[PATCH v2 2/2] nvme-tcp: parallelize I/O queue allocation and startup

Sagi Grimberg sagi at grimberg.me
Sat Sep 5 15:29:22 PDT 2026



On 04/09/2026 19:04, Surabhi Gogte (she/her) wrote:
> On Sun, Aug 30, 2026 at 3:59 PM Mohamed Khalfella
> <mkhalfella at purestorage.com> wrote:
>> On Mon 2026-08-31 01:44:32 +0300, Sagi Grimberg wrote:
>>>
>>> On 31/08/2026 1:25, Mohamed Khalfella wrote:
>>>> On Mon 2026-08-31 01:00:52 +0300, Sagi Grimberg wrote:
>>>>> On 25/08/2026 1:56, Surabhi Gogte wrote:
>>>>>> - ret = sock_create_kern(current->nsproxy->net_ns,
>>>>>> + ret = sock_create_kern(&init_net,
>>>>> I don't think we can just change this...
>>>> Why we can not change it? Today the error recovery code running in
>>>> worker thread will be using init_net anyway, no?
>>> See commit: 1be52169c3488ef98582ed553ab35cefa3978817
>> Thanks for the pointer. I was not aware of that. What about error
>> recovery case? I _think_ it will be using init_net, no?
>>
>  From what I gathered, reconnect and error recovery run on kworkers,
> so those paths were already resolving to init_net even before this change.
> If we pin the caller's netns to the ctrl, we can keep netns consistent
> throughout,
> if that is desired.

We probably should.



More information about the Linux-nvme mailing list