On 14/04/10 09:44, ext Herbert Xu wrote:
On Wed, Apr 14, 2010 at 09:37:47AM +0300, Dmitry Kasatkin wrote:
Like just with import/export.
Problems for hw:
1. To have a good performance with DMA we need to have large buffer.
Not just 64 bytes block. state becomes large
Sure. But it shouldn't be up to the driver to merge operations.
Higher layers (either the end-user or the crypto API) should perform
merging.
What do you mean by "merge operation".
request merging?
We don't put request merging into storage drivers for a reason.
2. supporting concurrent requests means switching HW state and it takes
a time.
Well that's just the way it is. I'm not saying that it's going
to occur frequently, but if it does, you need to support it.
All that burden significantly slowdown hw acceleration and increase CPU
load.
I guess my points are understandable.
I understand your concerns, but these are really limitations
common to all off-chip acceleration. If you have to write drivers
for them, then you'll have to deal with these issues.
Cheers,
--
To unsubscribe from this list: send the line "unsubscribe linux-crypto" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
