On 19-06-25 22:48:57, Weiping Zhang wrote:
> Minwoo Im <minwoo.im.dev@xxxxxxxxx> 于2019年6月25日周二 上午6:00写道：
> >
> > On 19-06-24 22:29:19, Weiping Zhang wrote:
> > > Now nvme support three type hardware queues, read, poll and default,
> > > this patch rename write_queues to read_queues to set the number of
> > > read queues more explicitly. This patch alos is prepared for nvme
> > > support WRR(weighted round robin) that we can get the number of
> > > each queue type easily.
> > >
> > > Signed-off-by: Weiping Zhang <zhangweiping@xxxxxxxxxxxxxx>
> >
> > Hello, Weiping.
> >
> > Thanks for making this patch as a separated one. Actually I'd like to
> > hear about if the origin purpose of this param can be changed or not.
> >
> > I can see a log from Jens when it gets added her:
> > Commit 3b6592f70ad7("nvme: utilize two queue maps, one for reads and
> > one for writes")
> > It says:
> > """
> > NVMe does round-robin between queues by default, which means that
> > sharing a queue map for both reads and writes can be problematic
> > in terms of read servicing. It's much easier to flood the queue
> > with writes and reduce the read servicing.
> > """
> >
> > So, I'd like to hear what other people think about this patch :)
> >
>
> This patch does not change its original behavior, if we set read_queue
> greater than 0, the read and write request will use different tagset map,
> so they will use different hardware queue.
Yes, that's why I want to hear some comments for this change from other
people. I'm not against this change, though.
