Re: [PATCH V11 3/5] i2c: tegra: Add DMA support

[Dmitry Osipenko <digetx@gmail.com>]

05.02.2019 11:56, Sowjanya Komatineni пишет:
> This patch adds DMA support for Tegra I2C.
> 
> Tegra I2C TX and RX FIFO depth is 8 words. PIO mode is used for
> transfer size of the max FIFO depth and DMA mode is used for
> transfer size higher than max FIFO depth to save CPU overhead.
> 
> PIO mode needs full intervention of CPU to fill or empty FIFO's
> and also need to service multiple data requests interrupt for the
> same transaction. This adds delay between data bytes of the same
> transfer when CPU is fully loaded and some slave devices has
> internal timeout for no bus activity and stops transaction to
> avoid bus hang. DMA mode is helpful in such cases.
> 
> DMA mode is also helpful for Large transfers during downloading or
> uploading FW over I2C to some external devices.
> 
> Signed-off-by: Sowjanya Komatineni <skomatineni@nvidia.com>
> ---
>  [V11] : Replaced deprecated dmaengine_terminate_all with dmaengine_termine_async
> 	from non-atomic context and dmaengine_terminate_sync from atomic context.
> 	Fixed to program fifo trigger levels properly when transfer falls back to
> 	pio mode in case of dma slave configuration failure and other minor fixes.
>  [V10] : APBDMA is replaced with GPCDMA on Tegra186 and Tegra194 designs.
> 	Added apbdma hw support flag to now allow Tegra186 and later use
> 	APBDMA driver.
> 	Added explicit flow control enable for DMA slave config and error handling.
> 	Moved releasing DMA resources to seperate function to reuse in
> 	multiple places.
> 	Updated to register tegra_i2c_driver from module level rather than subsys
> 	level.
> 	Other minor feedback
>  [V9] : Rebased to 5.0-rc4
> 	Removed dependency of APB DMA in Kconfig and added conditional check
> 	in I2C driver to decide on using DMA mode.
> 	Changed back the allocation of dma buffer during i2c probe.
> 	Fixed FIFO triggers depending on DMA Vs PIO.
>  [V8] : Moved back dma init to i2c probe, removed ALL_PACKETS_XFER_COMPLETE
> 	interrupt and using PACKETS_XFER_COMPLETE interrupt only and some
> 	other fixes
> 	Updated Kconfig for APB_DMA dependency
>  [V7] : Same as V6
>  [V6] : Updated for proper buffer allocation/freeing, channel release.
> 	Updated to use exact xfer size for syncing dma buffer.
>  [V5] : Same as V4
>  [V4] : Updated to allocate DMA buffer only when DMA mode.
> 	Updated to fall back to PIO mode when DMA channel request or
> 	buffer allocation fails.
>  [V3] : Updated without additional buffer allocation.
>  [V2] : Updated based on V1 review feedback along with code cleanup for
> 	proper implementation of DMA.
> 
>  drivers/i2c/busses/i2c-tegra.c | 412 ++++++++++++++++++++++++++++++++++++-----
>  1 file changed, 368 insertions(+), 44 deletions(-)
> 
> diff --git a/drivers/i2c/busses/i2c-tegra.c b/drivers/i2c/busses/i2c-tegra.c
> index 118b7023a0f4..267a6b3084bf 100644
> --- a/drivers/i2c/busses/i2c-tegra.c
> +++ b/drivers/i2c/busses/i2c-tegra.c
> @@ -8,6 +8,9 @@
>  
>  #include <linux/clk.h>
>  #include <linux/delay.h>
> +#include <linux/dmaengine.h>
> +#include <linux/dmapool.h>
> +#include <linux/dma-mapping.h>
>  #include <linux/err.h>
>  #include <linux/i2c.h>
>  #include <linux/init.h>
> @@ -44,6 +47,8 @@
>  #define I2C_FIFO_CONTROL_RX_FLUSH		BIT(0)
>  #define I2C_FIFO_CONTROL_TX_TRIG_SHIFT		5
>  #define I2C_FIFO_CONTROL_RX_TRIG_SHIFT		2
> +#define I2C_FIFO_CONTROL_TX_TRIG(x)		(((x) - 1) << 5)
> +#define I2C_FIFO_CONTROL_RX_TRIG(x)		(((x) - 1) << 2)
>  #define I2C_FIFO_STATUS				0x060
>  #define I2C_FIFO_STATUS_TX_MASK			0xF0
>  #define I2C_FIFO_STATUS_TX_SHIFT		4
> @@ -125,6 +130,19 @@
>  #define I2C_MST_FIFO_STATUS_TX_MASK		0xff0000
>  #define I2C_MST_FIFO_STATUS_TX_SHIFT		16
>  
> +/* Packet header size in bytes */
> +#define I2C_PACKET_HEADER_SIZE			12
> +
> +#define DATA_DMA_DIR_TX				(1 << 0)
> +#define DATA_DMA_DIR_RX				(1 << 1)

Please use "./scripts/checkpatch.pl --strict *.patch" and fix all its complains, but only those that really make sense. For example ignore the "CHECK: Lines should not end with a '('" warnings. 

Here checkpatch recommends to use the BIT() macro:

CHECK: Prefer using the BIT macro
#394: FILE: drivers/i2c/busses/i2c-tegra.c:136:
+#define DATA_DMA_DIR_TX                                (1 << 0)

CHECK: Prefer using the BIT macro
#395: FILE: drivers/i2c/busses/i2c-tegra.c:137:
+#define DATA_DMA_DIR_RX                                (1 << 1)

> +
> +/*
> + * Upto I2C_PIO_MODE_MAX_LEN bytes, controller will use PIO mode,
> + * above this, controller will use DMA to fill FIFO.
> + * MAX PIO len is 20 bytes excluding packet header.
> + */
> +#define I2C_PIO_MODE_MAX_LEN			32
> +
>  /*
>   * msg_end_type: The bus control which need to be send at end of transfer.
>   * @MSG_END_STOP: Send stop pulse at end of transfer.
> @@ -166,6 +184,7 @@ enum msg_end_type {
>   *		allowing 0 length transfers.
>   * @supports_bus_clear: Bus Clear support to recover from bus hang during
>   *		SDA stuck low from device for some unknown reasons.
> + * @has_apb_dma: Support of APBDMA on corresponding Tegra chip.
>   */
>  struct tegra_i2c_hw_feature {
>  	bool has_continue_xfer_support;
> @@ -180,6 +199,7 @@ struct tegra_i2c_hw_feature {
>  	bool has_mst_fifo;
>  	const struct i2c_adapter_quirks *quirks;
>  	bool supports_bus_clear;
> +	bool has_apb_dma;
>  };
>  
>  /**
> @@ -191,6 +211,7 @@ struct tegra_i2c_hw_feature {
>   * @fast_clk: clock reference for fast clock of I2C controller
>   * @rst: reset control for the I2C controller
>   * @base: ioremapped registers cookie
> + * @base_phys: Physical base address of the I2C controller
>   * @cont_id: I2C controller ID, used for packet header
>   * @irq: IRQ number of transfer complete interrupt
>   * @irq_disabled: used to track whether or not the interrupt is enabled
> @@ -204,6 +225,13 @@ struct tegra_i2c_hw_feature {
>   * @clk_divisor_non_hs_mode: clock divider for non-high-speed modes
>   * @is_multimaster_mode: track if I2C controller is in multi-master mode
>   * @xfer_lock: lock to serialize transfer submission and processing
> + * @tx_dma_chan: DMA transmit channel
> + * @rx_dma_chan: DMA receive channel
> + * @dma_phys: handle to DMA resources
> + * @dma_buf: pointer to allocated DMA buffer
> + * @dma_buf_size: DMA buffer size
> + * @is_curr_dma_xfer: indicates active DMA transfer
> + * @dma_complete: DMA completion notifier
>   */
>  struct tegra_i2c_dev {
>  	struct device *dev;
> @@ -213,6 +241,7 @@ struct tegra_i2c_dev {
>  	struct clk *fast_clk;
>  	struct reset_control *rst;
>  	void __iomem *base;
> +	phys_addr_t base_phys;
>  	int cont_id;
>  	int irq;
>  	bool irq_disabled;
> @@ -226,6 +255,13 @@ struct tegra_i2c_dev {
>  	u16 clk_divisor_non_hs_mode;
>  	bool is_multimaster_mode;
>  	spinlock_t xfer_lock;
> +	struct dma_chan *tx_dma_chan;
> +	struct dma_chan *rx_dma_chan;
> +	dma_addr_t dma_phys;
> +	u32 *dma_buf;
> +	unsigned int dma_buf_size;
> +	bool is_curr_dma_xfer;
> +	struct completion dma_complete;
>  };
>  
>  static void dvc_writel(struct tegra_i2c_dev *i2c_dev, u32 val,
> @@ -294,6 +330,109 @@ static void tegra_i2c_unmask_irq(struct tegra_i2c_dev *i2c_dev, u32 mask)
>  	i2c_writel(i2c_dev, int_mask, I2C_INT_MASK);
>  }
>  
> +static void tegra_i2c_dma_complete(void *args)
> +{
> +	struct tegra_i2c_dev *i2c_dev = args;
> +
> +	complete(&i2c_dev->dma_complete);
> +}
> +
> +static int tegra_i2c_dma_submit(struct tegra_i2c_dev *i2c_dev, size_t len)
> +{
> +	struct dma_async_tx_descriptor *dma_desc;
> +	enum dma_transfer_direction dir;
> +	struct dma_chan *chan;
> +
> +	dev_dbg(i2c_dev->dev, "starting DMA for length: %zu\n", len);
> +	reinit_completion(&i2c_dev->dma_complete);
> +	dir = i2c_dev->msg_read ? DMA_DEV_TO_MEM : DMA_MEM_TO_DEV;
> +	chan = i2c_dev->msg_read ? i2c_dev->rx_dma_chan : i2c_dev->tx_dma_chan;
> +	dma_desc = dmaengine_prep_slave_single(chan, i2c_dev->dma_phys,
> +					       len, dir, DMA_PREP_INTERRUPT |
> +					       DMA_CTRL_ACK);
> +	if (!dma_desc) {
> +		dev_err(i2c_dev->dev, "failed to get DMA descriptor\n");
> +		return -EIO;

Returning the -EIO is technically incorrect because there is no hardware failure here. The dmaengine_prep_slave_single() merely allocates the DMA descriptor, hence it should be either -EINVAL (preferably) or at least -ENOMEM.

Oh, another important moment is that physically contiguous dma_buf allocation isn't guaranteed by the DMA API. This may become a problem for T186+ that can transfer up to 64K. We need to enforce the contiguous-allocation requirement by using dma_alloc_attrs(DMA_ATTR_FORCE_CONTIGUOUS) instead of the dma_alloc_coherent(), please see my other comment below.

> +	}
> +
> +	dma_desc->callback = tegra_i2c_dma_complete;
> +	dma_desc->callback_param = i2c_dev;
> +	dmaengine_submit(dma_desc);
> +	dma_async_issue_pending(chan);
> +	return 0;
> +}
> +
> +static void tegra_i2c_release_dma(struct tegra_i2c_dev *i2c_dev)
> +{
> +	if (i2c_dev->dma_buf) {
> +		dma_free_coherent(i2c_dev->dev, i2c_dev->dma_buf_size,
> +				  i2c_dev->dma_buf, i2c_dev->dma_phys);

With the above comment in mind:

		dma_free_attrs(i2c_dev->dev, i2c_dev->dma_buf_size,
			       i2c_dev->dma_buf, i2c_dev->dma_phys,
			       DMA_ATTR_FORCE_CONTIGUOUS);

> +		i2c_dev->dma_buf = NULL;
> +	}
> +
> +	if (i2c_dev->tx_dma_chan) {
> +		dma_release_channel(i2c_dev->tx_dma_chan);
> +		i2c_dev->tx_dma_chan = NULL;
> +	}
> +
> +	if (i2c_dev->rx_dma_chan) {
> +		dma_release_channel(i2c_dev->rx_dma_chan);
> +		i2c_dev->rx_dma_chan = NULL;
> +	}
> +}
> +
> +static int tegra_i2c_init_dma(struct tegra_i2c_dev *i2c_dev)
> +{
> +	struct dma_chan *chan;
> +	u32 *dma_buf;
> +	dma_addr_t dma_phys;
> +	int err = 0;

No need to initialize "err" to 0 here.

> +
> +	if (!IS_ENABLED(CONFIG_TEGRA20_APB_DMA) ||
> +	    !i2c_dev->hw->has_apb_dma) {
> +		err = -ENODEV;
> +		goto err_out;
> +	}
> +
> +	chan = dma_request_slave_channel_reason(i2c_dev->dev, "rx");
> +	if (IS_ERR(chan)) {
> +		err = PTR_ERR(chan);
> +		goto err_out;
> +	}
> +
> +	i2c_dev->rx_dma_chan = chan;
> +
> +	chan = dma_request_slave_channel_reason(i2c_dev->dev, "tx");
> +	if (IS_ERR(chan)) {
> +		err = PTR_ERR(chan);
> +		goto err_out;
> +	}
> +
> +	i2c_dev->tx_dma_chan = chan;
> +
> +	dma_buf = dma_alloc_coherent(i2c_dev->dev, i2c_dev->dma_buf_size,
> +				     &dma_phys, GFP_KERNEL | __GFP_NOWARN);

Please use dma_alloc_attrs() instead of dma_alloc_coherent() because it could return a sparse allocation. This is especially troublesome for ARM64 platforms because IOMMU_DOMAIN_DMA is used by default there. We need to explicitly ask for the contiguous allocation:

	dma_buf = dma_alloc_attrs(i2c_dev->dev, i2c_dev->dma_buf_size,
				  &dma_phys, GFP_KERNEL,
				  DMA_ATTR_FORCE_CONTIGUOUS |
				  DMA_ATTR_NO_WARN);

> +	if (!dma_buf) {
> +		dev_err(i2c_dev->dev, "failed to allocate the DMA buffer\n");
> +		err = -ENOMEM;
> +		goto err_out;
> +	}
> +
> +	i2c_dev->dma_buf = dma_buf;
> +	i2c_dev->dma_phys = dma_phys;
> +	return 0;
> +
> +err_out:
> +	tegra_i2c_release_dma(i2c_dev);
> +	if (err != -EPROBE_DEFER) {
> +		dev_err(i2c_dev->dev, "can't use DMA, err: %d, using PIO\n",
> +			err);
> +		return 0;
> +	}
> +
> +	return err;
> +}
> +
>  static int tegra_i2c_flush_fifos(struct tegra_i2c_dev *i2c_dev)
>  {
>  	unsigned long timeout = jiffies + HZ;
> @@ -571,16 +710,6 @@ static int tegra_i2c_init(struct tegra_i2c_dev *i2c_dev)
>  		i2c_writel(i2c_dev, 0x00, I2C_SL_ADDR2);
>  	}
>  
> -	if (i2c_dev->hw->has_mst_fifo) {
> -		val = I2C_MST_FIFO_CONTROL_TX_TRIG(8) |
> -		      I2C_MST_FIFO_CONTROL_RX_TRIG(1);
> -		i2c_writel(i2c_dev, val, I2C_MST_FIFO_CONTROL);
> -	} else {
> -		val = 7 << I2C_FIFO_CONTROL_TX_TRIG_SHIFT |
> -			0 << I2C_FIFO_CONTROL_RX_TRIG_SHIFT;
> -		i2c_writel(i2c_dev, val, I2C_FIFO_CONTROL);
> -	}
> -
>  	err = tegra_i2c_flush_fifos(i2c_dev);
>  	if (err)
>  		goto err;
> @@ -660,25 +789,37 @@ static irqreturn_t tegra_i2c_isr(int irq, void *dev_id)
>  	if (i2c_dev->hw->supports_bus_clear && (status & I2C_INT_BUS_CLR_DONE))
>  		goto err;
>  
> -	if (i2c_dev->msg_read && (status & I2C_INT_RX_FIFO_DATA_REQ)) {
> -		if (i2c_dev->msg_buf_remaining)
> -			tegra_i2c_empty_rx_fifo(i2c_dev);
> -		else
> -			BUG();
> -	}
> +	if (!i2c_dev->is_curr_dma_xfer) {
> +		if (i2c_dev->msg_read && (status & I2C_INT_RX_FIFO_DATA_REQ)) {
> +			if (i2c_dev->msg_buf_remaining)
> +				tegra_i2c_empty_rx_fifo(i2c_dev);
> +			else
> +				BUG();
> +		}
>  
> -	if (!i2c_dev->msg_read && (status & I2C_INT_TX_FIFO_DATA_REQ)) {
> -		if (i2c_dev->msg_buf_remaining)
> -			tegra_i2c_fill_tx_fifo(i2c_dev);
> -		else
> -			tegra_i2c_mask_irq(i2c_dev, I2C_INT_TX_FIFO_DATA_REQ);
> +		if (!i2c_dev->msg_read && (status & I2C_INT_TX_FIFO_DATA_REQ)) {
> +			if (i2c_dev->msg_buf_remaining)
> +				tegra_i2c_fill_tx_fifo(i2c_dev);
> +			else
> +				tegra_i2c_mask_irq(i2c_dev,
> +						   I2C_INT_TX_FIFO_DATA_REQ);
> +		}
>  	}
>  
>  	i2c_writel(i2c_dev, status, I2C_INT_STATUS);
>  	if (i2c_dev->is_dvc)
>  		dvc_writel(i2c_dev, DVC_STATUS_I2C_DONE_INTR, DVC_STATUS);
>  
> +	/*
> +	 * During message read XFER_COMPLETE interrupt is triggered prior to
> +	 * DMA completion and during message write XFER_COMPLETE interrupt is
> +	 * triggered after DMA completion.
> +	 * PACKETS_XFER_COMPLETE indicates completion of all bytes of transfer.
> +	 * so forcing msg_buf_remaining to 0 in DMA mode.
> +	 */
>  	if (status & I2C_INT_PACKET_XFER_COMPLETE) {
> +		if (i2c_dev->is_curr_dma_xfer)
> +			i2c_dev->msg_buf_remaining = 0;
>  		BUG_ON(i2c_dev->msg_buf_remaining);
>  		complete(&i2c_dev->msg_complete);
>  	}
> @@ -694,12 +835,89 @@ static irqreturn_t tegra_i2c_isr(int irq, void *dev_id)
>  	if (i2c_dev->is_dvc)
>  		dvc_writel(i2c_dev, DVC_STATUS_I2C_DONE_INTR, DVC_STATUS);
>  
> +	if (i2c_dev->is_curr_dma_xfer) {
> +		if (i2c_dev->msg_read)
> +			dmaengine_terminate_async(i2c_dev->rx_dma_chan);
> +		else
> +			dmaengine_terminate_async(i2c_dev->tx_dma_chan);
> +
> +		complete(&i2c_dev->dma_complete);
> +	}
> +
>  	complete(&i2c_dev->msg_complete);
>  done:
>  	spin_unlock(&i2c_dev->xfer_lock);
>  	return IRQ_HANDLED;
>  }
>  
> +static int tegra_i2c_config_fifo_trig(struct tegra_i2c_dev *i2c_dev,
> +				      size_t len)
> +{
> +	u32 val, reg;
> +	u8 dma_burst = 0;
> +	struct dma_slave_config slv_config = {0};
> +	struct dma_chan *chan;
> +	int ret = 0;
> +
> +	if (i2c_dev->hw->has_mst_fifo)
> +		reg = I2C_MST_FIFO_CONTROL;
> +	else
> +		reg = I2C_FIFO_CONTROL;
> +	val = i2c_readl(i2c_dev, reg);
> +
> +	if (i2c_dev->is_curr_dma_xfer) {
> +		if (len & 0xF)
> +			dma_burst = 1;
> +		else if (len & 0x10)
> +			dma_burst = 4;
> +		else
> +			dma_burst = 8;
> +
> +		if (i2c_dev->msg_read) {
> +			chan = i2c_dev->rx_dma_chan;
> +			slv_config.src_addr = i2c_dev->base_phys + I2C_RX_FIFO;
> +			slv_config.src_addr_width = DMA_SLAVE_BUSWIDTH_4_BYTES;
> +			slv_config.src_maxburst = dma_burst;
> +
> +			if (i2c_dev->hw->has_mst_fifo)
> +				val |= I2C_MST_FIFO_CONTROL_RX_TRIG(dma_burst);
> +			else
> +				val |= I2C_FIFO_CONTROL_RX_TRIG(dma_burst);
> +		} else {
> +			chan = i2c_dev->tx_dma_chan;
> +			slv_config.dst_addr = i2c_dev->base_phys + I2C_TX_FIFO;
> +			slv_config.dst_addr_width = DMA_SLAVE_BUSWIDTH_4_BYTES;
> +			slv_config.dst_maxburst = dma_burst;
> +
> +			if (i2c_dev->hw->has_mst_fifo)
> +				val |= I2C_MST_FIFO_CONTROL_TX_TRIG(dma_burst);
> +			else
> +				val |= I2C_FIFO_CONTROL_TX_TRIG(dma_burst);
> +		}
> +
> +		slv_config.device_fc = true;
> +		ret = dmaengine_slave_config(chan, &slv_config);
> +		if (ret < 0) {
> +			dev_err(i2c_dev->dev,
> +				"DMA slave config failed, err: %d using PIO\n",
> +				ret);

Probably better to say here:

			dev_err(i2c_dev->dev, "DMA slave config failed: %d\n", ret);
			dev_info(i2c_dev->dev, "fallbacking to PIO\n");

> +			tegra_i2c_release_dma(i2c_dev);

tegra_i2c_config_fifo_trig() is called from atomic context (under spinlock) and dma_release_channel() uses the mutex(), hence tegra_i2c_release_dma() could sleep and this is a bug. Simplest solution will be to move tegra_i2c_config_fifo_trig() out from the spinlock:

--------------

diff --git a/drivers/i2c/busses/i2c-tegra.c b/drivers/i2c/busses/i2c-tegra.c
index 841cad6ccb57..968318d23fa7 100644
--- a/drivers/i2c/busses/i2c-tegra.c
+++ b/drivers/i2c/busses/i2c-tegra.c
@@ -1054,13 +1054,16 @@ static int tegra_i2c_xfer_msg(struct tegra_i2c_dev *i2c_dev,
        xfer_time += DIV_ROUND_CLOSEST(((xfer_size * 9) + 2) * MSEC_PER_SEC,
                                        i2c_dev->bus_clk_rate);
 
+       err = tegra_i2c_config_fifo_trig(i2c_dev, xfer_size);
+       if (err < 0) {
+               i2c_dev->is_curr_dma_xfer = false;
+               dma = false;
+       }
+
        spin_lock_irqsave(&i2c_dev->xfer_lock, flags);
 
        int_mask = I2C_INT_NO_ACK | I2C_INT_ARBITRATION_LOST;
        tegra_i2c_unmask_irq(i2c_dev, int_mask);
-       err = tegra_i2c_config_fifo_trig(i2c_dev, xfer_size);
-       if (err < 0)
-               i2c_dev->is_curr_dma_xfer = dma = false;

----------------

> +		} else {
> +			goto out;
> +		}
> +	}
> +
> +	if (i2c_dev->hw->has_mst_fifo)
> +		val = I2C_MST_FIFO_CONTROL_TX_TRIG(8) |
> +		      I2C_MST_FIFO_CONTROL_RX_TRIG(1);
> +	else
> +		val = I2C_FIFO_CONTROL_TX_TRIG(8) |
> +		      I2C_FIFO_CONTROL_RX_TRIG(1);
> +out:
> +	i2c_writel(i2c_dev, val, reg);
> +	return ret;
> +}
> +
>  static int tegra_i2c_issue_bus_clear(struct tegra_i2c_dev *i2c_dev)
>  {
>  	int err;
> @@ -744,6 +962,10 @@ static int tegra_i2c_xfer_msg(struct tegra_i2c_dev *i2c_dev,
>  	u32 int_mask;
>  	unsigned long time_left;
>  	unsigned long flags;
> +	size_t xfer_size;
> +	u32 *buffer = NULL;
> +	int err = 0;
> +	bool dma = false;
>  
>  	tegra_i2c_flush_fifos(i2c_dev);
>  
> @@ -753,19 +975,59 @@ static int tegra_i2c_xfer_msg(struct tegra_i2c_dev *i2c_dev,
>  	i2c_dev->msg_read = (msg->flags & I2C_M_RD);
>  	reinit_completion(&i2c_dev->msg_complete);
>  
> +	if (i2c_dev->msg_read)
> +		xfer_size = msg->len;
> +	else
> +		xfer_size = msg->len + I2C_PACKET_HEADER_SIZE;
> +
> +	xfer_size = ALIGN(xfer_size, BYTES_PER_FIFO_WORD);
> +	dma = (xfer_size > I2C_PIO_MODE_MAX_LEN) && i2c_dev->dma_buf;
> +	i2c_dev->is_curr_dma_xfer = dma;
> +
>  	spin_lock_irqsave(&i2c_dev->xfer_lock, flags);
>  
>  	int_mask = I2C_INT_NO_ACK | I2C_INT_ARBITRATION_LOST;
>  	tegra_i2c_unmask_irq(i2c_dev, int_mask);
> +	err = tegra_i2c_config_fifo_trig(i2c_dev, xfer_size);
> +	if (err < 0)
> +		i2c_dev->is_curr_dma_xfer = dma = false;

CHECK: multiple assignments should be avoided
#763: FILE: drivers/i2c/busses/i2c-tegra.c:993:
+               i2c_dev->is_curr_dma_xfer = dma = false

Please write this as:
	if (err < 0) {
		i2c_dev->is_curr_dma_xfer = false;
		dma = false;
	}

> +
> +	if (dma) {
> +		if (i2c_dev->msg_read) {
> +			dma_sync_single_for_device(i2c_dev->dev,
> +						   i2c_dev->dma_phys,
> +						   xfer_size,
> +						   DMA_FROM_DEVICE);
> +			err = tegra_i2c_dma_submit(i2c_dev, xfer_size);
> +			if (err < 0) {
> +				dev_err(i2c_dev->dev,
> +					"starting RX DMA failed, err %d\n",
> +					err);
> +				goto unlock;
> +			}
> +		} else {
> +			dma_sync_single_for_cpu(i2c_dev->dev,
> +						i2c_dev->dma_phys,
> +						xfer_size,
> +						DMA_TO_DEVICE);
> +			buffer = i2c_dev->dma_buf;
> +		}
> +	}
>  
>  	packet_header = (0 << PACKET_HEADER0_HEADER_SIZE_SHIFT) |
>  			PACKET_HEADER0_PROTOCOL_I2C |
>  			(i2c_dev->cont_id << PACKET_HEADER0_CONT_ID_SHIFT) |
>  			(1 << PACKET_HEADER0_PACKET_ID_SHIFT);
> -	i2c_writel(i2c_dev, packet_header, I2C_TX_FIFO);
> +	if (dma && !i2c_dev->msg_read)
> +		*buffer++ = packet_header;
> +	else
> +		i2c_writel(i2c_dev, packet_header, I2C_TX_FIFO);
>  
>  	packet_header = msg->len - 1;
> -	i2c_writel(i2c_dev, packet_header, I2C_TX_FIFO);
> +	if (dma && !i2c_dev->msg_read)
> +		*buffer++ = packet_header;
> +	else
> +		i2c_writel(i2c_dev, packet_header, I2C_TX_FIFO);
>  
>  	packet_header = I2C_HEADER_IE_ENABLE;
>  	if (end_state == MSG_END_CONTINUE)
> @@ -782,23 +1044,79 @@ static int tegra_i2c_xfer_msg(struct tegra_i2c_dev *i2c_dev,
>  		packet_header |= I2C_HEADER_CONT_ON_NAK;
>  	if (msg->flags & I2C_M_RD)
>  		packet_header |= I2C_HEADER_READ;
> -	i2c_writel(i2c_dev, packet_header, I2C_TX_FIFO);
> -
> -	if (!(msg->flags & I2C_M_RD))
> -		tegra_i2c_fill_tx_fifo(i2c_dev);
> +	if (dma && !i2c_dev->msg_read)
> +		*buffer++ = packet_header;
> +	else
> +		i2c_writel(i2c_dev, packet_header, I2C_TX_FIFO);
> +
> +	if (!msg->flags & I2C_M_RD) {
> +		if (dma) {
> +			memcpy(buffer, msg->buf, msg->len);
> +			dma_sync_single_for_device(i2c_dev->dev,
> +						   i2c_dev->dma_phys,
> +						   xfer_size,
> +						   DMA_TO_DEVICE);
> +			err = tegra_i2c_dma_submit(i2c_dev, xfer_size);
> +			if (err < 0) {
> +				dev_err(i2c_dev->dev,
> +					"starting TX DMA failed, err %d\n",
> +					err);
> +				goto unlock;
> +			}
> +		} else {
> +			tegra_i2c_fill_tx_fifo(i2c_dev);
> +		}
> +	}
>  
>  	if (i2c_dev->hw->has_per_pkt_xfer_complete_irq)
>  		int_mask |= I2C_INT_PACKET_XFER_COMPLETE;
> -	if (msg->flags & I2C_M_RD)
> -		int_mask |= I2C_INT_RX_FIFO_DATA_REQ;
> -	else if (i2c_dev->msg_buf_remaining)
> -		int_mask |= I2C_INT_TX_FIFO_DATA_REQ;
> +
> +	if (!dma) {
> +		if (msg->flags & I2C_M_RD)
> +			int_mask |= I2C_INT_RX_FIFO_DATA_REQ;
> +		else if (i2c_dev->msg_buf_remaining)
> +			int_mask |= I2C_INT_TX_FIFO_DATA_REQ;
> +	}
>  
>  	tegra_i2c_unmask_irq(i2c_dev, int_mask);
> -	spin_unlock_irqrestore(&i2c_dev->xfer_lock, flags);
>  	dev_dbg(i2c_dev->dev, "unmasked irq: %02x\n",
>  		i2c_readl(i2c_dev, I2C_INT_MASK));
>  
> +unlock:
> +	spin_unlock_irqrestore(&i2c_dev->xfer_lock, flags);
> +
> +	if (dma) {
> +		if (err)
> +			return err;
> +
> +		time_left = wait_for_completion_timeout(
> +						&i2c_dev->dma_complete,
> +						TEGRA_I2C_TIMEOUT);
> +
> +		if (time_left == 0) {
> +			dev_err(i2c_dev->dev, "DMA transfer timeout\n");
> +			dmaengine_terminate_sync(i2c_dev->msg_read ?
> +						  i2c_dev->rx_dma_chan :
> +						  i2c_dev->tx_dma_chan);
> +			tegra_i2c_init(i2c_dev);
> +			return -ETIMEDOUT;
> +		}
> +
> +		if (i2c_dev->msg_read && (i2c_dev->msg_err == I2C_ERR_NONE)) {

CHECK: Unnecessary parentheses around 'i2c_dev->msg_err == I2C_ERR_NONE'
#877: FILE: drivers/i2c/busses/i2c-tegra.c:1105:
+               if (i2c_dev->msg_read && (i2c_dev->msg_err == I2C_ERR_NONE)) {

> +			dma_sync_single_for_cpu(i2c_dev->dev,
> +						i2c_dev->dma_phys,
> +						xfer_size,
> +						DMA_FROM_DEVICE);
> +			memcpy(i2c_dev->msg_buf, i2c_dev->dma_buf,
> +				msg->len);

CHECK: Alignment should match open parenthesis
#883: FILE: drivers/i2c/busses/i2c-tegra.c:1111:
+                       memcpy(i2c_dev->msg_buf, i2c_dev->dma_buf,
+                               msg->len);

> +		}

Newline here, please.

> +		if (i2c_dev->msg_err != I2C_ERR_NONE) {
> +			dmaengine_synchronize(i2c_dev->msg_read ?
> +					      i2c_dev->rx_dma_chan :
> +					      i2c_dev->tx_dma_chan);
> +		}
> +	}
> +
>  	time_left = wait_for_completion_timeout(&i2c_dev->msg_complete,
>  						TEGRA_I2C_TIMEOUT);

[snip]