[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <460cb285-7537-4c02-8e79-2c5087331ccf@linux.ibm.com>
Date: Tue, 21 May 2024 15:52:55 +0530
From: Nilay Shroff <nilay@...ux.ibm.com>
To: Sagi Grimberg <sagi@...mberg.me>, John Meneghini <jmeneghi@...hat.com>,
kbusch@...nel.org, hch@....de, emilne@...hat.com
Cc: linux-nvme@...ts.infradead.org, linux-kernel@...r.kernel.org,
jrani@...estorage.com, randyj@...estorage.com, hare@...nel.org
Subject: Re: [PATCH v3 1/1] nvme: multipath: Implemented new iopolicy
"queue-depth"
On 5/21/24 15:45, Sagi Grimberg wrote:
>
>
> On 21/05/2024 13:11, Sagi Grimberg wrote:
>>
>>>> Don't think this matters because cancellation only happens when we
>>>> teardown the controller anyways...
>>>>
>>> I think in case if we reset the nvme controller then we don't teardown
>>> controller, isn't it? In this case we cancel all pending requests, and
>>> later restart the controller.
>>
>> Exactly, nvme_mpath_init_ctrl resets the counter.
>
> Except you're right, the counter reset needs to move to nvme_mpath_init_identify()
> or some place that is called on every controller reset.
Yeah that was my point. Unfortunately, nvme_mpath_init_ctrl() is not called when we do nvme
controller reset. So either we move nvme_mpath_init_ctrl() to some common place which
is called from regular controller init code as well as controller reset code path
or we may explicitly reset nr_active counter from nvme_timeout().
Thanks,
--Nilay
Powered by blists - more mailing lists