linux-kernel - Re: [PATCH v2] driver/nvme/host: Support duplicated nsid for the private ns

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <6ac83f8b-01e6-5394-0570-023ac6b1a71d@grimberg.me>
Date:   Tue, 15 Mar 2022 10:42:56 +0200
From:   Sagi Grimberg <sagi@...mberg.me>
To:     "hch@....de" <hch@....de>
Cc:     Sungup Moon <sungup.moon@...sung.com>,
        "kbusch@...nel.org" <kbusch@...nel.org>,
        "axboe@...com" <axboe@...com>,
        "linux-nvme@...ts.infradead.org" <linux-nvme@...ts.infradead.org>,
        "linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH v2] driver/nvme/host: Support duplicated nsid for the
 private ns


>>> +static struct nvme_ns_head *nvme_find_ns_head(struct nvme_ctrl *ctrl,
>>>    		unsigned nsid)
>>>    {
>>> +	struct nvme_subsystem *subsys = ctrl->subsys;
>>>    	struct nvme_ns_head *h;
>>>      	lockdep_assert_held(&subsys->lock);
>>
>> IMO it is a bit strange that we now don't pass in the subsystem but
>> require that the subsys->lock is taken...
> 
> We do things like that in various places, mostly because information
> needed that is subsystem-wide hangs of the nvme_ctrl structure, in
> this case the various feature bitmaps.  We could move them to the
> subsystem structure, which would be the right thing to do but a fair
> amout of churn for little savings.

Yea I understand its not universally enforced, just popped up.

>>> +++ b/drivers/nvme/host/multipath.c
>>> @@ -504,10 +504,11 @@ int nvme_mpath_alloc_disk(struct nvme_ctrl *ctrl, struct nvme_ns_head *head)
>>>      	/*
>>>    	 * Add a multipath node if the subsystems supports multiple controllers.
>>> -	 * We also do this for private namespaces as the namespace sharing data could
>>> -	 * change after a rescan.
>>> +	 * We also do this for private namespaces as the namespace sharing flag
>>> +	 * could change after a rescan.
>>
>> What happens in this case? we now have non-unique shared namespaces?
> 
> The non-uniqueue NSIDs can only happen for private namespaces.

But what happens if this changes upon a rescan as you commented?