netdev - Re: [PATCH net-next v2 12/15] net: page_pool: report when page pool was destroyed

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] [day] [month] [year] [list]

Message-ID: <8504074e-5cf0-4e47-a10d-517a1e848764@kernel.org>
Date: Wed, 22 Nov 2023 09:53:06 +0100
From: Jesper Dangaard Brouer <hawk@...nel.org>
To: Jakub Kicinski <kuba@...nel.org>
Cc: davem@...emloft.net, netdev@...r.kernel.org, edumazet@...gle.com,
 pabeni@...hat.com, almasrymina@...gle.com, ilias.apalodimas@...aro.org,
 dsahern@...il.com, dtatulea@...dia.com,
 kernel-team <kernel-team@...udflare.com>
Subject: Re: [PATCH net-next v2 12/15] net: page_pool: report when page pool
 was destroyed



On 11/21/23 22:49, Jakub Kicinski wrote:
> On Tue, 21 Nov 2023 21:45:57 +0100 Jesper Dangaard Brouer wrote:
>> Hmm, this is called when kernel could *NOT* destroy the PP, but have to
>> start a work-queue that will retry deleting this. Thus, I think naming
>> this "destroyed" is confusing as I then assumed was successfully
>> destroyed, but it is not, instead it is on "deathrow".
> 
> I wasn't sure what to call it so I called what the driver API is
> called...
> 
> "deathrow" does not sound very intuitive to me. How about "detached"
> or "removed"?

I like "detached".

> 
>> Could we place this PP instance on another list of PP instances about to
>> be deleted?
>>
>> (e.g. a deathrow or sched_destroy list)
> 
> Is there a need for that?
> 
> I mean - many interesting extensions to this API are possible.
> I don't think they should all can all be here from day 1..
> 
>> Perhaps this could also allow us to list those PP instances that
>> no-longer have a netdev associated?
> 
> The current implementation uses loopback for that, since it's naturally
> tied to a name space.

Okay, guess it makes sense to tie this to loopback.

--Jesper