Add AMD MDB Endpoint and non-LL mode Support

PATCH Superseded Devendra K Verma 2 participants 6 messages Mar 10 – Mar 11

Subsystems (3)

PCI dmaengine DesignWare edma core IP driver

lore

Share & tools

lore

Devendra K Verma

author

2026-03-10 12:31 UTC #0 lore raw

v12 cover Add AMD MDB Endpoint and non-LL mode Support

This series of patch support the following:

- AMD MDB Endpoint Support, as part of this patch following are added:
o AMD supported device ID and vendor ID (Xilinx)
o AMD MDB specific driver data
o AMD specific VSEC capabilities to retrieve the base of phys address of MDB side DDR
o Logic to assign the offsets to LL and data blocks if more number of channels are enabled than configured in the given pci_data struct.

- Addition of non-LL mode
o The IP supported non-LL mode functions
o Flexibility to choose non-LL mode via dma_slave_config param peripheral_config, by the client for all the vendors using HDMA IP.
o Allow IP utilization if LL mode is not available

Devendra K Verma (2):
dmaengine: dw-edma: Add AMD MDB Endpoint Support
dmaengine: dw-edma: Add non-LL mode

 drivers/dma/dw-edma/dw-edma-core.c    |  47 +++++-
 drivers/dma/dw-edma/dw-edma-core.h    |   1 +
 drivers/dma/dw-edma/dw-edma-pcie.c    | 220 +++++++++++++++++++++++---
 drivers/dma/dw-edma/dw-hdma-v0-core.c |  65 +++++++-
 drivers/dma/dw-edma/dw-hdma-v0-regs.h |   1 +
 include/linux/dma/edma.h              |   1 +
 6 files changed, 313 insertions(+), 22 deletions(-)

Signature · 1 line

2.43.0

Devendra K Verma

author

2026-03-10 12:31 UTC #1 lore raw

PATCH v12 1/2 dmaengine: dw-edma: Add AMD MDB Endpoint Support

AMD MDB PCIe endpoint support. For AMD specific support added the following
- AMD supported PCIe Device IDs and Vendor ID (Xilinx).
- AMD MDB specific driver data
- AMD MDB specific VSEC capability to retrieve the device DDR base address.

Signed-off-by Devendra K Verma <[email protected]>

---
Changes in v12:
No Changes

Changes in v11:
Replaced min_t() function with min().

Changes in v10:
For Xilinx VSEC function kept only HDMA map format as Xilinx only supports HDMA.

Changes in v9:
Moved Xilinx specific VSEC capability functions under the vendor ID condition.

Changes in v8:
Changed the contant names to includer product vendor. Moved the vendor specific code to vendor specific functions.

Changes in v7:
Introduced vendor specific functions to retrieve the vsec data.

Changes in v6:
Included "sizes.h" header and used the appropriate definitions instead of constants.

Changes in v5:
Added the definitions for Xilinx specific VSEC header id, revision, and register offsets.
Corrected the error type when no physical offset found for device side memory.
Corrected the order of variables.

Changes in v4:
Configured 8 read and 8 write channels for Xilinx vendor Added checks to validate vendor ID for vendor specific vsec id.
Added Xilinx specific vendor id for vsec specific to Xilinx Added the LL and data region offsets, size as input params to function dw_edma_set_chan_region_offset().
Moved the LL and data region offsets assignment to function for Xilinx specific case.
Corrected comments.

Changes in v3:
Corrected a typo when assigning AMD (Xilinx) vsec id macro and condition check.

Changes in v2:
Reverted the devmem_phys_off type to u64. Renamed the function appropriately to suit the functionality for setting the LL & data region offsets.

Changes in v1:
Removed the pci device id from pci_ids.h file.
Added the vendor id macro as per the suggested method. Changed the type of the newly added devmem_phys_off variable. Added to logic to assign offsets for LL and data region blocks in case more number of channels are enabled than given in amd_mdb_data struct.

---
 drivers/dma/dw-edma/dw-edma-pcie.c | 190 ++++++++++++++++++++++++++---
 1 file changed, 176 insertions(+), 14 deletions(-)

drivers/dma/dw-edma/dw-edma-pcie.c +176 ~~-14~~

index 3371e0a76d3c..b8208186a250 100644
@@ -14,14 +14,35 @@
 #include <linux/pci-epf.h>
 #include <linux/msi.h>
 #include <linux/bitfield.h>
+#include <linux/sizes.h>
 
 #include "dw-edma-core.h"
 
-#define DW_PCIE_VSEC_DMA_ID			0x6
-#define DW_PCIE_VSEC_DMA_BAR			GENMASK(10, 8)
-#define DW_PCIE_VSEC_DMA_MAP			GENMASK(2, 0)
-#define DW_PCIE_VSEC_DMA_WR_CH			GENMASK(9, 0)
-#define DW_PCIE_VSEC_DMA_RD_CH			GENMASK(25, 16)
+/* Synopsys */
+#define DW_PCIE_SYNOPSYS_VSEC_DMA_ID		0x6
+#define DW_PCIE_SYNOPSYS_VSEC_DMA_BAR		GENMASK(10, 8)
+#define DW_PCIE_SYNOPSYS_VSEC_DMA_MAP		GENMASK(2, 0)
+#define DW_PCIE_SYNOPSYS_VSEC_DMA_WR_CH		GENMASK(9, 0)
+#define DW_PCIE_SYNOPSYS_VSEC_DMA_RD_CH		GENMASK(25, 16)
+
+/* AMD MDB (Xilinx) specific defines */
+#define PCI_DEVICE_ID_XILINX_B054		0xb054
+
+#define DW_PCIE_XILINX_MDB_VSEC_DMA_ID		0x6
+#define DW_PCIE_XILINX_MDB_VSEC_ID		0x20
+#define DW_PCIE_XILINX_MDB_VSEC_DMA_BAR		GENMASK(10, 8)
+#define DW_PCIE_XILINX_MDB_VSEC_DMA_MAP		GENMASK(2, 0)
+#define DW_PCIE_XILINX_MDB_VSEC_DMA_WR_CH	GENMASK(9, 0)
+#define DW_PCIE_XILINX_MDB_VSEC_DMA_RD_CH	GENMASK(25, 16)
+
+#define DW_PCIE_XILINX_MDB_DEVMEM_OFF_REG_HIGH	0xc
+#define DW_PCIE_XILINX_MDB_DEVMEM_OFF_REG_LOW	0x8
+#define DW_PCIE_XILINX_MDB_INVALID_ADDR		(~0ULL)
+
+#define DW_PCIE_XILINX_MDB_LL_OFF_GAP		0x200000
+#define DW_PCIE_XILINX_MDB_LL_SIZE		0x800
+#define DW_PCIE_XILINX_MDB_DT_OFF_GAP		0x100000
+#define DW_PCIE_XILINX_MDB_DT_SIZE		0x800
 
 #define DW_BLOCK(a, b, c) \
 	{ \

@@ -50,6 +71,7 @@ struct dw_edma_pcie_data {
 	u8				irqs;
 	u16				wr_ch_cnt;
 	u16				rd_ch_cnt;
+	u64				devmem_phys_off;
 };
 
 static const struct dw_edma_pcie_data snps_edda_data = {

@@ -90,6 +112,64 @@ static const struct dw_edma_pcie_data snps_edda_data = {
 	.rd_ch_cnt			= 2,
 };
 
+static const struct dw_edma_pcie_data xilinx_mdb_data = {
+	/* MDB registers location */
+	.rg.bar				= BAR_0,
+	.rg.off				= SZ_4K,	/*  4 Kbytes */
+	.rg.sz				= SZ_8K,	/*  8 Kbytes */
+
+	/* Other */
+	.mf				= EDMA_MF_HDMA_NATIVE,
+	.irqs				= 1,
+	.wr_ch_cnt			= 8,
+	.rd_ch_cnt			= 8,
+};
+
+static void dw_edma_set_chan_region_offset(struct dw_edma_pcie_data *pdata,
+					   enum pci_barno bar, off_t start_off,
+					   off_t ll_off_gap, size_t ll_size,
+					   off_t dt_off_gap, size_t dt_size)
+{
+	u16 wr_ch = pdata->wr_ch_cnt;
+	u16 rd_ch = pdata->rd_ch_cnt;
+	off_t off;
+	u16 i;
+
+	off = start_off;
+
+	/* Write channel LL region */
+	for (i = 0; i < wr_ch; i++) {
+		pdata->ll_wr[i].bar = bar;
+		pdata->ll_wr[i].off = off;
+		pdata->ll_wr[i].sz = ll_size;
+		off += ll_off_gap;
+	}
+
+	/* Read channel LL region */
+	for (i = 0; i < rd_ch; i++) {
+		pdata->ll_rd[i].bar = bar;
+		pdata->ll_rd[i].off = off;
+		pdata->ll_rd[i].sz = ll_size;
+		off += ll_off_gap;
+	}
+
+	/* Write channel data region */
+	for (i = 0; i < wr_ch; i++) {
+		pdata->dt_wr[i].bar = bar;
+		pdata->dt_wr[i].off = off;
+		pdata->dt_wr[i].sz = dt_size;
+		off += dt_off_gap;
+	}
+
+	/* Read channel data region */
+	for (i = 0; i < rd_ch; i++) {
+		pdata->dt_rd[i].bar = bar;
+		pdata->dt_rd[i].off = off;
+		pdata->dt_rd[i].sz = dt_size;
+		off += dt_off_gap;
+	}
+}
+
 static int dw_edma_pcie_irq_vector(struct device *dev, unsigned int nr)
 {
 	return pci_irq_vector(to_pci_dev(dev), nr);

@@ -114,15 +194,15 @@ static const struct dw_edma_plat_ops dw_edma_pcie_plat_ops = {
 	.pci_address = dw_edma_pcie_address,
 };
 
-static void dw_edma_pcie_get_vsec_dma_data(struct pci_dev *pdev,
-					   struct dw_edma_pcie_data *pdata)
+static void dw_edma_pcie_get_synopsys_dma_data(struct pci_dev *pdev,
+					       struct dw_edma_pcie_data *pdata)
 {
 	u32 val, map;
 	u16 vsec;
 	u64 off;
 
 	vsec = pci_find_vsec_capability(pdev, PCI_VENDOR_ID_SYNOPSYS,
-					DW_PCIE_VSEC_DMA_ID);
+					DW_PCIE_SYNOPSYS_VSEC_DMA_ID);
 	if (!vsec)
 		return;

@@ -131,9 +211,9 @@ static void dw_edma_pcie_get_vsec_dma_data(struct pci_dev *pdev,
 	    PCI_VNDR_HEADER_LEN(val) != 0x18)
 		return;
 
-	pci_dbg(pdev, "Detected PCIe Vendor-Specific Extended Capability DMA\n");
+	pci_dbg(pdev, "Detected Synopsys PCIe Vendor-Specific Extended Capability DMA\n");
 	pci_read_config_dword(pdev, vsec + 0x8, &val);
-	map = FIELD_GET(DW_PCIE_VSEC_DMA_MAP, val);
+	map = FIELD_GET(DW_PCIE_SYNOPSYS_VSEC_DMA_MAP, val);
 	if (map != EDMA_MF_EDMA_LEGACY &&
 	    map != EDMA_MF_EDMA_UNROLL &&
 	    map != EDMA_MF_HDMA_COMPAT &&

@@ -141,13 +221,13 @@ static void dw_edma_pcie_get_vsec_dma_data(struct pci_dev *pdev,
 		return;
 
 	pdata->mf = map;
-	pdata->rg.bar = FIELD_GET(DW_PCIE_VSEC_DMA_BAR, val);
+	pdata->rg.bar = FIELD_GET(DW_PCIE_SYNOPSYS_VSEC_DMA_BAR, val);
 
 	pci_read_config_dword(pdev, vsec + 0xc, &val);
 	pdata->wr_ch_cnt = min_t(u16, pdata->wr_ch_cnt,
-				 FIELD_GET(DW_PCIE_VSEC_DMA_WR_CH, val));
+				 FIELD_GET(DW_PCIE_SYNOPSYS_VSEC_DMA_WR_CH, val));
 	pdata->rd_ch_cnt = min_t(u16, pdata->rd_ch_cnt,
-				 FIELD_GET(DW_PCIE_VSEC_DMA_RD_CH, val));
+				 FIELD_GET(DW_PCIE_SYNOPSYS_VSEC_DMA_RD_CH, val));
 
 	pci_read_config_dword(pdev, vsec + 0x14, &val);
 	off = val;

@@ -157,6 +237,64 @@ static void dw_edma_pcie_get_vsec_dma_data(struct pci_dev *pdev,
 	pdata->rg.off = off;
 }
 
+static void dw_edma_pcie_get_xilinx_dma_data(struct pci_dev *pdev,
+					     struct dw_edma_pcie_data *pdata)
+{
+	u32 val, map;
+	u16 vsec;
+	u64 off;
+
+	pdata->devmem_phys_off = DW_PCIE_XILINX_MDB_INVALID_ADDR;
+
+	vsec = pci_find_vsec_capability(pdev, PCI_VENDOR_ID_XILINX,
+					DW_PCIE_XILINX_MDB_VSEC_DMA_ID);
+	if (!vsec)
+		return;
+
+	pci_read_config_dword(pdev, vsec + PCI_VNDR_HEADER, &val);
+	if (PCI_VNDR_HEADER_REV(val) != 0x00 ||
+	    PCI_VNDR_HEADER_LEN(val) != 0x18)
+		return;
+
+	pci_dbg(pdev, "Detected Xilinx PCIe Vendor-Specific Extended Capability DMA\n");
+	pci_read_config_dword(pdev, vsec + 0x8, &val);
+	map = FIELD_GET(DW_PCIE_XILINX_MDB_VSEC_DMA_MAP, val);
+	if (map != EDMA_MF_HDMA_NATIVE)
+		return;
+
+	pdata->mf = map;
+	pdata->rg.bar = FIELD_GET(DW_PCIE_XILINX_MDB_VSEC_DMA_BAR, val);
+
+	pci_read_config_dword(pdev, vsec + 0xc, &val);
+	pdata->wr_ch_cnt = min(pdata->wr_ch_cnt,
+			       FIELD_GET(DW_PCIE_XILINX_MDB_VSEC_DMA_WR_CH, val));
+	pdata->rd_ch_cnt = min(pdata->rd_ch_cnt,
+			       FIELD_GET(DW_PCIE_XILINX_MDB_VSEC_DMA_RD_CH, val));
+
+	pci_read_config_dword(pdev, vsec + 0x14, &val);
+	off = val;
+	pci_read_config_dword(pdev, vsec + 0x10, &val);
+	off <<= 32;
+	off |= val;
+	pdata->rg.off = off;
+
+	vsec = pci_find_vsec_capability(pdev, PCI_VENDOR_ID_XILINX,
+					DW_PCIE_XILINX_MDB_VSEC_ID);
+	if (!vsec)
+		return;
+
+	pci_read_config_dword(pdev,
+			      vsec + DW_PCIE_XILINX_MDB_DEVMEM_OFF_REG_HIGH,
+			      &val);
+	off = val;
+	pci_read_config_dword(pdev,
+			      vsec + DW_PCIE_XILINX_MDB_DEVMEM_OFF_REG_LOW,
+			      &val);
+	off <<= 32;
+	off |= val;
+	pdata->devmem_phys_off = off;
+}
+
 static int dw_edma_pcie_probe(struct pci_dev *pdev,
 			      const struct pci_device_id *pid)
 {

@@ -184,7 +322,29 @@ static int dw_edma_pcie_probe(struct pci_dev *pdev,
 	 * Tries to find if exists a PCIe Vendor-Specific Extended Capability
 	 * for the DMA, if one exists, then reconfigures it.
 	 */
-	dw_edma_pcie_get_vsec_dma_data(pdev, vsec_data);
+	dw_edma_pcie_get_synopsys_dma_data(pdev, vsec_data);
+
+	if (pdev->vendor == PCI_VENDOR_ID_XILINX) {
+		dw_edma_pcie_get_xilinx_dma_data(pdev, vsec_data);
+
+		/*
+		 * There is no valid address found for the LL memory
+		 * space on the device side.
+		 */
+		if (vsec_data->devmem_phys_off == DW_PCIE_XILINX_MDB_INVALID_ADDR)
+			return -ENOMEM;
+
+		/*
+		 * Configure the channel LL and data blocks if number of
+		 * channels enabled in VSEC capability are more than the
+		 * channels configured in xilinx_mdb_data.
+		 */
+		dw_edma_set_chan_region_offset(vsec_data, BAR_2, 0,
+					       DW_PCIE_XILINX_MDB_LL_OFF_GAP,
+					       DW_PCIE_XILINX_MDB_LL_SIZE,
+					       DW_PCIE_XILINX_MDB_DT_OFF_GAP,
+					       DW_PCIE_XILINX_MDB_DT_SIZE);
+	}
 
 	/* Mapping PCI BAR regions */
 	mask = BIT(vsec_data->rg.bar);

@@ -367,6 +527,8 @@ static void dw_edma_pcie_remove(struct pci_dev *pdev)
 
 static const struct pci_device_id dw_edma_pcie_id_table[] = {
 	{ PCI_DEVICE_DATA(SYNOPSYS, EDDA, &snps_edda_data) },
+	{ PCI_VDEVICE(XILINX, PCI_DEVICE_ID_XILINX_B054),
+	  (kernel_ulong_t)&xilinx_mdb_data },
 	{ }
 };
 MODULE_DEVICE_TABLE(pci, dw_edma_pcie_id_table);

Signature · 1 line

2.43.0

Frank Li

Reviewed-by on v12 1/2

2026-03-10 18:59 UTC #2 lore raw

Re: [PATCH v12 1/2] dmaengine: dw-edma: Add AMD MDB Endpoint Support

Quoted 9 lines · Devendra K Verma

AMD MDB PCIe endpoint support. For AMD specific support added the following
- AMD supported PCIe Device IDs and Vendor ID (Xilinx).
- AMD MDB specific driver data
- AMD MDB specific VSEC capability to retrieve the device DDR base address.
Signed-off-by Devendra K Verma <[email protected]>
---

You missed collect my review by at v11

Reviewed-by Frank Li <[email protected]>

Quoted 338 lines

Changes in v12:
No Changes

Changes in v11:
Replaced min_t() function with min().

Changes in v10:
For Xilinx VSEC function kept only HDMA map format as Xilinx only supports HDMA.

Changes in v9:
Moved Xilinx specific VSEC capability functions under the vendor ID condition.

Changes in v8:
Changed the contant names to includer product vendor. Moved the vendor specific code to vendor specific functions.

Changes in v7:
Introduced vendor specific functions to retrieve the vsec data.

Changes in v6:
Included "sizes.h" header and used the appropriate definitions instead of constants.

Changes in v3:
Corrected a typo when assigning AMD (Xilinx) vsec id macro and condition check.

Changes in v2:
Reverted the devmem_phys_off type to u64. Renamed the function appropriately to suit the functionality for setting the LL & data region offsets.

---
 drivers/dma/dw-edma/dw-edma-pcie.c | 190 ++++++++++++++++++++++++++---
 1 file changed, 176 insertions(+), 14 deletions(-)

drivers/dma/dw-edma/dw-edma-pcie.c +176 ~~-15~~

index 3371e0a76d3c..b8208186a250 100644
@@ -14,14 +14,35 @@
 #include <linux/pci-epf.h>
 #include <linux/msi.h>
 #include <linux/bitfield.h>
+#include <linux/sizes.h>
 
 #include "dw-edma-core.h"
 
-#define DW_PCIE_VSEC_DMA_ID			0x6
-#define DW_PCIE_VSEC_DMA_BAR			GENMASK(10, 8)
-#define DW_PCIE_VSEC_DMA_MAP			GENMASK(2, 0)
-#define DW_PCIE_VSEC_DMA_WR_CH			GENMASK(9, 0)
-#define DW_PCIE_VSEC_DMA_RD_CH			GENMASK(25, 16)
+/* Synopsys */
+#define DW_PCIE_SYNOPSYS_VSEC_DMA_ID		0x6
+#define DW_PCIE_SYNOPSYS_VSEC_DMA_BAR		GENMASK(10, 8)
+#define DW_PCIE_SYNOPSYS_VSEC_DMA_MAP		GENMASK(2, 0)
+#define DW_PCIE_SYNOPSYS_VSEC_DMA_WR_CH		GENMASK(9, 0)
+#define DW_PCIE_SYNOPSYS_VSEC_DMA_RD_CH		GENMASK(25, 16)
+
+/* AMD MDB (Xilinx) specific defines */
+#define PCI_DEVICE_ID_XILINX_B054		0xb054
+
+#define DW_PCIE_XILINX_MDB_VSEC_DMA_ID		0x6
+#define DW_PCIE_XILINX_MDB_VSEC_ID		0x20
+#define DW_PCIE_XILINX_MDB_VSEC_DMA_BAR		GENMASK(10, 8)
+#define DW_PCIE_XILINX_MDB_VSEC_DMA_MAP		GENMASK(2, 0)
+#define DW_PCIE_XILINX_MDB_VSEC_DMA_WR_CH	GENMASK(9, 0)
+#define DW_PCIE_XILINX_MDB_VSEC_DMA_RD_CH	GENMASK(25, 16)
+
+#define DW_PCIE_XILINX_MDB_DEVMEM_OFF_REG_HIGH	0xc
+#define DW_PCIE_XILINX_MDB_DEVMEM_OFF_REG_LOW	0x8
+#define DW_PCIE_XILINX_MDB_INVALID_ADDR		(~0ULL)
+
+#define DW_PCIE_XILINX_MDB_LL_OFF_GAP		0x200000
+#define DW_PCIE_XILINX_MDB_LL_SIZE		0x800
+#define DW_PCIE_XILINX_MDB_DT_OFF_GAP		0x100000
+#define DW_PCIE_XILINX_MDB_DT_SIZE		0x800
 
 #define DW_BLOCK(a, b, c) \
 	{ \

@@ -50,6 +71,7 @@ struct dw_edma_pcie_data {
 	u8				irqs;
 	u16				wr_ch_cnt;
 	u16				rd_ch_cnt;
+	u64				devmem_phys_off;
 };
 
 static const struct dw_edma_pcie_data snps_edda_data = {

@@ -90,6 +112,64 @@ static const struct dw_edma_pcie_data snps_edda_data = {
 	.rd_ch_cnt			= 2,
 };
 
+static const struct dw_edma_pcie_data xilinx_mdb_data = {
+	/* MDB registers location */
+	.rg.bar				= BAR_0,
+	.rg.off				= SZ_4K,	/*  4 Kbytes */
+	.rg.sz				= SZ_8K,	/*  8 Kbytes */
+
+	/* Other */
+	.mf				= EDMA_MF_HDMA_NATIVE,
+	.irqs				= 1,
+	.wr_ch_cnt			= 8,
+	.rd_ch_cnt			= 8,
+};
+
+static void dw_edma_set_chan_region_offset(struct dw_edma_pcie_data *pdata,
+					   enum pci_barno bar, off_t start_off,
+					   off_t ll_off_gap, size_t ll_size,
+					   off_t dt_off_gap, size_t dt_size)
+{
+	u16 wr_ch = pdata->wr_ch_cnt;
+	u16 rd_ch = pdata->rd_ch_cnt;
+	off_t off;
+	u16 i;
+
+	off = start_off;
+
+	/* Write channel LL region */
+	for (i = 0; i < wr_ch; i++) {
+		pdata->ll_wr[i].bar = bar;
+		pdata->ll_wr[i].off = off;
+		pdata->ll_wr[i].sz = ll_size;
+		off += ll_off_gap;
+	}
+
+	/* Read channel LL region */
+	for (i = 0; i < rd_ch; i++) {
+		pdata->ll_rd[i].bar = bar;
+		pdata->ll_rd[i].off = off;
+		pdata->ll_rd[i].sz = ll_size;
+		off += ll_off_gap;
+	}
+
+	/* Write channel data region */
+	for (i = 0; i < wr_ch; i++) {
+		pdata->dt_wr[i].bar = bar;
+		pdata->dt_wr[i].off = off;
+		pdata->dt_wr[i].sz = dt_size;
+		off += dt_off_gap;
+	}
+
+	/* Read channel data region */
+	for (i = 0; i < rd_ch; i++) {
+		pdata->dt_rd[i].bar = bar;
+		pdata->dt_rd[i].off = off;
+		pdata->dt_rd[i].sz = dt_size;
+		off += dt_off_gap;
+	}
+}
+
 static int dw_edma_pcie_irq_vector(struct device *dev, unsigned int nr)
 {
 	return pci_irq_vector(to_pci_dev(dev), nr);

@@ -114,15 +194,15 @@ static const struct dw_edma_plat_ops dw_edma_pcie_plat_ops = {
 	.pci_address = dw_edma_pcie_address,
 };
 
-static void dw_edma_pcie_get_vsec_dma_data(struct pci_dev *pdev,
-					   struct dw_edma_pcie_data *pdata)
+static void dw_edma_pcie_get_synopsys_dma_data(struct pci_dev *pdev,
+					       struct dw_edma_pcie_data *pdata)
 {
 	u32 val, map;
 	u16 vsec;
 	u64 off;
 
 	vsec = pci_find_vsec_capability(pdev, PCI_VENDOR_ID_SYNOPSYS,
-					DW_PCIE_VSEC_DMA_ID);
+					DW_PCIE_SYNOPSYS_VSEC_DMA_ID);
 	if (!vsec)
 		return;

@@ -131,9 +211,9 @@ static void dw_edma_pcie_get_vsec_dma_data(struct pci_dev *pdev,
 	    PCI_VNDR_HEADER_LEN(val) != 0x18)
 		return;
 
-	pci_dbg(pdev, "Detected PCIe Vendor-Specific Extended Capability DMA\n");
+	pci_dbg(pdev, "Detected Synopsys PCIe Vendor-Specific Extended Capability DMA\n");
 	pci_read_config_dword(pdev, vsec + 0x8, &val);
-	map = FIELD_GET(DW_PCIE_VSEC_DMA_MAP, val);
+	map = FIELD_GET(DW_PCIE_SYNOPSYS_VSEC_DMA_MAP, val);
 	if (map != EDMA_MF_EDMA_LEGACY &&
 	    map != EDMA_MF_EDMA_UNROLL &&
 	    map != EDMA_MF_HDMA_COMPAT &&

@@ -141,13 +221,13 @@ static void dw_edma_pcie_get_vsec_dma_data(struct pci_dev *pdev,
 		return;
 
 	pdata->mf = map;
-	pdata->rg.bar = FIELD_GET(DW_PCIE_VSEC_DMA_BAR, val);
+	pdata->rg.bar = FIELD_GET(DW_PCIE_SYNOPSYS_VSEC_DMA_BAR, val);
 
 	pci_read_config_dword(pdev, vsec + 0xc, &val);
 	pdata->wr_ch_cnt = min_t(u16, pdata->wr_ch_cnt,
-				 FIELD_GET(DW_PCIE_VSEC_DMA_WR_CH, val));
+				 FIELD_GET(DW_PCIE_SYNOPSYS_VSEC_DMA_WR_CH, val));
 	pdata->rd_ch_cnt = min_t(u16, pdata->rd_ch_cnt,
-				 FIELD_GET(DW_PCIE_VSEC_DMA_RD_CH, val));
+				 FIELD_GET(DW_PCIE_SYNOPSYS_VSEC_DMA_RD_CH, val));
 
 	pci_read_config_dword(pdev, vsec + 0x14, &val);
 	off = val;

@@ -157,6 +237,64 @@ static void dw_edma_pcie_get_vsec_dma_data(struct pci_dev *pdev,
 	pdata->rg.off = off;
 }
 
+static void dw_edma_pcie_get_xilinx_dma_data(struct pci_dev *pdev,
+					     struct dw_edma_pcie_data *pdata)
+{
+	u32 val, map;
+	u16 vsec;
+	u64 off;
+
+	pdata->devmem_phys_off = DW_PCIE_XILINX_MDB_INVALID_ADDR;
+
+	vsec = pci_find_vsec_capability(pdev, PCI_VENDOR_ID_XILINX,
+					DW_PCIE_XILINX_MDB_VSEC_DMA_ID);
+	if (!vsec)
+		return;
+
+	pci_read_config_dword(pdev, vsec + PCI_VNDR_HEADER, &val);
+	if (PCI_VNDR_HEADER_REV(val) != 0x00 ||
+	    PCI_VNDR_HEADER_LEN(val) != 0x18)
+		return;
+
+	pci_dbg(pdev, "Detected Xilinx PCIe Vendor-Specific Extended Capability DMA\n");
+	pci_read_config_dword(pdev, vsec + 0x8, &val);
+	map = FIELD_GET(DW_PCIE_XILINX_MDB_VSEC_DMA_MAP, val);
+	if (map != EDMA_MF_HDMA_NATIVE)
+		return;
+
+	pdata->mf = map;
+	pdata->rg.bar = FIELD_GET(DW_PCIE_XILINX_MDB_VSEC_DMA_BAR, val);
+
+	pci_read_config_dword(pdev, vsec + 0xc, &val);
+	pdata->wr_ch_cnt = min(pdata->wr_ch_cnt,
+			       FIELD_GET(DW_PCIE_XILINX_MDB_VSEC_DMA_WR_CH, val));
+	pdata->rd_ch_cnt = min(pdata->rd_ch_cnt,
+			       FIELD_GET(DW_PCIE_XILINX_MDB_VSEC_DMA_RD_CH, val));
+
+	pci_read_config_dword(pdev, vsec + 0x14, &val);
+	off = val;
+	pci_read_config_dword(pdev, vsec + 0x10, &val);
+	off <<= 32;
+	off |= val;
+	pdata->rg.off = off;
+
+	vsec = pci_find_vsec_capability(pdev, PCI_VENDOR_ID_XILINX,
+					DW_PCIE_XILINX_MDB_VSEC_ID);
+	if (!vsec)
+		return;
+
+	pci_read_config_dword(pdev,
+			      vsec + DW_PCIE_XILINX_MDB_DEVMEM_OFF_REG_HIGH,
+			      &val);
+	off = val;
+	pci_read_config_dword(pdev,
+			      vsec + DW_PCIE_XILINX_MDB_DEVMEM_OFF_REG_LOW,
+			      &val);
+	off <<= 32;
+	off |= val;
+	pdata->devmem_phys_off = off;
+}
+
 static int dw_edma_pcie_probe(struct pci_dev *pdev,
 			      const struct pci_device_id *pid)
 {

@@ -184,7 +322,29 @@ static int dw_edma_pcie_probe(struct pci_dev *pdev,
 	 * Tries to find if exists a PCIe Vendor-Specific Extended Capability
 	 * for the DMA, if one exists, then reconfigures it.
 	 */
-	dw_edma_pcie_get_vsec_dma_data(pdev, vsec_data);
+	dw_edma_pcie_get_synopsys_dma_data(pdev, vsec_data);
+
+	if (pdev->vendor == PCI_VENDOR_ID_XILINX) {
+		dw_edma_pcie_get_xilinx_dma_data(pdev, vsec_data);
+
+		/*
+		 * There is no valid address found for the LL memory
+		 * space on the device side.
+		 */
+		if (vsec_data->devmem_phys_off == DW_PCIE_XILINX_MDB_INVALID_ADDR)
+			return -ENOMEM;
+
+		/*
+		 * Configure the channel LL and data blocks if number of
+		 * channels enabled in VSEC capability are more than the
+		 * channels configured in xilinx_mdb_data.
+		 */
+		dw_edma_set_chan_region_offset(vsec_data, BAR_2, 0,
+					       DW_PCIE_XILINX_MDB_LL_OFF_GAP,
+					       DW_PCIE_XILINX_MDB_LL_SIZE,
+					       DW_PCIE_XILINX_MDB_DT_OFF_GAP,
+					       DW_PCIE_XILINX_MDB_DT_SIZE);
+	}
 
 	/* Mapping PCI BAR regions */
 	mask = BIT(vsec_data->rg.bar);

@@ -367,6 +527,8 @@ static void dw_edma_pcie_remove(struct pci_dev *pdev)
 
 static const struct pci_device_id dw_edma_pcie_id_table[] = {
 	{ PCI_DEVICE_DATA(SYNOPSYS, EDDA, &snps_edda_data) },
+	{ PCI_VDEVICE(XILINX, PCI_DEVICE_ID_XILINX_B054),
+	  (kernel_ulong_t)&xilinx_mdb_data },
 	{ }
 };
 MODULE_DEVICE_TABLE(pci, dw_edma_pcie_id_table);
--

2.43.0

Devendra K Verma

author

2026-03-10 12:31 UTC #3 lore raw

PATCH v12 2/2 dmaengine: dw-edma: Add non-LL mode

In reply to Devendra K Verma

AMD MDB IP supports Linked List (LL) mode as well as non-LL mode. The current code does not have the mechanisms to enable the DMA transactions using the non-LL mode. The following two cases are added with this patch:
- For the AMD (Xilinx) only, when a valid physical base address of the device side DDR is not configured, then the IP can still be used in non-LL mode. For all the channels DMA transactions will be using the non-LL mode only. This, the default non-LL mode, is not applicable for Synopsys IP with the current code addition.

- If the default mode is LL-mode, for both AMD (Xilinx) and Synosys, and if user wants to use non-LL mode then user can do so via configuring the peripheral_config param of dma_slave_config.

Signed-off-by Devendra K Verma <[email protected]>

---
Changes in v12
Simplified the logic to set the channel non_ll flag

Changes in v11
Added error check for the following
If user enables LL mode when non-LL is default mode. If config_param provided by user for eDMA
Renamed the 'non_ll' flag in dw_edma_chip to 'cfg_non_ll'

Changes in v10
Added the peripheral_config check only for HDMA IP in dw_edma_device_config().
Replaced the loop with single entry retrieval for non-LL mode.
Addressed review comments and handled the burst allocation by defining 'bursts_max' as per suggestions.

Changes in v9
Fixed compilation errors related to macro name mismatch.

Changes in v8
Cosmetic change related to comment and code.

Changes in v7
No change

Changes in v6
Gave definition to bits used for channel configuration. Removed the comment related to doorbell.

Changes in v5
Variable name 'nollp' changed to 'non_ll'.
In the dw_edma_device_config() WARN_ON replaced with dev_err(). Comments follow the 80-column guideline.

Changes in v4
No change

Changes in v3
No change

Changes in v2
Reverted the function return type to u64 for
dw_edma_get_phys_addr().

Changes in v1
Changed the function return type for dw_edma_get_phys_addr(). Corrected the typo raised in review.

---
 drivers/dma/dw-edma/dw-edma-core.c    | 47 ++++++++++++++++++-
 drivers/dma/dw-edma/dw-edma-core.h    |  1 +
 drivers/dma/dw-edma/dw-edma-pcie.c    | 44 ++++++++++++------
 drivers/dma/dw-edma/dw-hdma-v0-core.c | 65 ++++++++++++++++++++++++++-
 drivers/dma/dw-edma/dw-hdma-v0-regs.h |  1 +
 include/linux/dma/edma.h              |  1 +
 6 files changed, 144 insertions(+), 15 deletions(-)

drivers/dma/dw-edma/dw-edma-core.c +46 -1

index b43255f914f3..1e3266daccc8 100644
@@ -223,6 +223,43 @@ static int dw_edma_device_config(struct dma_chan *dchan,
 				 struct dma_slave_config *config)
 {
 	struct dw_edma_chan *chan = dchan2dw_edma_chan(dchan);
+	bool cfg_non_ll;
+	int non_ll = 0;
+
+	chan->non_ll = false;
+	if (chan->dw->chip->mf == EDMA_MF_HDMA_NATIVE) {
+		if (config->peripheral_config &&
+		    config->peripheral_size != sizeof(int)) {
+			dev_err(dchan->device->dev,
+				"config param peripheral size mismatch\n");
+			return -EINVAL;
+		}
+
+		/*
+		 * When there is no valid LLP base address available then the
+		 * default DMA ops will use the non-LL mode.
+		 *
+		 * Cases where LL mode is enabled and client wants to use the
+		 * non-LL mode then also client can do so via providing the
+		 * peripheral_config param.
+		 */
+		cfg_non_ll = chan->dw->chip->cfg_non_ll;
+		if (config->peripheral_config) {
+			non_ll = *(int *)config->peripheral_config;
+
+			if (cfg_non_ll && !non_ll) {
+				dev_err(dchan->device->dev, "invalid configuration\n");
+				return -EINVAL;
+			}
+		}
+
+		if (cfg_non_ll || non_ll)
+			chan->non_ll = true;
+	} else if (config->peripheral_config) {
+		dev_err(dchan->device->dev,
+			"peripheral config param applicable only for HDMA\n");
+		return -EINVAL;
+	}
 
 	memcpy(&chan->config, config, sizeof(*config));
 	chan->configured = true;

@@ -358,6 +395,7 @@ dw_edma_device_transfer(struct dw_edma_transfer *xfer)
 	struct dw_edma_desc *desc;
 	u64 src_addr, dst_addr;
 	size_t fsz = 0;
+	u32 bursts_max;
 	u32 cnt = 0;
 	int i;

@@ -415,6 +453,13 @@ dw_edma_device_transfer(struct dw_edma_transfer *xfer)
 		return NULL;
 	}
 
+	/*
+	 * For non-LL mode, only a single burst can be handled
+	 * in a single chunk unlike LL mode where multiple bursts
+	 * can be configured in a single chunk.
+	 */
+	bursts_max = chan->non_ll ? 1 : chan->ll_max;
+
 	desc = dw_edma_alloc_desc(chan);
 	if (unlikely(!desc))
 		goto err_alloc;

@@ -450,7 +495,7 @@ dw_edma_device_transfer(struct dw_edma_transfer *xfer)
 		if (xfer->type == EDMA_XFER_SCATTER_GATHER && !sg)
 			break;
 
-		if (chunk->bursts_alloc == chan->ll_max) {
+		if (chunk->bursts_alloc == bursts_max) {
 			chunk = dw_edma_alloc_chunk(desc);
 			if (unlikely(!chunk))
 				goto err_alloc;

drivers/dma/dw-edma/dw-edma-core.h +1 -0

index 71894b9e0b15..c8e3d196a549 100644
@@ -86,6 +86,7 @@ struct dw_edma_chan {
 	u8				configured;
 
 	struct dma_slave_config		config;
+	bool				non_ll;
 };
 
 struct dw_edma_irq {

drivers/dma/dw-edma/dw-edma-pcie.c +31 ~~-13~~

110 lines not on this page · raw message

drivers/dma/dw-edma/dw-hdma-v0-core.c +64 -1

77 lines not on this page · raw message

drivers/dma/dw-edma/dw-hdma-v0-regs.h +1 -0

index eab5fd7177e5..7759ba9b4850 100644
@@ -12,6 +12,7 @@
 #include <linux/dmaengine.h>
 
 #define HDMA_V0_MAX_NR_CH			8
+#define HDMA_V0_CH_EN				BIT(0)
 #define HDMA_V0_LOCAL_ABORT_INT_EN		BIT(6)
 #define HDMA_V0_REMOTE_ABORT_INT_EN		BIT(5)
 #define HDMA_V0_LOCAL_STOP_INT_EN		BIT(4)

include/linux/dma/edma.h +1 -0

index 3080747689f6..bb88439bd091 100644
@@ -99,6 +99,7 @@ struct dw_edma_chip {
 	enum dw_edma_map_format	mf;
 
 	struct dw_edma		*dw;
+	bool			cfg_non_ll;
 };
 
 /* Export to the platform drivers */

Signature · 1 line

2.43.0

Frank Li

Reviewed-by on v12 2/2

2026-03-10 19:47 UTC #4 lore raw

Re: [PATCH v12 2/2] dmaengine: dw-edma: Add non-LL mode

Quoted 16 lines · Devendra K Verma

AMD MDB IP supports Linked List (LL) mode as well as non-LL mode. The current code does not have the mechanisms to enable the DMA transactions using the non-LL mode. The following two cases are added with this patch:
- For the AMD (Xilinx) only, when a valid physical base address of the device side DDR is not configured, then the IP can still be used in non-LL mode. For all the channels DMA transactions will be using the non-LL mode only. This, the default non-LL mode, is not applicable for Synopsys IP with the current code addition.
- If the default mode is LL-mode, for both AMD (Xilinx) and Synosys, and if user wants to use non-LL mode then user can do so via configuring the peripheral_config param of dma_slave_config.
Signed-off-by Devendra K Verma <[email protected]>
---

...

Quoted 7 lines

drivers/dma/dw-edma/dw-edma-pcie.c +0 -0

index b8208186a250..f538d728609f 100644
@@ -295,6 +295,15 @@ static void dw_edma_pcie_get_xilinx_dma_data(struct pci_dev *pdev,
 	pdata->devmem_phys_off = off;
 }

drivers/dma/dw-edma/dw-edma-pcie.c static void dw_edma_pcie_get_xilinx_dma_data(struct pci_dev *pdev, · line 306

Quoted 10 lines

drivers/dma/dw-edma/dw-edma-pcie.c +9 -0

 
+static u64 dw_edma_get_phys_addr(struct pci_dev *pdev,
+				 struct dw_edma_pcie_data *pdata,
+				 enum pci_barno bar)
+{
+	if (pdev->vendor == PCI_VENDOR_ID_XILINX)
+		return pdata->devmem_phys_off;
+	return pci_bus_address(pdev, bar);
+}
+

You missed my previous review feedback about create new patch for code restructure. But change related small.

Reviewed-by Frank Li <[email protected]>

Quoted 14 lines

drivers/dma/dw-edma/dw-edma-pcie.c +1 -1

-	for (i = 0; i < chip->ll_wr_cnt; i++) {
+	for (i = 0; i < chip->ll_wr_cnt && !non_ll; i++) {
 		struct dw_edma_region *ll_region = &chip->ll_region_wr[i];
 		struct dw_edma_region *dt_region = &chip->dt_region_wr[i];

drivers/dma/dw-edma/dw-edma-pcie.c +2 -1

 		struct dw_edma_block *ll_block = &vsec_data->ll_wr[i];

@@ -410,7 +424,8 @@ static int dw_edma_pcie_probe(struct pci_dev *pdev,
 			return -ENOMEM;
 
 		ll_region->vaddr.io += ll_block->off;
-		ll_region->paddr = pci_bus_address(pdev, ll_block->bar);
+		ll_region->paddr = dw_edma_get_phys_addr(pdev, vsec_data,
+							 ll_block->bar);

Verma, Devendra

author

2026-03-11 08:41 UTC #5 lore raw

RE: [PATCH v12 2/2] dmaengine: dw-edma: Add non-LL mode

[Public]

Quoted 52 lines

-----Original Message-----
From: Frank Li <[email protected]>
Sent: Wednesday, March 11, 2026 1:17 AM
To: Verma, Devendra <[email protected]>
Cc: [email protected]; [email protected]; [email protected]; [email protected]; [email protected]; linux- [email protected]; Simek, Michal <[email protected]>
Subject: Re: [PATCH v12 2/2] dmaengine: dw-edma: Add non-LL mode
Caution: This message originated from an External Source. Use proper caution when opening attachments, clicking links, or responding.
Quoted 16 lines · Devendra K Verma
AMD MDB IP supports Linked List (LL) mode as well as non-LL mode. The current code does not have the mechanisms to enable the DMA transactions using the non-LL mode. The following two cases are added with this patch:
- For the AMD (Xilinx) only, when a valid physical base address of the device side DDR is not configured, then the IP can still be used in non-LL mode. For all the channels DMA transactions will be using the non-LL mode only. This, the default non-LL mode, is not applicable for Synopsys IP with the current code addition.
- If the default mode is LL-mode, for both AMD (Xilinx) and Synosys, and if user wants to use non-LL mode then user can do so via configuring the peripheral_config param of dma_slave_config.
Signed-off-by Devendra K Verma <[email protected]>
---
...
drivers/dma/dw-edma/dw-edma-pcie.c static void
Quoted 6 lines
drivers/dma/dw-edma/dw-edma-pcie.c +0 -0
index b8208186a250..f538d728609f 100644
@@ -295,6 +295,15 @@ static void
dw_edma_pcie_get_xilinx_dma_data(struct pci_dev *pdev,
drivers/dma/dw-edma/dw-edma-pcie.c static void dw_edma_pcie_get_xilinx_dma_data(struct pci_dev *pdev, · line 306
Quoted 10 lines
drivers/dma/dw-edma/dw-edma-pcie.c +7 -0
      pdata->devmem_phys_off = off;
 }
 
+static u64 dw_edma_get_phys_addr(struct pci_dev *pdev,
+                              struct dw_edma_pcie_data *pdata,
+                              enum pci_barno bar) {
+     if (pdev->vendor == PCI_VENDOR_ID_XILINX)
+             return pdata->devmem_phys_off;
+     return pci_bus_address(pdev, bar); }
+
You missed my previous review feedback about create new patch for code restructure. But change related small.

Quoted 1 line

Reviewed-by Frank Li <[email protected]>

Thank you Frank!
I think you missed in the mails, I provided the explanation why this request was not taken up. But you can take a look at the discussion happen on the following link:
https://lore.kernel.org/all/aXe5ts7E6lUF7YRq@lizhi-Precision-Tower-5810/

Please check for the string "Thank you for your persevere."

I will issue another revision with the review tags collected. Thanks!

Quoted 16 lines

drivers/dma/dw-edma/dw-edma-pcie.c static int dw_edma_pcie_probe(struct pci_dev

Quoted 8 lines

drivers/dma/dw-edma/dw-edma-pcie.c +1 -1

-     for (i = 0; i < chip->ll_wr_cnt; i++) {
+     for (i = 0; i < chip->ll_wr_cnt && !non_ll; i++) {
              struct dw_edma_region *ll_region = &chip->ll_region_wr[i];
              struct dw_edma_region *dt_region = &chip->dt_region_wr[i];
              struct dw_edma_block *ll_block = &vsec_data->ll_wr[i];

@@ -410,7 +424,8 @@ static int dw_edma_pcie_probe(struct pci_dev

*pdev,

Quoted 7 lines

drivers/dma/dw-edma/dw-edma-pcie.c +2 -1

                      return -ENOMEM;
 
              ll_region->vaddr.io += ll_block->off;
-             ll_region->paddr = pci_bus_address(pdev, ll_block->bar);
+             ll_region->paddr = dw_edma_get_phys_addr(pdev, vsec_data,
+                                                      ll_block->bar);