cc *ecc, int offset)
+{
+ return (unsigned int)__raw_readl(ecc->base + offset);
+}
+
+static inline void edma_write(struct edma_cc *ecc, int offset, int val)
+{
+ __raw_writel(val, ecc->base + offset);
+}
+
+static inline void edma_modify(struct edma_cc *ecc, int offset, unsigned and,
+ unsigned or)
+{
+ unsigned val = edma_read(ecc, offset);
+
+ val &= and;
+ val |= or;
+ edma_write(ecc, offset, val);
+}
+
+static inline void edma_and(struct edma_cc *ecc, int offset, unsigned and)
+{
+ unsigned val = edma_read(ecc, offset);
+
+ val &= and;
+ edma_write(ecc, offset, val);
+}
+
+static inline void edma_or(struct edma_cc *ecc, int offset, unsigned or)
+{
+ unsigned val = edma_read(ecc, offset);
+
+ val |= or;
+ edma_write(ecc, offset, val);
+}
+
+static inline unsigned int edma_read_array(struct edma_cc *ecc, int offset,
+ int i)
+{
+ return edma_read(ecc, offset + (i << 2));
+}
+
+static inline void edma_write_array(struct edma_cc *ecc, int offset, int i,
+ unsigned val)
+{
+ edma_write(ecc, offset + (i << 2), val);
+}
+
+static inline void edma_modify_array(struct edma_cc *ecc, int offset, int i,
+ unsigned and, unsigned or)
+{
+ edma_modify(ecc, offset + (i << 2), and, or);
+}
+
+static inline void edma_or_array(struct edma_cc *ecc, int offset, int i,
+ unsigned or)
+{
+ edma_or(ecc, offset + (i << 2), or);
+}
+
+static inline void edma_or_array2(struct edma_cc *ecc, int offset, int i, int j,
+ unsigned or)
+{
+ edma_or(ecc, offset + ((i * 2 + j) << 2), or);
+}
+
+static inline void edma_write_array2(struct edma_cc *ecc, int offset, int i,
+ int j, unsigned val)
+{
+ edma_write(ecc, offset + ((i * 2 + j) << 2), val);
+}
+
+static inline unsigned int edma_shadow0_read(struct edma_cc *ecc, int offset)
+{
+ return edma_read(ecc, EDMA_SHADOW0 + offset);
+}
+
+static inline unsigned int edma_shadow0_read_array(struct edma_cc *ecc,
+ int offset, int i)
+{
+ return edma_read(ecc, EDMA_SHADOW0 + offset + (i << 2));
+}
+
+static inline void edma_shadow0_write(struct edma_cc *ecc, int offset,
+ unsigned val)
+{
+ edma_write(ecc, EDMA_SHADOW0 + offset, val);
+}
+
+static inline void edma_shadow0_write_array(struct edma_cc *ecc, int offset,
+ int i, unsigned val)
+{
+ edma_write(ecc, EDMA_SHADOW0 + offset + (i << 2), val);
+}
+
+static inline unsigned int edma_param_read(struct edma_cc *ecc, int offset,
+ int param_no)
+{
+ return edma_read(ecc, EDMA_PARM + offset + (param_no << 5));
+}
+
+static inline void edma_param_write(struct edma_cc *ecc, int offset,
+ int param_no, unsigned val)
+{
+ edma_write(ecc, EDMA_PARM + offset + (param_no << 5), val);
+}
+
+static inline void edma_param_modify(struct edma_cc *ecc, int offset,
+ int param_no, unsigned and, unsigned or)
+{
+ edma_modify(ecc, EDMA_PARM + offset + (param_no << 5), and, or);
+}
+
+static inline void edma_param_and(struct edma_cc *ecc, int offset, int param_no,
+ unsigned and)
+{
+ edma_and(ecc, EDMA_PARM + offset + (param_no << 5), and);
+}
+
+static inline void edma_param_or(struct edma_cc *ecc, int offset, int param_no,
+ unsigned or)
+{
+ edma_or(ecc, EDMA_PARM + offset + (param_no << 5), or);
+}
+
+static inline void edma_set_bits(int offset, int len, unsigned long *p)
+{
+ for (; len > 0; len--)
+ set_bit(offset + (len - 1), p);
+}
+
+static void edma_assign_priority_to_queue(struct edma_cc *ecc, int queue_no,
+ int priority)
+{
+ int bit = queue_no * 4;
+
+ edma_modify(ecc, EDMA_QUEPRI, ~(0x7 << bit), ((priority & 0x7) << bit));
+}
+
+static void edma_set_chmap(struct edma_chan *echan, int slot)
+{
+ struct edma_cc *ecc = echan->ecc;
+ int channel = EDMA_CHAN_SLOT(echan->ch_num);
+
+ if (ecc->chmap_exist) {
+ slot = EDMA_CHAN_SLOT(slot);
+ edma_write_array(ecc, EDMA_DCHMAP, channel, (slot << 5));
+ }
+}
+
+static void edma_setup_interrupt(struct edma_chan *echan, bool enable)
+{
+ struct edma_cc *ecc = echan->ecc;
+ int channel = EDMA_CHAN_SLOT(echan->ch_num);
+
+ if (enable) {
+ edma_shadow0_write_array(ecc, SH_ICR, channel >> 5,
+ BIT(channel & 0x1f));
+ edma_shadow0_write_array(ecc, SH_IESR, channel >> 5,
+ BIT(channel & 0x1f));
+ } else {
+ edma_shadow0_write_array(ecc, SH_IECR, channel >> 5,
+ BIT(channel & 0x1f));
+ }
+}
+
+/*
+ * paRAM slot management functions
+ */
+static void edma_write_slot(struct edma_cc *ecc, unsigned slot,
+ const struct edmacc_param *param)
+{
+ slot = EDMA_CHAN_SLOT(slot);
+ if (slot >= ecc->num_slots)
+ return;
+ memcpy_toio(ecc->base + PARM_OFFSET(slot), param, PARM_SIZE);
+}
+
+static int edma_read_slot(struct edma_cc *ecc, unsigned slot,
+ struct edmacc_param *param)
+{
+ slot = EDMA_CHAN_SLOT(slot);
+ if (slot >= ecc->num_slots)
+ return -EINVAL;
+ memcpy_fromio(param, ecc->base + PARM_OFFSET(slot), PARM_SIZE);
+
+ return 0;
+}
+
+/**
+ * edma_alloc_slot - allocate DMA parameter RAM
+ * @ecc: pointer to edma_cc struct
+ * @slot: specific slot to allocate; negative for "any unused slot"
+ *
+ * This allocates a parameter RAM slot, initializing it to hold a
+ * dummy transfer. Slots allocated using this routine have not been
+ * mapped to a hardware DMA channel, and will normally be used by
+ * linking to them from a slot associated with a DMA channel.
+ *
+ * Normal use is to pass EDMA_SLOT_ANY as the @slot, but specific
+ * slots may be allocated on behalf of DSP firmware.
+ *
+ * Returns the number of the slot, else negative errno.
+ */
+static int edma_alloc_slot(struct edma_cc *ecc, int slot)
+{
+ if (slot >= 0) {
+ slot = EDMA_CHAN_SLOT(slot);
+ /* Requesting entry paRAM slot for a HW triggered channel. */
+ if (ecc->chmap_exist && slot < ecc->num_channels)
+ slot = EDMA_SLOT_ANY;
+ }
+
+ if (slot < 0) {
+ if (ecc->chmap_exist)
+ slot = 0;
+ else
+ slot = ecc->num_channels;
+ for (;;) {
+ slot = find_next_zero_bit(ecc->slot_inuse,
+ ecc->num_slots,
+ slot);
+ if (slot == ecc->num_slots)
+ return -ENOMEM;
+ if (!test_and_set_bit(slot, ecc->slot_inuse))
+ break;
+ }
+ } else if (slot >= ecc->num_slots) {
+ return -EINVAL;
+ } else if (test_and_set_bit(slot, ecc->slot_inuse)) {
+ return -EBUSY;
+ }
+
+ edma_write_slot(ecc, slot, &dummy_paramset);
+
+ return EDMA_CTLR_CHAN(ecc->id, slot);
+}
+
+static void edma_free_slot(struct edma_cc *ecc, unsigned slot)
+{
+ slot = EDMA_CHAN_SLOT(slot);
+ if (slot >= ecc->num_slots)
+ return;
+
+ edma_write_slot(ecc, slot, &dummy_paramset);
+ clear_bit(slot, ecc->slot_inuse);
+}
+
+/**
+ * edma_link - link one parameter RAM slot to another
+ * @ecc: pointer to edma_cc struct
+ * @from: parameter RAM slot originating the link
+ * @to: parameter RAM slot which is the link target
+ *
+ * The originating slot should not be part of any active DMA transfer.
+ */
+static void edma_link(struct edma_cc *ecc, unsigned from, unsigned to)
+{
+ if (unlikely(EDMA_CTLR(from) != EDMA_CTLR(to)))
+ dev_warn(ecc->dev, "Ignoring eDMA instance for linking\n");
+
+ from = EDMA_CHAN_SLOT(from);
+ to = EDMA_CHAN_SLOT(to);
+ if (from >= ecc->num_slots || to >= ecc->num_slots)
+ return;
+
+ edma_param_modify(ecc, PARM_LINK_BCNTRLD, from, 0xffff0000,
+ PARM_OFFSET(to));
+}
+
+/**
+ * edma_get_position - returns the current transfer point
+ * @ecc: pointer to edma_cc struct
+ * @slot: parameter RAM slot being examined
+ * @dst: true selects the dest position, false the source
+ *
+ * Returns the position of the current active slot
+ */
+static dma_addr_t edma_get_position(struct edma_cc *ecc, unsigned slot,
+ bool dst)
+{
+ u32 offs;
+
+ slot = EDMA_CHAN_SLOT(slot);
+ offs = PARM_OFFSET(slot);
+ offs += dst ? PARM_DST : PARM_SRC;
+
+ return edma_read(ecc, offs);
+}
+
+/*
+ * Channels with event associations will be triggered by their hardware
+ * events, and channels without such associations will be triggered by
+ * software. (At this writing there is no interface for using software
+ * triggers except with channels that don't support hardware triggers.)
+ */
+static void edma_start(struct edma_chan *echan)
+{
+ struct edma_cc *ecc = echan->ecc;
+ int channel = EDMA_CHAN_SLOT(echan->ch_num);
+ int j = (channel >> 5);
+ unsigned int mask = BIT(channel & 0x1f);
+
+ if (!echan->hw_triggered) {
+ /* EDMA channels without event association */
+ dev_dbg(ecc->dev, "ESR%d %08x\n", j,
+ edma_shadow0_read_array(ecc, SH_ESR, j));
+ edma_shadow0_write_array(ecc, SH_ESR, j, mask);
+ } else {
+ /* EDMA channel with event association */
+ dev_dbg(ecc->dev, "ER%d %08x\n", j,
+ edma_shadow0_read_array(ecc, SH_ER, j));
+ /* Clear any pending event or error */
+ edma_write_array(ecc, EDMA_ECR, j, mask);
+ edma_write_array(ecc, EDMA_EMCR, j, mask);
+ /* Clear any SER */
+ edma_shadow0_write_array(ecc, SH_SECR, j, mask);
+ edma_shadow0_write_array(ecc, SH_EESR, j, mask);
+ dev_dbg(ecc->dev, "EER%d %08x\n", j,
+ edma_shadow0_read_array(ecc, SH_EER, j));
+ }
+}
+
+static void edma_stop(struct edma_chan *echan)
+{
+ struct edma_cc *ecc = echan->ecc;
+ int channel = EDMA_CHAN_SLOT(echan->ch_num);
+ int j = (channel >> 5);
+ unsigned int mask = BIT(channel & 0x1f);
+
+ edma_shadow0_write_array(ecc, SH_EECR, j, mask);
+ edma_shadow0_write_array(ecc, SH_ECR, j, mask);
+ edma_shadow0_write_array(ecc, SH_SECR, j, mask);
+ edma_write_array(ecc, EDMA_EMCR, j, mask);
+
+ /* clear possibly pending completion interrupt */
+ edma_shadow0_write_array(ecc, SH_ICR, j, mask);
+
+ dev_dbg(ecc->dev, "EER%d %08x\n", j,
+ edma_shadow0_read_array(ecc, SH_EER, j));
+
+ /* REVISIT: consider guarding against inappropriate event
+ * chaining by overwriting with dummy_paramset.
+ */
+}
+
+/*
+ * Temporarily disable EDMA hardware events on the specified channel,
+ * preventing them from triggering new transfers
+ */
+static void edma_pause(struct edma_chan *echan)
+{
+ int channel = EDMA_CHAN_SLOT(echan->ch_num);
+ unsigned int mask = BIT(channel & 0x1f);
+
+ edma_shadow0_write_array(echan->ecc, SH_EECR, channel >> 5, mask);
+}
+
+/* Re-enable EDMA hardware events on the specified channel. */
+static void edma_resume(struct edma_chan *echan)
+{
+ int channel = EDMA_CHAN_SLOT(echan->ch_num);
+ unsigned int mask = BIT(channel & 0x1f);
+
+ edma_shadow0_write_array(echan->ecc, SH_EESR, channel >> 5, mask);
+}
+
+static void edma_trigger_channel(struct edma_chan *echan)
+{
+ struct edma_cc *ecc = echan->ecc;
+ int channel = EDMA_CHAN_SLOT(echan->ch_num);
+ unsigned int mask = BIT(channel & 0x1f);
+
+ edma_shadow0_write_array(ecc, SH_ESR, (channel >> 5), mask);
+
+ dev_dbg(ecc->dev, "ESR%d %08x\n", (channel >> 5),
+ edma_shadow0_read_array(ecc, SH_ESR, (channel >> 5)));
+}
+
+static void edma_clean_channel(struct edma_chan *echan)
+{
+ struct edma_cc *ecc = echan->ecc;
+ int channel = EDMA_CHAN_SLOT(echan->ch_num);
+ int j = (channel >> 5);
+ unsigned int mask = BIT(channel & 0x1f);
+
+ dev_dbg(ecc->dev, "EMR%d %08x\n", j, edma_read_array(ecc, EDMA_EMR, j));
+ edma_shadow0_write_array(ecc, SH_ECR, j, mask);
+ /* Clear the corresponding EMR bits */
+ edma_write_array(ecc, EDMA_EMCR, j, mask);
+ /* Clear any SER */
+ edma_shadow0_write_array(ecc, SH_SECR, j, mask);
+ edma_write(ecc, EDMA_CCERRCLR, BIT(16) | BIT(1) | BIT(0));
+}
+
+/* Move channel to a specific event queue */
+static void edma_assign_channel_eventq(struct edma_chan *echan,
+ enum dma_event_q eventq_no)
+{
+ struct edma_cc *ecc = echan->ecc;
+ int channel = EDMA_CHAN_SLOT(echan->ch_num);
+ int bit = (channel & 0x7) * 4;
+
+ /* default to low priority queue */
+ if (eventq_no == EVENTQ_DEFAULT)
+ eventq_no = ecc->default_queue;
+ if (eventq_no >= ecc->num_tc)
+ return;
+
+ eventq_no &= 7;
+ edma_modify_array(ecc, EDMA_DMAQNUM, (channel >> 3), ~(0x7 << bit),
+ eventq_no << bit);
+}
+
+static int edma_alloc_channel(struct edma_chan *echan,
+ enum dma_event_q eventq_no)
+{
+ struct edma_cc *ecc = echan->ecc;
+ int channel = EDMA_CHAN_SLOT(echan->ch_num);
+
+ /* ensure access through shadow region 0 */
+ edma_or_array2(ecc, EDMA_DRAE, 0, channel >> 5, BIT(channel & 0x1f));
+
+ /* ensure no events are pending */
+ edma_stop(echan);
+
+ edma_setup_interrupt(echan, true);
+
+ edma_assign_channel_eventq(echan, eventq_no);
+
+ return 0;
+}
+
+static void edma_free_channel(struct edma_chan *echan)
+{
+ /* ensure no events are pending */
+ edma_stop(echan);
+ /* REVISIT should probably take out of shadow region 0 */
+ edma_setup_interrupt(echan, false);
+}
+
+static inline struct edma_cc *to_edma_cc(struct dma_device *d)
+{
+ return container_of(d, struct edma_cc, dma_slave);
+}
+
+static inline struct edma_chan *to_edma_chan(struct dma_chan *c)
+{
+ return container_of(c, struct edma_chan, vchan.chan);
+}
+
+static inline struct edma_desc *to_edma_desc(struct dma_async_tx_descriptor *tx)
+{
+ return container_of(tx, struct edma_desc, vdesc.tx);
+}
+
+static void edma_desc_free(struct virt_dma_desc *vdesc)
+{
+ kfree(container_of(vdesc, struct edma_desc, vdesc));
+}
+
+/* Dispatch a queued descriptor to the controller (caller holds lock) */
+static void edma_execute(struct edma_chan *echan)
+{
+ struct edma_cc *ecc = echan->ecc;
+ struct virt_dma_desc *vdesc;
+ struct edma_desc *edesc;
+ struct device *dev = echan->vchan.chan.device->dev;
+ int i, j, left, nslots;
+
+ if (!echan->edesc) {
+ /* Setup is needed for the first transfer */
+ vdesc = vchan_next_desc(&echan->vchan);
+ if (!vdesc)
+ return;
+ list_del(&vdesc->node);
+ echan->edesc = to_edma_desc(&vdesc->tx);
+ }
+
+ edesc = echan->edesc;
+
+ /* Find out how many left */
+ left = edesc->pset_nr - edesc->processed;
+ nslots = min(MAX_NR_SG, left);
+ edesc->sg_len = 0;
+
+ /* Write descriptor PaRAM set(s) */
+ for (i = 0; i < nslots; i++) {
+ j = i + edesc->processed;
+ edma_write_slot(ecc, echan->slot[i], &edesc->pset[j].param);
+ edesc->sg_len += edesc->pset[j].len;
+ dev_vdbg(dev,
+ "\n pset[%d]:\n"
+ " chnum\t%d\n"
+ " slot\t%d\n"
+ " opt\t%08x\n"
+ " src\t%08x\n"
+ " dst\t%08x\n"
+ " abcnt\t%08x\n"
+ " ccnt\t%08x\n"
+ " bidx\t%08x\n"
+ " cidx\t%08x\n"
+ " lkrld\t%08x\n",
+ j, echan->ch_num, echan->slot[i],
+ edesc->pset[j].param.opt,
+ edesc->pset[j].param.src,
+ edesc->pset[j].param.dst,
+ edesc->pset[j].param.a_b_cnt,
+ edesc->pset[j].param.ccnt,
+ edesc->pset[j].param.src_dst_bidx,
+ edesc->pset[j].param.src_dst_cidx,
+ edesc->pset[j].param.link_bcntrld);
+ /* Link to the previous slot if not the last set */
+ if (i != (nslots - 1))
+ edma_link(ecc, echan->slot[i], echan->slot[i + 1]);
+ }
+
+ edesc->processed += nslots;
+
+ /*
+ * If this is either the last set in a set of SG-list transactions
+ * then setup a link to the dummy slot, this results in all future
+ * events being absorbed and that's OK because we're done
+ */
+ if (edesc->processed == edesc->pset_nr) {
+ if (edesc->cyclic)
+ edma_link(ecc, echan->slot[nslots - 1], echan->slot[1]);
+ else
+ edma_link(ecc, echan->slot[nslots - 1],
+ echan->ecc->dummy_slot);
+ }
+
+ if (echan->missed) {
+ /*
+ * This happens due to setup times between intermediate
+ * transfers in long SG lists which have to be broken up into
+ * transfers of MAX_NR_SG
+ */
+ dev_dbg(dev, "missed event on channel %d\n", echan->ch_num);
+ edma_clean_channel(echan);
+ edma_stop(echan);
+ edma_start(echan);
+ edma_trigger_channel(echan);
+ echan->missed = 0;
+ } else if (edesc->processed <= MAX_NR_SG) {
+ dev_dbg(dev, "first transfer starting on channel %d\n",
+ echan->ch_num);
+ edma_start(echan);
+ } else {
+ dev_dbg(dev, "chan: %d: completed %d elements, resuming\n",
+ echan->ch_num, edesc->processed);
+ edma_resume(echan);
+ }
+}
+
+static int edma_terminate_all(struct dma_chan *chan)
+{
+ struct edma_chan *echan = to_edma_chan(chan);
+ unsigned long flags;
+ LIST_HEAD(head);
+
+ spin_lock_irqsave(&echan->vchan.lock, flags);
+
+ /*
+ * Stop DMA activity: we assume the callback will not be called
+ * after edma_dma() returns (even if it does, it will see
+ * echan->edesc is NULL and exit.)
+ */
+ if (echan->edesc) {
+ edma_stop(echan);
+ /* Move the cyclic channel back to default queue */
+ if (!echan->tc && echan->edesc->cyclic)
+ edma_assign_channel_eventq(echan, EVENTQ_DEFAULT);
+
+ vchan_terminate_vdesc(&echan->edesc->vdesc);
+ echan->edesc = NULL;
+ }
+
+ vchan_get_all_descriptors(&echan->vchan, &head);
+ spin_unlock_irqrestore(&echan->vchan.lock, flags);
+ vchan_dma_desc_free_list(&echan->vchan, &head);
+
+ return 0;
+}
+
+static void edma_synchronize(struct dma_chan *chan)
+{
+ struct edma_chan *echan = to_edma_chan(chan);
+
+ vchan_synchronize(&echan->vchan);
+}
+
+static int edma_slave_config(struct dma_chan *chan,
+ struct dma_slave_config *cfg)
+{
+ struct edma_chan *echan = to_edma_chan(chan);
+
+ if (cfg->src_addr_width == DMA_SLAVE_BUSWIDTH_8_BYTES ||
+ cfg->dst_addr_width == DMA_SLAVE_BUSWIDTH_8_BYTES)
+ return -EINVAL;
+
+ if (cfg->src_maxburst > chan->device->max_burst ||
+ cfg->dst_maxburst > chan->device->max_burst)
+ return -EINVAL;
+
+ memcpy(&echan->cfg, cfg, sizeof(echan->cfg));
+
+ return 0;
+}
+
+static int edma_dma_pause(struct dma_chan *chan)
+{
+ struct edma_chan *echan = to_edma_chan(chan);
+
+ if (!echan->edesc)
+ return -EINVAL;
+
+ edma_pause(echan);
+ return 0;
+}
+
+static int edma_dma_resume(struct dma_chan *chan)
+{
+ struct edma_chan *echan = to_edma_chan(chan);
+
+ edma_resume(echan);
+ return 0;
+}
+
+/*
+ * A PaRAM set configuration abstraction used by other modes
+ * @chan: Channel who's PaRAM set we're configuring
+ * @pset: PaRAM set to initialize and setup.
+ * @src_addr: Source address of the DMA
+ * @dst_addr: Destination address of the DMA
+ * @burst: In units of dev_width, how much to send
+ * @dev_width: How much is the dev_width
+ * @dma_length: Total length of the DMA transfer
+ * @direction: Direction of the transfer
+ */
+static int edma_config_pset(struct dma_chan *chan, struct edma_pset *epset,
+ dma_addr_t src_addr, dma_addr_t dst_addr, u32 burst,
+ unsigned int acnt, unsigned int dma_length,
+ enum dma_transfer_direction direction)
+{
+ struct edma_chan *echan = to_edma_chan(chan);
+ struct device *dev = chan->device->dev;
+ struct edmacc_param *param = &epset->param;
+ int bcnt, ccnt, cidx;
+ int src_bidx, dst_bidx, src_cidx, dst_cidx;
+ int absync;
+
+ /* src/dst_maxburst == 0 is the same case as src/dst_maxburst == 1 */
+ if (!burst)
+ burst = 1;
+ /*
+ * If the maxburst is equal to the fifo width, use
+ * A-synced transfers. This allows for large contiguous
+ * buffer transfers using only one PaRAM set.
+ */
+ if (burst == 1) {
+ /*
+ * For the A-sync case, bcnt and ccnt are the remainder
+ * and quotient respectively of the division of:
+ * (dma_length / acnt) by (SZ_64K -1). This is so
+ * that in case bcnt over flows, we have ccnt to use.
+ * Note: In A-sync tranfer only, bcntrld is used, but it
+ * only applies for sg_dma_len(sg) >= SZ_64K.
+ * In this case, the best way adopted is- bccnt for the
+ * first frame will be the remainder below. Then for
+ * every successive frame, bcnt will be SZ_64K-1. This
+ * is assured as bcntrld = 0xffff in end of function.
+ */
+ absync = false;
+ ccnt = dma_length / acnt / (SZ_64K - 1);
+ bcnt = dma_length / acnt - ccnt * (SZ_64K - 1);
+ /*
+ * If bcnt is non-zero, we have a remainder and hence an
+ * extra frame to transfer, so increment ccnt.
+ */
+ if (bcnt)
+ ccnt++;
+ else
+ bcnt = SZ_64K - 1;
+ cidx = acnt;
+ } else {
+ /*
+ * If maxburst is greater than the fifo address_width,
+ * use AB-synced transfers where A count is the fifo
+ * address_width and B count is the maxburst. In this
+ * case, we are limited to transfers of C count frames
+ * of (address_width * maxburst) where C count is limited
+ * to SZ_64K-1. This places an upper bound on the length
+ * of an SG segment that can be handled.
+ */
+ absync = true;
+ bcnt = burst;
+ ccnt = dma_length / (acnt * bcnt);
+ if (ccnt > (SZ_64K - 1)) {
+ dev_err(dev, "Exceeded max SG segment size\n");
+ return -EINVAL;
+ }
+ cidx = acnt * bcnt;
+ }
+
+ epset->len = dma_length;
+
+ if (direction == DMA_MEM_TO_DEV) {
+ src_bidx = acnt;
+ src_cidx = cidx;
+ dst_bidx = 0;
+ dst_cidx = 0;
+ epset->addr = src_addr;
+ } else if (direction == DMA_DEV_TO_MEM) {
+ src_bidx = 0;
+ src_cidx = 0;
+ dst_bidx = acnt;
+ dst_cidx = cidx;
+ epset->addr = dst_addr;
+ } else if (direction == DMA_MEM_TO_MEM) {
+ src_bidx = acnt;
+ src_cidx = cidx;
+ dst_bidx = acnt;
+ dst_cidx = cidx;
+ } else {
+ dev_err(dev, "%s: direction not implemented yet\n", __func__);
+ return -EINVAL;
+ }
+
+ param->opt = EDMA_TCC(EDMA_CHAN_SLOT(echan->ch_num));
+ /* Configure A or AB synchronized transfers */
+ if (absync)
+ param->opt |= SYNCDIM;
+
+ param->src = src_addr;
+ param->dst = dst_addr;
+
+ param->src_dst_bidx = (dst_bidx << 16) | src_bidx;
+ param->src_dst_cidx = (dst_cidx << 16) | src_cidx;
+
+ param->a_b_cnt = bcnt << 16 | acnt;
+ param->ccnt = ccnt;
+ /*
+ * Only time when (bcntrld) auto reload is required is for
+ * A-sync case, and in this case, a requirement of reload value
+ * of SZ_64K-1 only is assured. 'link' is initially set to NULL
+ * and then later will be populated by edma_execute.
+ */
+ param->link_bcntrld = 0xffffffff;
+ return absync;
+}
+
+static struct dma_async_tx_descriptor *edma_prep_slave_sg(
+ struct dma_chan *chan, struct scatterlist *sgl,
+ unsigned int sg_len, enum dma_transfer_direction direction,
+ unsigned long tx_flags, void *context)
+{
+ struct edma_chan *echan = to_edma_chan(chan);
+ struct device *dev = chan->device->dev;
+ struct edma_desc *edesc;
+ dma_addr_t src_addr = 0, dst_addr = 0;
+ enum dma_slave_buswidth dev_width;
+ u32 burst;
+ struct scatterlist *sg;
+ int i, nslots, ret;
+
+ if (unlikely(!echan || !sgl || !sg_len))
+ return NULL;
+
+ if (direction == DMA_DEV_TO_MEM) {
+ src_addr = echan->cfg.src_addr;
+ dev_width = echan->cfg.src_addr_width;
+ burst = echan->cfg.src_maxburst;
+ } else if (direction == DMA_MEM_TO_DEV) {
+ dst_addr = echan->cfg.dst_addr;
+ dev_width = echan->cfg.dst_addr_width;
+ burst = echan->cfg.dst_maxburst;
+ } else {
+ dev_err(dev, "%s: bad direction: %d\n", __func__, direction);
+ return NULL;
+ }
+
+ if (dev_width == DMA_SLAVE_BUSWIDTH_UNDEFINED) {
+ dev_err(dev, "%s: Undefined slave buswidth\n", __func__);
+ return NULL;
+ }
+
+ edesc = kzalloc(struct_size(edesc, pset, sg_len), GFP_ATOMIC);
+ if (!edesc)
+ return NULL;
+
+ edesc->pset_nr = sg_len;
+ edesc->residue = 0;
+ edesc->direction = direction;
+ edesc->echan = echan;
+
+ /* Allocate a PaRAM slot, if needed */
+ nslots = min_t(unsigned, MAX_NR_SG, sg_len);
+
+ for (i = 0; i < nslots; i++) {
+ if (echan->slot[i] < 0) {
+ echan->slot[i] =
+ edma_alloc_slot(echan->ecc, EDMA_SLOT_ANY);
+ if (echan->slot[i] < 0) {
+ kfree(edesc);
+ dev_err(dev, "%s: Failed to allocate slot\n",
+ __func__);
+ return NULL;
+ }
+ }
+ }
+
+ /* Configure PaRAM sets for each SG */
+ for_each_sg(sgl, sg, sg_len, i) {
+ /* Get address for each SG */
+ if (direction == DMA_DEV_TO_MEM)
+ dst_addr = sg_dma_address(sg);
+ else
+ src_addr = sg_dma_address(sg);
+
+ ret = edma_config_pset(chan, &edesc->pset[i], src_addr,
+ dst_addr, burst, dev_width,
+ sg_dma_len(sg), direction);
+ if (ret < 0) {
+ kfree(edesc);
+ return NULL;
+ }
+
+ edesc->absync = ret;
+ edesc->residue += sg_dma_len(sg);
+
+ if (i == sg_len - 1)
+ /* Enable completion interrupt */
+ edesc->pset[i].param.opt |= TCINTEN;
+ else if (!((i+1) % MAX_NR_SG))
+ /*
+ * Enable early completion interrupt for the
+ * intermediateset. In this case the driver will be
+ * notified when the paRAM set is submitted to TC. This
+ * will allow more time to set up the next set of slots.
+ */
+ edesc->pset[i].param.opt |= (TCINTEN | TCCMODE);
+ }
+ edesc->residue_stat = edesc->residue;
+
+ return vchan_tx_prep(&echan->vchan, &edesc->vdesc, tx_flags);
+}
+
+static struct dma_async_tx_descriptor *edma_prep_dma_memcpy(
+ struct dma_chan *chan, dma_addr_t dest, dma_addr_t src,
+ size_t len, unsigned long tx_flags)
+{
+ int ret, nslots;
+ struct edma_desc *edesc;
+ struct device *dev = chan->device->dev;
+ struct edma_chan *echan = to_edma_chan(chan);
+ unsigned int width, pset_len, array_size;
+
+ if (unlikely(!echan || !len))
+ return NULL;
+
+ /* Align the array size (acnt block) with the transfer properties */
+ switch (__ffs((src | dest | len))) {
+ case 0:
+ array_size = SZ_32K - 1;
+ break;
+ case 1:
+ array_size = SZ_32K - 2;
+ break;
+ default:
+ array_size = SZ_32K - 4;
+ break;
+ }
+
+ if (len < SZ_64K) {
+ /*
+ * Transfer size less than 64K can be handled with one paRAM
+ * slot and with one burst.
+ * ACNT = length
+ */
+ width = len;
+ pset_len = len;
+ nslots = 1;
+ } else {
+ /*
+ * Transfer size bigger than 64K will be handled with maximum of
+ * two paRAM slots.
+ * slot1: (full_length / 32767) times 32767 bytes bursts.
+ * ACNT = 32767, length1: (full_length / 32767) * 32767
+ * slot2: the remaining amount of data after slot1.
+ * ACNT = full_length - length1, length2 = ACNT
+ *
+ * When the full_length is multibple of 32767 one slot can be
+ * used to complete the transfer.
+ */
+ width = array_size;
+ pset_len = rounddown(len, width);
+ /* One slot is enough for lengths multiple of (SZ_32K -1) */
+ if (unlikely(pset_len == len))
+ nslots = 1;
+ else
+ nslots = 2;
+ }
+
+ edesc = kzalloc(struct_size(edesc, pset, nslots), GFP_ATOMIC);
+ if (!edesc)
+ return NULL;
+
+ edesc->pset_nr = nslots;
+ edesc->residue = edesc->residue_stat = len;
+ edesc->direction = DMA_MEM_TO_MEM;
+ edesc->echan = echan;
+
+ ret = edma_config_pset(chan, &edesc->pset[0], src, dest, 1,
+ width, pset_len, DMA_MEM_TO_MEM);
+ if (ret < 0) {
+ kfree(edesc);
+ return NULL;
+ }
+
+ edesc->absync = ret;
+
+ edesc->pset[0].param.opt |= ITCCHEN;
+ if (nslots == 1) {
+ /* Enable transfer complete interrupt */
+ edesc->pset[0].param.opt |= TCINTEN;
+ } else {
+ /* Enable transfer complete chaining for the first slot */
+ edesc->pset[0].param.opt |= TCCHEN;
+
+ if (echan->slot[1] < 0) {
+ echan->slot[1] = edma_alloc_slot(echan->ecc,
+ EDMA_SLOT_ANY);
+ if (echan->slot[1] < 0) {
+ kfree(edesc);
+ dev_err(dev, "%s: Failed to allocate slot\n",
+ __func__);
+ return NULL;
+ }
+ }
+ dest += pset_len;
+ src += pset_len;
+ pset_len = width = len % array_size;
+
+ ret = edma_config_pset(chan, &edesc->pset[1], src, dest, 1,
+ width, pset_len, DMA_MEM_TO_MEM);
+ if (ret < 0) {
+ kfree(edesc);
+ return NULL;
+ }
+
+ edesc->pset[1].param.opt |= ITCCHEN;
+ edesc->pset[1].param.opt |= TCINTEN;
+ }
+
+ return vchan_tx_prep(&echan->vchan, &edesc->vdesc, tx_flags);
+}
+
+static struct dma_async_tx_descriptor *edma_prep_dma_cyclic(
+ struct dma_chan *chan, dma_addr_t buf_addr, size_t buf_len,
+ size_t period_len, enum dma_transfer_direction direction,
+ unsigned long tx_flags)
+{
+ struct edma_chan *echan = to_edma_chan(chan);
+ struct device *dev = chan->device->dev;
+ struct edma_desc *edesc;
+ dma_addr_t src_addr, dst_addr;
+ enum dma_slave_buswidth dev_width;
+ bool use_intermediate = false;
+ u32 burst;
+ int i, ret, nslots;
+
+ if (unlikely(!echan || !buf_len || !period_len))
+ return NULL;
+
+ if (direction == DMA_DEV_TO_MEM) {
+ src_addr = echan->cfg.src_addr;
+ dst_addr = buf_addr;
+ dev_width = echan->cfg.src_addr_width;
+ burst = echan->cfg.src_maxburst;
+ } else if (direction == DMA_MEM_TO_DEV) {
+ src_addr = buf_addr;
+ dst_addr = echan->cfg.dst_addr;
+ dev_width = echan->cfg.dst_addr_width;
+ burst = echan->cfg.dst_maxburst;
+ } else {
+ dev_err(dev, "%s: bad direction: %d\n", __func__, direction);
+ return NULL;
+ }
+
+ if (dev_width == DMA_SLAVE_BUSWIDTH_UNDEFINED) {
+ dev_err(dev, "%s: Undefined slave buswidth\n", __func__);
+ return NULL;
+ }
+
+ if (unlikely(buf_len % period_len)) {
+ dev_err(dev, "Period should be multiple of Buffer length\n");
+ return NULL;
+ }
+
+ nslots = (buf_len / period_len) + 1;
+
+ /*
+ * Cyclic DMA users such as audio cannot tolerate delays introduced
+ * by cases where the number of periods is more than the maximum
+ * number of SGs the EDMA driver can handle at a time. For DMA types
+ * such as Slave SGs, such delays are tolerable and synchronized,
+ * but the synchronization is difficult to achieve with Cyclic and
+ * cannot be guaranteed, so we error out early.
+ */
+ if (nslots > MAX_NR_SG) {
+ /*
+ * If the burst and period sizes are the same, we can put
+ * the full buffer into a single period and activate
+ * intermediate interrupts. This will produce interrupts
+ * after each burst, which is also after each desired period.
+ */
+ if (burst == period_len) {
+ period_len = buf_len;
+ nslots = 2;
+ use_intermediate = true;
+ } else {
+ return NULL;
+ }
+ }
+
+ edesc = kzalloc(struct_size(edesc, pset, nslots), GFP_ATOMIC);
+ if (!edesc)
+ return NULL;
+
+ edesc->cyclic = 1;
+ edesc->pset_nr = nslots;
+ edesc->residue = edesc->residue_stat = buf_len;
+ edesc->direction = direction;
+ edesc->echan = echan;
+
+ dev_dbg(dev, "%s: channel=%d nslots=%d period_len=%zu buf_len=%zu\n",
+ __func__, echan->ch_num, nslots, period_len, buf_len);
+
+ for (i = 0; i < nslots; i++) {
+ /* Allocate a PaRAM slot, if needed */
+ if (echan->slot[i] < 0) {
+ echan->slot[i] =
+ edma_alloc_slot(echan->ecc, EDMA_SLOT_ANY);
+ if (echan->slot[i] < 0) {
+ kfree(edesc);
+ dev_err(dev, "%s: Failed to allocate slot\n",
+ __func__);
+ return NULL;
+ }
+ }
+
+ if (i == nslots - 1) {
+ memcpy(&edesc->pset[i], &edesc->pset[0],
+ sizeof(edesc->pset[0]));
+ break;
+ }
+
+ ret = edma_config_pset(chan, &edesc->pset[i], src_addr,
+ dst_addr, burst, dev_width, period_len,
+ direction);
+ if (ret < 0) {
+ kfree(edesc);
+ return NULL;
+ }
+
+ if (direction == DMA_DEV_TO_MEM)
+ dst_addr += period_len;
+ else
+ src_addr += period_len;
+
+ dev_vdbg(dev, "%s: Configure period %d of buf:\n", __func__, i);
+ dev_vdbg(dev,
+ "\n pset[%d]:\n"
+ " chnum\t%d\n"
+ " slot\t%d\n"
+ " opt\t%08x\n"
+ " src\t%08x\n"
+ " dst\t%08x\n"
+ " abcnt\t%08x\n"
+ " ccnt\t%08x\n"
+ " bidx\t%08x\n"
+ " cidx\t%08x\n"
+ " lkrld\t%08x\n",
+ i, echan->ch_num, echan->slot[i],
+ edesc->pset[i].param.opt,
+ edesc->pset[i].param.src,
+ edesc->pset[i].param.dst,
+ edesc->pset[i].param.a_b_cnt,
+ edesc->pset[i].param.ccnt,
+ edesc->pset[i].param.src_dst_bidx,
+ edesc->pset[i].param.src_dst_cidx,
+ edesc->pset[i].param.link_bcntrld);
+
+ edesc->absync = ret;
+
+ /*
+ * Enable period interrupt only if it is requested
+ */
+ if (tx_flags & DMA_PREP_INTERRUPT) {
+ edesc->pset[i].param.opt |= TCINTEN;
+
+ /* Also enable intermediate interrupts if necessary */
+ if (use_intermediate)
+ edesc->pset[i].param.opt |= ITCINTEN;
+ }
+ }
+
+ /* Place the cyclic channel to highest priority queue */
+ if (!echan->tc)
+ edma_assign_channel_eventq(echan, EVENTQ_0);
+
+ return vchan_tx_prep(&echan->vchan, &edesc->vdesc, tx_flags);
+}
+
+static void edma_completion_handler(struct edma_chan *echan)
+{
+ struct device *dev = echan->vchan.chan.device->dev;
+ struct edma_desc *edesc;
+
+ spin_lock(&echan->vchan.lock);
+ edesc = echan->edesc;
+ if (edesc) {
+ if (edesc->cyclic) {
+ vchan_cyclic_callback(&edesc->vdesc);
+ spin_unlock(&echan->vchan.lock);
+ return;
+ } else if (edesc->processed == edesc->pset_nr) {
+ edesc->residue = 0;
+ edma_stop(echan);
+ vchan_cookie_complete(&edesc->vdesc);
+ echan->edesc = NULL;
+
+ dev_dbg(dev, "Transfer completed on channel %d\n",
+ echan->ch_num);
+ } else {
+ dev_dbg(dev, "Sub transfer completed on channel %d\n",
+ echan->ch_num);
+
+ edma_pause(echan);
+
+ /* Update statistics for tx_status */
+ edesc->residue -= edesc->sg_len;
+ edesc->residue_stat = edesc->residue;
+ edesc->processed_stat = edesc->processed;
+ }
+ edma_execute(echan);
+ }
+
+ spin_unlock(&echan->vchan.lock);
+}
+
+/* eDMA interrupt handler */
+static irqreturn_t dma_irq_handler(int irq, void *data)
+{
+ struct edma_cc *ecc = data;
+ int ctlr;
+ u32 sh_ier;
+ u32 sh_ipr;
+ u32 bank;
+
+ ctlr = ecc->id;
+ if (ctlr < 0)
+ return IRQ_NONE;
+
+ dev_vdbg(ecc->dev, "dma_irq_handler\n");
+
+ sh_ipr = edma_shadow0_read_array(ecc, SH_IPR, 0);
+ if (!sh_ipr) {
+ sh_ipr = edma_shadow0_read_array(ecc, SH_IPR, 1);
+ if (!sh_ipr)
+ return IRQ_NONE;
+ sh_ier = edma_shadow0_read_array(ecc, SH_IER, 1);
+ bank = 1;
+ } else {
+ sh_ier = edma_shadow0_read_array(ecc, SH_IER, 0);
+ bank = 0;
+ }
+
+ do {
+ u32 slot;
+ u32 channel;
+
+ slot = __ffs(sh_ipr);
+ sh_ipr &= ~(BIT(slot));
+
+ if (sh_ier & BIT(slot)) {
+ channel = (bank << 5) | slot;
+ /* Clear the corresponding IPR bits */
+ edma_shadow0_write_array(ecc, SH_ICR, bank, BIT(slot));
+ edma_completion_handler(&ecc->slave_chans[channel]);
+ }
+ } while (sh_ipr);
+
+ edma_shadow0_write(ecc, SH_IEVAL, 1);
+ return IRQ_HANDLED;
+}
+
+static void edma_error_handler(struct edma_chan *echan)
+{
+ struct edma_cc *ecc = echan->ecc;
+ struct device *dev = echan->vchan.chan.device->dev;
+ struct edmacc_param p;
+ int err;
+
+ if (!echan->edesc)
+ return;
+
+ spin_lock(&echan->vchan.lock);
+
+ err = edma_read_slot(ecc, echan->slot[0], &p);
+
+ /*
+ * Issue later based on missed flag which will be sure
+ * to happen as:
+ * (1) we finished transmitting an intermediate slot and
+ * edma_execute is coming up.
+ * (2) or we finished current transfer and issue will
+ * call edma_execute.
+ *
+ * Important note: issuing can be dangerous here and
+ * lead to some nasty recursion when we are in a NULL
+ * slot. So we avoid doing so and set the missed flag.
+ */
+ if (err || (p.a_b_cnt == 0 && p.ccnt == 0)) {
+ dev_dbg(dev, "Error on null slot, setting miss\n");
+ echan->missed = 1;
+ } else {
+ /*
+ * The slot is already programmed but the event got
+ * missed, so its safe to issue it here.
+ */
+ dev_dbg(dev, "Missed event, TRIGGERING\n");
+ edma_clean_channel(echan);
+ edma_stop(echan);
+ edma_start(echan);
+ edma_trigger_channel(echan);
+ }
+ spin_unlock(&echan->vchan.lock);
+}
+
+static inline bool edma_error_pending(struct edma_cc *ecc)
+{
+ if (edma_read_array(ecc, EDMA_EMR, 0) ||
+ edma_read_array(ecc, EDMA_EMR, 1) ||
+ edma_read(ecc, EDMA_QEMR) || edma_read(ecc, EDMA_CCERR))
+ return true;
+
+ return false;
+}
+
+/* eDMA error interrupt handler */
+static irqreturn_t dma_ccerr_handler(int irq, void *data)
+{
+ struct edma_cc *ecc = data;
+ int i, j;
+ int ctlr;
+ unsigned int cnt = 0;
+ unsigned int val;
+
+ ctlr = ecc->id;
+ if (ctlr < 0)
+ return IRQ_NONE;
+
+ dev_vdbg(ecc->dev, "dma_ccerr_handler\n");
+
+ if (!edma_error_pending(ecc)) {
+ /*
+ * The registers indicate no pending error event but the irq
+ * handler has been called.
+ * Ask eDMA to re-evaluate the error registers.
+ */
+ dev_err(ecc->dev, "%s: Error interrupt without error event!\n",
+ __func__);
+ edma_write(ecc, EDMA_EEVAL, 1);
+ return IRQ_NONE;
+ }
+
+ while (1) {
+ /* Event missed register(s) */
+ for (j = 0; j < 2; j++) {
+ unsigned long emr;
+
+ val = edma_read_array(ecc, EDMA_EMR, j);
+ if (!val)
+ continue;
+
+ dev_dbg(ecc->dev, "EMR%d 0x%08x\n", j, val);
+ emr = val;
+ for (i = find_next_bit(&emr, 32, 0); i < 32;
+ i = find_next_bit(&emr, 32, i + 1)) {
+ int k = (j << 5) + i;
+
+ /* Clear the corresponding EMR bits */
+ edma_write_array(ecc, EDMA_EMCR, j, BIT(i));
+ /* Clear any SER */
+ edma_shadow0_write_array(ecc, SH_SECR, j,
+ BIT(i));
+ edma_error_handler(&ecc->slave_chans[k]);
+ }
+ }
+
+ val = edma_read(ecc, EDMA_QEMR);
+ if (val) {
+ dev_dbg(ecc->dev, "QEMR 0x%02x\n", val);
+ /* Not reported, just clear the interrupt reason. */
+ edma_write(ecc, EDMA_QEMCR, val);
+ edma_shadow0_write(ecc, SH_QSECR, val);
+ }
+
+ val = edma_read(ecc, EDMA_CCERR);
+ if (val) {
+ dev_warn(ecc->dev, "CCERR 0x%08x\n", val);
+ /* Not reported, just clear the interrupt reason. */
+ edma_write(ecc, EDMA_CCERRCLR, val);
+ }
+
+ if (!edma_error_pending(ecc))
+ break;
+ cnt++;
+ if (cnt > 10)
+ break;
+ }
+ edma_write(ecc, EDMA_EEVAL, 1);
+ return IRQ_HANDLED;
+}
+
+/* Alloc channel resources */
+static int edma_alloc_chan_resources(struct dma_chan *chan)
+{
+ struct edma_chan *echan = to_edma_chan(chan);
+ struct edma_cc *ecc = echan->ecc;
+ struct device *dev = ecc->dev;
+ enum dma_event_q eventq_no = EVENTQ_DEFAULT;
+ int ret;
+
+ if (echan->tc) {
+ eventq_no = echan->tc->id;
+ } else if (ecc->tc_list) {
+ /* memcpy channel */
+ echan->tc = &ecc->tc_list[ecc->info->default_queue];
+ eventq_no = echan->tc->id;
+ }
+
+ ret = edma_alloc_channel(echan, eventq_no);
+ if (ret)
+ return ret;
+
+ echan->slot[0] = edma_alloc_slot(ecc, echan->ch_num);
+ if (echan->slot[0] < 0) {
+ dev_err(dev, "Entry slot allocation failed for channel %u\n",
+ EDMA_CHAN_SLOT(echan->ch_num));
+ ret = echan->slot[0];
+ goto err_slot;
+ }
+
+ /* Set up channel -> slot mapping for the entry slot */
+ edma_set_chmap(echan, echan->slot[0]);
+ echan->alloced = true;
+
+ dev_dbg(dev, "Got eDMA channel %d for virt channel %d (%s trigger)\n",
+ EDMA_CHAN_SLOT(echan->ch_num), chan->chan_id,
+ echan->hw_triggered ? "HW" : "SW");
+
+ return 0;
+
+err_slot:
+ edma_free_channel(echan);
+ return ret;
+}
+
+/* Free channel resources */
+static void edma_free_chan_resources(struct dma_chan *chan)
+{
+ struct edma_chan *echan = to_edma_chan(chan);
+ struct device *dev = echan->ecc->dev;
+ int i;
+
+ /* Terminate transfers */
+ edma_stop(echan);
+
+ vchan_free_chan_resources(&echan->vchan);
+
+ /* Free EDMA PaRAM slots */
+ for (i = 0; i < EDMA_MAX_SLOTS; i++) {
+ if (echan->slot[i] >= 0) {
+ edma_free_slot(echan->ecc, echan->slot[i]);
+ echan->slot[i] = -1;
+ }
+ }
+
+ /* Set entry slot to the dummy slot */
+ edma_set_chmap(echan, echan->ecc->dummy_slot);
+
+ /* Free EDMA channel */
+ if (echan->alloced) {
+ edma_free_channel(echan);
+ echan->alloced = false;
+ }
+
+ echan->tc = NULL;
+ echan->hw_triggered = false;
+
+ dev_dbg(dev, "Free eDMA channel %d for virt channel %d\n",
+ EDMA_CHAN_SLOT(echan->ch_num), chan->chan_id);
+}
+
+/* Send pending descriptor to hardware */
+static void edma_issue_pending(struct dma_chan *chan)
+{
+ struct edma_chan *echan = to_edma_chan(chan);
+ unsigned long flags;
+
+ spin_lock_irqsave(&echan->vchan.lock, flags);
+ if (vchan_issue_pending(&echan->vchan) && !echan->edesc)
+ edma_execute(echan);
+ spin_unlock_irqrestore(&echan->vchan.lock, flags);
+}
+
+/*
+ * This limit exists to avoid a possible infinite loop when waiting for proof
+ * that a particular transfer is completed. This limit can be hit if there
+ * are large bursts to/from slow devices or the CPU is never able to catch
+ * the DMA hardware idle. On an AM335x transfering 48 bytes from the UART
+ * RX-FIFO, as many as 55 loops have been seen.
+ */
+#define EDMA_MAX_TR_WAIT_LOOPS 1000
+
+static u32 edma_residue(struct edma_desc *edesc)
+{
+ bool dst = edesc->direction == DMA_DEV_TO_MEM;
+ int loop_count = EDMA_MAX_TR_WAIT_LOOPS;
+ struct edma_chan *echan = edesc->echan;
+ struct edma_pset *pset = edesc->pset;
+ dma_addr_t done, pos;
+ int i;
+
+ /*
+ * We always read the dst/src position from the first RamPar
+ * pset. That's the one which is active now.
+ */
+ pos = edma_get_position(echan->ecc, echan->slot[0], dst);
+
+ /*
+ * "pos" may represent a transfer request that is still being
+ * processed by the EDMACC or EDMATC. We will busy wait until
+ * any one of the situations occurs:
+ * 1. the DMA hardware is idle
+ * 2. a new transfer request is setup
+ * 3. we hit the loop limit
+ */
+ while (edma_read(echan->ecc, EDMA_CCSTAT) & EDMA_CCSTAT_ACTV) {
+ /* check if a new transfer request is setup */
+ if (edma_get_position(echan->ecc,
+ echan->slot[0], dst) != pos) {
+ break;
+ }
+
+ if (!--loop_count) {
+ dev_dbg_ratelimited(echan->vchan.chan.device->dev,
+ "%s: timeout waiting for PaRAM update\n",
+ __func__);
+ break;
+ }
+
+ cpu_relax();
+ }
+
+ /*
+ * Cyclic is simple. Just subtract pset[0].addr from pos.
+ *
+ * We never update edesc->residue in the cyclic case, so we
+ * can tell the remaining room to the end of the circular
+ * buffer.
+ */
+ if (edesc->cyclic) {
+ done = pos - pset->addr;
+ edesc->residue_stat = edesc->residue - done;
+ return edesc->residue_stat;
+ }
+
+ /*
+ * For SG operation we catch up with the last processed
+ * status.
+ */
+ pset += edesc->processed_stat;
+
+ for (i = edesc->processed_stat; i < edesc->processed; i++, pset++) {
+ /*
+ * If we are inside this pset address range, we know
+ * this is the active one. Get the current delta and
+ * stop walking the psets.
+ */
+ if (pos >= pset->addr && pos < pset->addr + pset->len)
+ return edesc->residue_stat - (pos - pset->addr);
+
+ /* Otherwise mark it done and update residue_stat. */
+ edesc->processed_stat++;
+ edesc->residue_stat -= pset->len;
+ }
+ return edesc->residue_stat;
+}
+
+/* Check request completion status */
+static enum dma_status edma_tx_status(struct dma_chan *chan,
+ dma_cookie_t cookie,
+ struct dma_tx_state *txstate)
+{
+ struct edma_chan *echan = to_edma_chan(chan);
+ struct virt_dma_desc *vdesc;
+ enum dma_status ret;
+ unsigned long flags;
+
+ ret = dma_cookie_status(chan, cookie, txstate);
+ if (ret == DMA_COMPLETE || !txstate)
+ return ret;
+
+ spin_lock_irqsave(&echan->vchan.lock, flags);
+ if (echan->edesc && echan->edesc->vdesc.tx.cookie == cookie)
+ txstate->residue = edma_residue(echan->edesc);
+ else if ((vdesc = vchan_find_desc(&echan->vchan, cookie)))
+ txstate->residue = to_edma_desc(&vdesc->tx)->residue;
+ spin_unlock_irqrestore(&echan->vchan.lock, flags);
+
+ return ret;
+}
+
+static bool edma_is_memcpy_channel(int ch_num, s32 *memcpy_channels)
+{
+ if (!memcpy_channels)
+ return false;
+ while (*memcpy_channels != -1) {
+ if (*memcpy_channels == ch_num)
+ return true;
+ memcpy_channels++;
+ }
+ return false;
+}
+
+#define EDMA_DMA_BUSWIDTHS (BIT(DMA_SLAVE_BUSWIDTH_1_BYTE) | \
+ BIT(DMA_SLAVE_BUSWIDTH_2_BYTES) | \
+ BIT(DMA_SLAVE_BUSWIDTH_3_BYTES) | \
+ BIT(DMA_SLAVE_BUSWIDTH_4_BYTES))
+
+static void edma_dma_init(struct edma_cc *ecc, bool legacy_mode)
+{
+ struct dma_device *s_ddev = &ecc->dma_slave;
+ struct dma_device *m_ddev = NULL;
+ s32 *memcpy_channels = ecc->info->memcpy_channels;
+ int i, j;
+
+ dma_cap_zero(s_ddev->cap_mask);
+ dma_cap_set(DMA_SLAVE, s_ddev->cap_mask);
+ dma_cap_set(DMA_CYCLIC, s_ddev->cap_mask);
+ if (ecc->legacy_mode && !memcpy_channels) {
+ dev_warn(ecc->dev,
+ "Legacy memcpy is enabled, things might not work\n");
+
+ dma_cap_set(DMA_MEMCPY, s_ddev->cap_mask);
+ s_ddev->device_prep_dma_memcpy = edma_prep_dma_memcpy;
+ s_ddev->directions = BIT(DMA_MEM_TO_MEM);
+ }
+
+ s_ddev->device_prep_slave_sg = edma_prep_slave_sg;
+ s_ddev->device_prep_dma_cyclic = edma_prep_dma_cyclic;
+ s_ddev->device_alloc_chan_resources = edma_alloc_chan_resources;
+ s_ddev->device_free_chan_resources = edma_free_chan_resources;
+ s_ddev->device_issue_pending = edma_issue_pending;
+ s_ddev->device_tx_status = edma_tx_status;
+ s_ddev->device_config = edma_slave_config;
+ s_ddev->device_pause = edma_dma_pause;
+ s_ddev->device_resume = edma_dma_resume;
+ s_ddev->device_terminate_all = edma_terminate_all;
+ s_ddev->device_synchronize = edma_synchronize;
+
+ s_ddev->src_addr_widths = EDMA_DMA_BUSWIDTHS;
+ s_ddev->dst_addr_widths = EDMA_DMA_BUSWIDTHS;
+ s_ddev->directions |= (BIT(DMA_DEV_TO_MEM) | BIT(DMA_MEM_TO_DEV));
+ s_ddev->residue_granularity = DMA_RESIDUE_GRANULARITY_BURST;
+ s_ddev->max_burst = SZ_32K - 1; /* CIDX: 16bit signed */
+
+ s_ddev->dev = ecc->dev;
+ INIT_LIST_HEAD(&s_ddev->channels);
+
+ if (memcpy_channels) {
+ m_ddev = devm_kzalloc(ecc->dev, sizeof(*m_ddev), GFP_KERNEL);
+ if (!m_ddev) {
+ dev_warn(ecc->dev, "memcpy is disabled due to OoM\n");
+ memcpy_channels = NULL;
+ goto ch_setup;
+ }
+ ecc->dma_memcpy = m_ddev;
+
+ dma_cap_zero(m_ddev->cap_mask);
+ dma_cap_set(DMA_MEMCPY, m_ddev->cap_mask);
+
+ m_ddev->device_prep_dma_memcpy = edma_prep_dma_memcpy;
+ m_ddev->device_alloc_chan_resources = edma_alloc_chan_resources;
+ m_ddev->device_free_chan_resources = edma_free_chan_resources;
+ m_ddev->device_issue_pending = edma_issue_pending;
+ m_ddev->device_tx_status = edma_tx_status;
+ m_ddev->device_config = edma_slave_config;
+ m_ddev->device_pause = edma_dma_pause;
+ m_ddev->device_resume = edma_dma_resume;
+ m_ddev->device_terminate_all = edma_terminate_all;
+ m_ddev->device_synchronize = edma_synchronize;
+
+ m_ddev->src_addr_widths = EDMA_DMA_BUSWIDTHS;
+ m_ddev->dst_addr_widths = EDMA_DMA_BUSWIDTHS;
+ m_ddev->directions = BIT(DMA_MEM_TO_MEM);
+ m_ddev->residue_granularity = DMA_RESIDUE_GRANULARITY_BURST;
+
+ m_ddev->dev = ecc->dev;
+ INIT_LIST_HEAD(&m_ddev->channels);
+ } else if (!ecc->legacy_mode) {
+ dev_info(ecc->dev, "memcpy is disabled\n");
+ }
+
+ch_setup:
+ for (i = 0; i < ecc->num_channels; i++) {
+ struct edma_chan *echan = &ecc->slave_chans[i];
+ echan->ch_num = EDMA_CTLR_CHAN(ecc->id, i);
+ echan->ecc = ecc;
+ echan->vchan.desc_free = edma_desc_free;
+
+ if (m_ddev && edma_is_memcpy_channel(i, memcpy_channels))
+ vchan_init(&echan->vchan, m_ddev);
+ else
+ vchan_init(&echan->vchan, s_ddev);
+
+ INIT_LIST_HEAD(&echan->node);
+ for (j = 0; j < EDMA_MAX_SLOTS; j++)
+ echan->slot[j] = -1;
+ }
+}
+
+static int edma_setup_from_hw(struct device *dev, struct edma_soc_info *pdata,
+ struct edma_cc *ecc)
+{
+ int i;
+ u32 value, cccfg;
+ s8 (*queue_priority_map)[2];
+
+ /* Decode the eDMA3 configuration from CCCFG register */
+ cccfg = edma_read(ecc, EDMA_CCCFG);
+
+ value = GET_NUM_REGN(cccfg);
+ ecc->num_region = BIT(value);
+
+ value = GET_NUM_DMACH(cccfg);
+ ecc->num_channels = BIT(value + 1);
+
+ value = GET_NUM_QDMACH(cccfg);
+ ecc->num_qchannels = value * 2;
+
+ value = GET_NUM_PAENTRY(cccfg);
+ ecc->num_slots = BIT(value + 4);
+
+ value = GET_NUM_EVQUE(cccfg);
+ ecc->num_tc = value + 1;
+
+ ecc->chmap_exist = (cccfg & CHMAP_EXIST) ? true : false;
+
+ dev_dbg(dev, "eDMA3 CC HW configuration (cccfg: 0x%08x):\n", cccfg);
+ dev_dbg(dev, "num_region: %u\n", ecc->num_region);
+ dev_dbg(dev, "num_channels: %u\n", ecc->num_channels);
+ dev_dbg(dev, "num_qchannels: %u\n", ecc->num_qchannels);
+ dev_dbg(dev, "num_slots: %u\n", ecc->num_slots);
+ dev_dbg(dev, "num_tc: %u\n", ecc->num_tc);
+ dev_dbg(dev, "chmap_exist: %s\n", ecc->chmap_exist ? "yes" : "no");
+
+ /* Nothing need to be done if queue priority is provided */
+ if (pdata->queue_priority_mapping)
+ return 0;
+
+ /*
+ * Configure TC/queue priority as follows:
+ * Q0 - priority 0
+ * Q1 - priority 1
+ * Q2 - priority 2
+ * ...
+ * The meaning of priority numbers: 0 highest priority, 7 lowest
+ * priority. So Q0 is the highest priority queue and the last queue has
+ * the lowest priority.
+ */
+ queue_priority_map = devm_kcalloc(dev, ecc->num_tc + 1, sizeof(s8),
+ GFP_KERNEL);
+ if (!queue_priority_map)
+ return -ENOMEM;
+
+ for (i = 0; i < ecc->num_tc; i++) {
+ queue_priority_map[i][0] = i;
+ queue_priority_map[i][1] = i;
+ }
+ queue_priority_map[i][0] = -1;
+ queue_priority_map[i][1] = -1;
+
+ pdata->queue_priority_mapping = queue_priority_map;
+ /* Default queue has the lowest priority */
+ pdata->default_queue = i - 1;
+
+ return 0;
+}
+
+#if IS_ENABLED(CONFIG_OF)
+static int edma_xbar_event_map(struct device *dev, struct edma_soc_info *pdata,
+ size_t sz)
+{
+ const char pname[] = "ti,edma-xbar-event-map";
+ struct resource res;
+ void __iomem *xbar;
+ s16 (*xbar_chans)[2];
+ size_t nelm = sz / sizeof(s16);
+ u32 shift, offset, mux;
+ int ret, i;
+
+ xbar_chans = devm_kcalloc(dev, nelm + 2, sizeof(s16), GFP_KERNEL);
+ if (!xbar_chans)
+ return -ENOMEM;
+
+ ret = of_address_to_resource(dev->of_node, 1, &res);
+ if (ret)
+ return -ENOMEM;
+
+ xbar = devm_ioremap(dev, res.start, resource_size(&res));
+ if (!xbar)
+ return -ENOMEM;
+
+ ret = of_property_read_u16_array(dev->of_node, pname, (u16 *)xbar_chans,
+ nelm);
+ if (ret)
+ return -EIO;
+
+ /* Invalidate last entry for the other user of this mess */
+ nelm >>= 1;
+ xbar_chans[nelm][0] = -1;
+ xbar_chans[nelm][1] = -1;
+
+ for (i = 0; i < nelm; i++) {
+ shift = (xbar_chans[i][1] & 0x03) << 3;
+ offset = xbar_chans[i][1] & 0xfffffffc;
+ mux = readl(xbar + offset);
+ mux &= ~(0xff << shift);
+ mux |= xbar_chans[i][0] << shift;
+ writel(mux, (xbar + offset));
+ }
+
+ pdata->xbar_chans = (const s16 (*)[2]) xbar_chans;
+ return 0;
+}
+
+static struct edma_soc_info *edma_setup_info_from_dt(struct device *dev,
+ bool legacy_mode)
+{
+ struct edma_soc_info *info;
+ struct property *prop;
+ int sz, ret;
+
+ info = devm_kzalloc(dev, sizeof(struct edma_soc_info), GFP_KERNEL);
+ if (!info)
+ return ERR_PTR(-ENOMEM);
+
+ if (legacy_mode) {
+ prop = of_find_property(dev->of_node, "ti,edma-xbar-event-map",
+ &sz);
+ if (prop) {
+ ret = edma_xbar_event_map(dev, info, sz);
+ if (ret)
+ return ERR_PTR(ret);
+ }
+ return info;
+ }
+
+ /* Get the list of channels allocated to be used for memcpy */
+ prop = of_find_property(dev->of_node, "ti,edma-memcpy-channels", &sz);
+ if (prop) {
+ const char pname[] = "ti,edma-memcpy-channels";
+ size_t nelm = sz / sizeof(s32);
+ s32 *memcpy_ch;
+
+ memcpy_ch = devm_kcalloc(dev, nelm + 1, sizeof(s32),
+ GFP_KERNEL);
+ if (!memcpy_ch)
+ return ERR_PTR(-ENOMEM);
+
+ ret = of_property_read_u32_array(dev->of_node, pname,
+ (u32 *)memcpy_ch, nelm);
+ if (ret)
+ return ERR_PTR(ret);
+
+ memcpy_ch[nelm] = -1;
+ info->memcpy_channels = memcpy_ch;
+ }
+
+ prop = of_find_property(dev->of_node, "ti,edma-reserved-slot-ranges",
+ &sz);
+ if (prop) {
+ const char pname[] = "ti,edma-reserved-slot-ranges";
+ u32 (*tmp)[2];
+ s16 (*rsv_slots)[2];
+ size_t nelm = sz / sizeof(*tmp);
+ struct edma_rsv_info *rsv_info;
+ int i;
+
+ if (!nelm)
+ return info;
+
+ tmp = kcalloc(nelm, sizeof(*tmp), GFP_KERNEL);
+ if (!tmp)
+ return ERR_PTR(-ENOMEM);
+
+ rsv_info = devm_kzalloc(dev, sizeof(*rsv_info), GFP_KERNEL);
+ if (!rsv_info) {
+ kfree(tmp);
+ return ERR_PTR(-ENOMEM);
+ }
+
+ rsv_slots = devm_kcalloc(dev, nelm + 1, sizeof(*rsv_slots),
+ GFP_KERNEL);
+ if (!rsv_slots) {
+ kfree(tmp);
+ return ERR_PTR(-ENOMEM);
+ }
+
+ ret = of_property_read_u32_array(dev->of_node, pname,
+ (u32 *)tmp, nelm * 2);
+ if (ret) {
+ kfree(tmp);
+ return ERR_PTR(ret);
+ }
+
+ for (i = 0; i < nelm; i++) {
+ rsv_slots[i][0] = tmp[i][0];
+ rsv_slots[i][1] = tmp[i][1];
+ }
+ rsv_slots[nelm][0] = -1;
+ rsv_slots[nelm][1] = -1;
+
+ info->rsv = rsv_info;
+ info->rsv->rsv_slots = (const s16 (*)[2])rsv_slots;
+
+ kfree(tmp);
+ }
+
+ return info;
+}
+
+static struct dma_chan *of_edma_xlate(struct of_phandle_args *dma_spec,
+ struct of_dma *ofdma)
+{
+ struct edma_cc *ecc = ofdma->of_dma_data;
+ struct dma_chan *chan = NULL;
+ struct edma_chan *echan;
+ int i;
+
+ if (!ecc || dma_spec->args_count < 1)
+ return NULL;
+
+ for (i = 0; i < ecc->num_channels; i++) {
+ echan = &ecc->slave_chans[i];
+ if (echan->ch_num == dma_spec->args[0]) {
+ chan = &echan->vchan.chan;
+ break;
+ }
+ }
+
+ if (!chan)
+ return NULL;
+
+ if (echan->ecc->legacy_mode && dma_spec->args_count == 1)
+ goto out;
+
+ if (!echan->ecc->legacy_mode && dma_spec->args_count == 2 &&
+ dma_spec->args[1] < echan->ecc->num_tc) {
+ echan->tc = &echan->ecc->tc_list[dma_spec->args[1]];
+ goto out;
+ }
+
+ return NULL;
+out:
+ /* The channel is going to be used as HW synchronized */
+ echan->hw_triggered = true;
+ return dma_get_slave_channel(chan);
+}
+#else
+static struct edma_soc_info *edma_setup_info_from_dt(struct device *dev,
+ bool legacy_mode)
+{
+ return ERR_PTR(-EINVAL);
+}
+
+static struct dma_chan *of_edma_xlate(struct of_phandle_args *dma_spec,
+ struct of_dma *ofdma)
+{
+ return NULL;
+}
+#endif
+
+static int edma_probe(struct platform_device *pdev)
+{
+ struct edma_soc_info *info = pdev->dev.platform_data;
+ s8 (*queue_priority_mapping)[2];
+ int i, off, ln;
+ const s16 (*rsv_slots)[2];
+ const s16 (*xbar_chans)[2];
+ int irq;
+ char *irq_name;
+ struct resource *mem;
+ struct device_node *node = pdev->dev.of_node;
+ struct device *dev = &pdev->dev;
+ struct edma_cc *ecc;
+ bool legacy_mode = true;
+ int ret;
+
+ if (node) {
+ const struct of_device_id *match;
+
+ match = of_match_node(edma_of_ids, node);
+ if (match && (*(u32 *)match->data) == EDMA_BINDING_TPCC)
+ legacy_mode = false;
+
+ info = edma_setup_info_from_dt(dev, legacy_mode);
+ if (IS_ERR(info)) {
+ dev_err(dev, "failed to get DT data\n");
+ return PTR_ERR(info);
+ }
+ }
+
+ if (!info)
+ return -ENODEV;
+
+ pm_runtime_enable(dev);
+ ret = pm_runtime_get_sync(dev);
+ if (ret < 0) {
+ dev_err(dev, "pm_runtime_get_sync() failed\n");
+ return ret;
+ }
+
+ ret = dma_set_mask_and_coherent(dev, DMA_BIT_MASK(32));
+ if (ret)
+ return ret;
+
+ ecc = devm_kzalloc(dev, sizeof(*ecc), GFP_KERNEL);
+ if (!ecc)
+ return -ENOMEM;
+
+ ecc->dev = dev;
+ ecc->id = pdev->id;
+ ecc->legacy_mode = legacy_mode;
+ /* When booting with DT the pdev->id is -1 */
+ if (ecc->id < 0)
+ ecc->id = 0;
+
+ mem = platform_get_resource_byname(pdev, IORESOURCE_MEM, "edma3_cc");
+ if (!mem) {
+ dev_dbg(dev, "mem resource not found, using index 0\n");
+ mem = platform_get_resource(pdev, IORESOURCE_MEM, 0);
+ if (!mem) {
+ dev_err(dev, "no mem resource?\n");
+ return -ENODEV;
+ }
+ }
+ ecc->base = devm_ioremap_resource(dev, mem);
+ if (IS_ERR(ecc->base))
+ return PTR_ERR(ecc->base);
+
+ platform_set_drvdata(pdev, ecc);
+
+ /* Get eDMA3 configuration from IP */
+ ret = edma_setup_from_hw(dev, info, ecc);
+ if (ret)
+ return ret;
+
+ /* Allocate memory based on the information we got from the IP */
+ ecc->slave_chans = devm_kcalloc(dev, ecc->num_channels,
+ sizeof(*ecc->slave_chans), GFP_KERNEL);
+ if (!ecc->slave_chans)
+ return -ENOMEM;
+
+ ecc->slot_inuse = devm_kcalloc(dev, BITS_TO_LONGS(ecc->num_slots),
+ sizeof(unsigned long), GFP_KERNEL);
+ if (!ecc->slot_inuse)
+ return -ENOMEM;
+
+ ecc->default_queue = info->default_queue;
+
+ for (i = 0; i < ecc->num_slots; i++)
+ edma_write_slot(ecc, i, &dummy_paramset);
+
+ if (info->rsv) {
+ /* Set the reserved slots in inuse list */
+ rsv_slots = info->rsv->rsv_slots;
+ if (rsv_slots) {
+ for (i = 0; rsv_slots[i][0] != -1; i++) {
+ off = rsv_slots[i][0];
+ ln = rsv_slots[i][1];
+ edma_set_bits(off, ln, ecc->slot_inuse);
+ }
+ }
+ }
+
+ /* Clear the xbar mapped channels in unused list */
+ xbar_chans = info->xbar_chans;
+ if (xbar_chans) {
+ for (i = 0; xbar_chans[i][1] != -1; i++) {
+ off = xbar_chans[i][1];
+ }
+ }
+
+ irq = platform_get_irq_byname(pdev, "edma3_ccint");
+ if (irq < 0 && node)
+ irq = irq_of_parse_and_map(node, 0);
+
+ if (irq >= 0) {
+ irq_name = devm_kasprintf(dev, GFP_KERNEL, "%s_ccint",
+ dev_name(dev));
+ ret = devm_request_irq(dev, irq, dma_irq_handler, 0, irq_name,
+ ecc);
+ if (ret) {
+ dev_err(dev, "CCINT (%d) failed --> %d\n", irq, ret);
+ return ret;
+ }
+ ecc->ccint = irq;
+ }
+
+ irq = platform_get_irq_byname(pdev, "edma3_ccerrint");
+ if (irq < 0 && node)
+ irq = irq_of_parse_and_map(node, 2);
+
+ if (irq >= 0) {
+ irq_name = devm_kasprintf(dev, GFP_KERNEL, "%s_ccerrint",
+ dev_name(dev));
+ ret = devm_request_irq(dev, irq, dma_ccerr_handler, 0, irq_name,
+ ecc);
+ if (ret) {
+ dev_err(dev, "CCERRINT (%d) failed --> %d\n", irq, ret);
+ return ret;
+ }
+ ecc->ccerrint = irq;
+ }
+
+ ecc->dummy_slot = edma_alloc_slot(ecc, EDMA_SLOT_ANY);
+ if (ecc->dummy_slot < 0) {
+ dev_err(dev, "Can't allocate PaRAM dummy slot\n");
+ return ecc->dummy_slot;
+ }
+
+ queue_priority_mapping = info->queue_priority_mapping;
+
+ if (!ecc->legacy_mode) {
+ int lowest_priority = 0;
+ struct of_phandle_args tc_args;
+
+ ecc->tc_list = devm_kcalloc(dev, ecc->num_tc,
+ sizeof(*ecc->tc_list), GFP_KERNEL);
+ if (!ecc->tc_list)
+ return -ENOMEM;
+
+ for (i = 0;; i++) {
+ ret = of_parse_phandle_with_fixed_args(node, "ti,tptcs",
+ 1, i, &tc_args);
+ if (ret || i == ecc->num_tc)
+ break;
+
+ ecc->tc_list[i].node = tc_args.np;
+ ecc->tc_list[i].id = i;
+ queue_priority_mapping[i][1] = tc_args.args[0];
+ if (queue_priority_mapping[i][1] > lowest_priority) {
+ lowest_priority = queue_priority_mapping[i][1];
+ info->default_queue = i;
+ }
+ }
+ }
+
+ /* Event queue priority mapping */
+ for (i = 0; queue_priority_mapping[i][0] != -1; i++)
+ edma_assign_priority_to_queue(ecc, queue_priority_mapping[i][0],
+ queue_priority_mapping[i][1]);
+
+ for (i = 0; i < ecc->num_region; i++) {
+ edma_write_array2(ecc, EDMA_DRAE, i, 0, 0x0);
+ edma_write_array2(ecc, EDMA_DRAE, i, 1, 0x0);
+ edma_write_array(ecc, EDMA_QRAE, i, 0x0);
+ }
+ ecc->info = info;
+
+ /* Init the dma device and channels */
+ edma_dma_init(ecc, legacy_mode);
+
+ for (i = 0; i < ecc->num_channels; i++) {
+ /* Assign all channels to the default queue */
+ edma_assign_channel_eventq(&ecc->slave_chans[i],
+ info->default_queue);
+ /* Set entry slot to the dummy slot */
+ edma_set_chmap(&ecc->slave_chans[i], ecc->dummy_slot);
+ }
+
+ ecc->dma_slave.filter.map = info->slave_map;
+ ecc->dma_slave.filter.mapcnt = info->slavecnt;
+ ecc->dma_slave.filter.fn = edma_filter_fn;
+
+ ret = dma_async_device_register(&ecc->dma_slave);
+ if (ret) {
+ dev_err(dev, "slave ddev registration failed (%d)\n", ret);
+ goto err_reg1;
+ }
+
+ if (ecc->dma_memcpy) {
+ ret = dma_async_device_register(ecc->dma_memcpy);
+ if (ret) {
+ dev_err(dev, "memcpy ddev registration failed (%d)\n",
+ ret);
+ dma_async_device_unregister(&ecc->dma_slave);
+ goto err_reg1;
+ }
+ }
+
+ if (node)
+ of_dma_controller_register(node, of_edma_xlate, ecc);
+
+ dev_info(dev, "TI EDMA DMA engine driver\n");
+
+ return 0;
+
+err_reg1:
+ edma_free_slot(ecc, ecc->dummy_slot);
+ return ret;
+}
+
+static void edma_cleanupp_vchan(struct dma_device *dmadev)
+{
+ struct edma_chan *echan, *_echan;
+
+ list_for_each_entry_safe(echan, _echan,
+ &dmadev->channels, vchan.chan.device_node) {
+ list_del(&echan->vchan.chan.device_node);
+ tasklet_kill(&echan->vchan.task);
+ }
+}
+
+static int edma_remove(struct platform_device *pdev)
+{
+ struct device *dev = &pdev->dev;
+ struct edma_cc *ecc = dev_get_drvdata(dev);
+
+ devm_free_irq(dev, ecc->ccint, ecc);
+ devm_free_irq(dev, ecc->ccerrint, ecc);
+
+ edma_cleanupp_vchan(&ecc->dma_slave);
+
+ if (dev->of_node)
+ of_dma_controller_free(dev->of_node);
+ dma_async_device_unregister(&ecc->dma_slave);
+ if (ecc->dma_memcpy)
+ dma_async_device_unregister(ecc->dma_memcpy);
+ edma_free_slot(ecc, ecc->dummy_slot);
+
+ return 0;
+}
+
+#ifdef CONFIG_PM_SLEEP
+static int edma_pm_suspend(struct device *dev)
+{
+ struct edma_cc *ecc = dev_get_drvdata(dev);
+ struct edma_chan *echan = ecc->slave_chans;
+ int i;
+
+ for (i = 0; i < ecc->num_channels; i++) {
+ if (echan[i].alloced)
+ edma_setup_interrupt(&echan[i], false);
+ }
+
+ return 0;
+}
+
+static int edma_pm_resume(struct device *dev)
+{
+ struct edma_cc *ecc = dev_get_drvdata(dev);
+ struct edma_chan *echan = ecc->slave_chans;
+ int i;
+ s8 (*queue_priority_mapping)[2];
+
+ /* re initialize dummy slot to dummy param set */
+ edma_write_slot(ecc, ecc->dummy_slot, &dummy_paramset);
+
+ queue_priority_mapping = ecc->info->queue_priority_mapping;
+
+ /* Event queue priority mapping */
+ for (i = 0; queue_priority_mapping[i][0] != -1; i++)
+ edma_assign_priority_to_queue(ecc, queue_priority_mapping[i][0],
+ queue_priority_mapping[i][1]);
+
+ for (i = 0; i < ecc->num_channels; i++) {
+ if (echan[i].alloced) {
+ /* ensure access through shadow region 0 */
+ edma_or_array2(ecc, EDMA_DRAE, 0, i >> 5,
+ BIT(i & 0x1f));
+
+ edma_setup_interrupt(&echan[i], true);
+
+ /* Set up channel -> slot mapping for the entry slot */
+ edma_set_chmap(&echan[i], echan[i].slot[0]);
+ }
+ }
+
+ return 0;
+}
+#endif
+
+static const struct dev_pm_ops edma_pm_ops = {
+ SET_LATE_SYSTEM_SLEEP_PM_OPS(edma_pm_suspend, edma_pm_resume)
+};
+
+static struct platform_driver edma_driver = {
+ .probe = edma_probe,
+ .remove = edma_remove,
+ .driver = {
+ .name = "edma",
+ .pm = &edma_pm_ops,
+ .of_match_table = edma_of_ids,
+ },
+};
+
+static int edma_tptc_probe(struct platform_device *pdev)
+{
+ pm_runtime_enable(&pdev->dev);
+ return pm_runtime_get_sync(&pdev->dev);
+}
+
+static struct platform_driver edma_tptc_driver = {
+ .probe = edma_tptc_probe,
+ .driver = {
+ .name = "edma3-tptc",
+ .of_match_table = edma_tptc_of_ids,
+ },
+};
+
+bool edma_filter_fn(struct dma_chan *chan, void *param)
+{
+ bool match = false;
+
+ if (chan->device->dev->driver == &edma_driver.driver) {
+ struct edma_chan *echan = to_edma_chan(chan);
+ unsigned ch_req = *(unsigned *)param;
+ if (ch_req == echan->ch_num) {
+ /* The channel is going to be used as HW synchronized */
+ echan->hw_triggered = true;
+ match = true;
+ }
+ }
+ return match;
+}
+EXPORT_SYMBOL(edma_filter_fn);
+
+static int edma_init(void)
+{
+ int ret;
+
+ ret = platform_driver_register(&edma_tptc_driver);
+ if (ret)
+ return ret;
+
+ return platform_driver_register(&edma_driver);
+}
+subsys_initcall(edma_init);
+
+static void __exit edma_exit(void)
+{
+ platform_driver_unregister(&edma_driver);
+ platform_driver_unregister(&edma_tptc_driver);
+}
+module_exit(edma_exit);
+
+MODULE_AUTHOR("Matt Porter <matt.porter@linaro.org>");
+MODULE_DESCRIPTION("TI EDMA DMA engine driver");
+MODULE_LICENSE("GPL v2");
diff --git a/drivers/dma/ti/omap-dma.c b/drivers/dma/ti/omap-dma.c
new file mode 100644
index 0000000..a4a931d
--- /dev/null
+++ b/drivers/dma/ti/omap-dma.c
@@ -0,0 +1,1672 @@
+/*
+ * OMAP DMAengine support
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+#include <linux/delay.h>
+#include <linux/dmaengine.h>
+#include <linux/dma-mapping.h>
+#include <linux/dmapool.h>
+#include <linux/err.h>
+#include <linux/init.h>
+#include <linux/interrupt.h>
+#include <linux/list.h>
+#include <linux/module.h>
+#include <linux/omap-dma.h>
+#include <linux/platform_device.h>
+#include <linux/slab.h>
+#include <linux/spinlock.h>
+#include <linux/of_dma.h>
+#include <linux/of_device.h>
+
+#include "../virt-dma.h"
+
+#define OMAP_SDMA_REQUESTS 127
+#define OMAP_SDMA_CHANNELS 32
+
+struct omap_dmadev {
+ struct dma_device ddev;
+ spinlock_t lock;
+ void __iomem *base;
+ const struct omap_dma_reg *reg_map;
+ struct omap_system_dma_plat_info *plat;
+ bool legacy;
+ bool ll123_supported;
+ struct dma_pool *desc_pool;
+ unsigned dma_requests;
+ spinlock_t irq_lock;
+ uint32_t irq_enable_mask;
+ struct omap_chan **lch_map;
+};
+
+struct omap_chan {
+ struct virt_dma_chan vc;
+ void __iomem *channel_base;
+ const struct omap_dma_reg *reg_map;
+ uint32_t ccr;
+
+ struct dma_slave_config cfg;
+ unsigned dma_sig;
+ bool cyclic;
+ bool paused;
+ bool running;
+
+ int dma_ch;
+ struct omap_desc *desc;
+ unsigned sgidx;
+};
+
+#define DESC_NXT_SV_REFRESH (0x1 << 24)
+#define DESC_NXT_SV_REUSE (0x2 << 24)
+#define DESC_NXT_DV_REFRESH (0x1 << 26)
+#define DESC_NXT_DV_REUSE (0x2 << 26)
+#define DESC_NTYPE_TYPE2 (0x2 << 29)
+
+/* Type 2 descriptor with Source or Destination address update */
+struct omap_type2_desc {
+ uint32_t next_desc;
+ uint32_t en;
+ uint32_t addr; /* src or dst */
+ uint16_t fn;
+ uint16_t cicr;
+ int16_t cdei;
+ int16_t csei;
+ int32_t cdfi;
+ int32_t csfi;
+} __packed;
+
+struct omap_sg {
+ dma_addr_t addr;
+ uint32_t en; /* number of elements (24-bit) */
+ uint32_t fn; /* number of frames (16-bit) */
+ int32_t fi; /* for double indexing */
+ int16_t ei; /* for double indexing */
+
+ /* Linked list */
+ struct omap_type2_desc *t2_desc;
+ dma_addr_t t2_desc_paddr;
+};
+
+struct omap_desc {
+ struct virt_dma_desc vd;
+ bool using_ll;
+ enum dma_transfer_direction dir;
+ dma_addr_t dev_addr;
+
+ int32_t fi; /* for OMAP_DMA_SYNC_PACKET / double indexing */
+ int16_t ei; /* for double indexing */
+ uint8_t es; /* CSDP_DATA_TYPE_xxx */
+ uint32_t ccr; /* CCR value */
+ uint16_t clnk_ctrl; /* CLNK_CTRL value */
+ uint16_t cicr; /* CICR value */
+ uint32_t csdp; /* CSDP value */
+
+ unsigned sglen;
+ struct omap_sg sg[0];
+};
+
+enum {
+ CAPS_0_SUPPORT_LL123 = BIT(20), /* Linked List type1/2/3 */
+ CAPS_0_SUPPORT_LL4 = BIT(21), /* Linked List type4 */
+
+ CCR_FS = BIT(5),
+ CCR_READ_PRIORITY = BIT(6),
+ CCR_ENABLE = BIT(7),
+ CCR_AUTO_INIT = BIT(8), /* OMAP1 only */
+ CCR_REPEAT = BIT(9), /* OMAP1 only */
+ CCR_OMAP31_DISABLE = BIT(10), /* OMAP1 only */
+ CCR_SUSPEND_SENSITIVE = BIT(8), /* OMAP2+ only */
+ CCR_RD_ACTIVE = BIT(9), /* OMAP2+ only */
+ CCR_WR_ACTIVE = BIT(10), /* OMAP2+ only */
+ CCR_SRC_AMODE_CONSTANT = 0 << 12,
+ CCR_SRC_AMODE_POSTINC = 1 << 12,
+ CCR_SRC_AMODE_SGLIDX = 2 << 12,
+ CCR_SRC_AMODE_DBLIDX = 3 << 12,
+ CCR_DST_AMODE_CONSTANT = 0 << 14,
+ CCR_DST_AMODE_POSTINC = 1 << 14,
+ CCR_DST_AMODE_SGLIDX = 2 << 14,
+ CCR_DST_AMODE_DBLIDX = 3 << 14,
+ CCR_CONSTANT_FILL = BIT(16),
+ CCR_TRANSPARENT_COPY = BIT(17),
+ CCR_BS = BIT(18),
+ CCR_SUPERVISOR = BIT(22),
+ CCR_PREFETCH = BIT(23),
+ CCR_TRIGGER_SRC = BIT(24),
+ CCR_BUFFERING_DISABLE = BIT(25),
+ CCR_WRITE_PRIORITY = BIT(26),
+ CCR_SYNC_ELEMENT = 0,
+ CCR_SYNC_FRAME = CCR_FS,
+ CCR_SYNC_BLOCK = CCR_BS,
+ CCR_SYNC_PACKET = CCR_BS | CCR_FS,
+
+ CSDP_DATA_TYPE_8 = 0,
+ CSDP_DATA_TYPE_16 = 1,
+ CSDP_DATA_TYPE_32 = 2,
+ CSDP_SRC_PORT_EMIFF = 0 << 2, /* OMAP1 only */
+ CSDP_SRC_PORT_EMIFS = 1 << 2, /* OMAP1 only */
+ CSDP_SRC_PORT_OCP_T1 = 2 << 2, /* OMAP1 only */
+ CSDP_SRC_PORT_TIPB = 3 << 2, /* OMAP1 only */
+ CSDP_SRC_PORT_OCP_T2 = 4 << 2, /* OMAP1 only */
+ CSDP_SRC_PORT_MPUI = 5 << 2, /* OMAP1 only */
+ CSDP_SRC_PACKED = BIT(6),
+ CSDP_SRC_BURST_1 = 0 << 7,
+ CSDP_SRC_BURST_16 = 1 << 7,
+ CSDP_SRC_BURST_32 = 2 << 7,
+ CSDP_SRC_BURST_64 = 3 << 7,
+ CSDP_DST_PORT_EMIFF = 0 << 9, /* OMAP1 only */
+ CSDP_DST_PORT_EMIFS = 1 << 9, /* OMAP1 only */
+ CSDP_DST_PORT_OCP_T1 = 2 << 9, /* OMAP1 only */
+ CSDP_DST_PORT_TIPB = 3 << 9, /* OMAP1 only */
+ CSDP_DST_PORT_OCP_T2 = 4 << 9, /* OMAP1 only */
+ CSDP_DST_PORT_MPUI = 5 << 9, /* OMAP1 only */
+ CSDP_DST_PACKED = BIT(13),
+ CSDP_DST_BURST_1 = 0 << 14,
+ CSDP_DST_BURST_16 = 1 << 14,
+ CSDP_DST_BURST_32 = 2 << 14,
+ CSDP_DST_BURST_64 = 3 << 14,
+ CSDP_WRITE_NON_POSTED = 0 << 16,
+ CSDP_WRITE_POSTED = 1 << 16,
+ CSDP_WRITE_LAST_NON_POSTED = 2 << 16,
+
+ CICR_TOUT_IE = BIT(0), /* OMAP1 only */
+ CICR_DROP_IE = BIT(1),
+ CICR_HALF_IE = BIT(2),
+ CICR_FRAME_IE = BIT(3),
+ CICR_LAST_IE = BIT(4),
+ CICR_BLOCK_IE = BIT(5),
+ CICR_PKT_IE = BIT(7), /* OMAP2+ only */
+ CICR_TRANS_ERR_IE = BIT(8), /* OMAP2+ only */
+ CICR_SUPERVISOR_ERR_IE = BIT(10), /* OMAP2+ only */
+ CICR_MISALIGNED_ERR_IE = BIT(11), /* OMAP2+ only */
+ CICR_DRAIN_IE = BIT(12), /* OMAP2+ only */
+ CICR_SUPER_BLOCK_IE = BIT(14), /* OMAP2+ only */
+
+ CLNK_CTRL_ENABLE_LNK = BIT(15),
+
+ CDP_DST_VALID_INC = 0 << 0,
+ CDP_DST_VALID_RELOAD = 1 << 0,
+ CDP_DST_VALID_REUSE = 2 << 0,
+ CDP_SRC_VALID_INC = 0 << 2,
+ CDP_SRC_VALID_RELOAD = 1 << 2,
+ CDP_SRC_VALID_REUSE = 2 << 2,
+ CDP_NTYPE_TYPE1 = 1 << 4,
+ CDP_NTYPE_TYPE2 = 2 << 4,
+ CDP_NTYPE_TYPE3 = 3 << 4,
+ CDP_TMODE_NORMAL = 0 << 8,
+ CDP_TMODE_LLIST = 1 << 8,
+ CDP_FAST = BIT(10),
+};
+
+static const unsigned es_bytes[] = {
+ [CSDP_DATA_TYPE_8] = 1,
+ [CSDP_DATA_TYPE_16] = 2,
+ [CSDP_DATA_TYPE_32] = 4,
+};
+
+static struct of_dma_filter_info omap_dma_info = {
+ .filter_fn = omap_dma_filter_fn,
+};
+
+static inline struct omap_dmadev *to_omap_dma_dev(struct dma_device *d)
+{
+ return container_of(d, struct omap_dmadev, ddev);
+}
+
+static inline struct omap_chan *to_omap_dma_chan(struct dma_chan *c)
+{
+ return container_of(c, struct omap_chan, vc.chan);
+}
+
+static inline struct omap_desc *to_omap_dma_desc(struct dma_async_tx_descriptor *t)
+{
+ return container_of(t, struct omap_desc, vd.tx);
+}
+
+static void omap_dma_desc_free(struct virt_dma_desc *vd)
+{
+ struct omap_desc *d = to_omap_dma_desc(&vd->tx);
+
+ if (d->using_ll) {
+ struct omap_dmadev *od = to_omap_dma_dev(vd->tx.chan->device);
+ int i;
+
+ for (i = 0; i < d->sglen; i++) {
+ if (d->sg[i].t2_desc)
+ dma_pool_free(od->desc_pool, d->sg[i].t2_desc,
+ d->sg[i].t2_desc_paddr);
+ }
+ }
+
+ kfree(d);
+}
+
+static void omap_dma_fill_type2_desc(struct omap_desc *d, int idx,
+ enum dma_transfer_direction dir, bool last)
+{
+ struct omap_sg *sg = &d->sg[idx];
+ struct omap_type2_desc *t2_desc = sg->t2_desc;
+
+ if (idx)
+ d->sg[idx - 1].t2_desc->next_desc = sg->t2_desc_paddr;
+ if (last)
+ t2_desc->next_desc = 0xfffffffc;
+
+ t2_desc->en = sg->en;
+ t2_desc->addr = sg->addr;
+ t2_desc->fn = sg->fn & 0xffff;
+ t2_desc->cicr = d->cicr;
+ if (!last)
+ t2_desc->cicr &= ~CICR_BLOCK_IE;
+
+ switch (dir) {
+ case DMA_DEV_TO_MEM:
+ t2_desc->cdei = sg->ei;
+ t2_desc->csei = d->ei;
+ t2_desc->cdfi = sg->fi;
+ t2_desc->csfi = d->fi;
+
+ t2_desc->en |= DESC_NXT_DV_REFRESH;
+ t2_desc->en |= DESC_NXT_SV_REUSE;
+ break;
+ case DMA_MEM_TO_DEV:
+ t2_desc->cdei = d->ei;
+ t2_desc->csei = sg->ei;
+ t2_desc->cdfi = d->fi;
+ t2_desc->csfi = sg->fi;
+
+ t2_desc->en |= DESC_NXT_SV_REFRESH;
+ t2_desc->en |= DESC_NXT_DV_REUSE;
+ break;
+ default:
+ return;
+ }
+
+ t2_desc->en |= DESC_NTYPE_TYPE2;
+}
+
+static void omap_dma_write(uint32_t val, unsigned type, void __iomem *addr)
+{
+ switch (type) {
+ case OMAP_DMA_REG_16BIT:
+ writew_relaxed(val, addr);
+ break;
+ case OMAP_DMA_REG_2X16BIT:
+ writew_relaxed(val, addr);
+ writew_relaxed(val >> 16, addr + 2);
+ break;
+ case OMAP_DMA_REG_32BIT:
+ writel_relaxed(val, addr);
+ break;
+ default:
+ WARN_ON(1);
+ }
+}
+
+static unsigned omap_dma_read(unsigned type, void __iomem *addr)
+{
+ unsigned val;
+
+ switch (type) {
+ case OMAP_DMA_REG_16BIT:
+ val = readw_relaxed(addr);
+ break;
+ case OMAP_DMA_REG_2X16BIT:
+ val = readw_relaxed(addr);
+ val |= readw_relaxed(addr + 2) << 16;
+ break;
+ case OMAP_DMA_REG_32BIT:
+ val = readl_relaxed(addr);
+ break;
+ default:
+ WARN_ON(1);
+ val = 0;
+ }
+
+ return val;
+}
+
+static void omap_dma_glbl_write(struct omap_dmadev *od, unsigned reg, unsigned val)
+{
+ const struct omap_dma_reg *r = od->reg_map + reg;
+
+ WARN_ON(r->stride);
+
+ omap_dma_write(val, r->type, od->base + r->offset);
+}
+
+static unsigned omap_dma_glbl_read(struct omap_dmadev *od, unsigned reg)
+{
+ const struct omap_dma_reg *r = od->reg_map + reg;
+
+ WARN_ON(r->stride);
+
+ return omap_dma_read(r->type, od->base + r->offset);
+}
+
+static void omap_dma_chan_write(struct omap_chan *c, unsigned reg, unsigned val)
+{
+ const struct omap_dma_reg *r = c->reg_map + reg;
+
+ omap_dma_write(val, r->type, c->channel_base + r->offset);
+}
+
+static unsigned omap_dma_chan_read(struct omap_chan *c, unsigned reg)
+{
+ const struct omap_dma_reg *r = c->reg_map + reg;
+
+ return omap_dma_read(r->type, c->channel_base + r->offset);
+}
+
+static void omap_dma_clear_csr(struct omap_chan *c)
+{
+ if (dma_omap1())
+ omap_dma_chan_read(c, CSR);
+ else
+ omap_dma_chan_write(c, CSR, ~0);
+}
+
+static unsigned omap_dma_get_csr(struct omap_chan *c)
+{
+ unsigned val = omap_dma_chan_read(c, CSR);
+
+ if (!dma_omap1())
+ omap_dma_chan_write(c, CSR, val);
+
+ return val;
+}
+
+static void omap_dma_assign(struct omap_dmadev *od, struct omap_chan *c,
+ unsigned lch)
+{
+ c->channel_base = od->base + od->plat->channel_stride * lch;
+
+ od->lch_map[lch] = c;
+}
+
+static void omap_dma_start(struct omap_chan *c, struct omap_desc *d)
+{
+ struct omap_dmadev *od = to_omap_dma_dev(c->vc.chan.device);
+ uint16_t cicr = d->cicr;
+
+ if (__dma_omap15xx(od->plat->dma_attr))
+ omap_dma_chan_write(c, CPC, 0);
+ else
+ omap_dma_chan_write(c, CDAC, 0);
+
+ omap_dma_clear_csr(c);
+
+ if (d->using_ll) {
+ uint32_t cdp = CDP_TMODE_LLIST | CDP_NTYPE_TYPE2 | CDP_FAST;
+
+ if (d->dir == DMA_DEV_TO_MEM)
+ cdp |= (CDP_DST_VALID_RELOAD | CDP_SRC_VALID_REUSE);
+ else
+ cdp |= (CDP_DST_VALID_REUSE | CDP_SRC_VALID_RELOAD);
+ omap_dma_chan_write(c, CDP, cdp);
+
+ omap_dma_chan_write(c, CNDP, d->sg[0].t2_desc_paddr);
+ omap_dma_chan_write(c, CCDN, 0);
+ omap_dma_chan_write(c, CCFN, 0xffff);
+ omap_dma_chan_write(c, CCEN, 0xffffff);
+
+ cicr &= ~CICR_BLOCK_IE;
+ } else if (od->ll123_supported) {
+ omap_dma_chan_write(c, CDP, 0);
+ }
+
+ /* Enable interrupts */
+ omap_dma_chan_write(c, CICR, cicr);
+
+ /* Enable channel */
+ omap_dma_chan_write(c, CCR, d->ccr | CCR_ENABLE);
+
+ c->running = true;
+}
+
+static void omap_dma_drain_chan(struct omap_chan *c)
+{
+ int i;
+ u32 val;
+
+ /* Wait for sDMA FIFO to drain */
+ for (i = 0; ; i++) {
+ val = omap_dma_chan_read(c, CCR);
+ if (!(val & (CCR_RD_ACTIVE | CCR_WR_ACTIVE)))
+ break;
+
+ if (i > 100)
+ break;
+
+ udelay(5);
+ }
+
+ if (val & (CCR_RD_ACTIVE | CCR_WR_ACTIVE))
+ dev_err(c->vc.chan.device->dev,
+ "DMA drain did not complete on lch %d\n",
+ c->dma_ch);
+}
+
+static int omap_dma_stop(struct omap_chan *c)
+{
+ struct omap_dmadev *od = to_omap_dma_dev(c->vc.chan.device);
+ uint32_t val;
+
+ /* disable irq */
+ omap_dma_chan_write(c, CICR, 0);
+
+ omap_dma_clear_csr(c);
+
+ val = omap_dma_chan_read(c, CCR);
+ if (od->plat->errata & DMA_ERRATA_i541 && val & CCR_TRIGGER_SRC) {
+ uint32_t sysconfig;
+
+ sysconfig = omap_dma_glbl_read(od, OCP_SYSCONFIG);
+ val = sysconfig & ~DMA_SYSCONFIG_MIDLEMODE_MASK;
+ val |= DMA_SYSCONFIG_MIDLEMODE(DMA_IDLEMODE_NO_IDLE);
+ omap_dma_glbl_write(od, OCP_SYSCONFIG, val);
+
+ val = omap_dma_chan_read(c, CCR);
+ val &= ~CCR_ENABLE;
+ omap_dma_chan_write(c, CCR, val);
+
+ if (!(c->ccr & CCR_BUFFERING_DISABLE))
+ omap_dma_drain_chan(c);
+
+ omap_dma_glbl_write(od, OCP_SYSCONFIG, sysconfig);
+ } else {
+ if (!(val & CCR_ENABLE))
+ return -EINVAL;
+
+ val &= ~CCR_ENABLE;
+ omap_dma_chan_write(c, CCR, val);
+
+ if (!(c->ccr & CCR_BUFFERING_DISABLE))
+ omap_dma_drain_chan(c);
+ }
+
+ mb();
+
+ if (!__dma_omap15xx(od->plat->dma_attr) && c->cyclic) {
+ val = omap_dma_chan_read(c, CLNK_CTRL);
+
+ if (dma_omap1())
+ val |= 1 << 14; /* set the STOP_LNK bit */
+ else
+ val &= ~CLNK_CTRL_ENABLE_LNK;
+
+ omap_dma_chan_write(c, CLNK_CTRL, val);
+ }
+ c->running = false;
+ return 0;
+}
+
+static void omap_dma_start_sg(struct omap_chan *c, struct omap_desc *d)
+{
+ struct omap_sg *sg = d->sg + c->sgidx;
+ unsigned cxsa, cxei, cxfi;
+
+ if (d->dir == DMA_DEV_TO_MEM || d->dir == DMA_MEM_TO_MEM) {
+ cxsa = CDSA;
+ cxei = CDEI;
+ cxfi = CDFI;
+ } else {
+ cxsa = CSSA;
+ cxei = CSEI;
+ cxfi = CSFI;
+ }
+
+ omap_dma_chan_write(c, cxsa, sg->addr);
+ omap_dma_chan_write(c, cxei, sg->ei);
+ omap_dma_chan_write(c, cxfi, sg->fi);
+ omap_dma_chan_write(c, CEN, sg->en);
+ omap_dma_chan_write(c, CFN, sg->fn);
+
+ omap_dma_start(c, d);
+ c->sgidx++;
+}
+
+static void omap_dma_start_desc(struct omap_chan *c)
+{
+ struct virt_dma_desc *vd = vchan_next_desc(&c->vc);
+ struct omap_desc *d;
+ unsigned cxsa, cxei, cxfi;
+
+ if (!vd) {
+ c->desc = NULL;
+ return;
+ }
+
+ list_del(&vd->node);
+
+ c->desc = d = to_omap_dma_desc(&vd->tx);
+ c->sgidx = 0;
+
+ /*
+ * This provides the necessary barrier to ensure data held in
+ * DMA coherent memory is visible to the DMA engine prior to
+ * the transfer starting.
+ */
+ mb();
+
+ omap_dma_chan_write(c, CCR, d->ccr);
+ if (dma_omap1())
+ omap_dma_chan_write(c, CCR2, d->ccr >> 16);
+
+ if (d->dir == DMA_DEV_TO_MEM || d->dir == DMA_MEM_TO_MEM) {
+ cxsa = CSSA;
+ cxei = CSEI;
+ cxfi = CSFI;
+ } else {
+ cxsa = CDSA;
+ cxei = CDEI;
+ cxfi = CDFI;
+ }
+
+ omap_dma_chan_write(c, cxsa, d->dev_addr);
+ omap_dma_chan_write(c, cxei, d->ei);
+ omap_dma_chan_write(c, cxfi, d->fi);
+ omap_dma_chan_write(c, CSDP, d->csdp);
+ omap_dma_chan_write(c, CLNK_CTRL, d->clnk_ctrl);
+
+ omap_dma_start_sg(c, d);
+}
+
+static void omap_dma_callback(int ch, u16 status, void *data)
+{
+ struct omap_chan *c = data;
+ struct omap_desc *d;
+ unsigned long flags;
+
+ spin_lock_irqsave(&c->vc.lock, flags);
+ d = c->desc;
+ if (d) {
+ if (c->cyclic) {
+ vchan_cyclic_callback(&d->vd);
+ } else if (d->using_ll || c->sgidx == d->sglen) {
+ omap_dma_start_desc(c);
+ vchan_cookie_complete(&d->vd);
+ } else {
+ omap_dma_start_sg(c, d);
+ }
+ }
+ spin_unlock_irqrestore(&c->vc.lock, flags);
+}
+
+static irqreturn_t omap_dma_irq(int irq, void *devid)
+{
+ struct omap_dmadev *od = devid;
+ unsigned status, channel;
+
+ spin_lock(&od->irq_lock);
+
+ status = omap_dma_glbl_read(od, IRQSTATUS_L1);
+ status &= od->irq_enable_mask;
+ if (status == 0) {
+ spin_unlock(&od->irq_lock);
+ return IRQ_NONE;
+ }
+
+ while ((channel = ffs(status)) != 0) {
+ unsigned mask, csr;
+ struct omap_chan *c;
+
+ channel -= 1;
+ mask = BIT(channel);
+ status &= ~mask;
+
+ c = od->lch_map[channel];
+ if (c == NULL) {
+ /* This should never happen */
+ dev_err(od->ddev.dev, "invalid channel %u\n", channel);
+ continue;
+ }
+
+ csr = omap_dma_get_csr(c);
+ omap_dma_glbl_write(od, IRQSTATUS_L1, mask);
+
+ omap_dma_callback(channel, csr, c);
+ }
+
+ spin_unlock(&od->irq_lock);
+
+ return IRQ_HANDLED;
+}
+
+static int omap_dma_alloc_chan_resources(struct dma_chan *chan)
+{
+ struct omap_dmadev *od = to_omap_dma_dev(chan->device);
+ struct omap_chan *c = to_omap_dma_chan(chan);
+ struct device *dev = od->ddev.dev;
+ int ret;
+
+ if (od->legacy) {
+ ret = omap_request_dma(c->dma_sig, "DMA engine",
+ omap_dma_callback, c, &c->dma_ch);
+ } else {
+ ret = omap_request_dma(c->dma_sig, "DMA engine", NULL, NULL,
+ &c->dma_ch);
+ }
+
+ dev_dbg(dev, "allocating channel %u for %u\n", c->dma_ch, c->dma_sig);
+
+ if (ret >= 0) {
+ omap_dma_assign(od, c, c->dma_ch);
+
+ if (!od->legacy) {
+ unsigned val;
+
+ spin_lock_irq(&od->irq_lock);
+ val = BIT(c->dma_ch);
+ omap_dma_glbl_write(od, IRQSTATUS_L1, val);
+ od->irq_enable_mask |= val;
+ omap_dma_glbl_write(od, IRQENABLE_L1, od->irq_enable_mask);
+
+ val = omap_dma_glbl_read(od, IRQENABLE_L0);
+ val &= ~BIT(c->dma_ch);
+ omap_dma_glbl_write(od, IRQENABLE_L0, val);
+ spin_unlock_irq(&od->irq_lock);
+ }
+ }
+
+ if (dma_omap1()) {
+ if (__dma_omap16xx(od->plat->dma_attr)) {
+ c->ccr = CCR_OMAP31_DISABLE;
+ /* Duplicate what plat-omap/dma.c does */
+ c->ccr |= c->dma_ch + 1;
+ } else {
+ c->ccr = c->dma_sig & 0x1f;
+ }
+ } else {
+ c->ccr = c->dma_sig & 0x1f;
+ c->ccr |= (c->dma_sig & ~0x1f) << 14;
+ }
+ if (od->plat->errata & DMA_ERRATA_IFRAME_BUFFERING)
+ c->ccr |= CCR_BUFFERING_DISABLE;
+
+ return ret;
+}
+
+static void omap_dma_free_chan_resources(struct dma_chan *chan)
+{
+ struct omap_dmadev *od = to_omap_dma_dev(chan->device);
+ struct omap_chan *c = to_omap_dma_chan(chan);
+
+ if (!od->legacy) {
+ spin_lock_irq(&od->irq_lock);
+ od->irq_enable_mask &= ~BIT(c->dma_ch);
+ omap_dma_glbl_write(od, IRQENABLE_L1, od->irq_enable_mask);
+ spin_unlock_irq(&od->irq_lock);
+ }
+
+ c->channel_base = NULL;
+ od->lch_map[c->dma_ch] = NULL;
+ vchan_free_chan_resources(&c->vc);
+ omap_free_dma(c->dma_ch);
+
+ dev_dbg(od->ddev.dev, "freeing channel %u used for %u\n", c->dma_ch,
+ c->dma_sig);
+ c->dma_sig = 0;
+}
+
+static size_t omap_dma_sg_size(struct omap_sg *sg)
+{
+ return sg->en * sg->fn;
+}
+
+static size_t omap_dma_desc_size(struct omap_desc *d)
+{
+ unsigned i;
+ size_t size;
+
+ for (size = i = 0; i < d->sglen; i++)
+ size += omap_dma_sg_size(&d->sg[i]);
+
+ return size * es_bytes[d->es];
+}
+
+static size_t omap_dma_desc_size_pos(struct omap_desc *d, dma_addr_t addr)
+{
+ unsigned i;
+ size_t size, es_size = es_bytes[d->es];
+
+ for (size = i = 0; i < d->sglen; i++) {
+ size_t this_size = omap_dma_sg_size(&d->sg[i]) * es_size;
+
+ if (size)
+ size += this_size;
+ else if (addr >= d->sg[i].addr &&
+ addr < d->sg[i].addr + this_size)
+ size += d->sg[i].addr + this_size - addr;
+ }
+ return size;
+}
+
+/*
+ * OMAP 3.2/3.3 erratum: sometimes 0 is returned if CSAC/CDAC is
+ * read before the DMA controller finished disabling the channel.
+ */
+static uint32_t omap_dma_chan_read_3_3(struct omap_chan *c, unsigned reg)
+{
+ struct omap_dmadev *od = to_omap_dma_dev(c->vc.chan.device);
+ uint32_t val;
+
+ val = omap_dma_chan_read(c, reg);
+ if (val == 0 && od->plat->errata & DMA_ERRATA_3_3)
+ val = omap_dma_chan_read(c, reg);
+
+ return val;
+}
+
+static dma_addr_t omap_dma_get_src_pos(struct omap_chan *c)
+{
+ struct omap_dmadev *od = to_omap_dma_dev(c->vc.chan.device);
+ dma_addr_t addr, cdac;
+
+ if (__dma_omap15xx(od->plat->dma_attr)) {
+ addr = omap_dma_chan_read(c, CPC);
+ } else {
+ addr = omap_dma_chan_read_3_3(c, CSAC);
+ cdac = omap_dma_chan_read_3_3(c, CDAC);
+
+ /*
+ * CDAC == 0 indicates that the DMA transfer on the channel has
+ * not been started (no data has been transferred so far).
+ * Return the programmed source start address in this case.
+ */
+ if (cdac == 0)
+ addr = omap_dma_chan_read(c, CSSA);
+ }
+
+ if (dma_omap1())
+ addr |= omap_dma_chan_read(c, CSSA) & 0xffff0000;
+
+ return addr;
+}
+
+static dma_addr_t omap_dma_get_dst_pos(struct omap_chan *c)
+{
+ struct omap_dmadev *od = to_omap_dma_dev(c->vc.chan.device);
+ dma_addr_t addr;
+
+ if (__dma_omap15xx(od->plat->dma_attr)) {
+ addr = omap_dma_chan_read(c, CPC);
+ } else {
+ addr = omap_dma_chan_read_3_3(c, CDAC);
+
+ /*
+ * CDAC == 0 indicates that the DMA transfer on the channel
+ * has not been started (no data has been transferred so
+ * far). Return the programmed destination start address in
+ * this case.
+ */
+ if (addr == 0)
+ addr = omap_dma_chan_read(c, CDSA);
+ }
+
+ if (dma_omap1())
+ addr |= omap_dma_chan_read(c, CDSA) & 0xffff0000;
+
+ return addr;
+}
+
+static enum dma_status omap_dma_tx_status(struct dma_chan *chan,
+ dma_cookie_t cookie, struct dma_tx_state *txstate)
+{
+ struct omap_chan *c = to_omap_dma_chan(chan);
+ struct virt_dma_desc *vd;
+ enum dma_status ret;
+ unsigned long flags;
+
+ ret = dma_cookie_status(chan, cookie, txstate);
+
+ if (!c->paused && c->running) {
+ uint32_t ccr = omap_dma_chan_read(c, CCR);
+ /*
+ * The channel is no longer active, set the return value
+ * accordingly
+ */
+ if (!(ccr & CCR_ENABLE))
+ ret = DMA_COMPLETE;
+ }
+
+ if (ret == DMA_COMPLETE || !txstate)
+ return ret;
+
+ spin_lock_irqsave(&c->vc.lock, flags);
+ vd = vchan_find_desc(&c->vc, cookie);
+ if (vd) {
+ txstate->residue = omap_dma_desc_size(to_omap_dma_desc(&vd->tx));
+ } else if (c->desc && c->desc->vd.tx.cookie == cookie) {
+ struct omap_desc *d = c->desc;
+ dma_addr_t pos;
+
+ if (d->dir == DMA_MEM_TO_DEV)
+ pos = omap_dma_get_src_pos(c);
+ else if (d->dir == DMA_DEV_TO_MEM || d->dir == DMA_MEM_TO_MEM)
+ pos = omap_dma_get_dst_pos(c);
+ else
+ pos = 0;
+
+ txstate->residue = omap_dma_desc_size_pos(d, pos);
+ } else {
+ txstate->residue = 0;
+ }
+ if (ret == DMA_IN_PROGRESS && c->paused)
+ ret = DMA_PAUSED;
+ spin_unlock_irqrestore(&c->vc.lock, flags);
+
+ return ret;
+}
+
+static void omap_dma_issue_pending(struct dma_chan *chan)
+{
+ struct omap_chan *c = to_omap_dma_chan(chan);
+ unsigned long flags;
+
+ spin_lock_irqsave(&c->vc.lock, flags);
+ if (vchan_issue_pending(&c->vc) && !c->desc)
+ omap_dma_start_desc(c);
+ spin_unlock_irqrestore(&c->vc.lock, flags);
+}
+
+static struct dma_async_tx_descriptor *omap_dma_prep_slave_sg(
+ struct dma_chan *chan, struct scatterlist *sgl, unsigned sglen,
+ enum dma_transfer_direction dir, unsigned long tx_flags, void *context)
+{
+ struct omap_dmadev *od = to_omap_dma_dev(chan->device);
+ struct omap_chan *c = to_omap_dma_chan(chan);
+ enum dma_slave_buswidth dev_width;
+ struct scatterlist *sgent;
+ struct omap_desc *d;
+ dma_addr_t dev_addr;
+ unsigned i, es, en, frame_bytes;
+ bool ll_failed = false;
+ u32 burst;
+ u32 port_window, port_window_bytes;
+
+ if (dir == DMA_DEV_TO_MEM) {
+ dev_addr = c->cfg.src_addr;
+ dev_width = c->cfg.src_addr_width;
+ burst = c->cfg.src_maxburst;
+ port_window = c->cfg.src_port_window_size;
+ } else if (dir == DMA_MEM_TO_DEV) {
+ dev_addr = c->cfg.dst_addr;
+ dev_width = c->cfg.dst_addr_width;
+ burst = c->cfg.dst_maxburst;
+ port_window = c->cfg.dst_port_window_size;
+ } else {
+ dev_err(chan->device->dev, "%s: bad direction?\n", __func__);
+ return NULL;
+ }
+
+ /* Bus width translates to the element size (ES) */
+ switch (dev_width) {
+ case DMA_SLAVE_BUSWIDTH_1_BYTE:
+ es = CSDP_DATA_TYPE_8;
+ break;
+ case DMA_SLAVE_BUSWIDTH_2_BYTES:
+ es = CSDP_DATA_TYPE_16;
+ break;
+ case DMA_SLAVE_BUSWIDTH_4_BYTES:
+ es = CSDP_DATA_TYPE_32;
+ break;
+ default: /* not reached */
+ return NULL;
+ }
+
+ /* Now allocate and setup the descriptor. */
+ d = kzalloc(struct_size(d, sg, sglen), GFP_ATOMIC);
+ if (!d)
+ return NULL;
+
+ d->dir = dir;
+ d->dev_addr = dev_addr;
+ d->es = es;
+
+ /* When the port_window is used, one frame must cover the window */
+ if (port_window) {
+ burst = port_window;
+ port_window_bytes = port_window * es_bytes[es];
+
+ d->ei = 1;
+ /*
+ * One frame covers the port_window and by configure
+ * the source frame index to be -1 * (port_window - 1)
+ * we instruct the sDMA that after a frame is processed
+ * it should move back to the start of the window.
+ */
+ d->fi = -(port_window_bytes - 1);
+ }
+
+ d->ccr = c->ccr | CCR_SYNC_FRAME;
+ if (dir == DMA_DEV_TO_MEM) {
+ d->csdp = CSDP_DST_BURST_64 | CSDP_DST_PACKED;
+
+ d->ccr |= CCR_DST_AMODE_POSTINC;
+ if (port_window) {
+ d->ccr |= CCR_SRC_AMODE_DBLIDX;
+
+ if (port_window_bytes >= 64)
+ d->csdp |= CSDP_SRC_BURST_64;
+ else if (port_window_bytes >= 32)
+ d->csdp |= CSDP_SRC_BURST_32;
+ else if (port_window_bytes >= 16)
+ d->csdp |= CSDP_SRC_BURST_16;
+
+ } else {
+ d->ccr |= CCR_SRC_AMODE_CONSTANT;
+ }
+ } else {
+ d->csdp = CSDP_SRC_BURST_64 | CSDP_SRC_PACKED;
+
+ d->ccr |= CCR_SRC_AMODE_POSTINC;
+ if (port_window) {
+ d->ccr |= CCR_DST_AMODE_DBLIDX;
+
+ if (port_window_bytes >= 64)
+ d->csdp |= CSDP_DST_BURST_64;
+ else if (port_window_bytes >= 32)
+ d->csdp |= CSDP_DST_BURST_32;
+ else if (port_window_bytes >= 16)
+ d->csdp |= CSDP_DST_BURST_16;
+ } else {
+ d->ccr |= CCR_DST_AMODE_CONSTANT;
+ }
+ }
+
+ d->cicr = CICR_DROP_IE | CICR_BLOCK_IE;
+ d->csdp |= es;
+
+ if (dma_omap1()) {
+ d->cicr |= CICR_TOUT_IE;
+
+ if (dir == DMA_DEV_TO_MEM)
+ d->csdp |= CSDP_DST_PORT_EMIFF | CSDP_SRC_PORT_TIPB;
+ else
+ d->csdp |= CSDP_DST_PORT_TIPB | CSDP_SRC_PORT_EMIFF;
+ } else {
+ if (dir == DMA_DEV_TO_MEM)
+ d->ccr |= CCR_TRIGGER_SRC;
+
+ d->cicr |= CICR_MISALIGNED_ERR_IE | CICR_TRANS_ERR_IE;
+
+ if (port_window)
+ d->csdp |= CSDP_WRITE_LAST_NON_POSTED;
+ }
+ if (od->plat->errata & DMA_ERRATA_PARALLEL_CHANNELS)
+ d->clnk_ctrl = c->dma_ch;
+
+ /*
+ * Build our scatterlist entries: each contains the address,
+ * the number of elements (EN) in each frame, and the number of
+ * frames (FN). Number of bytes for this entry = ES * EN * FN.
+ *
+ * Burst size translates to number of elements with frame sync.
+ * Note: DMA engine defines burst to be the number of dev-width
+ * transfers.
+ */
+ en = burst;
+ frame_bytes = es_bytes[es] * en;
+
+ if (sglen >= 2)
+ d->using_ll = od->ll123_supported;
+
+ for_each_sg(sgl, sgent, sglen, i) {
+ struct omap_sg *osg = &d->sg[i];
+
+ osg->addr = sg_dma_address(sgent);
+ osg->en = en;
+ osg->fn = sg_dma_len(sgent) / frame_bytes;
+
+ if (d->using_ll) {
+ osg->t2_desc = dma_pool_alloc(od->desc_pool, GFP_ATOMIC,
+ &osg->t2_desc_paddr);
+ if (!osg->t2_desc) {
+ dev_err(chan->device->dev,
+ "t2_desc[%d] allocation failed\n", i);
+ ll_failed = true;
+ d->using_ll = false;
+ continue;
+ }
+
+ omap_dma_fill_type2_desc(d, i, dir, (i == sglen - 1));
+ }
+ }
+
+ d->sglen = sglen;
+
+ /* Release the dma_pool entries if one allocation failed */
+ if (ll_failed) {
+ for (i = 0; i < d->sglen; i++) {
+ struct omap_sg *osg = &d->sg[i];
+
+ if (osg->t2_desc) {
+ dma_pool_free(od->desc_pool, osg->t2_desc,
+ osg->t2_desc_paddr);
+ osg->t2_desc = NULL;
+ }
+ }
+ }
+
+ return vchan_tx_prep(&c->vc, &d->vd, tx_flags);
+}
+
+static struct dma_async_tx_descriptor *omap_dma_prep_dma_cyclic(
+ struct dma_chan *chan, dma_addr_t buf_addr, size_t buf_len,
+ size_t period_len, enum dma_transfer_direction dir, unsigned long flags)
+{
+ struct omap_dmadev *od = to_omap_dma_dev(chan->device);
+ struct omap_chan *c = to_omap_dma_chan(chan);
+ enum dma_slave_buswidth dev_width;
+ struct omap_desc *d;
+ dma_addr_t dev_addr;
+ unsigned es;
+ u32 burst;
+
+ if (dir == DMA_DEV_TO_MEM) {
+ dev_addr = c->cfg.src_addr;
+ dev_width = c->cfg.src_addr_width;
+ burst = c->cfg.src_maxburst;
+ } else if (dir == DMA_MEM_TO_DEV) {
+ dev_addr = c->cfg.dst_addr;
+ dev_width = c->cfg.dst_addr_width;
+ burst = c->cfg.dst_maxburst;
+ } else {
+ dev_err(chan->device->dev, "%s: bad direction?\n", __func__);
+ return NULL;
+ }
+
+ /* Bus width translates to the element size (ES) */
+ switch (dev_width) {
+ case DMA_SLAVE_BUSWIDTH_1_BYTE:
+ es = CSDP_DATA_TYPE_8;
+ break;
+ case DMA_SLAVE_BUSWIDTH_2_BYTES:
+ es = CSDP_DATA_TYPE_16;
+ break;
+ case DMA_SLAVE_BUSWIDTH_4_BYTES:
+ es = CSDP_DATA_TYPE_32;
+ break;
+ default: /* not reached */
+ return NULL;
+ }
+
+ /* Now allocate and setup the descriptor. */
+ d = kzalloc(sizeof(*d) + sizeof(d->sg[0]), GFP_ATOMIC);
+ if (!d)
+ return NULL;
+
+ d->dir = dir;
+ d->dev_addr = dev_addr;
+ d->fi = burst;
+ d->es = es;
+ d->sg[0].addr = buf_addr;
+ d->sg[0].en = period_len / es_bytes[es];
+ d->sg[0].fn = buf_len / period_len;
+ d->sglen = 1;
+
+ d->ccr = c->ccr;
+ if (dir == DMA_DEV_TO_MEM)
+ d->ccr |= CCR_DST_AMODE_POSTINC | CCR_SRC_AMODE_CONSTANT;
+ else
+ d->ccr |= CCR_DST_AMODE_CONSTANT | CCR_SRC_AMODE_POSTINC;
+
+ d->cicr = CICR_DROP_IE;
+ if (flags & DMA_PREP_INTERRUPT)
+ d->cicr |= CICR_FRAME_IE;
+
+ d->csdp = es;
+
+ if (dma_omap1()) {
+ d->cicr |= CICR_TOUT_IE;
+
+ if (dir == DMA_DEV_TO_MEM)
+ d->csdp |= CSDP_DST_PORT_EMIFF | CSDP_SRC_PORT_MPUI;
+ else
+ d->csdp |= CSDP_DST_PORT_MPUI | CSDP_SRC_PORT_EMIFF;
+ } else {
+ if (burst)
+ d->ccr |= CCR_SYNC_PACKET;
+ else
+ d->ccr |= CCR_SYNC_ELEMENT;
+
+ if (dir == DMA_DEV_TO_MEM) {
+ d->ccr |= CCR_TRIGGER_SRC;
+ d->csdp |= CSDP_DST_PACKED;
+ } else {
+ d->csdp |= CSDP_SRC_PACKED;
+ }
+
+ d->cicr |= CICR_MISALIGNED_ERR_IE | CICR_TRANS_ERR_IE;
+
+ d->csdp |= CSDP_DST_BURST_64 | CSDP_SRC_BURST_64;
+ }
+
+ if (__dma_omap15xx(od->plat->dma_attr))
+ d->ccr |= CCR_AUTO_INIT | CCR_REPEAT;
+ else
+ d->clnk_ctrl = c->dma_ch | CLNK_CTRL_ENABLE_LNK;
+
+ c->cyclic = true;
+
+ return vchan_tx_prep(&c->vc, &d->vd, flags);
+}
+
+static struct dma_async_tx_descriptor *omap_dma_prep_dma_memcpy(
+ struct dma_chan *chan, dma_addr_t dest, dma_addr_t src,
+ size_t len, unsigned long tx_flags)
+{
+ struct omap_chan *c = to_omap_dma_chan(chan);
+ struct omap_desc *d;
+ uint8_t data_type;
+
+ d = kzalloc(sizeof(*d) + sizeof(d->sg[0]), GFP_ATOMIC);
+ if (!d)
+ return NULL;
+
+ data_type = __ffs((src | dest | len));
+ if (data_type > CSDP_DATA_TYPE_32)
+ data_type = CSDP_DATA_TYPE_32;
+
+ d->dir = DMA_MEM_TO_MEM;
+ d->dev_addr = src;
+ d->fi = 0;
+ d->es = data_type;
+ d->sg[0].en = len / BIT(data_type);
+ d->sg[0].fn = 1;
+ d->sg[0].addr = dest;
+ d->sglen = 1;
+ d->ccr = c->ccr;
+ d->ccr |= CCR_DST_AMODE_POSTINC | CCR_SRC_AMODE_POSTINC;
+
+ d->cicr = CICR_DROP_IE | CICR_FRAME_IE;
+
+ d->csdp = data_type;
+
+ if (dma_omap1()) {
+ d->cicr |= CICR_TOUT_IE;
+ d->csdp |= CSDP_DST_PORT_EMIFF | CSDP_SRC_PORT_EMIFF;
+ } else {
+ d->csdp |= CSDP_DST_PACKED | CSDP_SRC_PACKED;
+ d->cicr |= CICR_MISALIGNED_ERR_IE | CICR_TRANS_ERR_IE;
+ d->csdp |= CSDP_DST_BURST_64 | CSDP_SRC_BURST_64;
+ }
+
+ return vchan_tx_prep(&c->vc, &d->vd, tx_flags);
+}
+
+static struct dma_async_tx_descriptor *omap_dma_prep_dma_interleaved(
+ struct dma_chan *chan, struct dma_interleaved_template *xt,
+ unsigned long flags)
+{
+ struct omap_chan *c = to_omap_dma_chan(chan);
+ struct omap_desc *d;
+ struct omap_sg *sg;
+ uint8_t data_type;
+ size_t src_icg, dst_icg;
+
+ /* Slave mode is not supported */
+ if (is_slave_direction(xt->dir))
+ return NULL;
+
+ if (xt->frame_size != 1 || xt->numf == 0)
+ return NULL;
+
+ d = kzalloc(sizeof(*d) + sizeof(d->sg[0]), GFP_ATOMIC);
+ if (!d)
+ return NULL;
+
+ data_type = __ffs((xt->src_start | xt->dst_start | xt->sgl[0].size));
+ if (data_type > CSDP_DATA_TYPE_32)
+ data_type = CSDP_DATA_TYPE_32;
+
+ sg = &d->sg[0];
+ d->dir = DMA_MEM_TO_MEM;
+ d->dev_addr = xt->src_start;
+ d->es = data_type;
+ sg->en = xt->sgl[0].size / BIT(data_type);
+ sg->fn = xt->numf;
+ sg->addr = xt->dst_start;
+ d->sglen = 1;
+ d->ccr = c->ccr;
+
+ src_icg = dmaengine_get_src_icg(xt, &xt->sgl[0]);
+ dst_icg = dmaengine_get_dst_icg(xt, &xt->sgl[0]);
+ if (src_icg) {
+ d->ccr |= CCR_SRC_AMODE_DBLIDX;
+ d->ei = 1;
+ d->fi = src_icg;
+ } else if (xt->src_inc) {
+ d->ccr |= CCR_SRC_AMODE_POSTINC;
+ d->fi = 0;
+ } else {
+ dev_err(chan->device->dev,
+ "%s: SRC constant addressing is not supported\n",
+ __func__);
+ kfree(d);
+ return NULL;
+ }
+
+ if (dst_icg) {
+ d->ccr |= CCR_DST_AMODE_DBLIDX;
+ sg->ei = 1;
+ sg->fi = dst_icg;
+ } else if (xt->dst_inc) {
+ d->ccr |= CCR_DST_AMODE_POSTINC;
+ sg->fi = 0;
+ } else {
+ dev_err(chan->device->dev,
+ "%s: DST constant addressing is not supported\n",
+ __func__);
+ kfree(d);
+ return NULL;
+ }
+
+ d->cicr = CICR_DROP_IE | CICR_FRAME_IE;
+
+ d->csdp = data_type;
+
+ if (dma_omap1()) {
+ d->cicr |= CICR_TOUT_IE;
+ d->csdp |= CSDP_DST_PORT_EMIFF | CSDP_SRC_PORT_EMIFF;
+ } else {
+ d->csdp |= CSDP_DST_PACKED | CSDP_SRC_PACKED;
+ d->cicr |= CICR_MISALIGNED_ERR_IE | CICR_TRANS_ERR_IE;
+ d->csdp |= CSDP_DST_BURST_64 | CSDP_SRC_BURST_64;
+ }
+
+ return vchan_tx_prep(&c->vc, &d->vd, flags);
+}
+
+static int omap_dma_slave_config(struct dma_chan *chan, struct dma_slave_config *cfg)
+{
+ struct omap_chan *c = to_omap_dma_chan(chan);
+
+ if (cfg->src_addr_width == DMA_SLAVE_BUSWIDTH_8_BYTES ||
+ cfg->dst_addr_width == DMA_SLAVE_BUSWIDTH_8_BYTES)
+ return -EINVAL;
+
+ if (cfg->src_maxburst > chan->device->max_burst ||
+ cfg->dst_maxburst > chan->device->max_burst)
+ return -EINVAL;
+
+ memcpy(&c->cfg, cfg, sizeof(c->cfg));
+
+ return 0;
+}
+
+static int omap_dma_terminate_all(struct dma_chan *chan)
+{
+ struct omap_chan *c = to_omap_dma_chan(chan);
+ unsigned long flags;
+ LIST_HEAD(head);
+
+ spin_lock_irqsave(&c->vc.lock, flags);
+
+ /*
+ * Stop DMA activity: we assume the callback will not be called
+ * after omap_dma_stop() returns (even if it does, it will see
+ * c->desc is NULL and exit.)
+ */
+ if (c->desc) {
+ vchan_terminate_vdesc(&c->desc->vd);
+ c->desc = NULL;
+ /* Avoid stopping the dma twice */
+ if (!c->paused)
+ omap_dma_stop(c);
+ }
+
+ c->cyclic = false;
+ c->paused = false;
+
+ vchan_get_all_descriptors(&c->vc, &head);
+ spin_unlock_irqrestore(&c->vc.lock, flags);
+ vchan_dma_desc_free_list(&c->vc, &head);
+
+ return 0;
+}
+
+static void omap_dma_synchronize(struct dma_chan *chan)
+{
+ struct omap_chan *c = to_omap_dma_chan(chan);
+
+ vchan_synchronize(&c->vc);
+}
+
+static int omap_dma_pause(struct dma_chan *chan)
+{
+ struct omap_chan *c = to_omap_dma_chan(chan);
+ struct omap_dmadev *od = to_omap_dma_dev(chan->device);
+ unsigned long flags;
+ int ret = -EINVAL;
+ bool can_pause = false;
+
+ spin_lock_irqsave(&od->irq_lock, flags);
+
+ if (!c->desc)
+ goto out;
+
+ if (c->cyclic)
+ can_pause = true;
+
+ /*
+ * We do not allow DMA_MEM_TO_DEV transfers to be paused.
+ * From the AM572x TRM, 16.1.4.18 Disabling a Channel During Transfer:
+ * "When a channel is disabled during a transfer, the channel undergoes
+ * an abort, unless it is hardware-source-synchronized …".
+ * A source-synchronised channel is one where the fetching of data is
+ * under control of the device. In other words, a device-to-memory
+ * transfer. So, a destination-synchronised channel (which would be a
+ * memory-to-device transfer) undergoes an abort if the the CCR_ENABLE
+ * bit is cleared.
+ * From 16.1.4.20.4.6.2 Abort: "If an abort trigger occurs, the channel
+ * aborts immediately after completion of current read/write
+ * transactions and then the FIFO is cleaned up." The term "cleaned up"
+ * is not defined. TI recommends to check that RD_ACTIVE and WR_ACTIVE
+ * are both clear _before_ disabling the channel, otherwise data loss
+ * will occur.
+ * The problem is that if the channel is active, then device activity
+ * can result in DMA activity starting between reading those as both
+ * clear and the write to DMA_CCR to clear the enable bit hitting the
+ * hardware. If the DMA hardware can't drain the data in its FIFO to the
+ * destination, then data loss "might" occur (say if we write to an UART
+ * and the UART is not accepting any further data).
+ */
+ else if (c->desc->dir == DMA_DEV_TO_MEM)
+ can_pause = true;
+
+ if (can_pause && !c->paused) {
+ ret = omap_dma_stop(c);
+ if (!ret)
+ c->paused = true;
+ }
+out:
+ spin_unlock_irqrestore(&od->irq_lock, flags);
+
+ return ret;
+}
+
+static int omap_dma_resume(struct dma_chan *chan)
+{
+ struct omap_chan *c = to_omap_dma_chan(chan);
+ struct omap_dmadev *od = to_omap_dma_dev(chan->device);
+ unsigned long flags;
+ int ret = -EINVAL;
+
+ spin_lock_irqsave(&od->irq_lock, flags);
+
+ if (c->paused && c->desc) {
+ mb();
+
+ /* Restore channel link register */
+ omap_dma_chan_write(c, CLNK_CTRL, c->desc->clnk_ctrl);
+
+ omap_dma_start(c, c->desc);
+ c->paused = false;
+ ret = 0;
+ }
+ spin_unlock_irqrestore(&od->irq_lock, flags);
+
+ return ret;
+}
+
+static int omap_dma_chan_init(struct omap_dmadev *od)
+{
+ struct omap_chan *c;
+
+ c = kzalloc(sizeof(*c), GFP_KERNEL);
+ if (!c)
+ return -ENOMEM;
+
+ c->reg_map = od->reg_map;
+ c->vc.desc_free = omap_dma_desc_free;
+ vchan_init(&c->vc, &od->ddev);
+
+ return 0;
+}
+
+static void omap_dma_free(struct omap_dmadev *od)
+{
+ while (!list_empty(&od->ddev.channels)) {
+ struct omap_chan *c = list_first_entry(&od->ddev.channels,
+ struct omap_chan, vc.chan.device_node);
+
+ list_del(&c->vc.chan.device_node);
+ tasklet_kill(&c->vc.task);
+ kfree(c);
+ }
+}
+
+#define OMAP_DMA_BUSWIDTHS (BIT(DMA_SLAVE_BUSWIDTH_1_BYTE) | \
+ BIT(DMA_SLAVE_BUSWIDTH_2_BYTES) | \
+ BIT(DMA_SLAVE_BUSWIDTH_4_BYTES))
+
+static int omap_dma_probe(struct platform_device *pdev)
+{
+ struct omap_dmadev *od;
+ struct resource *res;
+ int rc, i, irq;
+ u32 lch_count;
+
+ od = devm_kzalloc(&pdev->dev, sizeof(*od), GFP_KERNEL);
+ if (!od)
+ return -ENOMEM;
+
+ res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
+ od->base = devm_ioremap_resource(&pdev->dev, res);
+ if (IS_ERR(od->base))
+ return PTR_ERR(od->base);
+
+ od->plat = omap_get_plat_info();
+ if (!od->plat)
+ return -EPROBE_DEFER;
+
+ od->reg_map = od->plat->reg_map;
+
+ dma_cap_set(DMA_SLAVE, od->ddev.cap_mask);
+ dma_cap_set(DMA_CYCLIC, od->ddev.cap_mask);
+ dma_cap_set(DMA_MEMCPY, od->ddev.cap_mask);
+ dma_cap_set(DMA_INTERLEAVE, od->ddev.cap_mask);
+ od->ddev.device_alloc_chan_resources = omap_dma_alloc_chan_resources;
+ od->ddev.device_free_chan_resources = omap_dma_free_chan_resources;
+ od->ddev.device_tx_status = omap_dma_tx_status;
+ od->ddev.device_issue_pending = omap_dma_issue_pending;
+ od->ddev.device_prep_slave_sg = omap_dma_prep_slave_sg;
+ od->ddev.device_prep_dma_cyclic = omap_dma_prep_dma_cyclic;
+ od->ddev.device_prep_dma_memcpy = omap_dma_prep_dma_memcpy;
+ od->ddev.device_prep_interleaved_dma = omap_dma_prep_dma_interleaved;
+ od->ddev.device_config = omap_dma_slave_config;
+ od->ddev.device_pause = omap_dma_pause;
+ od->ddev.device_resume = omap_dma_resume;
+ od->ddev.device_terminate_all = omap_dma_terminate_all;
+ od->ddev.device_synchronize = omap_dma_synchronize;
+ od->ddev.src_addr_widths = OMAP_DMA_BUSWIDTHS;
+ od->ddev.dst_addr_widths = OMAP_DMA_BUSWIDTHS;
+ od->ddev.directions = BIT(DMA_DEV_TO_MEM) | BIT(DMA_MEM_TO_DEV);
+ if (__dma_omap15xx(od->plat->dma_attr))
+ od->ddev.residue_granularity =
+ DMA_RESIDUE_GRANULARITY_DESCRIPTOR;
+ else
+ od->ddev.residue_granularity = DMA_RESIDUE_GRANULARITY_BURST;
+ od->ddev.max_burst = SZ_16M - 1; /* CCEN: 24bit unsigned */
+ od->ddev.dev = &pdev->dev;
+ INIT_LIST_HEAD(&od->ddev.channels);
+ spin_lock_init(&od->lock);
+ spin_lock_init(&od->irq_lock);
+
+ /* Number of DMA requests */
+ od->dma_requests = OMAP_SDMA_REQUESTS;
+ if (pdev->dev.of_node && of_property_read_u32(pdev->dev.of_node,
+ "dma-requests",
+ &od->dma_requests)) {
+ dev_info(&pdev->dev,
+ "Missing dma-requests property, using %u.\n",
+ OMAP_SDMA_REQUESTS);
+ }
+
+ /* Number of available logical channels */
+ if (!pdev->dev.of_node) {
+ lch_count = od->plat->dma_attr->lch_count;
+ if (unlikely(!lch_count))
+ lch_count = OMAP_SDMA_CHANNELS;
+ } else if (of_property_read_u32(pdev->dev.of_node, "dma-channels",
+ &lch_count)) {
+ dev_info(&pdev->dev,
+ "Missing dma-channels property, using %u.\n",
+ OMAP_SDMA_CHANNELS);
+ lch_count = OMAP_SDMA_CHANNELS;
+ }
+
+ od->lch_map = devm_kcalloc(&pdev->dev, lch_count, sizeof(*od->lch_map),
+ GFP_KERNEL);
+ if (!od->lch_map)
+ return -ENOMEM;
+
+ for (i = 0; i < od->dma_requests; i++) {
+ rc = omap_dma_chan_init(od);
+ if (rc) {
+ omap_dma_free(od);
+ return rc;
+ }
+ }
+
+ irq = platform_get_irq(pdev, 1);
+ if (irq <= 0) {
+ dev_info(&pdev->dev, "failed to get L1 IRQ: %d\n", irq);
+ od->legacy = true;
+ } else {
+ /* Disable all interrupts */
+ od->irq_enable_mask = 0;
+ omap_dma_glbl_write(od, IRQENABLE_L1, 0);
+
+ rc = devm_request_irq(&pdev->dev, irq, omap_dma_irq,
+ IRQF_SHARED, "omap-dma-engine", od);
+ if (rc)
+ return rc;
+ }
+
+ if (omap_dma_glbl_read(od, CAPS_0) & CAPS_0_SUPPORT_LL123)
+ od->ll123_supported = true;
+
+ od->ddev.filter.map = od->plat->slave_map;
+ od->ddev.filter.mapcnt = od->plat->slavecnt;
+ od->ddev.filter.fn = omap_dma_filter_fn;
+
+ if (od->ll123_supported) {
+ od->desc_pool = dma_pool_create(dev_name(&pdev->dev),
+ &pdev->dev,
+ sizeof(struct omap_type2_desc),
+ 4, 0);
+ if (!od->desc_pool) {
+ dev_err(&pdev->dev,
+ "unable to allocate descriptor pool\n");
+ od->ll123_supported = false;
+ }
+ }
+
+ rc = dma_async_device_register(&od->ddev);
+ if (rc) {
+ pr_warn("OMAP-DMA: failed to register slave DMA engine device: %d\n",
+ rc);
+ omap_dma_free(od);
+ return rc;
+ }
+
+ platform_set_drvdata(pdev, od);
+
+ if (pdev->dev.of_node) {
+ omap_dma_info.dma_cap = od->ddev.cap_mask;
+
+ /* Device-tree DMA controller registration */
+ rc = of_dma_controller_register(pdev->dev.of_node,
+ of_dma_simple_xlate, &omap_dma_info);
+ if (rc) {
+ pr_warn("OMAP-DMA: failed to register DMA controller\n");
+ dma_async_device_unregister(&od->ddev);
+ omap_dma_free(od);
+ }
+ }
+
+ dev_info(&pdev->dev, "OMAP DMA engine driver%s\n",
+ od->ll123_supported ? " (LinkedList1/2/3 supported)" : "");
+
+ return rc;
+}
+
+static int omap_dma_remove(struct platform_device *pdev)
+{
+ struct omap_dmadev *od = platform_get_drvdata(pdev);
+ int irq;
+
+ if (pdev->dev.of_node)
+ of_dma_controller_free(pdev->dev.of_node);
+
+ irq = platform_get_irq(pdev, 1);
+ devm_free_irq(&pdev->dev, irq, od);
+
+ dma_async_device_unregister(&od->ddev);
+
+ if (!od->legacy) {
+ /* Disable all interrupts */
+ omap_dma_glbl_write(od, IRQENABLE_L0, 0);
+ }
+
+ if (od->ll123_supported)
+ dma_pool_destroy(od->desc_pool);
+
+ omap_dma_free(od);
+
+ return 0;
+}
+
+static const struct of_device_id omap_dma_match[] = {
+ { .compatible = "ti,omap2420-sdma", },
+ { .compatible = "ti,omap2430-sdma", },
+ { .compatible = "ti,omap3430-sdma", },
+ { .compatible = "ti,omap3630-sdma", },
+ { .compatible = "ti,omap4430-sdma", },
+ {},
+};
+MODULE_DEVICE_TABLE(of, omap_dma_match);
+
+static struct platform_driver omap_dma_driver = {
+ .probe = omap_dma_probe,
+ .remove = omap_dma_remove,
+ .driver = {
+ .name = "omap-dma-engine",
+ .of_match_table = of_match_ptr(omap_dma_match),
+ },
+};
+
+bool omap_dma_filter_fn(struct dma_chan *chan, void *param)
+{
+ if (chan->device->dev->driver == &omap_dma_driver.driver) {
+ struct omap_dmadev *od = to_omap_dma_dev(chan->device);
+ struct omap_chan *c = to_omap_dma_chan(chan);
+ unsigned req = *(unsigned *)param;
+
+ if (req <= od->dma_requests) {
+ c->dma_sig = req;
+ return true;
+ }
+ }
+ return false;
+}
+EXPORT_SYMBOL_GPL(omap_dma_filter_fn);
+
+static int omap_dma_init(void)
+{
+ return platform_driver_register(&omap_dma_driver);
+}
+subsys_initcall(omap_dma_init);
+
+static void __exit omap_dma_exit(void)
+{
+ platform_driver_unregister(&omap_dma_driver);
+}
+module_exit(omap_dma_exit);
+
+MODULE_AUTHOR("Russell King");
+MODULE_LICENSE("GPL");