Re: [Qemu-devel] [PATCH 02/13] Implement cpu_physical_memory_set()

[Anthony Liguori <anthony@codemonkey.ws>]

On 06/20/2012 04:30 PM, Benjamin Herrenschmidt wrote:
> On Wed, 2012-06-20 at 16:15 -0500, Anthony Liguori wrote:
>> On 06/19/2012 01:39 AM, Benjamin Herrenschmidt wrote:
>>> From: David Gibson<david@gibson.dropbear.id.au>
>>>
>>> This patch adds cpu_physical_memory_set() function.  This is equivalent to
>>> calling cpu_physical_memory_write() with a buffer filled with a character,
>>> ie, a memset of target memory.
>>>
>>> It uses a small temporary buffer on the stack.
>>>
>>> Signed-off-by: David Gibson<david@gibson.dropbear.id.au>
>>> Signed-off-by: Benjamin Herrenschmidt<benh@kernel.crashing.org>
>>
>> Why should this be in the core API?  Shouldn't this be a helper on top of the
>> DMA API?
>
> This comes from the original patch which hand implemented the "set" by
> reproducing the logic inside cpu_physical_memory_rw(). I turned into a
> wrapper on top of the latter based on (your ?) previous reviews on this
> list. I don't care enough to argue to keep it if you want it gone, we do
> have a "clear" accessors in the PAPR vio dma accessors which is handy
> but I could implement it locally.

I think it's better to just stick this with the other DMA helpers and not have a 
cpu_physical version.  People shouldn't use that API directly so not having it 
there encourages people to look elsewhere.

Regards,

Anthony Liguori

>
> Cheers,
> Ben.
>
>