[PATCHv9 2/2] dma: Add Freescale eDMA engine driver support

[Mark Rutland]

On Thu, Jan 16, 2014 at 02:30:00AM +0000, Jingchang Lu wrote:
> Add Freescale enhanced direct memory(eDMA) controller support.
> This module can be found on Vybrid and LS-1 SoCs.
> 
> Signed-off-by: Alison Wang <b18965 at freescale.com>
> Signed-off-by: Jingchang Lu <b35083 at freescale.com>
> ---
> changes in v9:
>  define endian's operating functions instead of macro definition.
>  remove the filter function, using dma_get_slave_channel instead.
> 
> changes in v8:
>  change the edma driver according eDMA dts change.
>  add big-endian and little-endian handling.
> 
>  no changes in v4 ~ v7.
> 
>  changes in v3:
>   add vf610 edma dt-bindings namespace with prefix VF610_*.
> 
>  changes in v2:
>   using generic dma-channels property instead of fsl,dma-channels.
> 
> 
>  Documentation/devicetree/bindings/dma/fsl-edma.txt |  66 ++
>  drivers/dma/Kconfig                                |  10 +
>  drivers/dma/Makefile                               |   1 +
>  drivers/dma/fsl-edma.c                             | 957 +++++++++++++++++++++
>  4 files changed, 1034 insertions(+)
>  create mode 100644 Documentation/devicetree/bindings/dma/fsl-edma.txt
>  create mode 100644 drivers/dma/fsl-edma.c
> 
> diff --git a/Documentation/devicetree/bindings/dma/fsl-edma.txt b/Documentation/devicetree/bindings/dma/fsl-edma.txt
> new file mode 100644
> index 0000000..5a5fb61
> --- /dev/null
> +++ b/Documentation/devicetree/bindings/dma/fsl-edma.txt
> @@ -0,0 +1,66 @@
> +* Freescale enhanced Direct Memory Access(eDMA) Controller
> +
> +  The eDMA channels have multiplex capability by programmble memory-mapped
> +registers. channels are split into two groups, called DMAMUX0 and DMAMUX1,
> +specific DMA request source can only be multiplexed by any channel of certain
> +group, DMAMUX0 or DMAMUX1, but not both.
> +
> +* eDMA Controller
> +Required properties:
> +- compatible :
> +       - "fsl,vf610-edma" for eDMA used similar to that on Vybrid vf610 SoC
> +- reg : Specifies base physical address(s) and size of the eDMA registers.
> +       The 1st region is eDMA control register's address and size.
> +       The 2nd and the 3rd regions are programmable channel multiplexing
> +       control register's address and size.
> +- interrupts : Should contain eDMA interrupt
> +- interrupt-names : Should be "edma-tx" for transmission interrupt and
> +       "edma-err" for error interrupt

The description of interrupts and interrupt-names is a bit odd. How
about:

- interrupts: a list of interrupt-specifiers, one for each entry in
  interrupt-names.

- interrupt names: should contain:
  * "edma-tx" - the transmission interrupt.
  * "edma-err" - the error interrupt.

> +- #dma-cells : Must be <2>.
> +       The 1st cell specifies the DMAMUX(0 for DMAMUX0 and 1 for DMAMUX1).
> +       Specific request source can only be multiplexed by specific channels
> +       group called DMAMUX.
> +       The 2nd cell specifies the request source(slot) ID.
> +       See the SoC's reference manual for all the supported request sources.
> +- dma-channels : Number of channels supported by the controller
> +- clock-names : The channel group block clock names
> +- clocks : eDMA module clock

The set of expected clock-names should be well defined, and the example
has more than one clock..

This should be something like:

- clocks: a list of phandle + clock-specifier pairs, one for each entry
  in clock-names.

- clock-names: a list of clock names. Should contain:
  * "dmamux0" - <insert description here>
  * "dmamux1" - <insert description here>

> +
> +
> +Examples:
> +
> +edma0: dma-controller at 40018000 {
> +       #dma-cells = <2>;
> +       compatible = "fsl,vf610-edma";
> +       reg = <0x40018000 0x2000>,
> +               <0x40024000 0x1000>,
> +               <0x40025000 0x1000>;
> +       interrupts = <0 8 IRQ_TYPE_LEVEL_HIGH>,
> +               <0 9 IRQ_TYPE_LEVEL_HIGH>;
> +       interrupt-names = "edma-tx", "edma-err";
> +       dma-channels = <32>;
> +       clock-names = "dmamux0", "dmamux1";
> +       clocks = <&clks VF610_CLK_DMAMUX0>,
> +               <&clks VF610_CLK_DMAMUX1>;
> +};

[...]

> +       for (i = 0; i < DMAMUX_NR; i++) {
> +               char clkname[32];
> +
> +               res = platform_get_resource(pdev, IORESOURCE_MEM, 1 + i);
> +               fsl_edma->muxbase[i] = devm_ioremap_resource(&pdev->dev, res);
> +               if (IS_ERR(fsl_edma->muxbase[i]))
> +                       return PTR_ERR(fsl_edma->muxbase[i]);
> +
> +               sprintf(clkname, "dmamux%d", i);
> +               fsl_edma->muxclk[i] = of_clk_get(np, i);
> +               fsl_edma->muxclk[i] = devm_clk_get(&pdev->dev, clkname);

The of_clk_get should disappear. It's always overwritten and
devm_clk_get will do what you want.

> +               if (IS_ERR(fsl_edma->muxclk[i])) {
> +                       dev_err(&pdev->dev, "Missing DMAMUX block clock.\n");
> +                       return PTR_ERR(fsl_edma->muxclk[i]);
> +               }
> +
> +               ret = clk_prepare_enable(fsl_edma->muxclk[i]);
> +               if (ret) {
> +                       dev_err(&pdev->dev, "DMAMUX clk block failed.\n");
> +                       return ret;
> +               }
> +
> +       }
> +
> +       ret = fsl_edma_irq_init(pdev, fsl_edma);
> +       if (ret)
> +               return ret;
> +
> +       fsl_edma->big_endian = of_property_read_bool(np, "big-endian");

This isn't in the binding.

Thanks,
Mark.