[PATCH V5 XDMA 0/2] xilinx XDMA driver

[PATCH V5 XDMA 0/2] xilinx XDMA driver

[Lizhi Hou]

Hello,

This V5 of patch series is to provide the platform driver to support the
Xilinx XDMA subsystem. The XDMA subsystem is used in conjunction with the
PCI Express IP block to provide high performance data transfer between host
memory and the card's DMA subsystem. It also provides up to 16 user
interrupt wires to user logic that generate interrupts to the host.

            +-------+       +-------+       +-----------+
   PCIe     |       |       |       |       |           |
   Tx/Rx    |       |       |       |  AXI  |           |
 <=======>  | PCIE  | <===> | XDMA  | <====>| User Logic|
            |       |       |       |       |           |
            +-------+       +-------+       +-----------+

The XDMA has been used for Xilinx Alveo PCIe devices.
And it is also integrated into Versal ACAP DMA and Bridge Subsystem.
    https://www.xilinx.com/products/boards-and-kits/alveo.html
    https://docs.xilinx.com/r/en-US/pg344-pcie-dma-versal/Introduction-to-the-DMA-and-Bridge-Subsystems

The device driver for any FPGA based PCIe device which leverages XDMA can
call the standard dmaengine APIs to discover and use the XDMA subsystem
without duplicating the XDMA driver code in its own driver.

Changes since v4:
- Modified user logic interrupt APIs.

Changes since v3:
- Added one patch to support user logic interrupt.

Changes since v2:
- Removed tasklet.
- Fixed regression bug introduced to V2.
- Test Robot warning.

Changes since v1:
- Moved filling hardware descriptor to xdma_prep_device_sg().
- Changed hardware descriptor enum to "struct xdma_hw_desc".
- Minor changes from code review comments.

Lizhi Hou (2):
  dmaengine: xilinx: xdma: Add xilinx xdma driver
  dmaengine: xilinx: xdma: Add user logic interrupt support

 MAINTAINERS                            |   11 +
 drivers/dma/Kconfig                    |   13 +
 drivers/dma/xilinx/Makefile            |    1 +
 drivers/dma/xilinx/xdma-regs.h         |  171 ++++
 drivers/dma/xilinx/xdma.c              | 1020 ++++++++++++++++++++++++
 include/linux/dma/amd_xdma.h           |   16 +
 include/linux/platform_data/amd_xdma.h |   34 +
 7 files changed, 1266 insertions(+)
 create mode 100644 drivers/dma/xilinx/xdma-regs.h
 create mode 100644 drivers/dma/xilinx/xdma.c
 create mode 100644 include/linux/dma/amd_xdma.h
 create mode 100644 include/linux/platform_data/amd_xdma.h

-- 
2.27.0

[PATCH V5 XDMA 1/2] dmaengine: xilinx: xdma: Add xilinx xdma driver

[Lizhi Hou]

Add driver to enable PCIe board which uses XDMA (the DMA/Bridge Subsystem
for PCI Express). For example, Xilinx Alveo PCIe devices.
    https://www.xilinx.com/products/boards-and-kits/alveo.html

The XDMA engine support up to 4 Host to Card (H2C) and 4 Card to Host (C2H)
channels. Memory transfers are specified on a per-channel basis in
descriptor linked lists, which the DMA fetches from host memory and
processes. Events such as descriptor completion and errors are signaled
using interrupts. The hardware detail is provided by
    https://docs.xilinx.com/r/en-US/pg195-pcie-dma/Introduction

This driver implements dmaengine APIs.
    - probe the available DMA channels
    - use dma_slave_map for channel lookup
    - use virtual channel to manage dmaengine tx descriptors
    - implement device_prep_slave_sg callback to handle host scatter gather
      list
    - implement device_config to config device address for DMA transfer

Signed-off-by: Lizhi Hou <lizhi.hou@amd.com>
Signed-off-by: Sonal Santan <sonal.santan@amd.com>
Signed-off-by: Max Zhen <max.zhen@amd.com>
Signed-off-by: Brian Xu <brian.xu@amd.com>
---
 MAINTAINERS                            |  10 +
 drivers/dma/Kconfig                    |  13 +
 drivers/dma/xilinx/Makefile            |   1 +
 drivers/dma/xilinx/xdma-regs.h         | 171 +++++
 drivers/dma/xilinx/xdma.c              | 955 +++++++++++++++++++++++++
 include/linux/platform_data/amd_xdma.h |  34 +
 6 files changed, 1184 insertions(+)
 create mode 100644 drivers/dma/xilinx/xdma-regs.h
 create mode 100644 drivers/dma/xilinx/xdma.c
 create mode 100644 include/linux/platform_data/amd_xdma.h

diff --git a/MAINTAINERS b/MAINTAINERS
index e8c52d0192a6..c1be0b2e378a 100644
--- a/MAINTAINERS
+++ b/MAINTAINERS
@@ -21683,6 +21683,16 @@ F:	Documentation/devicetree/bindings/media/xilinx/
 F:	drivers/media/platform/xilinx/
 F:	include/uapi/linux/xilinx-v4l2-controls.h
 
+XILINX XDMA DRIVER
+M:	Lizhi Hou <lizhi.hou@amd.com>
+M:	Brian Xu <brian.xu@amd.com>
+M:	Raj Kumar Rampelli <raj.kumar.rampelli@amd.com>
+L:	dmaengine@vger.kernel.org
+S:	Supported
+F:	drivers/dma/xilinx/xdma-regs.h
+F:	drivers/dma/xilinx/xdma.c
+F:	include/linux/platform_data/amd_xdma.h
+
 XILINX ZYNQMP DPDMA DRIVER
 M:	Hyun Kwon <hyun.kwon@xilinx.com>
 M:	Laurent Pinchart <laurent.pinchart@ideasonboard.com>
diff --git a/drivers/dma/Kconfig b/drivers/dma/Kconfig
index d5de3f77d3aa..4d90f2f51655 100644
--- a/drivers/dma/Kconfig
+++ b/drivers/dma/Kconfig
@@ -733,6 +733,19 @@ config XILINX_ZYNQMP_DPDMA
 	  driver provides the dmaengine required by the DisplayPort subsystem
 	  display driver.
 
+config XILINX_XDMA
+	tristate "Xilinx DMA/Bridge Subsystem DMA Engine"
+	select DMA_ENGINE
+	select DMA_VIRTUAL_CHANNELS
+	select REGMAP_MMIO
+	help
+	  Enable support for Xilinx DMA/Bridge Subsystem DMA engine. The DMA
+	  provides high performance block data movement between Host memory
+	  and the DMA subsystem. These direct memory transfers can be both in
+	  the Host to Card (H2C) and Card to Host (C2H) transfers.
+	  The core also provides up to 16 user interrupt wires that generate
+	  interrupts to the host.
+
 # driver files
 source "drivers/dma/bestcomm/Kconfig"
 
diff --git a/drivers/dma/xilinx/Makefile b/drivers/dma/xilinx/Makefile
index 767bb45f641f..c7a538a56643 100644
--- a/drivers/dma/xilinx/Makefile
+++ b/drivers/dma/xilinx/Makefile
@@ -2,3 +2,4 @@
 obj-$(CONFIG_XILINX_DMA) += xilinx_dma.o
 obj-$(CONFIG_XILINX_ZYNQMP_DMA) += zynqmp_dma.o
 obj-$(CONFIG_XILINX_ZYNQMP_DPDMA) += xilinx_dpdma.o
+obj-$(CONFIG_XILINX_XDMA) += xdma.o
diff --git a/drivers/dma/xilinx/xdma-regs.h b/drivers/dma/xilinx/xdma-regs.h
new file mode 100644
index 000000000000..4bcda2c52b3b
--- /dev/null
+++ b/drivers/dma/xilinx/xdma-regs.h
@@ -0,0 +1,171 @@
+/* SPDX-License-Identifier: GPL-2.0-or-later */
+/*
+ * Copyright (C) 2017-2020 Xilinx, Inc. All rights reserved.
+ * Copyright (C) 2022, Advanced Micro Devices, Inc.
+ */
+
+#ifndef __DMA_XDMA_REGS_H
+#define __DMA_XDMA_REGS_H
+
+/* The length of register space exposed to host */
+#define XDMA_REG_SPACE_LEN	65536
+
+/*
+ * maximum number of DMA channels for each direction:
+ * Host to Card (H2C) or Card to Host (C2H)
+ */
+#define XDMA_MAX_CHANNELS	4
+
+/* macros to get higher and lower 32-bit address */
+#define XDMA_HI_ADDR_MASK	GENMASK_ULL(63, 32)
+#define XDMA_LO_ADDR_MASK	GENMASK_ULL(31, 0)
+
+/*
+ * macros to define the number of descriptor blocks can be used in one
+ * DMA transfer request.
+ * the DMA engine uses a linked list of descriptor blocks that specify the
+ * source, destination, and length of the DMA transfers.
+ */
+#define XDMA_DESC_BLOCK_NUM		BIT(7)
+#define XDMA_DESC_BLOCK_MASK		(XDMA_DESC_BLOCK_NUM - 1)
+
+/* descriptor definitions */
+#define XDMA_DESC_ADJACENT		BIT(5)
+#define XDMA_DESC_ADJACENT_MASK		(XDMA_DESC_ADJACENT - 1)
+#define XDMA_DESC_MAGIC			0xad4bUL
+#define XDMA_DESC_MAGIC_SHIFT		16
+#define XDMA_DESC_ADJACENT_SHIFT	8
+#define XDMA_DESC_STOPPED		BIT(0)
+#define XDMA_DESC_COMPLETED		BIT(1)
+#define XDMA_DESC_BLEN_BITS		28
+#define XDMA_DESC_BLEN_MAX		(BIT(XDMA_DESC_BLEN_BITS) - PAGE_SIZE)
+
+/* macros to construct the descriptor control word */
+#define XDMA_DESC_CONTROL(adjacent, flag)				\
+	((XDMA_DESC_MAGIC << XDMA_DESC_MAGIC_SHIFT) |			\
+	 (((adjacent) - 1) << XDMA_DESC_ADJACENT_SHIFT) | (flag))
+#define XDMA_DESC_CONTROL_LAST						\
+	XDMA_DESC_CONTROL(1, XDMA_DESC_STOPPED | XDMA_DESC_COMPLETED)
+
+/*
+ * Descriptor for a single contiguous memory block transfer.
+ *
+ * Multiple descriptors are linked by means of the next pointer. An additional
+ * extra adjacent number gives the amount of extra contiguous descriptors.
+ *
+ * The descriptors are in root complex memory, and the bytes in the 32-bit
+ * words must be in little-endian byte ordering.
+ */
+struct xdma_hw_desc {
+	__le32		control;
+	__le32		bytes;
+	__le64		src_addr;
+	__le64		dst_addr;
+	__le64		next_desc;
+};
+
+#define XDMA_DESC_SIZE		sizeof(struct xdma_hw_desc)
+#define XDMA_DESC_BLOCK_SIZE	(XDMA_DESC_SIZE * XDMA_DESC_ADJACENT)
+#define XDMA_DESC_BLOCK_ALIGN	4096
+
+/*
+ * Channel registers
+ */
+#define XDMA_CHAN_IDENTIFIER		0x0
+#define XDMA_CHAN_CONTROL		0x4
+#define XDMA_CHAN_CONTROL_W1S		0x8
+#define XDMA_CHAN_CONTROL_W1C		0xc
+#define XDMA_CHAN_STATUS		0x40
+#define XDMA_CHAN_COMPLETED_DESC	0x48
+#define XDMA_CHAN_ALIGNMENTS		0x4c
+#define XDMA_CHAN_INTR_ENABLE		0x90
+#define XDMA_CHAN_INTR_ENABLE_W1S	0x94
+#define XDMA_CHAN_INTR_ENABLE_W1C	0x9c
+
+#define XDMA_CHAN_STRIDE	0x100
+#define XDMA_CHAN_H2C_OFFSET	0x0
+#define XDMA_CHAN_C2H_OFFSET	0x1000
+#define XDMA_CHAN_H2C_TARGET	0x0
+#define XDMA_CHAN_C2H_TARGET	0x1
+
+/* macro to check if channel is available */
+#define XDMA_CHAN_MAGIC		0x1fc0
+#define XDMA_CHAN_CHECK_TARGET(id, target)		\
+	(((u32)(id) >> 16) == XDMA_CHAN_MAGIC + (target))
+
+/* bits of the channel control register */
+#define CHAN_CTRL_RUN_STOP			BIT(0)
+#define CHAN_CTRL_IE_DESC_STOPPED		BIT(1)
+#define CHAN_CTRL_IE_DESC_COMPLETED		BIT(2)
+#define CHAN_CTRL_IE_DESC_ALIGN_MISMATCH	BIT(3)
+#define CHAN_CTRL_IE_MAGIC_STOPPED		BIT(4)
+#define CHAN_CTRL_IE_IDLE_STOPPED		BIT(6)
+#define CHAN_CTRL_IE_READ_ERROR			(0x1FUL << 9)
+#define CHAN_CTRL_IE_DESC_ERROR			(0x1FUL << 19)
+#define CHAN_CTRL_NON_INCR_ADDR			BIT(25)
+#define CHAN_CTRL_POLL_MODE_WB			BIT(26)
+
+#define CHAN_CTRL_START	(CHAN_CTRL_RUN_STOP |				\
+			 CHAN_CTRL_IE_DESC_STOPPED |			\
+			 CHAN_CTRL_IE_DESC_COMPLETED |			\
+			 CHAN_CTRL_IE_DESC_ALIGN_MISMATCH |		\
+			 CHAN_CTRL_IE_MAGIC_STOPPED |			\
+			 CHAN_CTRL_IE_READ_ERROR |			\
+			 CHAN_CTRL_IE_DESC_ERROR)
+
+/* bits of the channel interrupt enable mask */
+#define CHAN_IM_DESC_ERROR			BIT(19)
+#define CHAN_IM_READ_ERROR			BIT(9)
+#define CHAN_IM_IDLE_STOPPED			BIT(6)
+#define CHAN_IM_MAGIC_STOPPED			BIT(4)
+#define CHAN_IM_DESC_COMPLETED			BIT(2)
+#define CHAN_IM_DESC_STOPPED			BIT(1)
+
+#define CHAN_IM_ALL	(CHAN_IM_DESC_ERROR | CHAN_IM_READ_ERROR |	\
+			 CHAN_IM_IDLE_STOPPED | CHAN_IM_MAGIC_STOPPED | \
+			 CHAN_IM_DESC_COMPLETED | CHAN_IM_DESC_STOPPED)
+
+/*
+ * Channel SGDMA registers
+ */
+#define XDMA_SGDMA_IDENTIFIER	0x0
+#define XDMA_SGDMA_DESC_LO	0x80
+#define XDMA_SGDMA_DESC_HI	0x84
+#define XDMA_SGDMA_DESC_ADJ	0x88
+#define XDMA_SGDMA_DESC_CREDIT	0x8c
+
+#define XDMA_SGDMA_BASE(chan_base)	((chan_base) + 0x4000)
+
+/* bits of the SG DMA control register */
+#define XDMA_CTRL_RUN_STOP			BIT(0)
+#define XDMA_CTRL_IE_DESC_STOPPED		BIT(1)
+#define XDMA_CTRL_IE_DESC_COMPLETED		BIT(2)
+#define XDMA_CTRL_IE_DESC_ALIGN_MISMATCH	BIT(3)
+#define XDMA_CTRL_IE_MAGIC_STOPPED		BIT(4)
+#define XDMA_CTRL_IE_IDLE_STOPPED		BIT(6)
+#define XDMA_CTRL_IE_READ_ERROR			(0x1FUL << 9)
+#define XDMA_CTRL_IE_DESC_ERROR			(0x1FUL << 19)
+#define XDMA_CTRL_NON_INCR_ADDR			BIT(25)
+#define XDMA_CTRL_POLL_MODE_WB			BIT(26)
+
+/*
+ * interrupt registers
+ */
+#define XDMA_IRQ_IDENTIFIER		0x0
+#define XDMA_IRQ_USER_INT_EN		0x04
+#define XDMA_IRQ_USER_INT_EN_W1S	0x08
+#define XDMA_IRQ_USER_INT_EN_W1C	0x0c
+#define XDMA_IRQ_CHAN_INT_EN		0x10
+#define XDMA_IRQ_CHAN_INT_EN_W1S	0x14
+#define XDMA_IRQ_CHAN_INT_EN_W1C	0x18
+#define XDMA_IRQ_USER_INT_REQ		0x40
+#define XDMA_IRQ_CHAN_INT_REQ		0x44
+#define XDMA_IRQ_USER_INT_PEND		0x48
+#define XDMA_IRQ_CHAN_INT_PEND		0x4c
+#define XDMA_IRQ_USER_VEC_NUM		0x80
+#define XDMA_IRQ_CHAN_VEC_NUM		0xa0
+
+#define XDMA_IRQ_BASE			0x2000
+#define XDMA_IRQ_VEC_SHIFT		8
+
+#endif /* __DMA_XDMA_REGS_H */
diff --git a/drivers/dma/xilinx/xdma.c b/drivers/dma/xilinx/xdma.c
new file mode 100644
index 000000000000..58a57e03bef5
--- /dev/null
+++ b/drivers/dma/xilinx/xdma.c
@@ -0,0 +1,955 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ * DMA driver for Xilinx DMA/Bridge Subsystem
+ *
+ * Copyright (C) 2017-2020 Xilinx, Inc. All rights reserved.
+ * Copyright (C) 2022, Advanced Micro Devices, Inc.
+ */
+
+/*
+ * The DMA/Bridge Subsystem for PCI Express allows for the movement of data
+ * between Host memory and the DMA subsystem. It does this by operating on
+ * 'descriptors' that contain information about the source, destination and
+ * amount of data to transfer. These direct memory transfers can be both in
+ * the Host to Card (H2C) and Card to Host (C2H) transfers. The DMA can be
+ * configured to have a single AXI4 Master interface shared by all channels
+ * or one AXI4-Stream interface for each channel enabled. Memory transfers are
+ * specified on a per-channel basis in descriptor linked lists, which the DMA
+ * fetches from host memory and processes. Events such as descriptor completion
+ * and errors are signaled using interrupts. The core also provides up to 16
+ * user interrupt wires that generate interrupts to the host.
+ */
+
+#include <linux/mod_devicetable.h>
+#include <linux/bitfield.h>
+#include <linux/dmapool.h>
+#include <linux/regmap.h>
+#include <linux/dmaengine.h>
+#include <linux/platform_device.h>
+#include <linux/platform_data/amd_xdma.h>
+#include <linux/dma-mapping.h>
+#include "../virt-dma.h"
+#include "xdma-regs.h"
+
+/* mmio regmap config for all XDMA registers */
+static const struct regmap_config xdma_regmap_config = {
+	.reg_bits = 32,
+	.val_bits = 32,
+	.reg_stride = 4,
+	.max_register = XDMA_REG_SPACE_LEN,
+};
+
+/**
+ * struct xdma_desc_block - Descriptor block
+ * @virt_addr: Virtual address of block start
+ * @dma_addr: DMA address of block start
+ */
+struct xdma_desc_block {
+	void		*virt_addr;
+	dma_addr_t	dma_addr;
+};
+
+/**
+ * struct xdma_chan - Driver specific DMA channel structure
+ * @vchan: Virtual channel
+ * @xdev_hdl: Pointer to DMA device structure
+ * @base: Offset of channel registers
+ * @desc_pool: Descriptor pool
+ * @busy: Busy flag of the channel
+ * @dir: Transferring direction of the channel
+ * @cfg: Transferring config of the channel
+ * @irq: IRQ assigned to the channel
+ */
+struct xdma_chan {
+	struct virt_dma_chan		vchan;
+	void				*xdev_hdl;
+	u32				base;
+	struct dma_pool			*desc_pool;
+	bool				busy;
+	enum dma_transfer_direction	dir;
+	struct dma_slave_config		cfg;
+	u32				irq;
+};
+
+/**
+ * struct xdma_desc - DMA desc structure
+ * @vdesc: Virtual DMA descriptor
+ * @chan: DMA channel pointer
+ * @dir: Transferring direction of the request
+ * @dev_addr: Physical address on DMA device side
+ * @desc_blocks: Hardware descriptor blocks
+ * @dblk_num: Number of hardware descriptor blocks
+ * @desc_num: Number of hardware descriptors
+ * @completed_desc_num: Completed hardware descriptors
+ */
+struct xdma_desc {
+	struct virt_dma_desc		vdesc;
+	struct xdma_chan		*chan;
+	enum dma_transfer_direction	dir;
+	u64				dev_addr;
+	struct xdma_desc_block		*desc_blocks;
+	u32				dblk_num;
+	u32				desc_num;
+	u32				completed_desc_num;
+};
+
+#define XDMA_DEV_STATUS_REG_DMA		BIT(0)
+#define XDMA_DEV_STATUS_INIT_MSIX	BIT(1)
+
+/**
+ * struct xdma_device - DMA device structure
+ * @pdev: Platform device pointer
+ * @dma_dev: DMA device structure
+ * @regmap: MMIO regmap for DMA registers
+ * @h2c_chans: Host to Card channels
+ * @c2h_chans: Card to Host channels
+ * @h2c_chan_num: Number of H2C channels
+ * @c2h_chan_num: Number of C2H channels
+ * @irq_start: Start IRQ assigned to device
+ * @irq_num: Number of IRQ assigned to device
+ * @status: Initialization status
+ */
+struct xdma_device {
+	struct platform_device	*pdev;
+	struct dma_device	dma_dev;
+	struct regmap		*regmap;
+	struct xdma_chan	*h2c_chans;
+	struct xdma_chan	*c2h_chans;
+	u32			h2c_chan_num;
+	u32			c2h_chan_num;
+	u32			irq_start;
+	u32			irq_num;
+	u32			status;
+};
+
+#define xdma_err(xdev, fmt, args...)					\
+	dev_err(&(xdev)->pdev->dev, fmt, ##args)
+
+/* Read and Write DMA registers */
+static inline int xdma_read_reg(struct xdma_device *xdev, u32 base, u32 reg,
+				u32 *val)
+{
+	return regmap_read(xdev->regmap, base + reg, val);
+}
+
+static inline int xdma_write_reg(struct xdma_device *xdev, u32 base, u32 reg,
+				 u32 val)
+{
+	return regmap_write(xdev->regmap, base + reg, val);
+}
+
+/* Get the last desc in a desc block */
+static inline void *xdma_blk_last_desc(struct xdma_desc_block *block)
+{
+	return block->virt_addr + (XDMA_DESC_ADJACENT - 1) * XDMA_DESC_SIZE;
+}
+
+/**
+ * xdma_link_desc_blocks - Link descriptor blocks for DMA transfer
+ * @sw_desc: Tx descriptor pointer
+ */
+static void xdma_link_desc_blocks(struct xdma_desc *sw_desc)
+{
+	struct xdma_desc_block *block;
+	u32 last_blk_desc_num, desc_control;
+	struct xdma_hw_desc *desc;
+	int i;
+
+	desc_control = XDMA_DESC_CONTROL(XDMA_DESC_ADJACENT, 0);
+	for (i = 1; i < sw_desc->dblk_num; i++) {
+		block = &sw_desc->desc_blocks[i - 1];
+		desc = xdma_blk_last_desc(block);
+
+		if (!(i & XDMA_DESC_BLOCK_MASK)) {
+			desc->control = cpu_to_le32(XDMA_DESC_CONTROL_LAST);
+			continue;
+		}
+		desc->control = cpu_to_le32(desc_control);
+		desc->next_desc = cpu_to_le64(block[1].dma_addr);
+	}
+
+	/* update the last block */
+	last_blk_desc_num =
+		((sw_desc->desc_num - 1) & XDMA_DESC_ADJACENT_MASK) + 1;
+	if ((sw_desc->dblk_num & XDMA_DESC_BLOCK_MASK) > 1) {
+		block = &sw_desc->desc_blocks[sw_desc->dblk_num - 2];
+		desc = xdma_blk_last_desc(block);
+		desc_control = XDMA_DESC_CONTROL(last_blk_desc_num, 0);
+		desc->control = cpu_to_le32(desc_control);
+	}
+
+	block = &sw_desc->desc_blocks[sw_desc->dblk_num - 1];
+	desc = block->virt_addr + (last_blk_desc_num - 1) * XDMA_DESC_SIZE;
+	desc->control = cpu_to_le32(XDMA_DESC_CONTROL_LAST);
+}
+
+static inline struct xdma_chan *to_xdma_chan(struct dma_chan *chan)
+{
+	return container_of(chan, struct xdma_chan, vchan.chan);
+}
+
+static inline struct xdma_desc *to_xdma_desc(struct virt_dma_desc *vdesc)
+{
+	return container_of(vdesc, struct xdma_desc, vdesc);
+}
+
+static int xdma_enable_intr(struct xdma_device *xdev)
+{
+	int ret;
+
+	ret = xdma_write_reg(xdev, XDMA_IRQ_BASE, XDMA_IRQ_CHAN_INT_EN_W1S, ~0);
+	if (ret)
+		xdma_err(xdev, "enable channel intr failed: %d", ret);
+
+	return ret;
+}
+
+static int xdma_disable_intr(struct xdma_device *xdev)
+{
+	int ret;
+
+	ret = xdma_write_reg(xdev, XDMA_IRQ_BASE, XDMA_IRQ_CHAN_INT_EN_W1C, ~0);
+	if (ret)
+		xdma_err(xdev, "disable channel intr failed: %d", ret);
+
+	return ret;
+}
+
+/**
+ * xdma_channel_init - Initialize DMA channel registers
+ * @chan: DMA channel pointer
+ */
+static int xdma_channel_init(struct xdma_chan *chan)
+{
+	struct xdma_device *xdev = chan->xdev_hdl;
+	int ret;
+
+	ret = xdma_write_reg(xdev, chan->base, XDMA_CHAN_CONTROL_W1C,
+			     CHAN_CTRL_NON_INCR_ADDR);
+	if (ret) {
+		xdma_err(xdev, "clear non incr addr failed: %d", ret);
+		return ret;
+	}
+
+	ret = xdma_write_reg(xdev, chan->base, XDMA_CHAN_INTR_ENABLE,
+			     CHAN_IM_ALL);
+	if (ret) {
+		xdma_err(xdev, "failed to set interrupt mask: %d", ret);
+		return ret;
+	}
+
+	return 0;
+}
+
+/**
+ * xdma_free_desc - Free descriptor
+ * @vdesc: Virtual DMA descriptor
+ */
+static void xdma_free_desc(struct virt_dma_desc *vdesc)
+{
+	struct xdma_desc *sw_desc;
+	int i;
+
+	sw_desc = to_xdma_desc(vdesc);
+	for (i = 0; i < sw_desc->dblk_num; i++) {
+		if (!sw_desc->desc_blocks[i].virt_addr)
+			break;
+		dma_pool_free(sw_desc->chan->desc_pool,
+			      sw_desc->desc_blocks[i].virt_addr,
+			      sw_desc->desc_blocks[i].dma_addr);
+	}
+	kfree(sw_desc->desc_blocks);
+	kfree(sw_desc);
+}
+
+/**
+ * xdma_alloc_desc - Allocate descriptor
+ * @chan: DMA channel pointer
+ * @desc_num: Number of hardware descriptors
+ */
+static struct xdma_desc *
+xdma_alloc_desc(struct xdma_chan *chan, u32 desc_num)
+{
+	struct xdma_desc *sw_desc;
+	struct xdma_hw_desc *desc;
+	dma_addr_t dma_addr;
+	u32 dblk_num;
+	void *addr;
+	int i, j;
+
+	sw_desc = kzalloc(sizeof(*sw_desc), GFP_NOWAIT);
+	if (!sw_desc)
+		return NULL;
+
+	sw_desc->chan = chan;
+	sw_desc->desc_num = desc_num;
+	dblk_num = DIV_ROUND_UP(desc_num, XDMA_DESC_ADJACENT);
+	sw_desc->desc_blocks = kcalloc(dblk_num, sizeof(*sw_desc->desc_blocks),
+				       GFP_NOWAIT);
+	if (!sw_desc->desc_blocks)
+		goto failed;
+
+	sw_desc->dblk_num = dblk_num;
+	for (i = 0; i < sw_desc->dblk_num; i++) {
+		addr = dma_pool_alloc(chan->desc_pool, GFP_NOWAIT, &dma_addr);
+		if (!addr)
+			goto failed;
+
+		sw_desc->desc_blocks[i].virt_addr = addr;
+		sw_desc->desc_blocks[i].dma_addr = dma_addr;
+		for (j = 0, desc = addr; j < XDMA_DESC_ADJACENT; j++)
+			desc[j].control = cpu_to_le32(XDMA_DESC_CONTROL(1, 0));
+	}
+
+	xdma_link_desc_blocks(sw_desc);
+
+	return sw_desc;
+
+failed:
+	xdma_free_desc(&sw_desc->vdesc);
+	return NULL;
+}
+
+/**
+ * xdma_xfer_start - Start DMA transfer
+ * @xdma_chan: DMA channel pointer
+ */
+static int xdma_xfer_start(struct xdma_chan *xdma_chan)
+{
+	struct virt_dma_desc *vd = vchan_next_desc(&xdma_chan->vchan);
+	struct xdma_device *xdev = xdma_chan->xdev_hdl;
+	struct xdma_desc_block *block;
+	u32 val, completed_blocks;
+	struct xdma_desc *desc;
+	int ret;
+
+	/*
+	 * check if there is not any submitted descriptor or channel is busy.
+	 * vchan lock should be held where this function is called.
+	 */
+	if (!vd || xdma_chan->busy)
+		return -EINVAL;
+
+	/* clear run stop bit to get ready for transfer */
+	ret = xdma_write_reg(xdev, xdma_chan->base, XDMA_CHAN_CONTROL_W1C,
+			     CHAN_CTRL_RUN_STOP);
+	if (ret) {
+		xdma_err(xdev, "write control failed: %d", ret);
+		return ret;
+	}
+
+	desc = to_xdma_desc(vd);
+	if (desc->dir != xdma_chan->dir) {
+		xdma_err(xdev, "incorrect request direction");
+		return -EINVAL;
+	}
+
+	/* set DMA engine to the first descriptor block */
+	completed_blocks = desc->completed_desc_num / XDMA_DESC_ADJACENT;
+	block = &desc->desc_blocks[completed_blocks];
+	val = FIELD_GET(XDMA_LO_ADDR_MASK, block->dma_addr);
+	ret = xdma_write_reg(xdev, XDMA_SGDMA_BASE(xdma_chan->base),
+			     XDMA_SGDMA_DESC_LO, val);
+	if (ret) {
+		xdma_err(xdev, "write hi addr failed: %d", ret);
+		return ret;
+	}
+
+	val = FIELD_GET(XDMA_HI_ADDR_MASK, (u64)block->dma_addr);
+	ret = xdma_write_reg(xdev, XDMA_SGDMA_BASE(xdma_chan->base),
+			     XDMA_SGDMA_DESC_HI, val);
+	if (ret) {
+		xdma_err(xdev, "write lo addr failed: %d", ret);
+		return ret;
+	}
+
+	if (completed_blocks + 1 == desc->dblk_num)
+		val = (desc->desc_num - 1) & XDMA_DESC_ADJACENT_MASK;
+	else
+		val = XDMA_DESC_ADJACENT - 1;
+	ret = xdma_write_reg(xdev, XDMA_SGDMA_BASE(xdma_chan->base),
+			     XDMA_SGDMA_DESC_ADJ, val);
+	if (ret) {
+		xdma_err(xdev, "write adjacent failed: %d", ret);
+		return ret;
+	}
+
+	/* kick off DMA transfer */
+	ret = xdma_write_reg(xdev, xdma_chan->base, XDMA_CHAN_CONTROL,
+			     CHAN_CTRL_START);
+	if (ret) {
+		xdma_err(xdev, "write control failed: %d", ret);
+		return ret;
+	}
+
+	xdma_chan->busy = true;
+	return 0;
+}
+
+/**
+ * xdma_config_channels - Detect and config DMA channels
+ * @xdev: DMA device pointer
+ * @dir: Channel direction
+ */
+static int xdma_config_channels(struct xdma_device *xdev,
+				enum dma_transfer_direction dir)
+{
+	struct xdma_platdata *pdata = dev_get_platdata(&xdev->pdev->dev);
+	u32 base, identifier, target;
+	struct xdma_chan **chans;
+	u32 *chan_num;
+	int i, j, ret;
+
+	if (dir == DMA_MEM_TO_DEV) {
+		base = XDMA_CHAN_H2C_OFFSET;
+		target = XDMA_CHAN_H2C_TARGET;
+		chans = &xdev->h2c_chans;
+		chan_num = &xdev->h2c_chan_num;
+	} else if (dir == DMA_DEV_TO_MEM) {
+		base = XDMA_CHAN_C2H_OFFSET;
+		target = XDMA_CHAN_C2H_TARGET;
+		chans = &xdev->c2h_chans;
+		chan_num = &xdev->c2h_chan_num;
+	} else {
+		xdma_err(xdev, "invalid direction specified");
+		return -EINVAL;
+	}
+
+	/* detect number of available DMA channels */
+	for (i = 0, *chan_num = 0; i < pdata->max_dma_channels; i++) {
+		ret = xdma_read_reg(xdev, base + i * XDMA_CHAN_STRIDE,
+				    XDMA_CHAN_IDENTIFIER, &identifier);
+		if (ret) {
+			xdma_err(xdev, "failed to read channel id: %d", ret);
+			return ret;
+		}
+
+		/* check if it is available DMA channel */
+		if (XDMA_CHAN_CHECK_TARGET(identifier, target))
+			(*chan_num)++;
+	}
+
+	if (!*chan_num) {
+		xdma_err(xdev, "does not probe any channel");
+		return -EINVAL;
+	}
+
+	*chans = devm_kzalloc(&xdev->pdev->dev, sizeof(**chans) * (*chan_num),
+			      GFP_KERNEL);
+	if (!*chans)
+		return -ENOMEM;
+
+	for (i = 0, j = 0; i < pdata->max_dma_channels; i++) {
+		ret = xdma_read_reg(xdev, base + i * XDMA_CHAN_STRIDE,
+				    XDMA_CHAN_IDENTIFIER, &identifier);
+		if (ret) {
+			xdma_err(xdev, "failed to read channel id: %d", ret);
+			return ret;
+		}
+
+		if (!XDMA_CHAN_CHECK_TARGET(identifier, target))
+			continue;
+
+		if (j == *chan_num) {
+			xdma_err(xdev, "invalid channel number");
+			return -EIO;
+		}
+
+		/* init channel structure and hardware */
+		(*chans)[j].xdev_hdl = xdev;
+		(*chans)[j].base = base + i * XDMA_CHAN_STRIDE;
+		(*chans)[j].dir = dir;
+
+		ret = xdma_channel_init(&(*chans)[j]);
+		if (ret)
+			return ret;
+		(*chans)[j].vchan.desc_free = xdma_free_desc;
+		vchan_init(&(*chans)[j].vchan, &xdev->dma_dev);
+
+		j++;
+	}
+
+	dev_info(&xdev->pdev->dev, "configured %d %s channels", j,
+		 (dir == DMA_MEM_TO_DEV) ? "H2C" : "C2H");
+
+	return 0;
+}
+
+/**
+ * xdma_issue_pending - Issue pending transactions
+ * @chan: DMA channel pointer
+ */
+static void xdma_issue_pending(struct dma_chan *chan)
+{
+	struct xdma_chan *xdma_chan = to_xdma_chan(chan);
+	unsigned long flags;
+
+	spin_lock_irqsave(&xdma_chan->vchan.lock, flags);
+	if (vchan_issue_pending(&xdma_chan->vchan))
+		xdma_xfer_start(xdma_chan);
+	spin_unlock_irqrestore(&xdma_chan->vchan.lock, flags);
+}
+
+/**
+ * xdma_prep_device_sg - prepare a descriptor for a
+ *	DMA transaction
+ * @chan: DMA channel pointer
+ * @sgl: Transfer scatter gather list
+ * @sg_len: Length of scatter gather list
+ * @dir: Transfer direction
+ * @flags: transfer ack flags
+ * @context: APP words of the descriptor
+ */
+static struct dma_async_tx_descriptor *
+xdma_prep_device_sg(struct dma_chan *chan, struct scatterlist *sgl,
+		    unsigned int sg_len, enum dma_transfer_direction dir,
+		    unsigned long flags, void *context)
+{
+	struct xdma_chan *xdma_chan = to_xdma_chan(chan);
+	struct dma_async_tx_descriptor *tx_desc;
+	u32 desc_num = 0, i, len, rest;
+	struct xdma_desc_block *dblk;
+	struct xdma_hw_desc *desc;
+	struct xdma_desc *sw_desc;
+	u64 dev_addr, *src, *dst;
+	struct scatterlist *sg;
+	u64 addr;
+
+	for_each_sg(sgl, sg, sg_len, i)
+		desc_num += DIV_ROUND_UP(sg_dma_len(sg), XDMA_DESC_BLEN_MAX);
+
+	sw_desc = xdma_alloc_desc(xdma_chan, desc_num);
+	if (!sw_desc)
+		return NULL;
+	sw_desc->dir = dir;
+
+	if (dir == DMA_MEM_TO_DEV) {
+		dev_addr = xdma_chan->cfg.dst_addr;
+		src = &addr;
+		dst = &dev_addr;
+	} else {
+		dev_addr = xdma_chan->cfg.src_addr;
+		src = &dev_addr;
+		dst = &addr;
+	}
+
+	dblk = sw_desc->desc_blocks;
+	desc = dblk->virt_addr;
+	desc_num = 1;
+	for_each_sg(sgl, sg, sg_len, i) {
+		addr = sg_dma_address(sg);
+		rest = sg_dma_len(sg);
+
+		do {
+			len = min_t(u32, rest, XDMA_DESC_BLEN_MAX);
+			/* set hardware descriptor */
+			desc->bytes = cpu_to_le32(len);
+			desc->src_addr = cpu_to_le64(*src);
+			desc->dst_addr = cpu_to_le64(*dst);
+
+			if (!(desc_num & XDMA_DESC_ADJACENT_MASK)) {
+				dblk++;
+				desc = dblk->virt_addr;
+			} else {
+				desc++;
+			}
+
+			desc_num++;
+			dev_addr += len;
+			addr += len;
+			rest -= len;
+		} while (rest);
+	}
+
+	tx_desc = vchan_tx_prep(&xdma_chan->vchan, &sw_desc->vdesc, flags);
+	if (!tx_desc)
+		goto failed;
+
+	return tx_desc;
+
+failed:
+	xdma_free_desc(&sw_desc->vdesc);
+
+	return NULL;
+}
+
+/**
+ * xdma_device_config - Configure the DMA channel
+ * @chan: DMA channel
+ * @cfg: channel configuration
+ */
+static int xdma_device_config(struct dma_chan *chan,
+			      struct dma_slave_config *cfg)
+{
+	struct xdma_chan *xdma_chan = to_xdma_chan(chan);
+
+	memcpy(&xdma_chan->cfg, cfg, sizeof(*cfg));
+
+	return 0;
+}
+
+/**
+ * xdma_free_chan_resources - Free channel resources
+ * @chan: DMA channel
+ */
+static void xdma_free_chan_resources(struct dma_chan *chan)
+{
+	struct xdma_chan *xdma_chan = to_xdma_chan(chan);
+
+	vchan_free_chan_resources(&xdma_chan->vchan);
+	dma_pool_destroy(xdma_chan->desc_pool);
+	xdma_chan->desc_pool = NULL;
+}
+
+/**
+ * xdma_alloc_chan_resources - Allocate channel resources
+ * @chan: DMA channel
+ */
+static int xdma_alloc_chan_resources(struct dma_chan *chan)
+{
+	struct xdma_chan *xdma_chan = to_xdma_chan(chan);
+	struct xdma_device *xdev = xdma_chan->xdev_hdl;
+
+	xdma_chan->desc_pool = dma_pool_create(dma_chan_name(chan),
+					       xdev->dma_dev.dev,
+					       XDMA_DESC_BLOCK_SIZE,
+					       XDMA_DESC_BLOCK_ALIGN,
+					       0);
+	if (!xdma_chan->desc_pool) {
+		xdma_err(xdev, "unable to allocate descriptor pool");
+		return -ENOMEM;
+	}
+
+	return 0;
+}
+
+/**
+ * xdma_channel_isr - XDMA channel interrupt handler
+ * @irq: IRQ number
+ * @dev_id: Pointer to the DMA channel structure
+ */
+static irqreturn_t xdma_channel_isr(int irq, void *dev_id)
+{
+	struct xdma_chan *xdma_chan = dev_id;
+	u32 complete_desc_num = 0;
+	struct virt_dma_desc *vd;
+	struct xdma_desc *desc;
+	int ret;
+
+	spin_lock(&xdma_chan->vchan.lock);
+
+	/* get submitted request */
+	vd = vchan_next_desc(&xdma_chan->vchan);
+	if (!vd)
+		goto out;
+
+	xdma_chan->busy = false;
+	desc = to_xdma_desc(vd);
+
+	ret = xdma_read_reg(xdma_chan->xdev_hdl, xdma_chan->base,
+			    XDMA_CHAN_COMPLETED_DESC, &complete_desc_num);
+	if (ret)
+		goto out;
+
+	desc->completed_desc_num += complete_desc_num;
+	/*
+	 * if all data blocks are transferred, remove and complete the request
+	 */
+	if (desc->completed_desc_num == desc->desc_num) {
+		list_del(&vd->node);
+		vchan_cookie_complete(vd);
+		goto out;
+	}
+
+	if (desc->completed_desc_num > desc->desc_num ||
+	    complete_desc_num != XDMA_DESC_BLOCK_NUM * XDMA_DESC_ADJACENT)
+		goto out;
+
+	/* transfer the rest of data */
+	xdma_xfer_start(xdma_chan);
+
+out:
+	spin_unlock(&xdma_chan->vchan.lock);
+	return IRQ_HANDLED;
+}
+
+/**
+ * xdma_irq_fini - Uninitialize IRQ
+ * @xdev: DMA device pointer
+ */
+static void xdma_irq_fini(struct xdma_device *xdev)
+{
+	int ret, i;
+
+	/* disable interrupt */
+	ret = xdma_disable_intr(xdev);
+	if (ret)
+		xdma_err(xdev, "failed to disable interrupts: %d", ret);
+
+	/* free irq handler */
+	for (i = 0; i < xdev->h2c_chan_num; i++)
+		free_irq(xdev->h2c_chans[i].irq, &xdev->h2c_chans[i]);
+
+	for (i = 0; i < xdev->c2h_chan_num; i++)
+		free_irq(xdev->c2h_chans[i].irq, &xdev->c2h_chans[i]);
+}
+
+/**
+ * xdma_set_vector_reg - configure hardware IRQ registers
+ * @xdev: DMA device pointer
+ * @vec_tbl_start: Start of IRQ registers
+ * @irq_start: Start of IRQ
+ * @irq_num: Number of IRQ
+ */
+static int xdma_set_vector_reg(struct xdma_device *xdev, u32 vec_tbl_start,
+			       u32 irq_start, u32 irq_num)
+{
+	u32 shift, i, val = 0;
+	int ret;
+
+	/* Each IRQ register is 32 bit and contains 4 IRQs */
+	while (irq_num > 0) {
+		for (i = 0; i < 4; i++) {
+			shift = XDMA_IRQ_VEC_SHIFT * i;
+			val |= irq_start << shift;
+			irq_start++;
+			irq_num--;
+		}
+
+		/* write IRQ register */
+		ret = xdma_write_reg(xdev, XDMA_IRQ_BASE, vec_tbl_start, val);
+		if (ret) {
+			xdma_err(xdev, "failed to set vector: %d", ret);
+			return ret;
+		}
+		vec_tbl_start += sizeof(u32);
+		val = 0;
+	}
+
+	return 0;
+}
+
+/**
+ * xdma_irq_init - initialize IRQs
+ * @xdev: DMA device pointer
+ */
+static int xdma_irq_init(struct xdma_device *xdev)
+{
+	u32 irq = xdev->irq_start;
+	int i, j, ret;
+
+	/* return failure if there are not enough IRQs */
+	if (xdev->irq_num < xdev->h2c_chan_num + xdev->c2h_chan_num) {
+		xdma_err(xdev, "not enough irq");
+		return -EINVAL;
+	}
+
+	/* setup H2C interrupt handler */
+	for (i = 0; i < xdev->h2c_chan_num; i++) {
+		ret = request_irq(irq, xdma_channel_isr, 0,
+				  "xdma-h2c-channel", &xdev->h2c_chans[i]);
+		if (ret) {
+			xdma_err(xdev, "H2C channel%d request irq%d failed: %d",
+				 i, irq, ret);
+			for (j = 0; j < i; j++) {
+				free_irq(xdev->h2c_chans[j].irq,
+					 &xdev->h2c_chans[j]);
+			}
+			return ret;
+		}
+		xdev->h2c_chans[i].irq = irq;
+		irq++;
+	}
+
+	/* setup C2H interrupt handler */
+	for (i = 0; i < xdev->c2h_chan_num; i++) {
+		ret = request_irq(irq, xdma_channel_isr, 0,
+				  "xdma-c2h-channel", &xdev->c2h_chans[i]);
+		if (ret) {
+			xdma_err(xdev, "H2C channel%d request irq%d failed: %d",
+				 i, irq, ret);
+			for (j = 0; j < i; j++) {
+				free_irq(xdev->c2h_chans[j].irq,
+					 &xdev->c2h_chans[j]);
+			}
+			goto failed_init_c2h;
+		}
+		xdev->c2h_chans[i].irq = irq;
+		irq++;
+	}
+
+	/* config hardware IRQ registers */
+	ret = xdma_set_vector_reg(xdev, XDMA_IRQ_CHAN_VEC_NUM, 0,
+				  xdev->h2c_chan_num + xdev->c2h_chan_num);
+	if (ret) {
+		xdma_err(xdev, "failed to set channel vectors: %d", ret);
+		goto failed;
+	}
+
+	/* enable interrupt */
+	ret = xdma_enable_intr(xdev);
+	if (ret) {
+		xdma_err(xdev, "failed to enable interrupts: %d", ret);
+		goto failed;
+	}
+
+	return 0;
+
+failed:
+	for (i = 0; i < xdev->c2h_chan_num; i++)
+		free_irq(xdev->c2h_chans[i].irq, &xdev->c2h_chans[i]);
+failed_init_c2h:
+	for (i = 0; i < xdev->h2c_chan_num; i++)
+		free_irq(xdev->h2c_chans[i].irq, &xdev->h2c_chans[i]);
+
+	return ret;
+}
+
+static bool xdma_filter_fn(struct dma_chan *chan, void *param)
+{
+	struct xdma_chan *xdma_chan = to_xdma_chan(chan);
+	struct xdma_chan_info *chan_info = param;
+
+	if (chan_info->dir != xdma_chan->dir)
+		return false;
+
+	return true;
+}
+
+/**
+ * xdma_remove - Driver remove function
+ * @pdev: Pointer to the platform_device structure
+ */
+static int xdma_remove(struct platform_device *pdev)
+{
+	struct xdma_device *xdev = platform_get_drvdata(pdev);
+
+	if (xdev->status & XDMA_DEV_STATUS_INIT_MSIX)
+		xdma_irq_fini(xdev);
+
+	if (xdev->status & XDMA_DEV_STATUS_REG_DMA)
+		dma_async_device_unregister(&xdev->dma_dev);
+
+	return 0;
+}
+
+/**
+ * xdma_probe - Driver probe function
+ * @pdev: Pointer to the platform_device structure
+ */
+static int xdma_probe(struct platform_device *pdev)
+{
+	struct xdma_platdata *pdata = dev_get_platdata(&pdev->dev);
+	struct xdma_device *xdev;
+	void __iomem *reg_base;
+	struct resource *res;
+	int ret = -ENODEV;
+
+	if (pdata->max_dma_channels > XDMA_MAX_CHANNELS) {
+		dev_err(&pdev->dev, "invalid max dma channels %d",
+			pdata->max_dma_channels);
+		return -EINVAL;
+	}
+
+	xdev = devm_kzalloc(&pdev->dev, sizeof(*xdev), GFP_KERNEL);
+	if (!xdev)
+		return -ENOMEM;
+
+	platform_set_drvdata(pdev, xdev);
+	xdev->pdev = pdev;
+
+	res = platform_get_resource(pdev, IORESOURCE_IRQ, 0);
+	if (!res) {
+		xdma_err(xdev, "failed to get irq resource");
+		goto failed;
+	}
+	xdev->irq_start = res->start;
+	xdev->irq_num = res->end - res->start + 1;
+
+	res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
+	if (!res) {
+		xdma_err(xdev, "failed to get io resource");
+		goto failed;
+	}
+
+	reg_base = devm_ioremap_resource(&pdev->dev, res);
+	if (!reg_base) {
+		xdma_err(xdev, "ioremap failed");
+		goto failed;
+	}
+
+	xdev->regmap = devm_regmap_init_mmio(&pdev->dev, reg_base,
+					     &xdma_regmap_config);
+	if (!xdev->regmap) {
+		xdma_err(xdev, "config regmap failed: %d", ret);
+		goto failed;
+	}
+	INIT_LIST_HEAD(&xdev->dma_dev.channels);
+
+	ret = xdma_config_channels(xdev, DMA_MEM_TO_DEV);
+	if (ret) {
+		xdma_err(xdev, "config H2C channels failed: %d", ret);
+		goto failed;
+	}
+
+	ret = xdma_config_channels(xdev, DMA_DEV_TO_MEM);
+	if (ret) {
+		xdma_err(xdev, "config C2H channels failed: %d", ret);
+		goto failed;
+	}
+
+	dma_cap_set(DMA_SLAVE, xdev->dma_dev.cap_mask);
+	dma_cap_set(DMA_PRIVATE, xdev->dma_dev.cap_mask);
+
+	xdev->dma_dev.dev = &pdev->dev;
+	xdev->dma_dev.device_free_chan_resources = xdma_free_chan_resources;
+	xdev->dma_dev.device_alloc_chan_resources = xdma_alloc_chan_resources;
+	xdev->dma_dev.device_tx_status = dma_cookie_status;
+	xdev->dma_dev.device_prep_slave_sg = xdma_prep_device_sg;
+	xdev->dma_dev.device_config = xdma_device_config;
+	xdev->dma_dev.device_issue_pending = xdma_issue_pending;
+	xdev->dma_dev.filter.map = pdata->device_map;
+	xdev->dma_dev.filter.mapcnt = pdata->device_map_cnt;
+	xdev->dma_dev.filter.fn = xdma_filter_fn;
+
+	ret = dma_async_device_register(&xdev->dma_dev);
+	if (ret) {
+		xdma_err(xdev, "failed to register Xilinx XDMA: %d", ret);
+		goto failed;
+	}
+	xdev->status |= XDMA_DEV_STATUS_REG_DMA;
+
+	ret = xdma_irq_init(xdev);
+	if (ret) {
+		xdma_err(xdev, "failed to init msix: %d", ret);
+		goto failed;
+	}
+	xdev->status |= XDMA_DEV_STATUS_INIT_MSIX;
+
+	return 0;
+
+failed:
+	xdma_remove(pdev);
+
+	return ret;
+}
+
+static const struct platform_device_id xdma_id_table[] = {
+	{ "xdma", 0},
+	{ },
+};
+
+static struct platform_driver xdma_driver = {
+	.driver		= {
+		.name = "xdma",
+	},
+	.id_table	= xdma_id_table,
+	.probe		= xdma_probe,
+	.remove		= xdma_remove,
+};
+
+module_platform_driver(xdma_driver);
+
+MODULE_DESCRIPTION("AMD XDMA driver");
+MODULE_AUTHOR("XRT Team <runtimeca39d@amd.com>");
+MODULE_LICENSE("GPL");
diff --git a/include/linux/platform_data/amd_xdma.h b/include/linux/platform_data/amd_xdma.h
new file mode 100644
index 000000000000..b5e23e14bac8
--- /dev/null
+++ b/include/linux/platform_data/amd_xdma.h
@@ -0,0 +1,34 @@
+/* SPDX-License-Identifier: GPL-2.0-or-later */
+/*
+ * Copyright (C) 2022, Advanced Micro Devices, Inc.
+ */
+
+#ifndef _PLATDATA_AMD_XDMA_H
+#define _PLATDATA_AMD_XDMA_H
+
+#include <linux/dmaengine.h>
+
+/**
+ * struct xdma_chan_info - DMA channel information
+ *	This information is used to match channel when request dma channel
+ * @dir: Channel transfer direction
+ */
+struct xdma_chan_info {
+	enum dma_transfer_direction dir;
+};
+
+#define XDMA_FILTER_PARAM(chan_info)	((void *)(chan_info))
+
+struct dma_slave_map;
+
+/**
+ * struct xdma_platdata - platform specific data for XDMA engine
+ * @max_dma_channels: Maximum dma channels in each direction
+ */
+struct xdma_platdata {
+	u32 max_dma_channels;
+	u32 device_map_cnt;
+	struct dma_slave_map *device_map;
+};
+
+#endif /* _PLATDATA_AMD_XDMA_H */
-- 
2.27.0

[PATCH V5 XDMA 2/2] dmaengine: xilinx: xdma: Add user logic interrupt support

[Lizhi Hou]

The Xilinx DMA/Bridge Subsystem for PCIe (XDMA) provides up to 16 user
interrupt wires to user logic that generate interrupts to the host.
This patch adds APIs to enable/disable user logic interrupt for a given
interrupt wire index.

Signed-off-by: Lizhi Hou <lizhi.hou@amd.com>
Signed-off-by: Sonal Santan <sonal.santan@amd.com>
Signed-off-by: Max Zhen <max.zhen@amd.com>
Signed-off-by: Brian Xu <brian.xu@amd.com>
---
 MAINTAINERS                  |  1 +
 drivers/dma/xilinx/xdma.c    | 65 ++++++++++++++++++++++++++++++++++++
 include/linux/dma/amd_xdma.h | 16 +++++++++
 3 files changed, 82 insertions(+)
 create mode 100644 include/linux/dma/amd_xdma.h

diff --git a/MAINTAINERS b/MAINTAINERS
index c1be0b2e378a..019d84b2b086 100644
--- a/MAINTAINERS
+++ b/MAINTAINERS
@@ -21691,6 +21691,7 @@ L:	dmaengine@vger.kernel.org
 S:	Supported
 F:	drivers/dma/xilinx/xdma-regs.h
 F:	drivers/dma/xilinx/xdma.c
+F:	include/linux/dma/amd_xdma.h
 F:	include/linux/platform_data/amd_xdma.h
 
 XILINX ZYNQMP DPDMA DRIVER
diff --git a/drivers/dma/xilinx/xdma.c b/drivers/dma/xilinx/xdma.c
index 58a57e03bef5..13f627445c4a 100644
--- a/drivers/dma/xilinx/xdma.c
+++ b/drivers/dma/xilinx/xdma.c
@@ -25,6 +25,7 @@
 #include <linux/dmapool.h>
 #include <linux/regmap.h>
 #include <linux/dmaengine.h>
+#include <linux/dma/amd_xdma.h>
 #include <linux/platform_device.h>
 #include <linux/platform_data/amd_xdma.h>
 #include <linux/dma-mapping.h>
@@ -736,6 +737,7 @@ static int xdma_set_vector_reg(struct xdma_device *xdev, u32 vec_tbl_start,
 static int xdma_irq_init(struct xdma_device *xdev)
 {
 	u32 irq = xdev->irq_start;
+	u32 user_irq_start;
 	int i, j, ret;
 
 	/* return failure if there are not enough IRQs */
@@ -786,6 +788,18 @@ static int xdma_irq_init(struct xdma_device *xdev)
 		goto failed;
 	}
 
+	/* config user IRQ registers if needed */
+	user_irq_start = xdev->h2c_chan_num + xdev->c2h_chan_num;
+	if (xdev->irq_num > user_irq_start) {
+		ret = xdma_set_vector_reg(xdev, XDMA_IRQ_USER_VEC_NUM,
+					  user_irq_start,
+					  xdev->irq_num - user_irq_start);
+		if (ret) {
+			xdma_err(xdev, "failed to set user vectors: %d", ret);
+			goto failed;
+		}
+	}
+
 	/* enable interrupt */
 	ret = xdma_enable_intr(xdev);
 	if (ret) {
@@ -816,6 +830,57 @@ static bool xdma_filter_fn(struct dma_chan *chan, void *param)
 	return true;
 }
 
+/**
+ * xdma_disable_user_irq - Disable user interrupt
+ * @pdev: Pointer to the platform_device structure
+ * @user_irq_index: User IRQ index
+ */
+void xdma_disable_user_irq(struct platform_device *pdev, u32 user_irq_index)
+{
+	struct xdma_device *xdev = platform_get_drvdata(pdev);
+
+	if (xdev->h2c_chan_num + xdev->c2h_chan_num + user_irq_index >=
+	    xdev->irq_num) {
+		xdma_err(xdev, "invalid user irq index");
+		return;
+	}
+
+	xdma_write_reg(xdev, XDMA_IRQ_BASE, XDMA_IRQ_USER_INT_EN_W1C,
+		       (1 << user_irq_index));
+}
+EXPORT_SYMBOL(xdma_disable_user_irq);
+
+/**
+ * xdma_enable_user_irq - Enable user logic interrupt
+ * @pdev: Pointer to the platform_device structure
+ * @user_irq_index: User logic IRQ wire index
+ * @irq: Pointer to returning system IRQ number
+ */
+int xdma_enable_user_irq(struct platform_device *pdev, u32 user_irq_index,
+			 u32 *irq)
+{
+	struct xdma_device *xdev = platform_get_drvdata(pdev);
+	u32 user_irq;
+	int ret;
+
+	user_irq = xdev->h2c_chan_num + xdev->c2h_chan_num + user_irq_index;
+	if (user_irq >= xdev->irq_num) {
+		xdma_err(xdev, "invalid user irq index");
+		return -EINVAL;
+	}
+
+	ret = xdma_write_reg(xdev, XDMA_IRQ_BASE, XDMA_IRQ_USER_INT_EN_W1S,
+			     (1 << user_irq_index));
+	if (ret) {
+		xdma_err(xdev, "set user irq mask failed, %d", ret);
+		return ret;
+	}
+	*irq = user_irq + xdev->irq_start;
+
+	return 0;
+}
+EXPORT_SYMBOL(xdma_enable_user_irq);
+
 /**
  * xdma_remove - Driver remove function
  * @pdev: Pointer to the platform_device structure
diff --git a/include/linux/dma/amd_xdma.h b/include/linux/dma/amd_xdma.h
new file mode 100644
index 000000000000..91fb02ff93a7
--- /dev/null
+++ b/include/linux/dma/amd_xdma.h
@@ -0,0 +1,16 @@
+/* SPDX-License-Identifier: GPL-2.0-or-later */
+/*
+ * Copyright (C) 2022, Advanced Micro Devices, Inc.
+ */
+
+#ifndef _DMAENGINE_AMD_XDMA_H
+#define _DMAENGINE_AMD_XDMA_H
+
+#include <linux/interrupt.h>
+#include <linux/platform_device.h>
+
+int xdma_enable_user_irq(struct platform_device *pdev, u32 user_irq_index,
+			 u32 *irq);
+void xdma_disable_user_irq(struct platform_device *pdev, u32 user_irq_index);
+
+#endif /* _DMAENGINE_AMD_XDMA_H */
-- 
2.27.0

Re: [PATCH V5 XDMA 2/2] dmaengine: xilinx: xdma: Add user logic interrupt support

[Martin Tůma]

On 29. 09. 22 1:58, Lizhi Hou wrote:
> The Xilinx DMA/Bridge Subsystem for PCIe (XDMA) provides up to 16 user
> interrupt wires to user logic that generate interrupts to the host.
> This patch adds APIs to enable/disable user logic interrupt for a given
> interrupt wire index.
> 
> Signed-off-by: Lizhi Hou <lizhi.hou@amd.com>
> Signed-off-by: Sonal Santan <sonal.santan@amd.com>
> Signed-off-by: Max Zhen <max.zhen@amd.com>
> Signed-off-by: Brian Xu <brian.xu@amd.com>
> ---
>   MAINTAINERS                  |  1 +
>   drivers/dma/xilinx/xdma.c    | 65 ++++++++++++++++++++++++++++++++++++
>   include/linux/dma/amd_xdma.h | 16 +++++++++
>   3 files changed, 82 insertions(+)
>   create mode 100644 include/linux/dma/amd_xdma.h
> 
> diff --git a/MAINTAINERS b/MAINTAINERS
> index c1be0b2e378a..019d84b2b086 100644
> --- a/MAINTAINERS
> +++ b/MAINTAINERS
> @@ -21691,6 +21691,7 @@ L:	dmaengine@vger.kernel.org
>   S:	Supported
>   F:	drivers/dma/xilinx/xdma-regs.h
>   F:	drivers/dma/xilinx/xdma.c
> +F:	include/linux/dma/amd_xdma.h
>   F:	include/linux/platform_data/amd_xdma.h
>   
>   XILINX ZYNQMP DPDMA DRIVER
> diff --git a/drivers/dma/xilinx/xdma.c b/drivers/dma/xilinx/xdma.c
> index 58a57e03bef5..13f627445c4a 100644
> --- a/drivers/dma/xilinx/xdma.c
> +++ b/drivers/dma/xilinx/xdma.c
> @@ -25,6 +25,7 @@
>   #include <linux/dmapool.h>
>   #include <linux/regmap.h>
>   #include <linux/dmaengine.h>
> +#include <linux/dma/amd_xdma.h>
>   #include <linux/platform_device.h>
>   #include <linux/platform_data/amd_xdma.h>
>   #include <linux/dma-mapping.h>
> @@ -736,6 +737,7 @@ static int xdma_set_vector_reg(struct xdma_device *xdev, u32 vec_tbl_start,
>   static int xdma_irq_init(struct xdma_device *xdev)
>   {
>   	u32 irq = xdev->irq_start;
> +	u32 user_irq_start;
>   	int i, j, ret;
>   
>   	/* return failure if there are not enough IRQs */
> @@ -786,6 +788,18 @@ static int xdma_irq_init(struct xdma_device *xdev)
>   		goto failed;
>   	}
>   
> +	/* config user IRQ registers if needed */
> +	user_irq_start = xdev->h2c_chan_num + xdev->c2h_chan_num;
> +	if (xdev->irq_num > user_irq_start) {
> +		ret = xdma_set_vector_reg(xdev, XDMA_IRQ_USER_VEC_NUM,
> +					  user_irq_start,
> +					  xdev->irq_num - user_irq_start);
> +		if (ret) {
> +			xdma_err(xdev, "failed to set user vectors: %d", ret);
> +			goto failed;
> +		}
> +	}
> +
>   	/* enable interrupt */
>   	ret = xdma_enable_intr(xdev);
>   	if (ret) {
> @@ -816,6 +830,57 @@ static bool xdma_filter_fn(struct dma_chan *chan, void *param)
>   	return true;
>   }
>   
> +/**
> + * xdma_disable_user_irq - Disable user interrupt
> + * @pdev: Pointer to the platform_device structure
> + * @user_irq_index: User IRQ index
> + */
> +void xdma_disable_user_irq(struct platform_device *pdev, u32 user_irq_index)
> +{
> +	struct xdma_device *xdev = platform_get_drvdata(pdev);
> +
> +	if (xdev->h2c_chan_num + xdev->c2h_chan_num + user_irq_index >=
> +	    xdev->irq_num) {
> +		xdma_err(xdev, "invalid user irq index");
> +		return;
> +	}
> +
> +	xdma_write_reg(xdev, XDMA_IRQ_BASE, XDMA_IRQ_USER_INT_EN_W1C,
> +		       (1 << user_irq_index));
> +}
> +EXPORT_SYMBOL(xdma_disable_user_irq);
> +
> +/**
> + * xdma_enable_user_irq - Enable user logic interrupt
> + * @pdev: Pointer to the platform_device structure
> + * @user_irq_index: User logic IRQ wire index
> + * @irq: Pointer to returning system IRQ number
> + */
> +int xdma_enable_user_irq(struct platform_device *pdev, u32 user_irq_index,
> +			 u32 *irq)
> +{
> +	struct xdma_device *xdev = platform_get_drvdata(pdev);
> +	u32 user_irq;
> +	int ret;
> +
> +	user_irq = xdev->h2c_chan_num + xdev->c2h_chan_num + user_irq_index;
> +	if (user_irq >= xdev->irq_num) {
> +		xdma_err(xdev, "invalid user irq index");
> +		return -EINVAL;
> +	}
> +
> +	ret = xdma_write_reg(xdev, XDMA_IRQ_BASE, XDMA_IRQ_USER_INT_EN_W1S,
> +			     (1 << user_irq_index));
> +	if (ret) {
> +		xdma_err(xdev, "set user irq mask failed, %d", ret);
> +		return ret;
> +	}
> +	*irq = user_irq + xdev->irq_start;
> +
> +	return 0;
> +}
> +EXPORT_SYMBOL(xdma_enable_user_irq);
> +
>   /**
>    * xdma_remove - Driver remove function
>    * @pdev: Pointer to the platform_device structure
> diff --git a/include/linux/dma/amd_xdma.h b/include/linux/dma/amd_xdma.h
> new file mode 100644
> index 000000000000..91fb02ff93a7
> --- /dev/null
> +++ b/include/linux/dma/amd_xdma.h
> @@ -0,0 +1,16 @@
> +/* SPDX-License-Identifier: GPL-2.0-or-later */
> +/*
> + * Copyright (C) 2022, Advanced Micro Devices, Inc.
> + */
> +
> +#ifndef _DMAENGINE_AMD_XDMA_H
> +#define _DMAENGINE_AMD_XDMA_H
> +
> +#include <linux/interrupt.h>
> +#include <linux/platform_device.h>
> +
> +int xdma_enable_user_irq(struct platform_device *pdev, u32 user_irq_index,
> +			 u32 *irq);
> +void xdma_disable_user_irq(struct platform_device *pdev, u32 user_irq_index);
> +
> +#endif /* _DMAENGINE_AMD_XDMA_H */

Hi,
While rewriting our V4L2 driver to use your XDMA driver, I realized, 
that the API for the user interrupts is still not fully usable. If
the expected outcome is that the "parent" driver using the xdma driver 
knows nothing about the IRQ allocation in XDMA, then the 
xdma_enable_user_irq() function must be split into two separate functions:

int xdma_enable_user_irq(struct platform_device *pdev, u32 user_irq_index)

and

int xdma_get_system_irq(u32 user_irq_index)

that returns the system IRQ number for the given user_irq_index.
Because without it you can not get the system IRQ number without 
enabling the irq at the same time which is usually not what you want.

As a workaround, you can compute the system IRQ yourself with the 
knowledge of the XDMA internals (this is what I have to do at the moment 
in our driver), but then the "u32 *irq" parameter becomes useless and 
can be removed anyway.

M.

Re: [PATCH V5 XDMA 2/2] dmaengine: xilinx: xdma: Add user logic interrupt support

[Lizhi Hou]

On 10/3/22 06:59, Martin Tůma wrote:
> On 29. 09. 22 1:58, Lizhi Hou wrote:
>> The Xilinx DMA/Bridge Subsystem for PCIe (XDMA) provides up to 16 user
>> interrupt wires to user logic that generate interrupts to the host.
>> This patch adds APIs to enable/disable user logic interrupt for a given
>> interrupt wire index.
>>
>> Signed-off-by: Lizhi Hou <lizhi.hou@amd.com>
>> Signed-off-by: Sonal Santan <sonal.santan@amd.com>
>> Signed-off-by: Max Zhen <max.zhen@amd.com>
>> Signed-off-by: Brian Xu <brian.xu@amd.com>
>> ---
>>   MAINTAINERS                  |  1 +
>>   drivers/dma/xilinx/xdma.c    | 65 ++++++++++++++++++++++++++++++++++++
>>   include/linux/dma/amd_xdma.h | 16 +++++++++
>>   3 files changed, 82 insertions(+)
>>   create mode 100644 include/linux/dma/amd_xdma.h
>>
>> diff --git a/MAINTAINERS b/MAINTAINERS
>> index c1be0b2e378a..019d84b2b086 100644
>> --- a/MAINTAINERS
>> +++ b/MAINTAINERS
>> @@ -21691,6 +21691,7 @@ L:    dmaengine@vger.kernel.org
>>   S:    Supported
>>   F:    drivers/dma/xilinx/xdma-regs.h
>>   F:    drivers/dma/xilinx/xdma.c
>> +F:    include/linux/dma/amd_xdma.h
>>   F:    include/linux/platform_data/amd_xdma.h
>>     XILINX ZYNQMP DPDMA DRIVER
>> diff --git a/drivers/dma/xilinx/xdma.c b/drivers/dma/xilinx/xdma.c
>> index 58a57e03bef5..13f627445c4a 100644
>> --- a/drivers/dma/xilinx/xdma.c
>> +++ b/drivers/dma/xilinx/xdma.c
>> @@ -25,6 +25,7 @@
>>   #include <linux/dmapool.h>
>>   #include <linux/regmap.h>
>>   #include <linux/dmaengine.h>
>> +#include <linux/dma/amd_xdma.h>
>>   #include <linux/platform_device.h>
>>   #include <linux/platform_data/amd_xdma.h>
>>   #include <linux/dma-mapping.h>
>> @@ -736,6 +737,7 @@ static int xdma_set_vector_reg(struct xdma_device 
>> *xdev, u32 vec_tbl_start,
>>   static int xdma_irq_init(struct xdma_device *xdev)
>>   {
>>       u32 irq = xdev->irq_start;
>> +    u32 user_irq_start;
>>       int i, j, ret;
>>         /* return failure if there are not enough IRQs */
>> @@ -786,6 +788,18 @@ static int xdma_irq_init(struct xdma_device *xdev)
>>           goto failed;
>>       }
>>   +    /* config user IRQ registers if needed */
>> +    user_irq_start = xdev->h2c_chan_num + xdev->c2h_chan_num;
>> +    if (xdev->irq_num > user_irq_start) {
>> +        ret = xdma_set_vector_reg(xdev, XDMA_IRQ_USER_VEC_NUM,
>> +                      user_irq_start,
>> +                      xdev->irq_num - user_irq_start);
>> +        if (ret) {
>> +            xdma_err(xdev, "failed to set user vectors: %d", ret);
>> +            goto failed;
>> +        }
>> +    }
>> +
>>       /* enable interrupt */
>>       ret = xdma_enable_intr(xdev);
>>       if (ret) {
>> @@ -816,6 +830,57 @@ static bool xdma_filter_fn(struct dma_chan 
>> *chan, void *param)
>>       return true;
>>   }
>>   +/**
>> + * xdma_disable_user_irq - Disable user interrupt
>> + * @pdev: Pointer to the platform_device structure
>> + * @user_irq_index: User IRQ index
>> + */
>> +void xdma_disable_user_irq(struct platform_device *pdev, u32 
>> user_irq_index)
>> +{
>> +    struct xdma_device *xdev = platform_get_drvdata(pdev);
>> +
>> +    if (xdev->h2c_chan_num + xdev->c2h_chan_num + user_irq_index >=
>> +        xdev->irq_num) {
>> +        xdma_err(xdev, "invalid user irq index");
>> +        return;
>> +    }
>> +
>> +    xdma_write_reg(xdev, XDMA_IRQ_BASE, XDMA_IRQ_USER_INT_EN_W1C,
>> +               (1 << user_irq_index));
>> +}
>> +EXPORT_SYMBOL(xdma_disable_user_irq);
>> +
>> +/**
>> + * xdma_enable_user_irq - Enable user logic interrupt
>> + * @pdev: Pointer to the platform_device structure
>> + * @user_irq_index: User logic IRQ wire index
>> + * @irq: Pointer to returning system IRQ number
>> + */
>> +int xdma_enable_user_irq(struct platform_device *pdev, u32 
>> user_irq_index,
>> +             u32 *irq)
>> +{
>> +    struct xdma_device *xdev = platform_get_drvdata(pdev);
>> +    u32 user_irq;
>> +    int ret;
>> +
>> +    user_irq = xdev->h2c_chan_num + xdev->c2h_chan_num + 
>> user_irq_index;
>> +    if (user_irq >= xdev->irq_num) {
>> +        xdma_err(xdev, "invalid user irq index");
>> +        return -EINVAL;
>> +    }
>> +
>> +    ret = xdma_write_reg(xdev, XDMA_IRQ_BASE, XDMA_IRQ_USER_INT_EN_W1S,
>> +                 (1 << user_irq_index));
>> +    if (ret) {
>> +        xdma_err(xdev, "set user irq mask failed, %d", ret);
>> +        return ret;
>> +    }
>> +    *irq = user_irq + xdev->irq_start;
>> +
>> +    return 0;
>> +}
>> +EXPORT_SYMBOL(xdma_enable_user_irq);
>> +
>>   /**
>>    * xdma_remove - Driver remove function
>>    * @pdev: Pointer to the platform_device structure
>> diff --git a/include/linux/dma/amd_xdma.h b/include/linux/dma/amd_xdma.h
>> new file mode 100644
>> index 000000000000..91fb02ff93a7
>> --- /dev/null
>> +++ b/include/linux/dma/amd_xdma.h
>> @@ -0,0 +1,16 @@
>> +/* SPDX-License-Identifier: GPL-2.0-or-later */
>> +/*
>> + * Copyright (C) 2022, Advanced Micro Devices, Inc.
>> + */
>> +
>> +#ifndef _DMAENGINE_AMD_XDMA_H
>> +#define _DMAENGINE_AMD_XDMA_H
>> +
>> +#include <linux/interrupt.h>
>> +#include <linux/platform_device.h>
>> +
>> +int xdma_enable_user_irq(struct platform_device *pdev, u32 
>> user_irq_index,
>> +             u32 *irq);
>> +void xdma_disable_user_irq(struct platform_device *pdev, u32 
>> user_irq_index);
>> +
>> +#endif /* _DMAENGINE_AMD_XDMA_H */
>
> Hi,
> While rewriting our V4L2 driver to use your XDMA driver, I realized, 
> that the API for the user interrupts is still not fully usable. If
> the expected outcome is that the "parent" driver using the xdma driver 
> knows nothing about the IRQ allocation in XDMA, then the 
> xdma_enable_user_irq() function must be split into two separate 
> functions:
>
> int xdma_enable_user_irq(struct platform_device *pdev, u32 
> user_irq_index)
>
> and
>
> int xdma_get_system_irq(u32 user_irq_index)
>
> that returns the system IRQ number for the given user_irq_index.
> Because without it you can not get the system IRQ number without 
> enabling the irq at the same time which is usually not what you want.

Is this because the user logic you are using does not have 
disable/enable interrupt by itself?  I thought that the user logic IP 
would have its own register to enable/disable interrupt.

And xdma_enable_user_irq() will only be called once for creating the 
user logic platform device. Then, user logic IP driver would use its own 
register to control interrupts. The existing platform driver does not 
call any xdma_* to control interrupt.


Thanks,

Lizhi

>
> As a workaround, you can compute the system IRQ yourself with the 
> knowledge of the XDMA internals (this is what I have to do at the 
> moment in our driver), but then the "u32 *irq" parameter becomes 
> useless and can be removed anyway.
>
> M.

Re: [PATCH V5 XDMA 1/2] dmaengine: xilinx: xdma: Add xilinx xdma driver

[Ilpo Järvinen]

On Wed, 28 Sep 2022, Lizhi Hou wrote:

> Add driver to enable PCIe board which uses XDMA (the DMA/Bridge Subsystem
> for PCI Express). For example, Xilinx Alveo PCIe devices.
>     https://www.xilinx.com/products/boards-and-kits/alveo.html
> 
> The XDMA engine support up to 4 Host to Card (H2C) and 4 Card to Host (C2H)
> channels. Memory transfers are specified on a per-channel basis in
> descriptor linked lists, which the DMA fetches from host memory and
> processes. Events such as descriptor completion and errors are signaled
> using interrupts. The hardware detail is provided by
>     https://docs.xilinx.com/r/en-US/pg195-pcie-dma/Introduction
> 
> This driver implements dmaengine APIs.
>     - probe the available DMA channels
>     - use dma_slave_map for channel lookup
>     - use virtual channel to manage dmaengine tx descriptors
>     - implement device_prep_slave_sg callback to handle host scatter gather
>       list
>     - implement device_config to config device address for DMA transfer
> 
> Signed-off-by: Lizhi Hou <lizhi.hou@amd.com>
> Signed-off-by: Sonal Santan <sonal.santan@amd.com>
> Signed-off-by: Max Zhen <max.zhen@amd.com>
> Signed-off-by: Brian Xu <brian.xu@amd.com>
> ---

> +/* bits of the channel control register */
> +#define CHAN_CTRL_RUN_STOP			BIT(0)
> +#define CHAN_CTRL_IE_DESC_STOPPED		BIT(1)
> +#define CHAN_CTRL_IE_DESC_COMPLETED		BIT(2)
> +#define CHAN_CTRL_IE_DESC_ALIGN_MISMATCH	BIT(3)
> +#define CHAN_CTRL_IE_MAGIC_STOPPED		BIT(4)
> +#define CHAN_CTRL_IE_IDLE_STOPPED		BIT(6)
> +#define CHAN_CTRL_IE_READ_ERROR			(0x1FUL << 9)
> +#define CHAN_CTRL_IE_DESC_ERROR			(0x1FUL << 19)

Looks GENMASK()

> +/* bits of the SG DMA control register */
> +#define XDMA_CTRL_RUN_STOP			BIT(0)
> +#define XDMA_CTRL_IE_DESC_STOPPED		BIT(1)
> +#define XDMA_CTRL_IE_DESC_COMPLETED		BIT(2)
> +#define XDMA_CTRL_IE_DESC_ALIGN_MISMATCH	BIT(3)
> +#define XDMA_CTRL_IE_MAGIC_STOPPED		BIT(4)
> +#define XDMA_CTRL_IE_IDLE_STOPPED		BIT(6)
> +#define XDMA_CTRL_IE_READ_ERROR			(0x1FUL << 9)
> +#define XDMA_CTRL_IE_DESC_ERROR			(0x1FUL << 19)

Ditto.

> +/**
> + * xdma_config_channels - Detect and config DMA channels
> + * @xdev: DMA device pointer
> + * @dir: Channel direction
> + */
> +static int xdma_config_channels(struct xdma_device *xdev,
> +				enum dma_transfer_direction dir)
> +{
> +	struct xdma_platdata *pdata = dev_get_platdata(&xdev->pdev->dev);
> +	u32 base, identifier, target;
> +	struct xdma_chan **chans;
> +	u32 *chan_num;
> +	int i, j, ret;
> +
> +	if (dir == DMA_MEM_TO_DEV) {
> +		base = XDMA_CHAN_H2C_OFFSET;
> +		target = XDMA_CHAN_H2C_TARGET;
> +		chans = &xdev->h2c_chans;
> +		chan_num = &xdev->h2c_chan_num;
> +	} else if (dir == DMA_DEV_TO_MEM) {
> +		base = XDMA_CHAN_C2H_OFFSET;
> +		target = XDMA_CHAN_C2H_TARGET;
> +		chans = &xdev->c2h_chans;
> +		chan_num = &xdev->c2h_chan_num;
> +	} else {
> +		xdma_err(xdev, "invalid direction specified");
> +		return -EINVAL;
> +	}
> +
> +	/* detect number of available DMA channels */
> +	for (i = 0, *chan_num = 0; i < pdata->max_dma_channels; i++) {
> +		ret = xdma_read_reg(xdev, base + i * XDMA_CHAN_STRIDE,
> +				    XDMA_CHAN_IDENTIFIER, &identifier);
> +		if (ret) {
> +			xdma_err(xdev, "failed to read channel id: %d", ret);
> +			return ret;
> +		}
> +
> +		/* check if it is available DMA channel */
> +		if (XDMA_CHAN_CHECK_TARGET(identifier, target))
> +			(*chan_num)++;
> +	}
> +
> +	if (!*chan_num) {
> +		xdma_err(xdev, "does not probe any channel");
> +		return -EINVAL;
> +	}
> +
> +	*chans = devm_kzalloc(&xdev->pdev->dev, sizeof(**chans) * (*chan_num),
> +			      GFP_KERNEL);
> +	if (!*chans)
> +		return -ENOMEM;
> +
> +	for (i = 0, j = 0; i < pdata->max_dma_channels; i++) {
> +		ret = xdma_read_reg(xdev, base + i * XDMA_CHAN_STRIDE,
> +				    XDMA_CHAN_IDENTIFIER, &identifier);
> +		if (ret) {
> +			xdma_err(xdev, "failed to read channel id: %d", ret);
> +			return ret;
> +		}

Is it ok to not rollback the allocation in case of an error occurs?

> +
> +		if (!XDMA_CHAN_CHECK_TARGET(identifier, target))
> +			continue;
> +
> +		if (j == *chan_num) {
> +			xdma_err(xdev, "invalid channel number");
> +			return -EIO;
> +		}

Same here?

> +
> +		/* init channel structure and hardware */
> +		(*chans)[j].xdev_hdl = xdev;
> +		(*chans)[j].base = base + i * XDMA_CHAN_STRIDE;
> +		(*chans)[j].dir = dir;
> +
> +		ret = xdma_channel_init(&(*chans)[j]);
> +		if (ret)
> +			return ret;

And here.

> +		(*chans)[j].vchan.desc_free = xdma_free_desc;
> +		vchan_init(&(*chans)[j].vchan, &xdev->dma_dev);
> +
> +		j++;
> +	}
> +
> +	dev_info(&xdev->pdev->dev, "configured %d %s channels", j,
> +		 (dir == DMA_MEM_TO_DEV) ? "H2C" : "C2H");
> +
> +	return 0;
> +}


> +static int xdma_irq_init(struct xdma_device *xdev)
> +{
> +	u32 irq = xdev->irq_start;
> +	int i, j, ret;
> +
> +	/* return failure if there are not enough IRQs */
> +	if (xdev->irq_num < xdev->h2c_chan_num + xdev->c2h_chan_num) {
> +		xdma_err(xdev, "not enough irq");
> +		return -EINVAL;
> +	}
> +
> +	/* setup H2C interrupt handler */
> +	for (i = 0; i < xdev->h2c_chan_num; i++) {
> +		ret = request_irq(irq, xdma_channel_isr, 0,
> +				  "xdma-h2c-channel", &xdev->h2c_chans[i]);
> +		if (ret) {
> +			xdma_err(xdev, "H2C channel%d request irq%d failed: %d",
> +				 i, irq, ret);
> +			for (j = 0; j < i; j++) {
> +				free_irq(xdev->h2c_chans[j].irq,
> +					 &xdev->h2c_chans[j]);
> +			}
> +			return ret;

Remove freeing for loop and just do
			goto failed_init_c2h;
And reverse the iteration order down on the error path (from i to zero) by 
using
	while (i--)

> +		}
> +		xdev->h2c_chans[i].irq = irq;
> +		irq++;
> +	}
> +
> +	/* setup C2H interrupt handler */
> +	for (i = 0; i < xdev->c2h_chan_num; i++) {
> +		ret = request_irq(irq, xdma_channel_isr, 0,
> +				  "xdma-c2h-channel", &xdev->c2h_chans[i]);
> +		if (ret) {
> +			xdma_err(xdev, "H2C channel%d request irq%d failed: %d",
> +				 i, irq, ret);
> +			for (j = 0; j < i; j++) {
> +				free_irq(xdev->c2h_chans[j].irq,
> +					 &xdev->c2h_chans[j]);
> +			}
> +			goto failed_init_c2h;

Ditto. But use j for this whole for loop so you can just do while (j--) on 
the error path as the index variable won't clash with the previous loop.

> +		}
> +		xdev->c2h_chans[i].irq = irq;
> +		irq++;
> +	}
> +
> +	/* config hardware IRQ registers */
> +	ret = xdma_set_vector_reg(xdev, XDMA_IRQ_CHAN_VEC_NUM, 0,
> +				  xdev->h2c_chan_num + xdev->c2h_chan_num);
> +	if (ret) {
> +		xdma_err(xdev, "failed to set channel vectors: %d", ret);
> +		goto failed;
> +	}
> +
> +	/* enable interrupt */
> +	ret = xdma_enable_intr(xdev);
> +	if (ret) {
> +		xdma_err(xdev, "failed to enable interrupts: %d", ret);
> +		goto failed;
> +	}
> +
> +	return 0;
> +
> +failed:
> +	for (i = 0; i < xdev->c2h_chan_num; i++)

As mentioned above:

while (j--)

> +		free_irq(xdev->c2h_chans[i].irq, &xdev->c2h_chans[i]);
> +failed_init_c2h:
> +	for (i = 0; i < xdev->h2c_chan_num; i++)

while (i--)

> +		free_irq(xdev->h2c_chans[i].irq, &xdev->h2c_chans[i]);
> +
> +	return ret;
> +}
> +
> +static bool xdma_filter_fn(struct dma_chan *chan, void *param)
> +{
> +	struct xdma_chan *xdma_chan = to_xdma_chan(chan);
> +	struct xdma_chan_info *chan_info = param;
> +
> +	if (chan_info->dir != xdma_chan->dir)
> +		return false;
> +
> +	return true;

return chan_info->dir == xdma_chan->dir;


-- 
 i.

Re: [PATCH V5 XDMA 2/2] dmaengine: xilinx: xdma: Add user logic interrupt support

[Martin Tůma]

On 03. 10. 22 17:52, Lizhi Hou wrote:
>
> On 10/3/22 06:59, Martin Tůma wrote:
>> On 29. 09. 22 1:58, Lizhi Hou wrote:
>>> The Xilinx DMA/Bridge Subsystem for PCIe (XDMA) provides up to 16 user
>>> interrupt wires to user logic that generate interrupts to the host.
>>> This patch adds APIs to enable/disable user logic interrupt for a given
>>> interrupt wire index.
>>>
>>> Signed-off-by: Lizhi Hou <lizhi.hou@amd.com>
>>> Signed-off-by: Sonal Santan <sonal.santan@amd.com>
>>> Signed-off-by: Max Zhen <max.zhen@amd.com>
>>> Signed-off-by: Brian Xu <brian.xu@amd.com>
>>> ---
>>>   MAINTAINERS                  |  1 +
>>>   drivers/dma/xilinx/xdma.c    | 65 
>>> ++++++++++++++++++++++++++++++++++++
>>>   include/linux/dma/amd_xdma.h | 16 +++++++++
>>>   3 files changed, 82 insertions(+)
>>>   create mode 100644 include/linux/dma/amd_xdma.h
>>>
>>> diff --git a/MAINTAINERS b/MAINTAINERS
>>> index c1be0b2e378a..019d84b2b086 100644
>>> --- a/MAINTAINERS
>>> +++ b/MAINTAINERS
>>> @@ -21691,6 +21691,7 @@ L:    dmaengine@vger.kernel.org
>>>   S:    Supported
>>>   F:    drivers/dma/xilinx/xdma-regs.h
>>>   F:    drivers/dma/xilinx/xdma.c
>>> +F:    include/linux/dma/amd_xdma.h
>>>   F:    include/linux/platform_data/amd_xdma.h
>>>     XILINX ZYNQMP DPDMA DRIVER
>>> diff --git a/drivers/dma/xilinx/xdma.c b/drivers/dma/xilinx/xdma.c
>>> index 58a57e03bef5..13f627445c4a 100644
>>> --- a/drivers/dma/xilinx/xdma.c
>>> +++ b/drivers/dma/xilinx/xdma.c
>>> @@ -25,6 +25,7 @@
>>>   #include <linux/dmapool.h>
>>>   #include <linux/regmap.h>
>>>   #include <linux/dmaengine.h>
>>> +#include <linux/dma/amd_xdma.h>
>>>   #include <linux/platform_device.h>
>>>   #include <linux/platform_data/amd_xdma.h>
>>>   #include <linux/dma-mapping.h>
>>> @@ -736,6 +737,7 @@ static int xdma_set_vector_reg(struct 
>>> xdma_device *xdev, u32 vec_tbl_start,
>>>   static int xdma_irq_init(struct xdma_device *xdev)
>>>   {
>>>       u32 irq = xdev->irq_start;
>>> +    u32 user_irq_start;
>>>       int i, j, ret;
>>>         /* return failure if there are not enough IRQs */
>>> @@ -786,6 +788,18 @@ static int xdma_irq_init(struct xdma_device *xdev)
>>>           goto failed;
>>>       }
>>>   +    /* config user IRQ registers if needed */
>>> +    user_irq_start = xdev->h2c_chan_num + xdev->c2h_chan_num;
>>> +    if (xdev->irq_num > user_irq_start) {
>>> +        ret = xdma_set_vector_reg(xdev, XDMA_IRQ_USER_VEC_NUM,
>>> +                      user_irq_start,
>>> +                      xdev->irq_num - user_irq_start);
>>> +        if (ret) {
>>> +            xdma_err(xdev, "failed to set user vectors: %d", ret);
>>> +            goto failed;
>>> +        }
>>> +    }
>>> +
>>>       /* enable interrupt */
>>>       ret = xdma_enable_intr(xdev);
>>>       if (ret) {
>>> @@ -816,6 +830,57 @@ static bool xdma_filter_fn(struct dma_chan 
>>> *chan, void *param)
>>>       return true;
>>>   }
>>>   +/**
>>> + * xdma_disable_user_irq - Disable user interrupt
>>> + * @pdev: Pointer to the platform_device structure
>>> + * @user_irq_index: User IRQ index
>>> + */
>>> +void xdma_disable_user_irq(struct platform_device *pdev, u32 
>>> user_irq_index)
>>> +{
>>> +    struct xdma_device *xdev = platform_get_drvdata(pdev);
>>> +
>>> +    if (xdev->h2c_chan_num + xdev->c2h_chan_num + user_irq_index >=
>>> +        xdev->irq_num) {
>>> +        xdma_err(xdev, "invalid user irq index");
>>> +        return;
>>> +    }
>>> +
>>> +    xdma_write_reg(xdev, XDMA_IRQ_BASE, XDMA_IRQ_USER_INT_EN_W1C,
>>> +               (1 << user_irq_index));
>>> +}
>>> +EXPORT_SYMBOL(xdma_disable_user_irq);
>>> +
>>> +/**
>>> + * xdma_enable_user_irq - Enable user logic interrupt
>>> + * @pdev: Pointer to the platform_device structure
>>> + * @user_irq_index: User logic IRQ wire index
>>> + * @irq: Pointer to returning system IRQ number
>>> + */
>>> +int xdma_enable_user_irq(struct platform_device *pdev, u32 
>>> user_irq_index,
>>> +             u32 *irq)
>>> +{
>>> +    struct xdma_device *xdev = platform_get_drvdata(pdev);
>>> +    u32 user_irq;
>>> +    int ret;
>>> +
>>> +    user_irq = xdev->h2c_chan_num + xdev->c2h_chan_num + 
>>> user_irq_index;
>>> +    if (user_irq >= xdev->irq_num) {
>>> +        xdma_err(xdev, "invalid user irq index");
>>> +        return -EINVAL;
>>> +    }
>>> +
>>> +    ret = xdma_write_reg(xdev, XDMA_IRQ_BASE, 
>>> XDMA_IRQ_USER_INT_EN_W1S,
>>> +                 (1 << user_irq_index));
>>> +    if (ret) {
>>> +        xdma_err(xdev, "set user irq mask failed, %d", ret);
>>> +        return ret;
>>> +    }
>>> +    *irq = user_irq + xdev->irq_start;
>>> +
>>> +    return 0;
>>> +}
>>> +EXPORT_SYMBOL(xdma_enable_user_irq);
>>> +
>>>   /**
>>>    * xdma_remove - Driver remove function
>>>    * @pdev: Pointer to the platform_device structure
>>> diff --git a/include/linux/dma/amd_xdma.h 
>>> b/include/linux/dma/amd_xdma.h
>>> new file mode 100644
>>> index 000000000000..91fb02ff93a7
>>> --- /dev/null
>>> +++ b/include/linux/dma/amd_xdma.h
>>> @@ -0,0 +1,16 @@
>>> +/* SPDX-License-Identifier: GPL-2.0-or-later */
>>> +/*
>>> + * Copyright (C) 2022, Advanced Micro Devices, Inc.
>>> + */
>>> +
>>> +#ifndef _DMAENGINE_AMD_XDMA_H
>>> +#define _DMAENGINE_AMD_XDMA_H
>>> +
>>> +#include <linux/interrupt.h>
>>> +#include <linux/platform_device.h>
>>> +
>>> +int xdma_enable_user_irq(struct platform_device *pdev, u32 
>>> user_irq_index,
>>> +             u32 *irq);
>>> +void xdma_disable_user_irq(struct platform_device *pdev, u32 
>>> user_irq_index);
>>> +
>>> +#endif /* _DMAENGINE_AMD_XDMA_H */
>>
>> Hi,
>> While rewriting our V4L2 driver to use your XDMA driver, I realized, 
>> that the API for the user interrupts is still not fully usable. If
>> the expected outcome is that the "parent" driver using the xdma 
>> driver knows nothing about the IRQ allocation in XDMA, then the 
>> xdma_enable_user_irq() function must be split into two separate 
>> functions:
>>
>> int xdma_enable_user_irq(struct platform_device *pdev, u32 
>> user_irq_index)
>>
>> and
>>
>> int xdma_get_system_irq(u32 user_irq_index)
>>
>> that returns the system IRQ number for the given user_irq_index.
>> Because without it you can not get the system IRQ number without 
>> enabling the irq at the same time which is usually not what you want.
>
> Is this because the user logic you are using does not have 
> disable/enable interrupt by itself?  I thought that the user logic IP 
> would have its own register to enable/disable interrupt.
>
Yes. Our HW has no "own" registers to disable the IRQs. As such 
registers are already there in XDMA, there was no need for them.


> And xdma_enable_user_irq() will only be called once for creating the 
> user logic platform device. Then, user logic IP driver would use its 
> own register to control interrupts. The existing platform driver does 
> not call any xdma_* to control interrupt.
>
As you can see, there is HW that was designed to use the XDMA registers 
during normal operation and separate enable/disable and irq mapping 
functions are required. Maybe it is not the best HW design, but it can 
evidently happen and then the driver author is screwed with the merged API.

M.

>
> Thanks,
>
> Lizhi
>
>>
>> As a workaround, you can compute the system IRQ yourself with the 
>> knowledge of the XDMA internals (this is what I have to do at the 
>> moment in our driver), but then the "u32 *irq" parameter becomes 
>> useless and can be removed anyway.
>>
>> M.

Re: [PATCH V5 XDMA 1/2] dmaengine: xilinx: xdma: Add xilinx xdma driver

[Lizhi Hou]

On 10/4/22 01:18, Ilpo Järvinen wrote:
> On Wed, 28 Sep 2022, Lizhi Hou wrote:
>
>> Add driver to enable PCIe board which uses XDMA (the DMA/Bridge Subsystem
>> for PCI Express). For example, Xilinx Alveo PCIe devices.
>>      https://www.xilinx.com/products/boards-and-kits/alveo.html
>>
>> The XDMA engine support up to 4 Host to Card (H2C) and 4 Card to Host (C2H)
>> channels. Memory transfers are specified on a per-channel basis in
>> descriptor linked lists, which the DMA fetches from host memory and
>> processes. Events such as descriptor completion and errors are signaled
>> using interrupts. The hardware detail is provided by
>>      https://docs.xilinx.com/r/en-US/pg195-pcie-dma/Introduction
>>
>> This driver implements dmaengine APIs.
>>      - probe the available DMA channels
>>      - use dma_slave_map for channel lookup
>>      - use virtual channel to manage dmaengine tx descriptors
>>      - implement device_prep_slave_sg callback to handle host scatter gather
>>        list
>>      - implement device_config to config device address for DMA transfer
>>
>> Signed-off-by: Lizhi Hou <lizhi.hou@amd.com>
>> Signed-off-by: Sonal Santan <sonal.santan@amd.com>
>> Signed-off-by: Max Zhen <max.zhen@amd.com>
>> Signed-off-by: Brian Xu <brian.xu@amd.com>
>> ---
>> +/* bits of the channel control register */
>> +#define CHAN_CTRL_RUN_STOP			BIT(0)
>> +#define CHAN_CTRL_IE_DESC_STOPPED		BIT(1)
>> +#define CHAN_CTRL_IE_DESC_COMPLETED		BIT(2)
>> +#define CHAN_CTRL_IE_DESC_ALIGN_MISMATCH	BIT(3)
>> +#define CHAN_CTRL_IE_MAGIC_STOPPED		BIT(4)
>> +#define CHAN_CTRL_IE_IDLE_STOPPED		BIT(6)
>> +#define CHAN_CTRL_IE_READ_ERROR			(0x1FUL << 9)
>> +#define CHAN_CTRL_IE_DESC_ERROR			(0x1FUL << 19)
> Looks GENMASK()
Sure. Thanks for pointing it out.
>
>> +/* bits of the SG DMA control register */
>> +#define XDMA_CTRL_RUN_STOP			BIT(0)
>> +#define XDMA_CTRL_IE_DESC_STOPPED		BIT(1)
>> +#define XDMA_CTRL_IE_DESC_COMPLETED		BIT(2)
>> +#define XDMA_CTRL_IE_DESC_ALIGN_MISMATCH	BIT(3)
>> +#define XDMA_CTRL_IE_MAGIC_STOPPED		BIT(4)
>> +#define XDMA_CTRL_IE_IDLE_STOPPED		BIT(6)
>> +#define XDMA_CTRL_IE_READ_ERROR			(0x1FUL << 9)
>> +#define XDMA_CTRL_IE_DESC_ERROR			(0x1FUL << 19)
> Ditto.
>
>> +/**
>> + * xdma_config_channels - Detect and config DMA channels
>> + * @xdev: DMA device pointer
>> + * @dir: Channel direction
>> + */
>> +static int xdma_config_channels(struct xdma_device *xdev,
>> +				enum dma_transfer_direction dir)
>> +{
>> +	struct xdma_platdata *pdata = dev_get_platdata(&xdev->pdev->dev);
>> +	u32 base, identifier, target;
>> +	struct xdma_chan **chans;
>> +	u32 *chan_num;
>> +	int i, j, ret;
>> +
>> +	if (dir == DMA_MEM_TO_DEV) {
>> +		base = XDMA_CHAN_H2C_OFFSET;
>> +		target = XDMA_CHAN_H2C_TARGET;
>> +		chans = &xdev->h2c_chans;
>> +		chan_num = &xdev->h2c_chan_num;
>> +	} else if (dir == DMA_DEV_TO_MEM) {
>> +		base = XDMA_CHAN_C2H_OFFSET;
>> +		target = XDMA_CHAN_C2H_TARGET;
>> +		chans = &xdev->c2h_chans;
>> +		chan_num = &xdev->c2h_chan_num;
>> +	} else {
>> +		xdma_err(xdev, "invalid direction specified");
>> +		return -EINVAL;
>> +	}
>> +
>> +	/* detect number of available DMA channels */
>> +	for (i = 0, *chan_num = 0; i < pdata->max_dma_channels; i++) {
>> +		ret = xdma_read_reg(xdev, base + i * XDMA_CHAN_STRIDE,
>> +				    XDMA_CHAN_IDENTIFIER, &identifier);
>> +		if (ret) {
>> +			xdma_err(xdev, "failed to read channel id: %d", ret);
>> +			return ret;
>> +		}
>> +
>> +		/* check if it is available DMA channel */
>> +		if (XDMA_CHAN_CHECK_TARGET(identifier, target))
>> +			(*chan_num)++;
>> +	}
>> +
>> +	if (!*chan_num) {
>> +		xdma_err(xdev, "does not probe any channel");
>> +		return -EINVAL;
>> +	}
>> +
>> +	*chans = devm_kzalloc(&xdev->pdev->dev, sizeof(**chans) * (*chan_num),
>> +			      GFP_KERNEL);
>> +	if (!*chans)
>> +		return -ENOMEM;
>> +
>> +	for (i = 0, j = 0; i < pdata->max_dma_channels; i++) {
>> +		ret = xdma_read_reg(xdev, base + i * XDMA_CHAN_STRIDE,
>> +				    XDMA_CHAN_IDENTIFIER, &identifier);
>> +		if (ret) {
>> +			xdma_err(xdev, "failed to read channel id: %d", ret);
>> +			return ret;
>> +		}
> Is it ok to not rollback the allocation in case of an error occurs?

In this loop, the failures are returned by read/write registers. The 
read/write register failure indicates serious hardware issue and the 
hardware may not be rollback in this situation.

>
>> +
>> +		if (!XDMA_CHAN_CHECK_TARGET(identifier, target))
>> +			continue;
>> +
>> +		if (j == *chan_num) {
>> +			xdma_err(xdev, "invalid channel number");
>> +			return -EIO;
>> +		}
> Same here?
>
>> +
>> +		/* init channel structure and hardware */
>> +		(*chans)[j].xdev_hdl = xdev;
>> +		(*chans)[j].base = base + i * XDMA_CHAN_STRIDE;
>> +		(*chans)[j].dir = dir;
>> +
>> +		ret = xdma_channel_init(&(*chans)[j]);
>> +		if (ret)
>> +			return ret;
> And here.
>
>> +		(*chans)[j].vchan.desc_free = xdma_free_desc;
>> +		vchan_init(&(*chans)[j].vchan, &xdev->dma_dev);
>> +
>> +		j++;
>> +	}
>> +
>> +	dev_info(&xdev->pdev->dev, "configured %d %s channels", j,
>> +		 (dir == DMA_MEM_TO_DEV) ? "H2C" : "C2H");
>> +
>> +	return 0;
>> +}
>
>> +static int xdma_irq_init(struct xdma_device *xdev)
>> +{
>> +	u32 irq = xdev->irq_start;
>> +	int i, j, ret;
>> +
>> +	/* return failure if there are not enough IRQs */
>> +	if (xdev->irq_num < xdev->h2c_chan_num + xdev->c2h_chan_num) {
>> +		xdma_err(xdev, "not enough irq");
>> +		return -EINVAL;
>> +	}
>> +
>> +	/* setup H2C interrupt handler */
>> +	for (i = 0; i < xdev->h2c_chan_num; i++) {
>> +		ret = request_irq(irq, xdma_channel_isr, 0,
>> +				  "xdma-h2c-channel", &xdev->h2c_chans[i]);
>> +		if (ret) {
>> +			xdma_err(xdev, "H2C channel%d request irq%d failed: %d",
>> +				 i, irq, ret);
>> +			for (j = 0; j < i; j++) {
>> +				free_irq(xdev->h2c_chans[j].irq,
>> +					 &xdev->h2c_chans[j]);
>> +			}
>> +			return ret;
> Remove freeing for loop and just do
> 			goto failed_init_c2h;
> And reverse the iteration order down on the error path (from i to zero) by
> using
> 	while (i--)
Thanks. This looks much cleaner than before. :)
>
>> +		}
>> +		xdev->h2c_chans[i].irq = irq;
>> +		irq++;
>> +	}
>> +
>> +	/* setup C2H interrupt handler */
>> +	for (i = 0; i < xdev->c2h_chan_num; i++) {
>> +		ret = request_irq(irq, xdma_channel_isr, 0,
>> +				  "xdma-c2h-channel", &xdev->c2h_chans[i]);
>> +		if (ret) {
>> +			xdma_err(xdev, "H2C channel%d request irq%d failed: %d",
>> +				 i, irq, ret);
>> +			for (j = 0; j < i; j++) {
>> +				free_irq(xdev->c2h_chans[j].irq,
>> +					 &xdev->c2h_chans[j]);
>> +			}
>> +			goto failed_init_c2h;
> Ditto. But use j for this whole for loop so you can just do while (j--) on
> the error path as the index variable won't clash with the previous loop.
>
>> +		}
>> +		xdev->c2h_chans[i].irq = irq;
>> +		irq++;
>> +	}
>> +
>> +	/* config hardware IRQ registers */
>> +	ret = xdma_set_vector_reg(xdev, XDMA_IRQ_CHAN_VEC_NUM, 0,
>> +				  xdev->h2c_chan_num + xdev->c2h_chan_num);
>> +	if (ret) {
>> +		xdma_err(xdev, "failed to set channel vectors: %d", ret);
>> +		goto failed;
>> +	}
>> +
>> +	/* enable interrupt */
>> +	ret = xdma_enable_intr(xdev);
>> +	if (ret) {
>> +		xdma_err(xdev, "failed to enable interrupts: %d", ret);
>> +		goto failed;
>> +	}
>> +
>> +	return 0;
>> +
>> +failed:
>> +	for (i = 0; i < xdev->c2h_chan_num; i++)
> As mentioned above:
>
> while (j--)
>
>> +		free_irq(xdev->c2h_chans[i].irq, &xdev->c2h_chans[i]);
>> +failed_init_c2h:
>> +	for (i = 0; i < xdev->h2c_chan_num; i++)
> while (i--)
>
>> +		free_irq(xdev->h2c_chans[i].irq, &xdev->h2c_chans[i]);
>> +
>> +	return ret;
>> +}
>> +
>> +static bool xdma_filter_fn(struct dma_chan *chan, void *param)
>> +{
>> +	struct xdma_chan *xdma_chan = to_xdma_chan(chan);
>> +	struct xdma_chan_info *chan_info = param;
>> +
>> +	if (chan_info->dir != xdma_chan->dir)
>> +		return false;
>> +
>> +	return true;
> return chan_info->dir == xdma_chan->dir;

Sure. I will change this.


Thanks,

Lizhi

>
>

Re: [PATCH V5 XDMA 2/2] dmaengine: xilinx: xdma: Add user logic interrupt support

[Lizhi Hou]

On 10/4/22 03:26, Martin Tůma wrote:
>
> On 03. 10. 22 17:52, Lizhi Hou wrote:
>>
>> On 10/3/22 06:59, Martin Tůma wrote:
>>> On 29. 09. 22 1:58, Lizhi Hou wrote:
>>>> The Xilinx DMA/Bridge Subsystem for PCIe (XDMA) provides up to 16 user
>>>> interrupt wires to user logic that generate interrupts to the host.
>>>> This patch adds APIs to enable/disable user logic interrupt for a 
>>>> given
>>>> interrupt wire index.
>>>>
>>>> Signed-off-by: Lizhi Hou <lizhi.hou@amd.com>
>>>> Signed-off-by: Sonal Santan <sonal.santan@amd.com>
>>>> Signed-off-by: Max Zhen <max.zhen@amd.com>
>>>> Signed-off-by: Brian Xu <brian.xu@amd.com>
>>>> ---
>>>>   MAINTAINERS                  |  1 +
>>>>   drivers/dma/xilinx/xdma.c    | 65 
>>>> ++++++++++++++++++++++++++++++++++++
>>>>   include/linux/dma/amd_xdma.h | 16 +++++++++
>>>>   3 files changed, 82 insertions(+)
>>>>   create mode 100644 include/linux/dma/amd_xdma.h
>>>>
>>>> diff --git a/MAINTAINERS b/MAINTAINERS
>>>> index c1be0b2e378a..019d84b2b086 100644
>>>> --- a/MAINTAINERS
>>>> +++ b/MAINTAINERS
>>>> @@ -21691,6 +21691,7 @@ L:    dmaengine@vger.kernel.org
>>>>   S:    Supported
>>>>   F:    drivers/dma/xilinx/xdma-regs.h
>>>>   F:    drivers/dma/xilinx/xdma.c
>>>> +F:    include/linux/dma/amd_xdma.h
>>>>   F:    include/linux/platform_data/amd_xdma.h
>>>>     XILINX ZYNQMP DPDMA DRIVER
>>>> diff --git a/drivers/dma/xilinx/xdma.c b/drivers/dma/xilinx/xdma.c
>>>> index 58a57e03bef5..13f627445c4a 100644
>>>> --- a/drivers/dma/xilinx/xdma.c
>>>> +++ b/drivers/dma/xilinx/xdma.c
>>>> @@ -25,6 +25,7 @@
>>>>   #include <linux/dmapool.h>
>>>>   #include <linux/regmap.h>
>>>>   #include <linux/dmaengine.h>
>>>> +#include <linux/dma/amd_xdma.h>
>>>>   #include <linux/platform_device.h>
>>>>   #include <linux/platform_data/amd_xdma.h>
>>>>   #include <linux/dma-mapping.h>
>>>> @@ -736,6 +737,7 @@ static int xdma_set_vector_reg(struct 
>>>> xdma_device *xdev, u32 vec_tbl_start,
>>>>   static int xdma_irq_init(struct xdma_device *xdev)
>>>>   {
>>>>       u32 irq = xdev->irq_start;
>>>> +    u32 user_irq_start;
>>>>       int i, j, ret;
>>>>         /* return failure if there are not enough IRQs */
>>>> @@ -786,6 +788,18 @@ static int xdma_irq_init(struct xdma_device 
>>>> *xdev)
>>>>           goto failed;
>>>>       }
>>>>   +    /* config user IRQ registers if needed */
>>>> +    user_irq_start = xdev->h2c_chan_num + xdev->c2h_chan_num;
>>>> +    if (xdev->irq_num > user_irq_start) {
>>>> +        ret = xdma_set_vector_reg(xdev, XDMA_IRQ_USER_VEC_NUM,
>>>> +                      user_irq_start,
>>>> +                      xdev->irq_num - user_irq_start);
>>>> +        if (ret) {
>>>> +            xdma_err(xdev, "failed to set user vectors: %d", ret);
>>>> +            goto failed;
>>>> +        }
>>>> +    }
>>>> +
>>>>       /* enable interrupt */
>>>>       ret = xdma_enable_intr(xdev);
>>>>       if (ret) {
>>>> @@ -816,6 +830,57 @@ static bool xdma_filter_fn(struct dma_chan 
>>>> *chan, void *param)
>>>>       return true;
>>>>   }
>>>>   +/**
>>>> + * xdma_disable_user_irq - Disable user interrupt
>>>> + * @pdev: Pointer to the platform_device structure
>>>> + * @user_irq_index: User IRQ index
>>>> + */
>>>> +void xdma_disable_user_irq(struct platform_device *pdev, u32 
>>>> user_irq_index)
>>>> +{
>>>> +    struct xdma_device *xdev = platform_get_drvdata(pdev);
>>>> +
>>>> +    if (xdev->h2c_chan_num + xdev->c2h_chan_num + user_irq_index >=
>>>> +        xdev->irq_num) {
>>>> +        xdma_err(xdev, "invalid user irq index");
>>>> +        return;
>>>> +    }
>>>> +
>>>> +    xdma_write_reg(xdev, XDMA_IRQ_BASE, XDMA_IRQ_USER_INT_EN_W1C,
>>>> +               (1 << user_irq_index));
>>>> +}
>>>> +EXPORT_SYMBOL(xdma_disable_user_irq);
>>>> +
>>>> +/**
>>>> + * xdma_enable_user_irq - Enable user logic interrupt
>>>> + * @pdev: Pointer to the platform_device structure
>>>> + * @user_irq_index: User logic IRQ wire index
>>>> + * @irq: Pointer to returning system IRQ number
>>>> + */
>>>> +int xdma_enable_user_irq(struct platform_device *pdev, u32 
>>>> user_irq_index,
>>>> +             u32 *irq)
>>>> +{
>>>> +    struct xdma_device *xdev = platform_get_drvdata(pdev);
>>>> +    u32 user_irq;
>>>> +    int ret;
>>>> +
>>>> +    user_irq = xdev->h2c_chan_num + xdev->c2h_chan_num + 
>>>> user_irq_index;
>>>> +    if (user_irq >= xdev->irq_num) {
>>>> +        xdma_err(xdev, "invalid user irq index");
>>>> +        return -EINVAL;
>>>> +    }
>>>> +
>>>> +    ret = xdma_write_reg(xdev, XDMA_IRQ_BASE, 
>>>> XDMA_IRQ_USER_INT_EN_W1S,
>>>> +                 (1 << user_irq_index));
>>>> +    if (ret) {
>>>> +        xdma_err(xdev, "set user irq mask failed, %d", ret);
>>>> +        return ret;
>>>> +    }
>>>> +    *irq = user_irq + xdev->irq_start;
>>>> +
>>>> +    return 0;
>>>> +}
>>>> +EXPORT_SYMBOL(xdma_enable_user_irq);
>>>> +
>>>>   /**
>>>>    * xdma_remove - Driver remove function
>>>>    * @pdev: Pointer to the platform_device structure
>>>> diff --git a/include/linux/dma/amd_xdma.h 
>>>> b/include/linux/dma/amd_xdma.h
>>>> new file mode 100644
>>>> index 000000000000..91fb02ff93a7
>>>> --- /dev/null
>>>> +++ b/include/linux/dma/amd_xdma.h
>>>> @@ -0,0 +1,16 @@
>>>> +/* SPDX-License-Identifier: GPL-2.0-or-later */
>>>> +/*
>>>> + * Copyright (C) 2022, Advanced Micro Devices, Inc.
>>>> + */
>>>> +
>>>> +#ifndef _DMAENGINE_AMD_XDMA_H
>>>> +#define _DMAENGINE_AMD_XDMA_H
>>>> +
>>>> +#include <linux/interrupt.h>
>>>> +#include <linux/platform_device.h>
>>>> +
>>>> +int xdma_enable_user_irq(struct platform_device *pdev, u32 
>>>> user_irq_index,
>>>> +             u32 *irq);
>>>> +void xdma_disable_user_irq(struct platform_device *pdev, u32 
>>>> user_irq_index);
>>>> +
>>>> +#endif /* _DMAENGINE_AMD_XDMA_H */
>>>
>>> Hi,
>>> While rewriting our V4L2 driver to use your XDMA driver, I realized, 
>>> that the API for the user interrupts is still not fully usable. If
>>> the expected outcome is that the "parent" driver using the xdma 
>>> driver knows nothing about the IRQ allocation in XDMA, then the 
>>> xdma_enable_user_irq() function must be split into two separate 
>>> functions:
>>>
>>> int xdma_enable_user_irq(struct platform_device *pdev, u32 
>>> user_irq_index)
>>>
>>> and
>>>
>>> int xdma_get_system_irq(u32 user_irq_index)
>>>
>>> that returns the system IRQ number for the given user_irq_index.
>>> Because without it you can not get the system IRQ number without 
>>> enabling the irq at the same time which is usually not what you want.
>>
>> Is this because the user logic you are using does not have 
>> disable/enable interrupt by itself?  I thought that the user logic IP 
>> would have its own register to enable/disable interrupt.
>>
> Yes. Our HW has no "own" registers to disable the IRQs. As such 
> registers are already there in XDMA, there was no need for them.
>
>
>> And xdma_enable_user_irq() will only be called once for creating the 
>> user logic platform device. Then, user logic IP driver would use its 
>> own register to control interrupts. The existing platform driver does 
>> not call any xdma_* to control interrupt.
>>
> As you can see, there is HW that was designed to use the XDMA 
> registers during normal operation and separate enable/disable and irq 
> mapping functions are required. Maybe it is not the best HW design, 
> but it can evidently happen and then the driver author is screwed with 
> the merged API.

Ok. To support this case, I will separate the API as you mentioned.


Thanks,

Lizhi

>
> M.
>
>>
>> Thanks,
>>
>> Lizhi
>>
>>>
>>> As a workaround, you can compute the system IRQ yourself with the 
>>> knowledge of the XDMA internals (this is what I have to do at the 
>>> moment in our driver), but then the "u32 *irq" parameter becomes 
>>> useless and can be removed anyway.
>>>
>>> M.

Re: [PATCH V5 XDMA 1/2] dmaengine: xilinx: xdma: Add xilinx xdma driver

[Ilpo Järvinen]

[-- Attachment #1: Type: text/plain, Size: 3149 bytes --]

On Tue, 4 Oct 2022, Lizhi Hou wrote:

> 
> On 10/4/22 01:18, Ilpo Järvinen wrote:
> > On Wed, 28 Sep 2022, Lizhi Hou wrote:
> > 
> > > Add driver to enable PCIe board which uses XDMA (the DMA/Bridge Subsystem
> > > for PCI Express). For example, Xilinx Alveo PCIe devices.
> > >      https://www.xilinx.com/products/boards-and-kits/alveo.html
> > > 
> > > The XDMA engine support up to 4 Host to Card (H2C) and 4 Card to Host
> > > (C2H)
> > > channels. Memory transfers are specified on a per-channel basis in
> > > descriptor linked lists, which the DMA fetches from host memory and
> > > processes. Events such as descriptor completion and errors are signaled
> > > using interrupts. The hardware detail is provided by
> > >      https://docs.xilinx.com/r/en-US/pg195-pcie-dma/Introduction
> > > 
> > > This driver implements dmaengine APIs.
> > >      - probe the available DMA channels
> > >      - use dma_slave_map for channel lookup
> > >      - use virtual channel to manage dmaengine tx descriptors
> > >      - implement device_prep_slave_sg callback to handle host scatter
> > > gather
> > >        list
> > >      - implement device_config to config device address for DMA transfer
> > > 
> > > Signed-off-by: Lizhi Hou <lizhi.hou@amd.com>
> > > Signed-off-by: Sonal Santan <sonal.santan@amd.com>
> > > Signed-off-by: Max Zhen <max.zhen@amd.com>
> > > Signed-off-by: Brian Xu <brian.xu@amd.com>
> > > ---

> > > +	*chans = devm_kzalloc(&xdev->pdev->dev, sizeof(**chans) * (*chan_num),
> > > +			      GFP_KERNEL);
> > > +	if (!*chans)
> > > +		return -ENOMEM;
> > > +
> > > +	for (i = 0, j = 0; i < pdata->max_dma_channels; i++) {
> > > +		ret = xdma_read_reg(xdev, base + i * XDMA_CHAN_STRIDE,
> > > +				    XDMA_CHAN_IDENTIFIER, &identifier);
> > > +		if (ret) {
> > > +			xdma_err(xdev, "failed to read channel id: %d", ret);
> > > +			return ret;
> > > +		}
> > Is it ok to not rollback the allocation in case of an error occurs?
> 
> In this loop, the failures are returned by read/write registers. The
> read/write register failure indicates serious hardware issue and the hardware
> may not be rollback in this situation.

What I meant is that you allocated memory above (to *chans, see above). 
Shouldn't that memory be free in case the hw is not working before you 
return the error from this function?

Check also the other returns below for the same problemx.

> > > +
> > > +		if (j == *chan_num) {
> > > +			xdma_err(xdev, "invalid channel number");
> > > +			return -EIO;
> > > +		}
> > Same here?
> > 
> > > +
> > > +		/* init channel structure and hardware */
> > > +		(*chans)[j].xdev_hdl = xdev;
> > > +		(*chans)[j].base = base + i * XDMA_CHAN_STRIDE;
> > > +		(*chans)[j].dir = dir;
> > > +
> > > +		ret = xdma_channel_init(&(*chans)[j]);
> > > +		if (ret)
> > > +			return ret;
> > And here.
> > 
> > > +		(*chans)[j].vchan.desc_free = xdma_free_desc;
> > > +		vchan_init(&(*chans)[j].vchan, &xdev->dma_dev);
> > > +
> > > +		j++;
> > > +	}
> > > +
> > > +	dev_info(&xdev->pdev->dev, "configured %d %s channels", j,
> > > +		 (dir == DMA_MEM_TO_DEV) ? "H2C" : "C2H");
> > > +
> > > +	return 0;
> > > +}


-- 
 i.

Re: [PATCH V5 XDMA 1/2] dmaengine: xilinx: xdma: Add xilinx xdma driver

[Lizhi Hou]

On 10/4/22 09:43, Ilpo Järvinen wrote:
> On Tue, 4 Oct 2022, Lizhi Hou wrote:
>
>> On 10/4/22 01:18, Ilpo Järvinen wrote:
>>> On Wed, 28 Sep 2022, Lizhi Hou wrote:
>>>
>>>> Add driver to enable PCIe board which uses XDMA (the DMA/Bridge Subsystem
>>>> for PCI Express). For example, Xilinx Alveo PCIe devices.
>>>>       https://www.xilinx.com/products/boards-and-kits/alveo.html
>>>>
>>>> The XDMA engine support up to 4 Host to Card (H2C) and 4 Card to Host
>>>> (C2H)
>>>> channels. Memory transfers are specified on a per-channel basis in
>>>> descriptor linked lists, which the DMA fetches from host memory and
>>>> processes. Events such as descriptor completion and errors are signaled
>>>> using interrupts. The hardware detail is provided by
>>>>       https://docs.xilinx.com/r/en-US/pg195-pcie-dma/Introduction
>>>>
>>>> This driver implements dmaengine APIs.
>>>>       - probe the available DMA channels
>>>>       - use dma_slave_map for channel lookup
>>>>       - use virtual channel to manage dmaengine tx descriptors
>>>>       - implement device_prep_slave_sg callback to handle host scatter
>>>> gather
>>>>         list
>>>>       - implement device_config to config device address for DMA transfer
>>>>
>>>> Signed-off-by: Lizhi Hou <lizhi.hou@amd.com>
>>>> Signed-off-by: Sonal Santan <sonal.santan@amd.com>
>>>> Signed-off-by: Max Zhen <max.zhen@amd.com>
>>>> Signed-off-by: Brian Xu <brian.xu@amd.com>
>>>> ---
>>>> +	*chans = devm_kzalloc(&xdev->pdev->dev, sizeof(**chans) * (*chan_num),
>>>> +			      GFP_KERNEL);
>>>> +	if (!*chans)
>>>> +		return -ENOMEM;
>>>> +
>>>> +	for (i = 0, j = 0; i < pdata->max_dma_channels; i++) {
>>>> +		ret = xdma_read_reg(xdev, base + i * XDMA_CHAN_STRIDE,
>>>> +				    XDMA_CHAN_IDENTIFIER, &identifier);
>>>> +		if (ret) {
>>>> +			xdma_err(xdev, "failed to read channel id: %d", ret);
>>>> +			return ret;
>>>> +		}
>>> Is it ok to not rollback the allocation in case of an error occurs?
>> In this loop, the failures are returned by read/write registers. The
>> read/write register failure indicates serious hardware issue and the hardware
>> may not be rollback in this situation.
> What I meant is that you allocated memory above (to *chans, see above).
> Shouldn't that memory be free in case the hw is not working before you
> return the error from this function?
>
> Check also the other returns below for the same problemx.

The memory does not need to be freed immediately. And it should not be 
memory leak because devm_* is used.

Thanks,

Lizhi

>>>> +
>>>> +		if (j == *chan_num) {
>>>> +			xdma_err(xdev, "invalid channel number");
>>>> +			return -EIO;
>>>> +		}
>>> Same here?
>>>
>>>> +
>>>> +		/* init channel structure and hardware */
>>>> +		(*chans)[j].xdev_hdl = xdev;
>>>> +		(*chans)[j].base = base + i * XDMA_CHAN_STRIDE;
>>>> +		(*chans)[j].dir = dir;
>>>> +
>>>> +		ret = xdma_channel_init(&(*chans)[j]);
>>>> +		if (ret)
>>>> +			return ret;
>>> And here.
>>>
>>>> +		(*chans)[j].vchan.desc_free = xdma_free_desc;
>>>> +		vchan_init(&(*chans)[j].vchan, &xdev->dma_dev);
>>>> +
>>>> +		j++;
>>>> +	}
>>>> +
>>>> +	dev_info(&xdev->pdev->dev, "configured %d %s channels", j,
>>>> +		 (dir == DMA_MEM_TO_DEV) ? "H2C" : "C2H");
>>>> +
>>>> +	return 0;
>>>> +}
>

Re: [PATCH V5 XDMA 1/2] dmaengine: xilinx: xdma: Add xilinx xdma driver

[Ilpo Järvinen]

[-- Attachment #1: Type: text/plain, Size: 2971 bytes --]

On Tue, 4 Oct 2022, Lizhi Hou wrote:

> 
> On 10/4/22 09:43, Ilpo Järvinen wrote:
> > On Tue, 4 Oct 2022, Lizhi Hou wrote:
> > 
> > > On 10/4/22 01:18, Ilpo Järvinen wrote:
> > > > On Wed, 28 Sep 2022, Lizhi Hou wrote:
> > > > 
> > > > > Add driver to enable PCIe board which uses XDMA (the DMA/Bridge
> > > > > Subsystem
> > > > > for PCI Express). For example, Xilinx Alveo PCIe devices.
> > > > >       https://www.xilinx.com/products/boards-and-kits/alveo.html
> > > > > 
> > > > > The XDMA engine support up to 4 Host to Card (H2C) and 4 Card to Host
> > > > > (C2H)
> > > > > channels. Memory transfers are specified on a per-channel basis in
> > > > > descriptor linked lists, which the DMA fetches from host memory and
> > > > > processes. Events such as descriptor completion and errors are
> > > > > signaled
> > > > > using interrupts. The hardware detail is provided by
> > > > >       https://docs.xilinx.com/r/en-US/pg195-pcie-dma/Introduction
> > > > > 
> > > > > This driver implements dmaengine APIs.
> > > > >       - probe the available DMA channels
> > > > >       - use dma_slave_map for channel lookup
> > > > >       - use virtual channel to manage dmaengine tx descriptors
> > > > >       - implement device_prep_slave_sg callback to handle host scatter
> > > > > gather
> > > > >         list
> > > > >       - implement device_config to config device address for DMA
> > > > > transfer
> > > > > 
> > > > > Signed-off-by: Lizhi Hou <lizhi.hou@amd.com>
> > > > > Signed-off-by: Sonal Santan <sonal.santan@amd.com>
> > > > > Signed-off-by: Max Zhen <max.zhen@amd.com>
> > > > > Signed-off-by: Brian Xu <brian.xu@amd.com>
> > > > > ---
> > > > > +	*chans = devm_kzalloc(&xdev->pdev->dev, sizeof(**chans) *
> > > > > (*chan_num),
> > > > > +			      GFP_KERNEL);
> > > > > +	if (!*chans)
> > > > > +		return -ENOMEM;
> > > > > +
> > > > > +	for (i = 0, j = 0; i < pdata->max_dma_channels; i++) {
> > > > > +		ret = xdma_read_reg(xdev, base + i * XDMA_CHAN_STRIDE,
> > > > > +				    XDMA_CHAN_IDENTIFIER,
> > > > > &identifier);
> > > > > +		if (ret) {
> > > > > +			xdma_err(xdev, "failed to read channel id:
> > > > > %d", ret);
> > > > > +			return ret;
> > > > > +		}
> > > > Is it ok to not rollback the allocation in case of an error occurs?
> > > In this loop, the failures are returned by read/write registers. The
> > > read/write register failure indicates serious hardware issue and the
> > > hardware
> > > may not be rollback in this situation.
> > What I meant is that you allocated memory above (to *chans, see above).
> > Shouldn't that memory be free in case the hw is not working before you
> > return the error from this function?
> > 
> > Check also the other returns below for the same problemx.
> 
> The memory does not need to be freed immediately. And it should not be memory
> leak because devm_* is used.

Ah, sorry. I think I checked exactly that there wasn't m in it but 
clearly there is now that I recheck.


-- 
 i.