linux-kernel - Re: [RFC net-next v2 1/2] devlink: add whole device devlink instance

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: <b0857058-98bd-4de5-8e8b-e1e52d66bfe9@intel.com>
Date: Mon, 24 Feb 2025 14:12:29 -0800
From: Jacob Keller <jacob.e.keller@...el.com>
To: Jiri Pirko <jiri@...nulli.us>, Przemek Kitszel
	<przemyslaw.kitszel@...el.com>
CC: <intel-wired-lan@...ts.osuosl.org>, Tony Nguyen
	<anthony.l.nguyen@...el.com>, Jakub Kicinski <kuba@...nel.org>, Cosmin Ratiu
	<cratiu@...dia.com>, Tariq Toukan <tariqt@...dia.com>,
	<netdev@...r.kernel.org>, Konrad Knitter <konrad.knitter@...el.com>,
	<davem@...emloft.net>, Eric Dumazet <edumazet@...gle.com>, Paolo Abeni
	<pabeni@...hat.com>, Andrew Lunn <andrew@...n.ch>,
	<linux-kernel@...r.kernel.org>, ITP Upstream
	<nxne.cnse.osdt.itp.upstreaming@...el.com>, Carolina Jubran
	<cjubran@...dia.com>
Subject: Re: [RFC net-next v2 1/2] devlink: add whole device devlink instance



On 2/24/2025 8:14 AM, Jiri Pirko wrote:
> 
> The shared parent entity for PFs (and other Fs) is always reference
> counted, first creates, last removes. I feel like this is something
> missing in PCI spec. If such beast would exist, very easy to implement
> this in devlink. We have all we need in place already.
> 
> 

This is basically what I was thinking too. It does feel like it fits in
the PCI layer better than in the devlink layer.. but Przemek also
mentioned the following from reading the other lists:

> 
> I get that the general sentiment is to "leave the complexities to the
> driver/other layers", but it was based on reading only limited amount
> of internal (non networking) mailing lists.


Which makes me think going the PCI route might be tricky. I am not sure
if there is another way to get that model without though...


There was also something I saw recently merge, faux_bus? Not sure if
that would be just as ugly as your other 3 options though...