mmc: dw_mmc: Add external dma interface support

DesignWare MMC Controller can supports two types of DMA
mode: external dma and internal dma. We get a RK312x platform
integrated dw_mmc and ARM pl330 dma controller. This patch add
edmac ops to support these platforms. I've tested it on RK31xx
platform with edmac mode and RK3288 platform with idmac mode.

Signed-off-by: Shawn Lin <shawn.lin@rock-chips.com>
Signed-off-by: Jaehoon Chung <jh80.chung@samsung.com>
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
diff --git a/drivers/mmc/host/Kconfig b/drivers/mmc/host/Kconfig
index 56f5b148..ef54084 100644
--- a/drivers/mmc/host/Kconfig
+++ b/drivers/mmc/host/Kconfig
@@ -615,15 +615,7 @@
 	help
 	  This selects support for the Synopsys DesignWare Mobile Storage IP
 	  block, this provides host support for SD and MMC interfaces, in both
-	  PIO and external DMA modes.
-
-config MMC_DW_IDMAC
-	bool "Internal DMAC interface"
-	depends on MMC_DW
-	help
-	  This selects support for the internal DMAC block within the Synopsys
-	  Designware Mobile Storage IP block. This disables the external DMA
-	  interface.
+	  PIO, internal DMA mode and external DMA mode.
 
 config MMC_DW_PLTFM
 	tristate "Synopsys Designware MCI Support as platform device"
@@ -652,7 +644,6 @@
 	tristate "K3 specific extensions for Synopsys DW Memory Card Interface"
 	depends on MMC_DW
 	select MMC_DW_PLTFM
-	select MMC_DW_IDMAC
 	help
 	  This selects support for Hisilicon K3 SoC specific extensions to the
 	  Synopsys DesignWare Memory Card Interface driver. Select this option
diff --git a/drivers/mmc/host/dw_mmc-pltfm.c b/drivers/mmc/host/dw_mmc-pltfm.c
index ec6dbcd..7e1d13b 100644
--- a/drivers/mmc/host/dw_mmc-pltfm.c
+++ b/drivers/mmc/host/dw_mmc-pltfm.c
@@ -59,6 +59,8 @@
 	host->pdata = pdev->dev.platform_data;
 
 	regs = platform_get_resource(pdev, IORESOURCE_MEM, 0);
+	/* Get registers' physical base address */
+	host->phy_regs = (void *)(regs->start);
 	host->regs = devm_ioremap_resource(&pdev->dev, regs);
 	if (IS_ERR(host->regs))
 		return PTR_ERR(host->regs);
diff --git a/drivers/mmc/host/dw_mmc.c b/drivers/mmc/host/dw_mmc.c
index b1b7e7f..7fe0315 100644
--- a/drivers/mmc/host/dw_mmc.c
+++ b/drivers/mmc/host/dw_mmc.c
@@ -56,7 +56,6 @@
 #define DW_MCI_FREQ_MAX	200000000	/* unit: HZ */
 #define DW_MCI_FREQ_MIN	400000		/* unit: HZ */
 
-#ifdef CONFIG_MMC_DW_IDMAC
 #define IDMAC_INT_CLR		(SDMMC_IDMAC_INT_AI | SDMMC_IDMAC_INT_NI | \
 				 SDMMC_IDMAC_INT_CES | SDMMC_IDMAC_INT_DU | \
 				 SDMMC_IDMAC_INT_FBE | SDMMC_IDMAC_INT_RI | \
@@ -102,7 +101,6 @@
 
 /* Each descriptor can transfer up to 4KB of data in chained mode */
 #define DW_MCI_DESC_DATA_LENGTH	0x1000
-#endif /* CONFIG_MMC_DW_IDMAC */
 
 static bool dw_mci_reset(struct dw_mci *host);
 static bool dw_mci_ctrl_reset(struct dw_mci *host, u32 reset);
@@ -407,7 +405,6 @@
 		return DMA_FROM_DEVICE;
 }
 
-#ifdef CONFIG_MMC_DW_IDMAC
 static void dw_mci_dma_cleanup(struct dw_mci *host)
 {
 	struct mmc_data *data = host->data;
@@ -445,12 +442,21 @@
 	mci_writel(host, BMOD, temp);
 }
 
-static void dw_mci_idmac_complete_dma(struct dw_mci *host)
+static void dw_mci_dmac_complete_dma(void *arg)
 {
+	struct dw_mci *host = arg;
 	struct mmc_data *data = host->data;
 
 	dev_vdbg(host->dev, "DMA complete\n");
 
+	if ((host->use_dma == TRANS_MODE_EDMAC) &&
+	    data && (data->flags & MMC_DATA_READ))
+		/* Invalidate cache after read */
+		dma_sync_sg_for_cpu(mmc_dev(host->cur_slot->mmc),
+				    data->sg,
+				    data->sg_len,
+				    DMA_FROM_DEVICE);
+
 	host->dma_ops->cleanup(host);
 
 	/*
@@ -564,7 +570,7 @@
 	wmb(); /* drain writebuffer */
 }
 
-static void dw_mci_idmac_start_dma(struct dw_mci *host, unsigned int sg_len)
+static int dw_mci_idmac_start_dma(struct dw_mci *host, unsigned int sg_len)
 {
 	u32 temp;
 
@@ -589,6 +595,8 @@
 
 	/* Start it running */
 	mci_writel(host, PLDMND, 1);
+
+	return 0;
 }
 
 static int dw_mci_idmac_init(struct dw_mci *host)
@@ -669,10 +677,112 @@
 	.init = dw_mci_idmac_init,
 	.start = dw_mci_idmac_start_dma,
 	.stop = dw_mci_idmac_stop_dma,
-	.complete = dw_mci_idmac_complete_dma,
+	.complete = dw_mci_dmac_complete_dma,
 	.cleanup = dw_mci_dma_cleanup,
 };
-#endif /* CONFIG_MMC_DW_IDMAC */
+
+static void dw_mci_edmac_stop_dma(struct dw_mci *host)
+{
+	dmaengine_terminate_all(host->dms->ch);
+}
+
+static int dw_mci_edmac_start_dma(struct dw_mci *host,
+					    unsigned int sg_len)
+{
+	struct dma_slave_config cfg;
+	struct dma_async_tx_descriptor *desc = NULL;
+	struct scatterlist *sgl = host->data->sg;
+	const u32 mszs[] = {1, 4, 8, 16, 32, 64, 128, 256};
+	u32 sg_elems = host->data->sg_len;
+	u32 fifoth_val;
+	u32 fifo_offset = host->fifo_reg - host->regs;
+	int ret = 0;
+
+	/* Set external dma config: burst size, burst width */
+	cfg.dst_addr = (dma_addr_t)(host->phy_regs + fifo_offset);
+	cfg.src_addr = cfg.dst_addr;
+	cfg.dst_addr_width = DMA_SLAVE_BUSWIDTH_4_BYTES;
+	cfg.src_addr_width = DMA_SLAVE_BUSWIDTH_4_BYTES;
+
+	/* Match burst msize with external dma config */
+	fifoth_val = mci_readl(host, FIFOTH);
+	cfg.dst_maxburst = mszs[(fifoth_val >> 28) & 0x7];
+	cfg.src_maxburst = cfg.dst_maxburst;
+
+	if (host->data->flags & MMC_DATA_WRITE)
+		cfg.direction = DMA_MEM_TO_DEV;
+	else
+		cfg.direction = DMA_DEV_TO_MEM;
+
+	ret = dmaengine_slave_config(host->dms->ch, &cfg);
+	if (ret) {
+		dev_err(host->dev, "Failed to config edmac.\n");
+		return -EBUSY;
+	}
+
+	desc = dmaengine_prep_slave_sg(host->dms->ch, sgl,
+				       sg_len, cfg.direction,
+				       DMA_PREP_INTERRUPT | DMA_CTRL_ACK);
+	if (!desc) {
+		dev_err(host->dev, "Can't prepare slave sg.\n");
+		return -EBUSY;
+	}
+
+	/* Set dw_mci_dmac_complete_dma as callback */
+	desc->callback = dw_mci_dmac_complete_dma;
+	desc->callback_param = (void *)host;
+	dmaengine_submit(desc);
+
+	/* Flush cache before write */
+	if (host->data->flags & MMC_DATA_WRITE)
+		dma_sync_sg_for_device(mmc_dev(host->cur_slot->mmc), sgl,
+				       sg_elems, DMA_TO_DEVICE);
+
+	dma_async_issue_pending(host->dms->ch);
+
+	return 0;
+}
+
+static int dw_mci_edmac_init(struct dw_mci *host)
+{
+	/* Request external dma channel */
+	host->dms = kzalloc(sizeof(struct dw_mci_dma_slave), GFP_KERNEL);
+	if (!host->dms)
+		return -ENOMEM;
+
+	host->dms->ch = dma_request_slave_channel(host->dev, "rx-tx");
+	if (!host->dms->ch) {
+		dev_err(host->dev,
+			"Failed to get external DMA channel %d\n",
+			host->dms->ch->chan_id);
+		kfree(host->dms);
+		host->dms = NULL;
+		return -ENXIO;
+	}
+
+	return 0;
+}
+
+static void dw_mci_edmac_exit(struct dw_mci *host)
+{
+	if (host->dms) {
+		if (host->dms->ch) {
+			dma_release_channel(host->dms->ch);
+			host->dms->ch = NULL;
+		}
+		kfree(host->dms);
+		host->dms = NULL;
+	}
+}
+
+static const struct dw_mci_dma_ops dw_mci_edmac_ops = {
+	.init = dw_mci_edmac_init,
+	.exit = dw_mci_edmac_exit,
+	.start = dw_mci_edmac_start_dma,
+	.stop = dw_mci_edmac_stop_dma,
+	.complete = dw_mci_dmac_complete_dma,
+	.cleanup = dw_mci_dma_cleanup,
+};
 
 static int dw_mci_pre_dma_transfer(struct dw_mci *host,
 				   struct mmc_data *data,
@@ -752,7 +862,6 @@
 
 static void dw_mci_adjust_fifoth(struct dw_mci *host, struct mmc_data *data)
 {
-#ifdef CONFIG_MMC_DW_IDMAC
 	unsigned int blksz = data->blksz;
 	const u32 mszs[] = {1, 4, 8, 16, 32, 64, 128, 256};
 	u32 fifo_width = 1 << host->data_shift;
@@ -760,6 +869,10 @@
 	u32 msize = 0, rx_wmark = 1, tx_wmark, tx_wmark_invers;
 	int idx = ARRAY_SIZE(mszs) - 1;
 
+	/* pio should ship this scenario */
+	if (!host->use_dma)
+		return;
+
 	tx_wmark = (host->fifo_depth) / 2;
 	tx_wmark_invers = host->fifo_depth - tx_wmark;
 
@@ -788,7 +901,6 @@
 done:
 	fifoth_val = SDMMC_SET_FIFOTH(msize, rx_wmark, tx_wmark);
 	mci_writel(host, FIFOTH, fifoth_val);
-#endif
 }
 
 static void dw_mci_ctrl_rd_thld(struct dw_mci *host, struct mmc_data *data)
@@ -850,10 +962,12 @@
 
 	host->using_dma = 1;
 
-	dev_vdbg(host->dev,
-		 "sd sg_cpu: %#lx sg_dma: %#lx sg_len: %d\n",
-		 (unsigned long)host->sg_cpu, (unsigned long)host->sg_dma,
-		 sg_len);
+	if (host->use_dma == TRANS_MODE_IDMAC)
+		dev_vdbg(host->dev,
+			 "sd sg_cpu: %#lx sg_dma: %#lx sg_len: %d\n",
+			 (unsigned long)host->sg_cpu,
+			 (unsigned long)host->sg_dma,
+			 sg_len);
 
 	/*
 	 * Decide the MSIZE and RX/TX Watermark.
@@ -875,7 +989,11 @@
 	mci_writel(host, INTMASK, temp);
 	spin_unlock_irqrestore(&host->irq_lock, irqflags);
 
-	host->dma_ops->start(host, sg_len);
+	if (host->dma_ops->start(host, sg_len)) {
+		/* We can't do DMA */
+		dev_err(host->dev, "%s: failed to start DMA.\n", __func__);
+		return -ENODEV;
+	}
 
 	return 0;
 }
@@ -2338,15 +2456,17 @@
 
 	}
 
-#ifdef CONFIG_MMC_DW_IDMAC
-	/* Handle DMA interrupts */
+	if (host->use_dma != TRANS_MODE_IDMAC)
+		return IRQ_HANDLED;
+
+	/* Handle IDMA interrupts */
 	if (host->dma_64bit_address == 1) {
 		pending = mci_readl(host, IDSTS64);
 		if (pending & (SDMMC_IDMAC_INT_TI | SDMMC_IDMAC_INT_RI)) {
 			mci_writel(host, IDSTS64, SDMMC_IDMAC_INT_TI |
 							SDMMC_IDMAC_INT_RI);
 			mci_writel(host, IDSTS64, SDMMC_IDMAC_INT_NI);
-			host->dma_ops->complete(host);
+			host->dma_ops->complete((void *)host);
 		}
 	} else {
 		pending = mci_readl(host, IDSTS);
@@ -2354,10 +2474,9 @@
 			mci_writel(host, IDSTS, SDMMC_IDMAC_INT_TI |
 							SDMMC_IDMAC_INT_RI);
 			mci_writel(host, IDSTS, SDMMC_IDMAC_INT_NI);
-			host->dma_ops->complete(host);
+			host->dma_ops->complete((void *)host);
 		}
 	}
-#endif
 
 	return IRQ_HANDLED;
 }
@@ -2466,13 +2585,21 @@
 		goto err_host_allocated;
 
 	/* Useful defaults if platform data is unset. */
-	if (host->use_dma) {
+	if (host->use_dma == TRANS_MODE_IDMAC) {
 		mmc->max_segs = host->ring_size;
 		mmc->max_blk_size = 65536;
 		mmc->max_seg_size = 0x1000;
 		mmc->max_req_size = mmc->max_seg_size * host->ring_size;
 		mmc->max_blk_count = mmc->max_req_size / 512;
+	} else if (host->use_dma == TRANS_MODE_EDMAC) {
+		mmc->max_segs = 64;
+		mmc->max_blk_size = 65536;
+		mmc->max_blk_count = 65535;
+		mmc->max_req_size =
+				mmc->max_blk_size * mmc->max_blk_count;
+		mmc->max_seg_size = mmc->max_req_size;
 	} else {
+		/* TRANS_MODE_PIO */
 		mmc->max_segs = 64;
 		mmc->max_blk_size = 65536; /* BLKSIZ is 16 bits */
 		mmc->max_blk_count = 512;
@@ -2512,38 +2639,74 @@
 static void dw_mci_init_dma(struct dw_mci *host)
 {
 	int addr_config;
-	/* Check ADDR_CONFIG bit in HCON to find IDMAC address bus width */
-	addr_config = (mci_readl(host, HCON) >> 27) & 0x01;
+	struct device *dev = host->dev;
+	struct device_node *np = dev->of_node;
 
-	if (addr_config == 1) {
-		/* host supports IDMAC in 64-bit address mode */
-		host->dma_64bit_address = 1;
-		dev_info(host->dev, "IDMAC supports 64-bit address mode.\n");
-		if (!dma_set_mask(host->dev, DMA_BIT_MASK(64)))
-			dma_set_coherent_mask(host->dev, DMA_BIT_MASK(64));
+	/*
+	* Check tansfer mode from HCON[17:16]
+	* Clear the ambiguous description of dw_mmc databook:
+	* 2b'00: No DMA Interface -> Actually means using Internal DMA block
+	* 2b'01: DesignWare DMA Interface -> Synopsys DW-DMA block
+	* 2b'10: Generic DMA Interface -> non-Synopsys generic DMA block
+	* 2b'11: Non DW DMA Interface -> pio only
+	* Compared to DesignWare DMA Interface, Generic DMA Interface has a
+	* simpler request/acknowledge handshake mechanism and both of them
+	* are regarded as external dma master for dw_mmc.
+	*/
+	host->use_dma = SDMMC_GET_TRANS_MODE(mci_readl(host, HCON));
+	if (host->use_dma == DMA_INTERFACE_IDMA) {
+		host->use_dma = TRANS_MODE_IDMAC;
+	} else if (host->use_dma == DMA_INTERFACE_DWDMA ||
+		   host->use_dma == DMA_INTERFACE_GDMA) {
+		host->use_dma = TRANS_MODE_EDMAC;
 	} else {
-		/* host supports IDMAC in 32-bit address mode */
-		host->dma_64bit_address = 0;
-		dev_info(host->dev, "IDMAC supports 32-bit address mode.\n");
-	}
-
-	/* Alloc memory for sg translation */
-	host->sg_cpu = dmam_alloc_coherent(host->dev, PAGE_SIZE,
-					  &host->sg_dma, GFP_KERNEL);
-	if (!host->sg_cpu) {
-		dev_err(host->dev, "%s: could not alloc DMA memory\n",
-			__func__);
 		goto no_dma;
 	}
 
 	/* Determine which DMA interface to use */
-#ifdef CONFIG_MMC_DW_IDMAC
-	host->dma_ops = &dw_mci_idmac_ops;
-	dev_info(host->dev, "Using internal DMA controller.\n");
-#endif
+	if (host->use_dma == TRANS_MODE_IDMAC) {
+		/*
+		* Check ADDR_CONFIG bit in HCON to find
+		* IDMAC address bus width
+		*/
+		addr_config = (mci_readl(host, HCON) >> 27) & 0x01;
 
-	if (!host->dma_ops)
-		goto no_dma;
+		if (addr_config == 1) {
+			/* host supports IDMAC in 64-bit address mode */
+			host->dma_64bit_address = 1;
+			dev_info(host->dev,
+				 "IDMAC supports 64-bit address mode.\n");
+			if (!dma_set_mask(host->dev, DMA_BIT_MASK(64)))
+				dma_set_coherent_mask(host->dev,
+						      DMA_BIT_MASK(64));
+		} else {
+			/* host supports IDMAC in 32-bit address mode */
+			host->dma_64bit_address = 0;
+			dev_info(host->dev,
+				 "IDMAC supports 32-bit address mode.\n");
+		}
+
+		/* Alloc memory for sg translation */
+		host->sg_cpu = dmam_alloc_coherent(host->dev, PAGE_SIZE,
+						   &host->sg_dma, GFP_KERNEL);
+		if (!host->sg_cpu) {
+			dev_err(host->dev,
+				"%s: could not alloc DMA memory\n",
+				__func__);
+			goto no_dma;
+		}
+
+		host->dma_ops = &dw_mci_idmac_ops;
+		dev_info(host->dev, "Using internal DMA controller.\n");
+	} else {
+		/* TRANS_MODE_EDMAC: check dma bindings again */
+		if ((of_property_count_strings(np, "dma-names") < 0) ||
+		    (!of_find_property(np, "dmas", NULL))) {
+			goto no_dma;
+		}
+		host->dma_ops = &dw_mci_edmac_ops;
+		dev_info(host->dev, "Using external DMA controller.\n");
+	}
 
 	if (host->dma_ops->init && host->dma_ops->start &&
 	    host->dma_ops->stop && host->dma_ops->cleanup) {
@@ -2557,12 +2720,11 @@
 		goto no_dma;
 	}
 
-	host->use_dma = 1;
 	return;
 
 no_dma:
 	dev_info(host->dev, "Using PIO mode.\n");
-	host->use_dma = 0;
+	host->use_dma = TRANS_MODE_PIO;
 }
 
 static bool dw_mci_ctrl_reset(struct dw_mci *host, u32 reset)
@@ -2645,10 +2807,9 @@
 		}
 	}
 
-#if IS_ENABLED(CONFIG_MMC_DW_IDMAC)
-	/* It is also recommended that we reset and reprogram idmac */
-	dw_mci_idmac_reset(host);
-#endif
+	if (host->use_dma == TRANS_MODE_IDMAC)
+		/* It is also recommended that we reset and reprogram idmac */
+		dw_mci_idmac_reset(host);
 
 	ret = true;
 
@@ -3062,6 +3223,9 @@
  */
 int dw_mci_suspend(struct dw_mci *host)
 {
+	if (host->use_dma && host->dma_ops->exit)
+		host->dma_ops->exit(host);
+
 	return 0;
 }
 EXPORT_SYMBOL(dw_mci_suspend);
diff --git a/drivers/mmc/host/dw_mmc.h b/drivers/mmc/host/dw_mmc.h
index 8ce4674..811d467 100644
--- a/drivers/mmc/host/dw_mmc.h
+++ b/drivers/mmc/host/dw_mmc.h
@@ -148,6 +148,12 @@
 #define SDMMC_SET_FIFOTH(m, r, t)	(((m) & 0x7) << 28 | \
 					 ((r) & 0xFFF) << 16 | \
 					 ((t) & 0xFFF))
+/* HCON register defines */
+#define DMA_INTERFACE_IDMA		(0x0)
+#define DMA_INTERFACE_DWDMA		(0x1)
+#define DMA_INTERFACE_GDMA		(0x2)
+#define DMA_INTERFACE_NODMA		(0x3)
+#define SDMMC_GET_TRANS_MODE(x)		(((x)>>16) & 0x3)
 /* Internal DMAC interrupt defines */
 #define SDMMC_IDMAC_INT_AI		BIT(9)
 #define SDMMC_IDMAC_INT_NI		BIT(8)