Re: [PATCH 4/4] block: move plug rq alloc into helper and ensure queue match

On 11/4/21 11:33 AM, Christoph Hellwig wrote:

On Thu, Nov 04, 2021 at 05:35:10AM -0600, Jens Axboe wrote:

quoted

I tend to prefer having logic flow from the expected conditions. And
since we need to check if the plug is valid anyway, I prefer the current
logic.

Well, for 99% of the I/O not using a cached request is the expected
condition.

That may very well be the case right now, but as more things plug into
propagating expected number of requests, that's likely to change.

That said, I did reflow it for you in the updated version.

-- 
Jens Axboe