On Wed, Nov 22, 2023 at 11:59:47AM +0300, Nikita Shubin wrote:
> - drop subsys_initcall code
> - drop platform probe
> - add OF ID match table with data
> - add of_probe for device tree
> - add xlate for m2m/m2p
> - drop platform structs usage
It's not the best commit message (e.g., unaligned with verb "add"
in the Subject).
...
> + edmac->clk = of_clk_get(np, i);
Why devm_clk_get() can't be used?
> if (IS_ERR(edmac->clk)) {
> + dev_warn(&pdev->dev, "failed to get clock\n");
> continue;
> }
...
> + if (direction != DMA_MEM_TO_DEV && direction != DMA_DEV_TO_MEM)
> + return NULL;
is_slave_direction() ?
...
> + dev_info(edma->dma_dev.dev, "%s: port=%d", __func__, port);
info level?! Wouldn't be noisy a bit?
...
> + if (direction != DMA_MEM_TO_DEV && direction != DMA_DEV_TO_MEM)
> + return NULL;
As per above.
...
> + dev_info(dma_dev->dev, "EP93xx M2%s DMA ready\n",
> + edma->m2m ? "M" : "P");
One line?
...
> + if (!IS_ERR_OR_NULL(edmac->clk))
> + clk_put(edmac->clk);
CLK framework is at least NULL aware. Perhaps you can make sure it's never
IS_ERR() and drop this conditional altogether.
...
> -
> +module_platform_driver(ep93xx_dma_driver);
+ blank line.
> MODULE_AUTHOR("Mika Westerberg <mika.westerberg@xxxxxx>");
> MODULE_DESCRIPTION("EP93xx DMA driver");
...
> --- a/include/linux/platform_data/dma-ep93xx.h
> +++ b/include/linux/platform_data/dma-ep93xx.h
> @@ -5,6 +5,7 @@
> #include <linux/types.h>
> #include <linux/dmaengine.h>
> #include <linux/dma-mapping.h>
> +#include <linux/property.h>
Can this be a bit more ordered, like put before types.h (at least from the
context I see here)?
Also you missing device.h and string.h according to the new function
implementation.
--
With Best Regards,
Andy Shevchenko
