Determine the number of DMA channels by 'dma-channels' property

PATCH Applied Zong Li 5 participants 24 messages Mar 28, 2022 – Apr 21, 2022

Subsystems (5)

Device tree bindings RISC-V dmaengine SiFive drivers +1 more subsystems

lore

Share & tools

lore

Zong Li

author

2022-03-28 09:52 UTC #0 lore raw

v8 cover Determine the number of DMA channels by 'dma-channels' property

The PDMA driver currently assumes there are four channels by default, it might cause the error if there is actually less than four channels. Change that by getting number of channel dynamically from device tree. For backwards-compatible, it uses the default value (i.e. 4) when there is no 'dma-channels' information in dts.

This patch set contains the dts and dt-bindings change.

Changed in v8:
- Rebase on master
- Remove modification of microchip-mpfs.dtsi
- Rename DMA node name of fu540-c000.dtsi

Changed in v7:
- Rebase on tag v5.17-rc7
- Modify the subject of patch

Changed in v6:
- Rebase on tag v5.17-rc6
- Change sf_pdma.chans[] to a flexible array member.

Changed in v5:
- Rebase on tag v5.17-rc3
- Fix typo in dt-bindings and commit message
- Add PDMA versioning scheme for compatible

Changed in v4:
- Remove cflags of debug use reported-by: kernel test robot <[email protected]>

Changed in v3:
- Fix allocating wrong size
- Return error if 'dma-channels' is larger than maximum

Changed in v2:
- Rebase on tag v5.16
- Use 4 as default value of dma-channels

Zong Li (4):
dt-bindings: dma-engine: sifive,fu540: Add dma-channels property and
modify compatible
riscv: dts: Add dma-channels property and modify compatible
riscv: dts: rename the node name of dma
dmaengine: sf-pdma: Get number of channel by device tree

 .../bindings/dma/sifive,fu540-c000-pdma.yaml  | 19 +++++++++++++--
 arch/riscv/boot/dts/sifive/fu540-c000.dtsi    |  5 ++--
 drivers/dma/sf-pdma/sf-pdma.c                 | 24 ++++++++++++-------
 drivers/dma/sf-pdma/sf-pdma.h                 |  8 ++-----
 4 files changed, 38 insertions(+), 18 deletions(-)

Signature · 1 line

2.35.1

Zong Li

author

2022-03-28 09:52 UTC #1 lore raw

PATCH v8 1/4 dt-bindings: dma-engine: sifive,fu540: Add dma-channels property and modify compatible

Add dma-channels property, then we can determine how many channels there by device tree, rather than statically defining it in PDMA driver. In addition, we also modify the compatible for PDMA versioning scheme.

Signed-off-by Zong Li <[email protected]>

Reviewed-by Rob Herring <[email protected]>

Reviewed-by Palmer Dabbelt <[email protected]>

Acked-by Palmer Dabbelt <[email protected]>

Acked-by Krzysztof Kozlowski <[email protected]>

---
 .../bindings/dma/sifive,fu540-c000-pdma.yaml  | 19 +++++++++++++++++--
 1 file changed, 17 insertions(+), 2 deletions(-)

Documentation/devicetree/bindings/dma/sifive,fu540-c000-pdma.yaml +17 -2

index 47c46af25536..3271755787b4 100644
@@ -28,7 +28,15 @@ allOf:
 properties:
   compatible:
     items:
-      - const: sifive,fu540-c000-pdma
+      - enum:
+          - sifive,fu540-c000-pdma
+      - const: sifive,pdma0
+    description:
+      Should be "sifive,<chip>-pdma" and "sifive,pdma<version>".
+      Supported compatible strings are -
+      "sifive,fu540-c000-pdma" for the SiFive PDMA v0 as integrated onto the
+      SiFive FU540 chip resp and "sifive,pdma0" for the SiFive PDMA v0 IP block
+      with no chip integration tweaks.
 
   reg:
     maxItems: 1

@@ -37,6 +45,12 @@ properties:
     minItems: 1
     maxItems: 8
 
+  dma-channels:
+    description: For backwards-compatibility, the default value is 4
+    minimum: 1
+    maximum: 4
+    default: 4
+
   '#dma-cells':
     const: 1

@@ -50,8 +64,9 @@ unevaluatedProperties: false
 examples:
   - |
     dma-controller@3000000 {
-      compatible = "sifive,fu540-c000-pdma";
+      compatible = "sifive,fu540-c000-pdma", "sifive,pdma0";
       reg = <0x3000000 0x8000>;
+      dma-channels = <4>;
       interrupts = <23>, <24>, <25>, <26>, <27>, <28>, <29>, <30>;
       #dma-cells = <1>;
     };

Signature · 1 line

2.35.1

Bin Meng

Reviewed-by on v8 1/4

2022-03-29 01:35 UTC #2 lore raw

Re: [PATCH v8 1/4] dt-bindings: dma-engine: sifive,fu540: Add dma-channels property and modify compatible

Quoted 11 lines · Zong Li

Add dma-channels property, then we can determine how many channels there by device tree, rather than statically defining it in PDMA driver. In addition, we also modify the compatible for PDMA versioning scheme.
Signed-off-by Zong Li <[email protected]>
Reviewed-by Rob Herring <[email protected]>
Suggested-by Palmer Dabbelt <[email protected]>
Reviewed-by Palmer Dabbelt <[email protected]>
Acked-by Palmer Dabbelt <[email protected]>
Acked-by Krzysztof Kozlowski <[email protected]>

Quoted 4 lines

---
 .../bindings/dma/sifive,fu540-c000-pdma.yaml  | 19 +++++++++++++++++--
 1 file changed, 17 insertions(+), 2 deletions(-)

Reviewed-by Bin Meng <[email protected]>

Zong Li

author

2022-03-28 09:52 UTC #3 lore raw

PATCH v8 2/4 riscv: dts: Add dma-channels property and modify compatible

In reply to Zong Li

Add dma-channels property, then we can determine how many channels there by device tree, in addition, we add the pdma versioning scheme for compatible.

Signed-off-by Zong Li <[email protected]>

Reviewed-by Palmer Dabbelt <[email protected]>

Acked-by Palmer Dabbelt <[email protected]>

Acked-by Conor Dooley <[email protected]>

---
 arch/riscv/boot/dts/sifive/fu540-c000.dtsi | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

arch/riscv/boot/dts/sifive/fu540-c000.dtsi +2 -1

index 3eef52b1a59b..6a3011180846 100644
@@ -168,11 +168,12 @@ uart0: serial@10010000 {
 			status = "disabled";
 		};
 		dma: dma@3000000 {
-			compatible = "sifive,fu540-c000-pdma";
+			compatible = "sifive,fu540-c000-pdma", "sifive,pdma0";
 			reg = <0x0 0x3000000 0x0 0x8000>;
 			interrupt-parent = <&plic0>;
 			interrupts = <23>, <24>, <25>, <26>, <27>, <28>, <29>,
 				     <30>;
+			dma-channels = <4>;
 			#dma-cells = <1>;
 		};
 		uart1: serial@10011000 {

Signature · 1 line

2.35.1

Bin Meng

Reviewed-by on v8 2/4

2022-03-29 01:35 UTC #4 lore raw

Re: [PATCH v8 2/4] riscv: dts: Add dma-channels property and modify compatible

Quoted 5 lines · Zong Li

Add dma-channels property, then we can determine how many channels there by device tree, in addition, we add the pdma versioning scheme for compatible.

Quoted 8 lines

Signed-off-by Zong Li <[email protected]>
Reviewed-by Palmer Dabbelt <[email protected]>
Acked-by Palmer Dabbelt <[email protected]>
Acked-by Conor Dooley <[email protected]>
---
 arch/riscv/boot/dts/sifive/fu540-c000.dtsi | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

Reviewed-by Bin Meng <[email protected]>

Zong Li

author

2022-03-28 09:52 UTC #5 lore raw

PATCH v8 3/4 riscv: dts: rename the node name of dma

In reply to Zong Li

Rename the node name by the generic DMA naming

Signed-off-by Zong Li <[email protected]>

CC Vinod Koul <[email protected]>

---
 arch/riscv/boot/dts/sifive/fu540-c000.dtsi | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

arch/riscv/boot/dts/sifive/fu540-c000.dtsi +1 -1

index 6a3011180846..6eebe34c2851 100644
@@ -167,7 +167,7 @@ uart0: serial@10010000 {
 			clocks = <&prci PRCI_CLK_TLCLK>;
 			status = "disabled";
 		};
-		dma: dma@3000000 {
+		dma: dma-controller@3000000 {
 			compatible = "sifive,fu540-c000-pdma", "sifive,pdma0";
 			reg = <0x0 0x3000000 0x0 0x8000>;
 			interrupt-parent = <&plic0>;

Signature · 1 line

2.35.1

Bin Meng

Reviewed-by on v8 3/4

2022-03-29 01:36 UTC #6 lore raw

Re: [PATCH v8 3/4] riscv: dts: rename the node name of dma

Quoted 9 lines · Zong Li

Rename the node name by the generic DMA naming
Signed-off-by Zong Li <[email protected]>
CC Vinod Koul <[email protected]>
---
 arch/riscv/boot/dts/sifive/fu540-c000.dtsi | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

Reviewed-by Bin Meng <[email protected]>

Zong Li

author

2022-03-28 09:52 UTC #7 lore raw

PATCH v8 4/4 dmaengine: sf-pdma: Get number of channel by device tree

In reply to Zong Li

It currently assumes that there are always four channels, it would cause the error if there is actually less than four channels. Change that by getting number of channel from device tree.

For backwards-compatibility, it uses the default value (i.e. 4) when there is no 'dma-channels' information in dts.

Signed-off-by Zong Li <[email protected]>

Acked-by Palmer Dabbelt <[email protected]>

---
 drivers/dma/sf-pdma/sf-pdma.c | 24 ++++++++++++++++--------
 drivers/dma/sf-pdma/sf-pdma.h |  8 ++------
 2 files changed, 18 insertions(+), 14 deletions(-)

drivers/dma/sf-pdma/sf-pdma.c +16 -8

index f12606aeff87..db5a4ef76077 100644
@@ -482,23 +482,30 @@ static void sf_pdma_setup_chans(struct sf_pdma *pdma)
 static int sf_pdma_probe(struct platform_device *pdev)
 {
 	struct sf_pdma *pdma;
-	struct sf_pdma_chan *chan;
 	struct resource *res;
-	int len, chans;
-	int ret;
+	int ret, n_chans;
 	const enum dma_slave_buswidth widths =
 		DMA_SLAVE_BUSWIDTH_1_BYTE | DMA_SLAVE_BUSWIDTH_2_BYTES |
 		DMA_SLAVE_BUSWIDTH_4_BYTES | DMA_SLAVE_BUSWIDTH_8_BYTES |
 		DMA_SLAVE_BUSWIDTH_16_BYTES | DMA_SLAVE_BUSWIDTH_32_BYTES |
 		DMA_SLAVE_BUSWIDTH_64_BYTES;
 
-	chans = PDMA_NR_CH;
-	len = sizeof(*pdma) + sizeof(*chan) * chans;
-	pdma = devm_kzalloc(&pdev->dev, len, GFP_KERNEL);
+	ret = of_property_read_u32(pdev->dev.of_node, "dma-channels", &n_chans);
+	if (ret) {
+		/* backwards-compatibility for no dma-channels property */
+		dev_dbg(&pdev->dev, "set number of channels to default value: 4\n");
+		n_chans = PDMA_MAX_NR_CH;
+	} else if (n_chans > PDMA_MAX_NR_CH) {
+		dev_err(&pdev->dev, "the number of channels exceeds the maximum\n");
+		return -EINVAL;
+	}
+
+	pdma = devm_kzalloc(&pdev->dev, struct_size(pdma, chans, n_chans),
+			    GFP_KERNEL);
 	if (!pdma)
 		return -ENOMEM;
 
-	pdma->n_chans = chans;
+	pdma->n_chans = n_chans;
 
 	res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
 	pdma->membase = devm_ioremap_resource(&pdev->dev, res);

@@ -556,7 +563,7 @@ static int sf_pdma_remove(struct platform_device *pdev)
 	struct sf_pdma_chan *ch;
 	int i;
 
-	for (i = 0; i < PDMA_NR_CH; i++) {
+	for (i = 0; i < pdma->n_chans; i++) {
 		ch = &pdma->chans[i];
 
 		devm_free_irq(&pdev->dev, ch->txirq, ch);

@@ -574,6 +581,7 @@ static int sf_pdma_remove(struct platform_device *pdev)
 
 static const struct of_device_id sf_pdma_dt_ids[] = {
 	{ .compatible = "sifive,fu540-c000-pdma" },
+	{ .compatible = "sifive,pdma0" },
 	{},
 };
 MODULE_DEVICE_TABLE(of, sf_pdma_dt_ids);

drivers/dma/sf-pdma/sf-pdma.h +2 -6

index 0c20167b097d..dcb3687bd5da 100644
@@ -22,11 +22,7 @@
 #include "../dmaengine.h"
 #include "../virt-dma.h"
 
-#define PDMA_NR_CH					4
-
-#if (PDMA_NR_CH != 4)
-#error "Please define PDMA_NR_CH to 4"
-#endif
+#define PDMA_MAX_NR_CH					4
 
 #define PDMA_BASE_ADDR					0x3000000
 #define PDMA_CHAN_OFFSET				0x1000

@@ -118,7 +114,7 @@ struct sf_pdma {
 	void __iomem            *membase;
 	void __iomem            *mappedbase;
 	u32			n_chans;
-	struct sf_pdma_chan	chans[PDMA_NR_CH];
+	struct sf_pdma_chan	chans[];
 };
 
 #endif /* _SF_PDMA_H */

Signature · 1 line

2.35.1

Bin Meng

Reviewed-by on v8 4/4

2022-03-29 01:36 UTC #8 lore raw

Re: [PATCH v8 4/4] dmaengine: sf-pdma: Get number of channel by device tree

Quoted 8 lines · Zong Li

It currently assumes that there are always four channels, it would cause the error if there is actually less than four channels. Change that by getting number of channel from device tree.
For backwards-compatibility, it uses the default value (i.e. 4) when there is no 'dma-channels' information in dts.

Quoted 7 lines

Signed-off-by Zong Li <[email protected]>
Acked-by Palmer Dabbelt <[email protected]>
---
 drivers/dma/sf-pdma/sf-pdma.c | 24 ++++++++++++++++--------
 drivers/dma/sf-pdma/sf-pdma.h |  8 ++------
 2 files changed, 18 insertions(+), 14 deletions(-)

Reviewed-by Bin Meng <[email protected]>

_______________________________________________
linux-riscv mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-riscv

Zong Li

author

2022-03-31 01:06 UTC #9 lore raw

In reply to Zong Li

Hi Vinod and Palmer,
This patchset got some reviewed-by and acked-by, if it is good to you as well, are you willing to help me to pick them? Thanks.

Quoted 54 lines · Zong Li

The PDMA driver currently assumes there are four channels by default, it might cause the error if there is actually less than four channels. Change that by getting number of channel dynamically from device tree. For backwards-compatible, it uses the default value (i.e. 4) when there is no 'dma-channels' information in dts.
This patch set contains the dts and dt-bindings change.
Changed in v8:
- Rebase on master
- Remove modification of microchip-mpfs.dtsi
- Rename DMA node name of fu540-c000.dtsi
Changed in v7:
- Rebase on tag v5.17-rc7
- Modify the subject of patch
Changed in v6:
- Rebase on tag v5.17-rc6
- Change sf_pdma.chans[] to a flexible array member.
Changed in v5:
- Rebase on tag v5.17-rc3
- Fix typo in dt-bindings and commit message
- Add PDMA versioning scheme for compatible
Changed in v4:
- Remove cflags of debug use reported-by: kernel test robot <[email protected]>
Changed in v3:
- Fix allocating wrong size
- Return error if 'dma-channels' is larger than maximum
Changed in v2:
- Rebase on tag v5.16
- Use 4 as default value of dma-channels
Zong Li (4):
dt-bindings: dma-engine: sifive,fu540: Add dma-channels property and
modify compatible
riscv: dts: Add dma-channels property and modify compatible
riscv: dts: rename the node name of dma
dmaengine: sf-pdma: Get number of channel by device tree
 .../bindings/dma/sifive,fu540-c000-pdma.yaml  | 19 +++++++++++++--
 arch/riscv/boot/dts/sifive/fu540-c000.dtsi    |  5 ++--
 drivers/dma/sf-pdma/sf-pdma.c                 | 24 ++++++++++++-------
 drivers/dma/sf-pdma/sf-pdma.h                 |  8 ++-----
 4 files changed, 38 insertions(+), 18 deletions(-)
--
2.35.1

Palmer Dabbelt

maintainer

2022-03-31 03:16 UTC #10 lore raw

In reply to Zong Li

Quoted 44 lines · On Mon, 28 Mar 2022 02:52:21 PDT (-0700), [email protected] wrote:

The PDMA driver currently assumes there are four channels by default, it might cause the error if there is actually less than four channels. Change that by getting number of channel dynamically from device tree. For backwards-compatible, it uses the default value (i.e. 4) when there is no 'dma-channels' information in dts.
This patch set contains the dts and dt-bindings change.
Changed in v8:
- Rebase on master
- Remove modification of microchip-mpfs.dtsi
- Rename DMA node name of fu540-c000.dtsi
Changed in v7:
- Rebase on tag v5.17-rc7
- Modify the subject of patch
Changed in v6:
- Rebase on tag v5.17-rc6
- Change sf_pdma.chans[] to a flexible array member.
Changed in v5:
- Rebase on tag v5.17-rc3
- Fix typo in dt-bindings and commit message
- Add PDMA versioning scheme for compatible
Changed in v4:
- Remove cflags of debug use reported-by: kernel test robot <[email protected]>
Changed in v3:
- Fix allocating wrong size
- Return error if 'dma-channels' is larger than maximum
Changed in v2:
- Rebase on tag v5.16
- Use 4 as default value of dma-channels
Zong Li (4):
dt-bindings: dma-engine: sifive,fu540: Add dma-channels property and
modify compatible
riscv: dts: Add dma-channels property and modify compatible
riscv: dts: rename the node name of dma
dmaengine: sf-pdma: Get number of channel by device tree

Quoted 5 lines

 .../bindings/dma/sifive,fu540-c000-pdma.yaml  | 19 +++++++++++++--
 arch/riscv/boot/dts/sifive/fu540-c000.dtsi    |  5 ++--
 drivers/dma/sf-pdma/sf-pdma.c                 | 24 ++++++++++++-------
 drivers/dma/sf-pdma/sf-pdma.h                 |  8 ++-----
 4 files changed, 38 insertions(+), 18 deletions(-)

Thanks, these are on for-next.

_______________________________________________
linux-riscv mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/linux-riscv

Vinod Koul

maintainer

2022-03-31 05:22 UTC #11 lore raw

Quoted 45 lines · Palmer Dabbelt

Quoted 44 lines · On Mon, 28 Mar 2022 02:52:21 PDT (-0700), [email protected] wrote:
The PDMA driver currently assumes there are four channels by default, it might cause the error if there is actually less than four channels. Change that by getting number of channel dynamically from device tree. For backwards-compatible, it uses the default value (i.e. 4) when there is no 'dma-channels' information in dts.
This patch set contains the dts and dt-bindings change.
Changed in v8:
- Rebase on master
- Remove modification of microchip-mpfs.dtsi
- Rename DMA node name of fu540-c000.dtsi
Changed in v7:
- Rebase on tag v5.17-rc7
- Modify the subject of patch
Changed in v6:
- Rebase on tag v5.17-rc6
- Change sf_pdma.chans[] to a flexible array member.
Changed in v5:
- Rebase on tag v5.17-rc3
- Fix typo in dt-bindings and commit message
- Add PDMA versioning scheme for compatible
Changed in v4:
- Remove cflags of debug use reported-by: kernel test robot <[email protected]>
Changed in v3:
- Fix allocating wrong size
- Return error if 'dma-channels' is larger than maximum
Changed in v2:
- Rebase on tag v5.16
- Use 4 as default value of dma-channels
Zong Li (4):
dt-bindings: dma-engine: sifive,fu540: Add dma-channels property and
modify compatible
riscv: dts: Add dma-channels property and modify compatible
riscv: dts: rename the node name of dma
dmaengine: sf-pdma: Get number of channel by device tree

Quoted 7 lines

Quoted 5 lines

 .../bindings/dma/sifive,fu540-c000-pdma.yaml  | 19 +++++++++++++--
 arch/riscv/boot/dts/sifive/fu540-c000.dtsi    |  5 ++--
 drivers/dma/sf-pdma/sf-pdma.c                 | 24 ++++++++++++-------
 drivers/dma/sf-pdma/sf-pdma.h                 |  8 ++-----
 4 files changed, 38 insertions(+), 18 deletions(-)

Thanks, these are on for-next.

The drivers/dma/ should go thru dmaengine tree. During merge window I dont apply the patches

Signature · 1 line

~Vinod

Palmer Dabbelt

maintainer

2022-03-31 05:54 UTC #12 lore raw

Quoted 52 lines · On Wed, 30 Mar 2022 22:21:56 PDT (-0700), [email protected] wrote:

Quoted 51 lines · Palmer Dabbelt
Quoted 49 lines · On Mon, 28 Mar 2022 02:52:21 PDT (-0700), [email protected] wrote:
The PDMA driver currently assumes there are four channels by default, it might cause the error if there is actually less than four channels. Change that by getting number of channel dynamically from device tree. For backwards-compatible, it uses the default value (i.e. 4) when there is no 'dma-channels' information in dts.
This patch set contains the dts and dt-bindings change.
Changed in v8:
- Rebase on master
- Remove modification of microchip-mpfs.dtsi
- Rename DMA node name of fu540-c000.dtsi
Changed in v7:
- Rebase on tag v5.17-rc7
- Modify the subject of patch
Changed in v6:
- Rebase on tag v5.17-rc6
- Change sf_pdma.chans[] to a flexible array member.
Changed in v5:
- Rebase on tag v5.17-rc3
- Fix typo in dt-bindings and commit message
- Add PDMA versioning scheme for compatible
Changed in v4:
- Remove cflags of debug use reported-by: kernel test robot <[email protected]>
Changed in v3:
- Fix allocating wrong size
- Return error if 'dma-channels' is larger than maximum
Changed in v2:
- Rebase on tag v5.16
- Use 4 as default value of dma-channels
Zong Li (4):
dt-bindings: dma-engine: sifive,fu540: Add dma-channels property and
modify compatible
riscv: dts: Add dma-channels property and modify compatible
riscv: dts: rename the node name of dma
dmaengine: sf-pdma: Get number of channel by device tree
 .../bindings/dma/sifive,fu540-c000-pdma.yaml  | 19 +++++++++++++--
 arch/riscv/boot/dts/sifive/fu540-c000.dtsi    |  5 ++--
 drivers/dma/sf-pdma/sf-pdma.c                 | 24 ++++++++++++-------
 drivers/dma/sf-pdma/sf-pdma.h                 |  8 ++-----
 4 files changed, 38 insertions(+), 18 deletions(-)

Quoted 4 lines

Quoted 1 line
Thanks, these are on for-next.
The drivers/dma/ should go thru dmaengine tree. During merge window I dont apply the patches

OK, I can drop this from my tree if you'd like?

Palmer Dabbelt

maintainer

2022-03-31 22:42 UTC #13 lore raw

Quoted 53 lines · Palmer Dabbelt

Quoted 52 lines · On Wed, 30 Mar 2022 22:21:56 PDT (-0700), [email protected] wrote:
Quoted 51 lines · Palmer Dabbelt
Quoted 49 lines · On Mon, 28 Mar 2022 02:52:21 PDT (-0700), [email protected] wrote:
The PDMA driver currently assumes there are four channels by default, it might cause the error if there is actually less than four channels. Change that by getting number of channel dynamically from device tree. For backwards-compatible, it uses the default value (i.e. 4) when there is no 'dma-channels' information in dts.
This patch set contains the dts and dt-bindings change.
Changed in v8:
- Rebase on master
- Remove modification of microchip-mpfs.dtsi
- Rename DMA node name of fu540-c000.dtsi
Changed in v7:
- Rebase on tag v5.17-rc7
- Modify the subject of patch
Changed in v6:
- Rebase on tag v5.17-rc6
- Change sf_pdma.chans[] to a flexible array member.
Changed in v5:
- Rebase on tag v5.17-rc3
- Fix typo in dt-bindings and commit message
- Add PDMA versioning scheme for compatible
Changed in v4:
- Remove cflags of debug use reported-by: kernel test robot <[email protected]>
Changed in v3:
- Fix allocating wrong size
- Return error if 'dma-channels' is larger than maximum
Changed in v2:
- Rebase on tag v5.16
- Use 4 as default value of dma-channels
Zong Li (4):
dt-bindings: dma-engine: sifive,fu540: Add dma-channels property and
modify compatible
riscv: dts: Add dma-channels property and modify compatible
riscv: dts: rename the node name of dma
dmaengine: sf-pdma: Get number of channel by device tree
 .../bindings/dma/sifive,fu540-c000-pdma.yaml  | 19 +++++++++++++--
 arch/riscv/boot/dts/sifive/fu540-c000.dtsi    |  5 ++--
 drivers/dma/sf-pdma/sf-pdma.c                 | 24 ++++++++++++-------
 drivers/dma/sf-pdma/sf-pdma.h                 |  8 ++-----
 4 files changed, 38 insertions(+), 18 deletions(-)

Quoted 6 lines

Quoted 4 lines
Quoted 1 line
Thanks, these are on for-next.
The drivers/dma/ should go thru dmaengine tree. During merge window I dont apply the patches
OK, I can drop this from my tree if you'd like?

Just to follow up from IRC: I'm dropping these from my tree.

Zong Li

author

2022-04-01 05:51 UTC #14 lore raw

Quoted 55 lines · Palmer Dabbelt

Quoted 53 lines · Palmer Dabbelt
Quoted 52 lines · On Wed, 30 Mar 2022 22:21:56 PDT (-0700), [email protected] wrote:
Quoted 51 lines · Palmer Dabbelt
Quoted 49 lines · On Mon, 28 Mar 2022 02:52:21 PDT (-0700), [email protected] wrote:
The PDMA driver currently assumes there are four channels by default, it might cause the error if there is actually less than four channels. Change that by getting number of channel dynamically from device tree. For backwards-compatible, it uses the default value (i.e. 4) when there is no 'dma-channels' information in dts.
This patch set contains the dts and dt-bindings change.
Changed in v8:
- Rebase on master
- Remove modification of microchip-mpfs.dtsi
- Rename DMA node name of fu540-c000.dtsi
Changed in v7:
- Rebase on tag v5.17-rc7
- Modify the subject of patch
Changed in v6:
- Rebase on tag v5.17-rc6
- Change sf_pdma.chans[] to a flexible array member.
Changed in v5:
- Rebase on tag v5.17-rc3
- Fix typo in dt-bindings and commit message
- Add PDMA versioning scheme for compatible
Changed in v4:
- Remove cflags of debug use reported-by: kernel test robot <[email protected]>
Changed in v3:
- Fix allocating wrong size
- Return error if 'dma-channels' is larger than maximum
Changed in v2:
- Rebase on tag v5.16
- Use 4 as default value of dma-channels
Zong Li (4):
dt-bindings: dma-engine: sifive,fu540: Add dma-channels property and
modify compatible
riscv: dts: Add dma-channels property and modify compatible
riscv: dts: rename the node name of dma
dmaengine: sf-pdma: Get number of channel by device tree
 .../bindings/dma/sifive,fu540-c000-pdma.yaml  | 19 +++++++++++++--
 arch/riscv/boot/dts/sifive/fu540-c000.dtsi    |  5 ++--
 drivers/dma/sf-pdma/sf-pdma.c                 | 24 ++++++++++++-------
 drivers/dma/sf-pdma/sf-pdma.h                 |  8 ++-----
 4 files changed, 38 insertions(+), 18 deletions(-)

Quoted 8 lines

Quoted 6 lines
Quoted 4 lines
Quoted 1 line
Thanks, these are on for-next.
The drivers/dma/ should go thru dmaengine tree. During merge window I dont apply the patches
OK, I can drop this from my tree if you'd like?
Just to follow up from IRC: I'm dropping these from my tree.

Hi Palmer and Vinod,
Many thanks for your help. As all you suggested, let's go this patchset by the dmaengine tree. Thanks a lot!

Zong Li

author

2022-04-08 08:03 UTC #15 lore raw

In reply to Vinod Koul

Quoted 53 lines · Vinod Koul

Quoted 51 lines · Palmer Dabbelt
Quoted 49 lines · On Mon, 28 Mar 2022 02:52:21 PDT (-0700), [email protected] wrote:
The PDMA driver currently assumes there are four channels by default, it might cause the error if there is actually less than four channels. Change that by getting number of channel dynamically from device tree. For backwards-compatible, it uses the default value (i.e. 4) when there is no 'dma-channels' information in dts.
This patch set contains the dts and dt-bindings change.
Changed in v8:
- Rebase on master
- Remove modification of microchip-mpfs.dtsi
- Rename DMA node name of fu540-c000.dtsi
Changed in v7:
- Rebase on tag v5.17-rc7
- Modify the subject of patch
Changed in v6:
- Rebase on tag v5.17-rc6
- Change sf_pdma.chans[] to a flexible array member.
Changed in v5:
- Rebase on tag v5.17-rc3
- Fix typo in dt-bindings and commit message
- Add PDMA versioning scheme for compatible
Changed in v4:
- Remove cflags of debug use reported-by: kernel test robot <[email protected]>
Changed in v3:
- Fix allocating wrong size
- Return error if 'dma-channels' is larger than maximum
Changed in v2:
- Rebase on tag v5.16
- Use 4 as default value of dma-channels
Zong Li (4):
dt-bindings: dma-engine: sifive,fu540: Add dma-channels property and
modify compatible
riscv: dts: Add dma-channels property and modify compatible
riscv: dts: rename the node name of dma
dmaengine: sf-pdma: Get number of channel by device tree
 .../bindings/dma/sifive,fu540-c000-pdma.yaml  | 19 +++++++++++++--
 arch/riscv/boot/dts/sifive/fu540-c000.dtsi    |  5 ++--
 drivers/dma/sf-pdma/sf-pdma.c                 | 24 ++++++++++++-------
 drivers/dma/sf-pdma/sf-pdma.h                 |  8 ++-----
 4 files changed, 38 insertions(+), 18 deletions(-)

Quoted 7 lines

Quoted 1 line
Thanks, these are on for-next.
The drivers/dma/ should go thru dmaengine tree. During merge window I dont apply the patches
--
~Vinod

Hi Vinod,
Many thanks for considering this patchset :) Could I know if there is an opportunity to pick this into the following few -rc kernels?

Vinod Koul

maintainer Applied

2022-04-08 13:13 UTC #16 lore raw

In reply to Zong Li

Quoted 5 lines · Zong Li

The PDMA driver currently assumes there are four channels by default, it might cause the error if there is actually less than four channels. Change that by getting number of channel dynamically from device tree. For backwards-compatible, it uses the default value (i.e. 4) when there is no 'dma-channels' information in dts.

Applied patch 1 & 4 to dmaengine-next, thanks

Signature · 1 line

~Vinod

Zong Li

author

2022-04-11 02:52 UTC #17 lore raw

Quoted 9 lines · Vinod Koul

Quoted 5 lines · Zong Li
The PDMA driver currently assumes there are four channels by default, it might cause the error if there is actually less than four channels. Change that by getting number of channel dynamically from device tree. For backwards-compatible, it uses the default value (i.e. 4) when there is no 'dma-channels' information in dts.
Applied patch 1 & 4 to dmaengine-next, thanks

Hi Vinod,
Thanks for your help and review. For patch 2 and 3, does it mean that we should go through the riscv tree?

Quoted 3 lines

--
~Vinod

Vinod Koul

maintainer

2022-04-11 10:49 UTC #18 lore raw

Quoted 9 lines · Zong Li

Quoted 8 lines · Vinod Koul
Quoted 5 lines · Zong Li
The PDMA driver currently assumes there are four channels by default, it might cause the error if there is actually less than four channels. Change that by getting number of channel dynamically from device tree. For backwards-compatible, it uses the default value (i.e. 4) when there is no 'dma-channels' information in dts.

Quoted 5 lines

Quoted 1 line
Applied patch 1 & 4 to dmaengine-next, thanks
Hi Vinod,
Thanks for your help and review. For patch 2 and 3, does it mean that we should go through the riscv tree?

Yes

Signature · 1 line

~Vinod

Zong Li

author

2022-04-11 11:43 UTC #19 lore raw

Quoted 11 lines · Vinod Koul

Quoted 9 lines · Zong Li
Quoted 8 lines · Vinod Koul
Quoted 5 lines · Zong Li
The PDMA driver currently assumes there are four channels by default, it might cause the error if there is actually less than four channels. Change that by getting number of channel dynamically from device tree. For backwards-compatible, it uses the default value (i.e. 4) when there is no 'dma-channels' information in dts.

Quoted 8 lines

Quoted 5 lines
Quoted 1 line
Applied patch 1 & 4 to dmaengine-next, thanks
Hi Vinod,
Thanks for your help and review. For patch 2 and 3, does it mean that we should go through the riscv tree?
Yes

Hi Palmer,
Could you please help me to pick up the patch 2 and 3. Thanks :)

Quoted 2 lines

--
~Vinod

Palmer Dabbelt

maintainer

2022-04-20 00:18 UTC #20 lore raw

Quoted 18 lines · On Mon, 11 Apr 2022 04:43:35 PDT (-0700), [email protected] wrote:

Quoted 17 lines · Vinod Koul
Quoted 14 lines · Zong Li
Quoted 9 lines · Vinod Koul
Quoted 5 lines · Zong Li
The PDMA driver currently assumes there are four channels by default, it might cause the error if there is actually less than four channels. Change that by getting number of channel dynamically from device tree. For backwards-compatible, it uses the default value (i.e. 4) when there is no 'dma-channels' information in dts.
Applied patch 1 & 4 to dmaengine-next, thanks
Hi Vinod,
Thanks for your help and review. For patch 2 and 3, does it mean that we should go through the riscv tree?

Quoted 5 lines

Quoted 2 lines
Yes
Hi Palmer,
Could you please help me to pick up the patch 2 and 3. Thanks :)

Sorry about that, I forgot about this one. I just put them on for-next, there was a minor merge conflict but it looks pretty simple.

Quoted 3 lines

Quoted 3 lines
--
~Vinod

Rob Herring

2022-04-20 14:26 UTC #21 lore raw

Quoted 20 lines · Palmer Dabbelt

Quoted 18 lines · On Mon, 11 Apr 2022 04:43:35 PDT (-0700), [email protected] wrote:
Quoted 17 lines · Vinod Koul
Quoted 14 lines · Zong Li
Quoted 9 lines · Vinod Koul
Quoted 5 lines · Zong Li
The PDMA driver currently assumes there are four channels by default, it might cause the error if there is actually less than four channels. Change that by getting number of channel dynamically from device tree. For backwards-compatible, it uses the default value (i.e. 4) when there is no 'dma-channels' information in dts.
Applied patch 1 & 4 to dmaengine-next, thanks
Hi Vinod,
Thanks for your help and review. For patch 2 and 3, does it mean that we should go through the riscv tree?

Quoted 8 lines

Quoted 5 lines
Quoted 2 lines
Yes
Hi Palmer,
Could you please help me to pick up the patch 2 and 3. Thanks :)
Sorry about that, I forgot about this one. I just put them on for-next, there was a minor merge conflict but it looks pretty simple.

Looks like you applied patch 1 too which Vinod already applied to the dmaengine tree. And you changed the 1st line removing the "# " before the SPDX tag which results in:

make[1]: *** Deleting file 'Documentation/devicetree/bindings/dma/sifive,fu540-c000-pdma.example.dts' Traceback (most recent call last):
File "/usr/local/bin/dt-extract-example", line 52, in <module>
binding = yaml.load(open(args.yamlfile, encoding='utf-8').read())
File "/usr/local/lib/python3.8/dist-packages/ruamel/yaml/main.py",
line 434, in load
return constructor.get_single_data()
File "/usr/local/lib/python3.8/dist-packages/ruamel/yaml/constructor.py",
line 119, in get_single_data
node = self.composer.get_single_node()
File "_ruamel_yaml.pyx", line 718, in _ruamel_yaml.CParser.get_single_node
ruamel.yaml.composer.ComposerError: expected a single document in the stream
in "<unicode string>", line 1, column 1
but found another document
in "<unicode string>", line 2, column 1
make[1]: *** [Documentation/devicetree/bindings/Makefile:26:
Documentation/devicetree/bindings/dma/sifive,fu540-c000-pdma.example.dts] Error 1 ./Documentation/devicetree/bindings/dma/sifive,fu540-c000-pdma.yaml:1:1:
[error] missing document start "---" (document-start)

Rob

Zong Li

author

2022-04-21 01:08 UTC #22 lore raw

Quoted 28 lines · Rob Herring

Quoted 26 lines · Palmer Dabbelt
Quoted 23 lines · On Mon, 11 Apr 2022 04:43:35 PDT (-0700), [email protected] wrote:
Quoted 11 lines · Vinod Koul
Quoted 9 lines · Zong Li
Quoted 8 lines · Vinod Koul
Quoted 5 lines · Zong Li
The PDMA driver currently assumes there are four channels by default, it might cause the error if there is actually less than four channels. Change that by getting number of channel dynamically from device tree. For backwards-compatible, it uses the default value (i.e. 4) when there is no 'dma-channels' information in dts.
Quoted 8 lines
Quoted 5 lines
Quoted 1 line
Applied patch 1 & 4 to dmaengine-next, thanks
Hi Vinod,
Thanks for your help and review. For patch 2 and 3, does it mean that we should go through the riscv tree?
Yes
Hi Palmer,
Could you please help me to pick up the patch 2 and 3. Thanks :)

Quoted 6 lines

Quoted 2 lines
Sorry about that, I forgot about this one. I just put them on for-next, there was a minor merge conflict but it looks pretty simple.
Looks like you applied patch 1 too which Vinod already applied to the dmaengine tree. And you changed the 1st line removing the "# " before the SPDX tag which results in:

Hi Palmer,
Many thanks for helping me to pick them into riscv-tree, It seems like we need to pick patch 2 and 3 in riscv tree, instead of patch 1 and 2. :)

Quoted 25 lines

make[1]: *** Deleting file 'Documentation/devicetree/bindings/dma/sifive,fu540-c000-pdma.example.dts' Traceback (most recent call last):
File "/usr/local/bin/dt-extract-example", line 52, in <module>
binding = yaml.load(open(args.yamlfile, encoding='utf-8').read())
File "/usr/local/lib/python3.8/dist-packages/ruamel/yaml/main.py",
line 434, in load
return constructor.get_single_data()
File "/usr/local/lib/python3.8/dist-packages/ruamel/yaml/constructor.py",
line 119, in get_single_data
node = self.composer.get_single_node()
File "_ruamel_yaml.pyx", line 718, in _ruamel_yaml.CParser.get_single_node
ruamel.yaml.composer.ComposerError: expected a single document in the stream
in "<unicode string>", line 1, column 1
but found another document
in "<unicode string>", line 2, column 1
make[1]: *** [Documentation/devicetree/bindings/Makefile:26:
Documentation/devicetree/bindings/dma/sifive,fu540-c000-pdma.example.dts] Error 1 ./Documentation/devicetree/bindings/dma/sifive,fu540-c000-pdma.yaml:1:1:
[error] missing document start "---" (document-start)
Rob

Palmer Dabbelt

maintainer

2022-04-21 02:49 UTC #23 lore raw

In reply to Rob Herring

Quoted 47 lines · On Wed, 20 Apr 2022 07:25:53 PDT (-0700), [email protected] wrote:

Quoted 20 lines · Palmer Dabbelt
Quoted 18 lines · On Mon, 11 Apr 2022 04:43:35 PDT (-0700), [email protected] wrote:
Quoted 17 lines · Vinod Koul
Quoted 14 lines · Zong Li
Quoted 9 lines · Vinod Koul
Quoted 5 lines · Zong Li
The PDMA driver currently assumes there are four channels by default, it might cause the error if there is actually less than four channels. Change that by getting number of channel dynamically from device tree. For backwards-compatible, it uses the default value (i.e. 4) when there is no 'dma-channels' information in dts.
Applied patch 1 & 4 to dmaengine-next, thanks
Hi Vinod,
Thanks for your help and review. For patch 2 and 3, does it mean that we should go through the riscv tree?
Quoted 8 lines
Quoted 5 lines
Quoted 2 lines
Yes
Hi Palmer,
Could you please help me to pick up the patch 2 and 3. Thanks :)
Sorry about that, I forgot about this one. I just put them on for-next, there was a minor merge conflict but it looks pretty simple.
Looks like you applied patch 1 too which Vinod already applied to the dmaengine tree. And you changed the 1st line removing the "# " before the SPDX tag which results in:
make[1]: *** Deleting file 'Documentation/devicetree/bindings/dma/sifive,fu540-c000-pdma.example.dts' Traceback (most recent call last):
File "/usr/local/bin/dt-extract-example", line 52, in <module>
binding = yaml.load(open(args.yamlfile, encoding='utf-8').read())
File "/usr/local/lib/python3.8/dist-packages/ruamel/yaml/main.py",
line 434, in load
return constructor.get_single_data()
File "/usr/local/lib/python3.8/dist-packages/ruamel/yaml/constructor.py",
line 119, in get_single_data
node = self.composer.get_single_node()
File "_ruamel_yaml.pyx", line 718, in _ruamel_yaml.CParser.get_single_node
ruamel.yaml.composer.ComposerError: expected a single document in the stream

Quoted 8 lines

in "<unicode string>", line 1, column 1
but found another document
in "<unicode string>", line 2, column 1
make[1]: *** [Documentation/devicetree/bindings/Makefile:26:
Documentation/devicetree/bindings/dma/sifive,fu540-c000-pdma.example.dts] Error 1 ./Documentation/devicetree/bindings/dma/sifive,fu540-c000-pdma.yaml:1:1:
[error] missing document start "---" (document-start)

Sorry about that, I cherry-picked them from my working repo which I assumed was the same but I guess had some nastiness (including the patch reordering). Then I also ran check in the wrong working repo, so I didn't notice the mess.

This should all be fixed.