Re: [PATCH 3/7] dmaengine: xilinx_vdma: Add Support for Xilinx AXI Direct Memory Access Engine

[Laurent Pinchart]

Hello,

Thank you for the patch.

On Tuesday 15 March 2016 22:53:08 Kedareswara rao Appana wrote:
> This patch adds support for the AXI Direct Memory Access (AXI DMA)
> core, which is a soft Xilinx IP core that provides high-
> bandwidth direct memory access between memory and AXI4-Stream
> type target peripherals.
> 
> Signed-off-by: Kedareswara rao Appana <appanad@xxxxxxxxxx>
> ---
>  drivers/dma/xilinx/xilinx_vdma.c | 385 +++++++++++++++++++++++++++++++-----
>  1 file changed, 341 insertions(+), 44 deletions(-)
> 
> diff --git a/drivers/dma/xilinx/xilinx_vdma.c
> b/drivers/dma/xilinx/xilinx_vdma.c index f682bef..87525a9 100644
> --- a/drivers/dma/xilinx/xilinx_vdma.c
> +++ b/drivers/dma/xilinx/xilinx_vdma.c
> @@ -16,6 +16,12 @@
>   * video device (S2MM). Initialization, status, interrupt and management
>   * registers are accessed through an AXI4-Lite slave interface.
>   *
> + *  The AXI DMA, is a soft IP, which provides high-bandwidth Direct Memory
> + *  Access between memory and AXI4-Stream-type target peripherals. It can
> be
> + *  configured to have one channel or two channels and if configured as two
> + *  channels, one is to transmit data from memory to a device and another
> is
> + *  to receive from a device.

Let's try to be both descriptive and consistent.

"The AXI Direct Memory Access (AXI DMA) core is a soft Xilinx IP core the 
provides high-bandwidth one dimensional direct memory access between memory 
and AXI4-Stream target peripherals. It supports one receive and one transmit 
channel, both of them optional at synthesis time."

> + *
>   * 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, either version 2 of the License, or
> @@ -140,6 +146,20 @@
>  #define XILINX_VDMA_LOOP_COUNT		1000000
> 
>  #define AXIVDMA_SUPPORT		BIT(0)
> +#define AXIDMA_SUPPORT		BIT(1)
> +
> +/* AXI DMA Specific Registers/Offsets */
> +#define XILINX_DMA_REG_SRCDSTADDR	0x18
> +#define XILINX_DMA_REG_DSTADDR		0x20
> +#define XILINX_DMA_REG_BTT		0x28
> +
> +#define XILINX_DMA_MAX_TRANS_LEN	GENMASK(22, 0)
> +#define XILINX_DMA_CR_COALESCE_MAX	GENMASK(23, 16)
> +#define XILINX_DMA_CR_COALESCE_SHIFT	16
> +#define XILINX_DMA_BD_SOP		BIT(27)
> +#define XILINX_DMA_BD_EOP		BIT(26)
> +#define XILINX_DMA_COALESCE_MAX		255
> +#define XILINX_DMA_NUM_APP_WORDS	5
> 
>  /**
>   * struct xilinx_vdma_desc_hw - Hardware Descriptor
> @@ -147,19 +167,23 @@
>   * @pad1: Reserved @0x04
>   * @buf_addr: Buffer address @0x08
>   * @pad2: Reserved @0x0C
> - * @vsize: Vertical Size @0x10
> + * @dstaddr_vsize: Vertical Size @0x10
>   * @hsize: Horizontal Size @0x14
> - * @stride: Number of bytes between the first
> + * @control_stride: Number of bytes between the first
>   *	    pixels of each horizontal line @0x18
> + * @status: Status field @0x1C
> + * @app: APP Fields @0x20 - 0x30

As the descriptors are not identical for the DMA and VDMA cores, please define 
two structures instead of abusing this one.

>   */
>  struct xilinx_vdma_desc_hw {
>  	u32 next_desc;
>  	u32 pad1;
>  	u32 buf_addr;
>  	u32 pad2;
> -	u32 vsize;
> +	u32 dstaddr_vsize;
>  	u32 hsize;
> -	u32 stride;
> +	u32 control_stride;
> +	u32 status;
> +	u32 app[XILINX_DMA_NUM_APP_WORDS];
>  } __aligned(64);
> 
>  /**
> @@ -209,6 +233,9 @@ struct xilinx_vdma_tx_descriptor {
>   * @config: Device configuration info
>   * @flush_on_fsync: Flush on Frame sync
>   * @desc_pendingcount: Descriptor pending count
> + * @residue: Residue for AXI DMA
> + * @seg_v: Statically allocated segments base
> + * @start_transfer: Differentiate b/w DMA IP's transfer

Please clearly document which fields are common and which fields are specific 
to one DMA engine type.

>   */
>  struct xilinx_vdma_chan {
>  	struct xilinx_vdma_device *xdev;
> @@ -232,6 +259,9 @@ struct xilinx_vdma_chan {
>  	struct xilinx_vdma_config config;
>  	bool flush_on_fsync;
>  	u32 desc_pendingcount;
> +	u32 residue;
> +	struct xilinx_vdma_tx_segment *seg_v;
> +	void   (*start_transfer)(struct xilinx_vdma_chan *chan);

One space after void is enough.

>  };
> 
>  /**
> @@ -436,6 +466,7 @@ static void xilinx_vdma_free_chan_resources(struct
> dma_chan *dchan) dev_dbg(chan->dev, "Free all channel resources.\n");
> 
>  	xilinx_vdma_free_descriptors(chan);
> +	xilinx_vdma_free_tx_segment(chan, chan->seg_v);
>  	dma_pool_destroy(chan->desc_pool);
>  	chan->desc_pool = NULL;
>  }
> @@ -515,7 +546,12 @@ static int xilinx_vdma_alloc_chan_resources(struct
> dma_chan *dchan) return -ENOMEM;
>  	}
> 
> +	chan->seg_v = xilinx_vdma_alloc_tx_segment(chan);

This is a bit scary. You allocate a segment here and free it in 
xilinx_vdma_free_chan_resources, but seg_v is reassigned in 
xilinx_dma_start_transfer(). A comment explaining how seg_v is used and why 
this is safe would be nice.

>  	dma_cookie_init(dchan);
> +
> +	/* Enable interrupts */
> +	vdma_ctrl_set(chan, XILINX_VDMA_REG_DMACR,
> +		      XILINX_VDMA_DMAXR_ALL_IRQ_MASK);

Why is this needed here ?

>  	return 0;
>  }
> 
> @@ -531,7 +567,37 @@ static enum dma_status xilinx_vdma_tx_status(struct
> dma_chan *dchan, dma_cookie_t cookie,
>  					struct dma_tx_state *txstate)
>  {
> -	return dma_cookie_status(dchan, cookie, txstate);
> +	struct xilinx_vdma_chan *chan = to_xilinx_chan(dchan);
> +	struct xilinx_vdma_tx_descriptor *desc;
> +	struct xilinx_vdma_tx_segment *segment;
> +	struct xilinx_vdma_desc_hw *hw;
> +	enum dma_status ret;
> +	unsigned long flags;
> +	u32 residue = 0;
> +
> +	ret = dma_cookie_status(dchan, cookie, txstate);
> +	if (ret == DMA_COMPLETE || !txstate)
> +		return ret;
> +
> +	if (chan->xdev->quirks & AXIDMA_SUPPORT) {
> +		desc = list_last_entry(&chan->active_list,
> +				       struct xilinx_vdma_tx_descriptor, node);

Doesn't this need to be protected against race conditions ?

> +
> +		spin_lock_irqsave(&chan->lock, flags);
> +		if (chan->has_sg) {
> +			list_for_each_entry(segment, &desc->segments, node) {
> +				hw = &segment->hw;
> +				residue += (hw->control_stride - hw->status) &
> +					   XILINX_DMA_MAX_TRANS_LEN;
> +			}
> +		}
> +		spin_unlock_irqrestore(&chan->lock, flags);
> +
> +		chan->residue = residue;
> +		dma_set_residue(txstate, chan->residue);
> +	}

Is this really specific to the DMA engine type, or is it applicable to the 
VDMA and CDMA engines as well ?

> +	return ret;
>  }
> 
>  /**
> @@ -713,8 +779,9 @@ static void xilinx_vdma_start_transfer(struct
> xilinx_vdma_chan *chan) /* HW expects these parameters to be same for one
> transaction */ vdma_desc_write(chan, XILINX_VDMA_REG_HSIZE,
> last->hw.hsize);
>  		vdma_desc_write(chan, XILINX_VDMA_REG_FRMDLY_STRIDE,
> -				last->hw.stride);
> -		vdma_desc_write(chan, XILINX_VDMA_REG_VSIZE, last->hw.vsize);
> +				last->hw.control_stride);
> +		vdma_desc_write(chan, XILINX_VDMA_REG_VSIZE,
> +				last->hw.dstaddr_vsize);
>  	}
> 
>  	list_splice_tail_init(&chan->pending_list, &chan->active_list);
> @@ -736,6 +803,105 @@ static void xilinx_vdma_issue_pending(struct dma_chan
> *dchan) }
> 
>  /**
> + * xilinx_dma_start_transfer - Starts DMA transfer
> + * @chan: Driver specific channel struct pointer
> + */
> +static void xilinx_dma_start_transfer(struct xilinx_vdma_chan *chan)
> +{
> +	struct xilinx_vdma_tx_descriptor *head_desc, *tail_desc;
> +	struct xilinx_vdma_tx_segment *tail_segment, *old_head, *new_head;
> +	u32 reg;
> +
> +	/* This function was invoked with lock held */

If you want this to be mentioned in a comment please add it to the comment 
block before the function. I'd write it as "This function must be called with 
the channel lock held.".

> +	if (chan->err)
> +		return;
> +
> +	if (list_empty(&chan->pending_list))
> +		return;
> +
> +	head_desc = list_first_entry(&chan->pending_list,
> +				     struct xilinx_vdma_tx_descriptor, node);
> +	tail_desc = list_last_entry(&chan->pending_list,
> +				    struct xilinx_vdma_tx_descriptor, node);
> +	tail_segment = list_last_entry(&tail_desc->segments,
> +				       struct xilinx_vdma_tx_segment, node);
> +
> +	old_head = list_first_entry(&head_desc->segments,
> +				struct xilinx_vdma_tx_segment, node);
> +	new_head = chan->seg_v;
> +	/* Copy Buffer Descriptor fields. */
> +	new_head->hw = old_head->hw;
> +
> +	/* Swap and save new reserve */
> +	list_replace_init(&old_head->node, &new_head->node);
> +	chan->seg_v = old_head;
> +
> +	tail_segment->hw.next_desc = chan->seg_v->phys;
> +	head_desc->async_tx.phys = new_head->phys;
> +
> +	/* If it is SG mode and hardware is busy, cannot submit */
> +	if (chan->has_sg && xilinx_vdma_is_running(chan) &&
> +	    !xilinx_vdma_is_idle(chan)) {
> +		dev_dbg(chan->dev, "DMA controller still busy\n");
> +		return;

Shouldn't this be checked before modifying the descriptors above ?

> +	}
> +
> +	reg = vdma_ctrl_read(chan, XILINX_VDMA_REG_DMACR);
> +
> +	if (chan->desc_pendingcount <= XILINX_DMA_COALESCE_MAX) {
> +		reg &= ~XILINX_DMA_CR_COALESCE_MAX;
> +		reg |= chan->desc_pendingcount <<
> +				  XILINX_DMA_CR_COALESCE_SHIFT;
> +		vdma_ctrl_write(chan, XILINX_VDMA_REG_DMACR, reg);
> +	}
> +
> +	if (chan->has_sg)
> +		vdma_ctrl_write(chan, XILINX_VDMA_REG_CURDESC,
> +				head_desc->async_tx.phys);
> +
> +	xilinx_vdma_start(chan);
> +
> +	if (chan->err)
> +		return;
> +
> +	/* Start the transfer */
> +	if (chan->has_sg) {
> +		vdma_ctrl_write(chan, XILINX_VDMA_REG_TAILDESC,
> +			       tail_segment->phys);
> +	} else {
> +		struct xilinx_vdma_tx_segment *segment;
> +		struct xilinx_vdma_desc_hw *hw;
> +
> +		segment = list_first_entry(&head_desc->segments,
> +					   struct xilinx_vdma_tx_segment, node);
> +		hw = &segment->hw;
> +
> +		vdma_ctrl_write(chan, XILINX_DMA_REG_SRCDSTADDR, hw->buf_addr);
> +
> +		/* Start the transfer */
> +		vdma_ctrl_write(chan, XILINX_DMA_REG_BTT,
> +			       hw->control_stride & XILINX_DMA_MAX_TRANS_LEN);
> +	}
> +
> +	list_splice_tail_init(&chan->pending_list, &chan->active_list);
> +	chan->desc_pendingcount = 0;
> +}
> +
> +/**
> + * xilinx_dma_issue_pending - Issue pending transactions
> + * @dchan: DMA channel
> + */
> +static void xilinx_dma_issue_pending(struct dma_chan *dchan)
> +{
> +	struct xilinx_vdma_chan *chan = to_xilinx_chan(dchan);
> +	unsigned long flags;
> +
> +	spin_lock_irqsave(&chan->lock, flags);
> +	xilinx_dma_start_transfer(chan);

If you write it as chan->start_transfer(chan) you won't have to duplicate the 
issue_pending function.

> +	spin_unlock_irqrestore(&chan->lock, flags);
> +}
> +
> +/**
>   * xilinx_vdma_complete_descriptor - Mark the active descriptor as complete
> * @chan : xilinx DMA channel
>   *
> @@ -841,15 +1007,12 @@ static irqreturn_t xilinx_vdma_irq_handler(int irq,
> void *data) vdma_ctrl_write(chan, XILINX_VDMA_REG_DMASR,
>  				errors & XILINX_VDMA_DMASR_ERR_RECOVER_MASK);
> 
> -		if (!chan->flush_on_fsync ||
> -		    (errors & ~XILINX_VDMA_DMASR_ERR_RECOVER_MASK)) {
> -			dev_err(chan->dev,
> -				"Channel %p has errors %x, cdr %x tdr %x\n",
> -				chan, errors,
> -				vdma_ctrl_read(chan, XILINX_VDMA_REG_CURDESC),
> -				vdma_ctrl_read(chan, XILINX_VDMA_REG_TAILDESC));
> -			chan->err = true;
> -		}
> +		dev_err(chan->dev,
> +			"Channel %p has errors %x, cdr %x tdr %x\n",
> +			chan, errors,
> +			vdma_ctrl_read(chan, XILINX_VDMA_REG_CURDESC),
> +			vdma_ctrl_read(chan, XILINX_VDMA_REG_TAILDESC));
> +		chan->err = true;

You're removing the ability to recover when C_FLUSH_ON_FSYNC is set. Why is 
that ?

>  	}
> 
>  	if (status & XILINX_VDMA_DMASR_DLY_CNT_IRQ) {
> @@ -863,7 +1026,7 @@ static irqreturn_t xilinx_vdma_irq_handler(int irq,
> void *data) if (status & XILINX_VDMA_DMASR_FRM_CNT_IRQ) {
>  		spin_lock(&chan->lock);
>  		xilinx_vdma_complete_descriptor(chan);
> -		xilinx_vdma_start_transfer(chan);
> +		chan->start_transfer(chan);
>  		spin_unlock(&chan->lock);
>  	}
> 
> @@ -903,7 +1066,8 @@ append:
>  	list_add_tail(&desc->node, &chan->pending_list);
>  	chan->desc_pendingcount++;
> 
> -	if (unlikely(chan->desc_pendingcount > chan->num_frms)) {
> +	if ((unlikely(chan->desc_pendingcount > chan->num_frms)) &&

Parenthesis around unlikely are not needed.

> +		(chan->xdev->quirks & AXIVDMA_SUPPORT)) {
>  		dev_dbg(chan->dev, "desc pendingcount is too high\n");
>  		chan->desc_pendingcount = chan->num_frms;
>  	}
> @@ -989,11 +1153,11 @@ xilinx_vdma_dma_prep_interleaved(struct dma_chan
> *dchan,
> 
>  	/* Fill in the hardware descriptor */
>  	hw = &segment->hw;
> -	hw->vsize = xt->numf;
> +	hw->dstaddr_vsize = xt->numf;
>  	hw->hsize = xt->sgl[0].size;
> -	hw->stride = (xt->sgl[0].icg + xt->sgl[0].size) <<
> +	hw->control_stride = (xt->sgl[0].icg + xt->sgl[0].size) <<
>  			XILINX_VDMA_FRMDLY_STRIDE_STRIDE_SHIFT;
> -	hw->stride |= chan->config.frm_dly <<
> +	hw->control_stride |= chan->config.frm_dly <<
>  			XILINX_VDMA_FRMDLY_STRIDE_FRMDLY_SHIFT;
> 
>  	if (xt->dir != DMA_MEM_TO_DEV)
> @@ -1019,6 +1183,108 @@ error:
>  }
> 
>  /**
> + * xilinx_dma_prep_slave_sg - prepare descriptors for a DMA_SLAVE
> transaction + * @dchan: DMA channel
> + * @sgl: scatterlist to transfer to/from
> + * @sg_len: number of entries in @scatterlist
> + * @direction: DMA direction
> + * @flags: transfer ack flags
> + * @context: APP words of the descriptor
> + *
> + * Return: Async transaction descriptor on success and NULL on failure
> + */
> +static struct dma_async_tx_descriptor *xilinx_dma_prep_slave_sg(
> +	struct dma_chan *dchan, struct scatterlist *sgl, unsigned int sg_len,
> +	enum dma_transfer_direction direction, unsigned long flags,
> +	void *context)
> +{
> +	struct xilinx_vdma_chan *chan = to_xilinx_chan(dchan);
> +	struct xilinx_vdma_tx_descriptor *desc;
> +	struct xilinx_vdma_tx_segment *segment = NULL, *prev = NULL;
> +	u32 *app_w = (u32 *)context;
> +	struct scatterlist *sg;
> +	size_t copy, sg_used;

Please don't declare multiple unrelated variables on a single line.

> +	int i;

i will never be negative, you can make it an unsigned int.

> +
> +	if (!is_slave_direction(direction))
> +		return NULL;
> +
> +	/* Allocate a transaction descriptor. */
> +	desc = xilinx_vdma_alloc_tx_descriptor(chan);
> +	if (!desc)
> +		return NULL;
> +
> +	dma_async_tx_descriptor_init(&desc->async_tx, &chan->common);
> +	desc->async_tx.tx_submit = xilinx_vdma_tx_submit;
> +
> +	/* Build transactions using information in the scatter gather list */
> +	for_each_sg(sgl, sg, sg_len, i) {
> +		sg_used = 0;
> +
> +		/* Loop until the entire scatterlist entry is used */
> +		while (sg_used < sg_dma_len(sg)) {
> +			struct xilinx_vdma_desc_hw *hw;
> +
> +			/* Get a free segment */
> +			segment = xilinx_vdma_alloc_tx_segment(chan);
> +			if (!segment)
> +				goto error;
> +
> +			/*
> +			 * Calculate the maximum number of bytes to transfer,
> +			 * making sure it is less than the hw limit
> +			 */
> +			copy = min_t(size_t, sg_dma_len(sg) - sg_used,
> +				     XILINX_DMA_MAX_TRANS_LEN);
> +			hw = &segment->hw;
> +
> +			/* Fill in the descriptor */
> +			hw->buf_addr = sg_dma_address(sg) + sg_used;
> +
> +			hw->control_stride = copy;
> +
> +			if (chan->direction == DMA_MEM_TO_DEV) {
> +				if (app_w)
> +					memcpy(hw->app, app_w, sizeof(u32) *
> +					       XILINX_DMA_NUM_APP_WORDS);
> +			}
> +
> +			if (prev)
> +				prev->hw.next_desc = segment->phys;
> +
> +			prev = segment;
> +			sg_used += copy;
> +
> +			/*
> +			 * Insert the segment into the descriptor segments
> +			 * list.
> +			 */
> +			list_add_tail(&segment->node, &desc->segments);
> +		}
> +	}
> +
> +	segment = list_first_entry(&desc->segments,
> +				   struct xilinx_vdma_tx_segment, node);
> +	desc->async_tx.phys = segment->phys;
> +	prev->hw.next_desc = segment->phys;
> +
> +	/* For the last DMA_MEM_TO_DEV transfer, set EOP */
> +	if (chan->direction == DMA_MEM_TO_DEV) {
> +		segment->hw.control_stride |= XILINX_DMA_BD_SOP;
> +		segment = list_last_entry(&desc->segments,
> +					  struct xilinx_vdma_tx_segment,
> +					  node);
> +		segment->hw.control_stride |= XILINX_DMA_BD_EOP;
> +	}
> +
> +	return &desc->async_tx;
> +
> +error:
> +	xilinx_vdma_free_tx_descriptor(chan, desc);
> +	return NULL;
> +}
> +
> +/**
>   * xilinx_vdma_terminate_all - Halt the channel and free descriptors
>   * @chan: Driver specific VDMA Channel pointer
>   */
> @@ -1179,27 +1445,36 @@ static int xilinx_vdma_chan_probe(struct
> xilinx_vdma_device *xdev, chan->id = 0;
> 
>  		chan->ctrl_offset = XILINX_VDMA_MM2S_CTRL_OFFSET;
> -		chan->desc_offset = XILINX_VDMA_MM2S_DESC_OFFSET;
> +		if (xdev->quirks & AXIVDMA_SUPPORT) {
> +			chan->desc_offset = XILINX_VDMA_MM2S_DESC_OFFSET;
> 
> -		if (xdev->flush_on_fsync == XILINX_VDMA_FLUSH_BOTH ||
> -		    xdev->flush_on_fsync == XILINX_VDMA_FLUSH_MM2S)
> -			chan->flush_on_fsync = true;
> +			if (xdev->flush_on_fsync == XILINX_VDMA_FLUSH_BOTH ||
> +			    xdev->flush_on_fsync == XILINX_VDMA_FLUSH_MM2S)
> +				chan->flush_on_fsync = true;
> +		}
>  	} else if (of_device_is_compatible(node,
>  					    "xlnx,axi-vdma-s2mm-channel")) {
>  		chan->direction = DMA_DEV_TO_MEM;
>  		chan->id = 1;
> 
>  		chan->ctrl_offset = XILINX_VDMA_S2MM_CTRL_OFFSET;
> -		chan->desc_offset = XILINX_VDMA_S2MM_DESC_OFFSET;
> +		if (xdev->quirks & AXIVDMA_SUPPORT) {
> +			chan->desc_offset = XILINX_VDMA_S2MM_DESC_OFFSET;
> 
> -		if (xdev->flush_on_fsync == XILINX_VDMA_FLUSH_BOTH ||
> -		    xdev->flush_on_fsync == XILINX_VDMA_FLUSH_S2MM)
> -			chan->flush_on_fsync = true;
> +			if (xdev->flush_on_fsync == XILINX_VDMA_FLUSH_BOTH ||
> +			    xdev->flush_on_fsync == XILINX_VDMA_FLUSH_S2MM)
> +				chan->flush_on_fsync = true;
> +		}
>  	} else {
>  		dev_err(xdev->dev, "Invalid channel compatible node\n");
>  		return -EINVAL;
>  	}
> 
> +	if (xdev->quirks & AXIVDMA_SUPPORT)
> +		chan->start_transfer = xilinx_vdma_start_transfer;
> +	else
> +		chan->start_transfer = xilinx_dma_start_transfer;
> +
>  	/* Request the interrupt */
>  	chan->irq = irq_of_parse_and_map(node, 0);
>  	err = request_irq(chan->irq, xilinx_vdma_irq_handler, IRQF_SHARED,
> @@ -1255,8 +1530,13 @@ static const struct xdma_platform_data xvdma_def = {
>  	.quirks = AXIVDMA_SUPPORT,
>  };
> 
> +static const struct xdma_platform_data xdma_def = {
> +	.quirks = AXIDMA_SUPPORT,
> +};
> +
>  static const struct of_device_id xilinx_vdma_of_ids[] = {
>  	{ .compatible = "xlnx,axi-vdma-1.00.a", .data = &xvdma_def},
> +	{ .compatible = "xlnx,axi-dma-1.00.a", .data = &xdma_def},

Please keep the entries alphabetically sorted.

>  	{}
>  };
>  MODULE_DEVICE_TABLE(of, xilinx_vdma_of_ids);
> @@ -1300,16 +1580,22 @@ static int xilinx_vdma_probe(struct platform_device
> *pdev) /* Retrieve the DMA engine properties from the device tree */
>  	xdev->has_sg = of_property_read_bool(node, "xlnx,include-sg");
> 
> -	err = of_property_read_u32(node, "xlnx,num-fstores", &num_frames);
> -	if (err < 0) {
> -		dev_err(xdev->dev, "missing xlnx,num-fstores property\n");
> -		return err;
> -	}
> +	if ((xdev->quirks & AXIVDMA_SUPPORT)) {

Extra parenthesis. Furthermore, as every DMA IP implements one and only one 
type, please replace the _SUPPORT bitmask macros with an enum and test for 
equality.

> 
> -	err = of_property_read_u32(node, "xlnx,flush-fsync",
> -					&xdev->flush_on_fsync);
> -	if (err < 0)
> -		dev_warn(xdev->dev, "missing xlnx,flush-fsync property\n");
> +		err = of_property_read_u32(node, "xlnx,num-fstores",
> +					   &num_frames);
> +		if (err < 0) {
> +			dev_err(xdev->dev,
> +				"missing xlnx,num-fstores property\n");
> +			return err;
> +		}
> +
> +		err = of_property_read_u32(node, "xlnx,flush-fsync",
> +						&xdev->flush_on_fsync);

Too many spaces, please align &xdev under node.

> +		if (err < 0)
> +			dev_warn(xdev->dev,
> +				 "missing xlnx,flush-fsync property\n");
> +	}
> 
>  	/* Initialize the DMA engine */
>  	xdev->common.dev = &pdev->dev;
> @@ -1322,11 +1608,20 @@ static int xilinx_vdma_probe(struct platform_device
> *pdev) xilinx_vdma_alloc_chan_resources;
>  	xdev->common.device_free_chan_resources =
>  				xilinx_vdma_free_chan_resources;
> -	xdev->common.device_prep_interleaved_dma =
> -				xilinx_vdma_dma_prep_interleaved;
>  	xdev->common.device_terminate_all = xilinx_vdma_terminate_all;
>  	xdev->common.device_tx_status = xilinx_vdma_tx_status;
> -	xdev->common.device_issue_pending = xilinx_vdma_issue_pending;
> +	if (xdev->quirks & AXIVDMA_SUPPORT) {
> +		xdev->common.device_issue_pending = xilinx_vdma_issue_pending;
> +		xdev->common.device_prep_interleaved_dma =
> +				xilinx_vdma_dma_prep_interleaved;

Shouldn't the VDMA also set directions and residue granularity ? Please add 
that as an additional patch before this one.

> +	} else {
> +		xdev->common.device_prep_slave_sg = xilinx_dma_prep_slave_sg;
> +		xdev->common.device_issue_pending = xilinx_dma_issue_pending;

I would use the same initialization order in both branches of the if, it will 
make the code easier to read.

> +		xdev->common.directions = BIT(DMA_DEV_TO_MEM) |
> +					  BIT(DMA_MEM_TO_DEV);

Shouldn't that depend on which channels have been synthesized (and thus 
declared in DT) ? The code to set the directions field can probably be made 
for all three DMA engine types.

> +		xdev->common.residue_granularity =
> +					  DMA_RESIDUE_GRANULARITY_SEGMENT;
> +	}
> 
>  	platform_set_drvdata(pdev, xdev);
> 
> @@ -1337,9 +1632,11 @@ static int xilinx_vdma_probe(struct platform_device
> *pdev) goto error;
>  	}
> 
> -	for (i = 0; i < XILINX_VDMA_MAX_CHANS_PER_DEVICE; i++)
> -		if (xdev->chan[i])
> -			xdev->chan[i]->num_frms = num_frames;
> +	if (xdev->quirks & AXIVDMA_SUPPORT) {
> +		for (i = 0; i < XILINX_VDMA_MAX_CHANS_PER_DEVICE; i++)
> +			if (xdev->chan[i])
> +				xdev->chan[i]->num_frms = num_frames;
> +	}
> 
>  	/* Register the DMA engine with the core */
>  	dma_async_device_register(&xdev->common);

-- 
Regards,

Laurent Pinchart
: