[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-Id: <200905071238.05110.arnd@arndb.de>
Date: Thu, 7 May 2009 12:38:04 +0200
From: Arnd Bergmann <arnd@...db.de>
To: FUJITA Tomonori <fujita.tomonori@....ntt.co.jp>
Cc: linux-kernel@...r.kernel.org, mingo@...e.hu, tony.luck@...el.com,
linux-ia64@...r.kernel.org
Subject: Re: [PATCH -tip 1/5] add asm-generic/dma-mappig.h
On Thursday 07 May 2009, FUJITA Tomonori wrote:
> This header file provides some generic dma mapping function
> definitions for the users of struct dma_map_ops.
>
> Signed-off-by: FUJITA Tomonori <fujita.tomonori@....ntt.co.jp>
I'm currently trying to add asm-generic variants for all of
the common headers. One of my goals there is to make them
as complete as possible, i.e. if a header file in asm-generic
has the same name as one in any of the architectures, it should
IMHO provide the same (full) API.
I would therefore suggest that you either rename the file to
something like dma-mapping-common.h or add some template
functions that can be overriden per architecture, like
#ifndef dma_get_ops
static inline struct dma_map_ops *get_dma_ops(struct device *dev)
{
return dma_ops;
}
#endif
#ifndef dma_alloc_coherent
/*
* to be defined out of line
*/
extern void *dma_alloc_coherent(struct device *dev, size_t size,
dma_addr_t *dma_handle, gfp_t gfp);
#endif
Then an architecture can opt-in to do things like:
#define dma_get_ops dma_get_ops
static inline struct dma_map_ops *get_dma_ops(struct device *dev)
{
return dev->archdata.dma_ops;
}
before the #include <asm-generic/dma-mapping.h>
> +static inline dma_addr_t dma_map_single_attrs(struct device *dev, void *ptr,
> + size_t size,
> + enum dma_data_direction dir,
> + struct dma_attrs *attrs)
> +{
> + struct dma_map_ops *ops = get_dma_ops(dev);
> + dma_addr_t addr;
> +
> + kmemcheck_mark_initialized(ptr, size);
> + BUG_ON(!valid_dma_direction(dir));
> + addr = ops->map_page(dev, virt_to_page(ptr),
> + (unsigned long)ptr & ~PAGE_MASK, size,
> + dir, NULL);
> + debug_dma_map_page(dev, virt_to_page(ptr),
> + (unsigned long)ptr & ~PAGE_MASK, size,
> + dir, addr, true);
> + return addr;
> +}
You are not passing the dma_attrs down here, which I think you need to,
if any of the architectures using this file want to support DMA attributes.
> +
> +static inline void dma_unmap_single_attrs(struct device *dev, dma_addr_t addr,
> + size_t size,
> + enum dma_data_direction dir,
> + struct dma_attrs *attrs)
> +{
> + struct dma_map_ops *ops = get_dma_ops(dev);
> +
> + BUG_ON(!valid_dma_direction(dir));
> + if (ops->unmap_page)
> + ops->unmap_page(dev, addr, size, dir, NULL);
> + debug_dma_unmap_page(dev, addr, size, dir, true);
> +}
same here and in the next few functions.
Arnd <><
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
Powered by blists - more mailing lists