On 4 September 2017 at 09:06, Adrian Hunter [off-list ref] wrote:

On 01/09/17 16:28, Adrian Hunter wrote:

quoted

On 01/09/17 15:58, Ulf Hansson wrote:

quoted

+ Christoph

On 1 September 2017 at 13:42, Adrian Hunter [off-list ref] wrote:

quoted

On 31/08/17 14:56, Adrian Hunter wrote:

quoted

Here is V7 of the hardware command queue patches without the software
command queue patches, now using blk-mq.

HW CMDQ offers 25% - 50% better random multi-threaded I/O.  I see a slight
2% drop in sequential read speed but no change to sequential write.

Any comments?

A couple of overall comments, for now.

To make sure we don't overlook something when converting to mq, I
would prefer that we first convert the existing mmc block code to mq,
then we add CMDQ on top.

That doesn't make sense.  This patch set is not converting the legacy driver
to mq therefore it cannot overlook anything for converting to mq.

And then you go silent again.

We have weekends in Sweden - and I also work on other things than mmc :-).

I do however admit, that I could have been reviewing a bit faster
throughout the re-spins. Apologize for that, but I am only doing my
best.

I can send blk-mq support for legacy requests in a few days if you like, but
I want to hear a better explanation of why you are delaying CQE support.

That would be very nice, however be aware of that we are in the merge
window, so I am not picking new material for 4.14 from this point. I
assume you understand why.

Still, big changes is always nice to queue up early for a release
cycle. Let's aim for that!

Moreover, I am not delaying CQE, but really want it to be merged asap!
However, I am also having the role as a maintainer and the things that
comes with it. For example, I would like the community to reach
consensus around how to move forward with CQE, before I decide to pick
it up.

Kind regards
Uffe