Re: [PATCH][RFC] asm-generic:remove calling flush_write_buffers() in dma_sync_*_for_cpu

[Joerg Roedel <joerg.roedel@amd.com>]

On Mon, Jun 29, 2009 at 09:51:34PM +0800, Ming Lei wrote:
> 2009/6/29 Joerg Roedel <joerg.roedel@amd.com>:
> > On Sun, Jun 28, 2009 at 03:34:35PM +0000, Arnd Bergmann wrote:
> >> On Sunday 28 June 2009 14:39:19 tom.leiming@gmail.com wrote:
> >> > From: Ming Lei <tom.leiming@gmail.com>
> >> >
> >> > dma_sync_*_for_cpu() is introduced to make cpu access dma buffers safely when
> >> > dma transfer is over, it seems there is nothing to do with cpu write buffer,
> >> > so remove it.
> >> >
> >> > Signed-off-by: Ming Lei <tom.leiming@gmail.com>
> >>
> >> Right, this looks correct. On a related note, flush_write_buffers is
> >> architecture specific right now: only x86 and frv implement it at all,
> >> though and with slightly different semantics.
> >
> > This doen't look correct to me. The sync functions may do bounce buffering
> > which is all about copying data from one place in main memory to another. So we
> > need these flush_write_buffer() calls in the _for_cpu path too.
> 
> IMHO, even we do not call flush_write_buffer(), CPU can read correct
> data from the
> dma buffer since write buffer can't affect cache, right?

flush_write_buffer is not about cache flushing. It is about read/write
reordering in the CPU. Think of it as a memory barrier. On most x86
systems this function is therefore a nop. Cache flushing for
architectures without cache-coherent DMA is typically handled in their
low-level dma-api code (I have seen that at least in parisc32).

	Joerg

-- 
           | Advanced Micro Devices GmbH
 Operating | Karl-Hammerschmidt-Str. 34, 85609 Dornach bei München
 System    | 
 Research  | Geschäftsführer: Thomas M. McCoy, Giuliano Meroni
 Center    | Sitz: Dornach, Gemeinde Aschheim, Landkreis München
           | Registergericht München, HRB Nr. 43632