netdev - Re: [PATCH net-next 1/2] net: renesas: rswitch: use per-port irq handlers

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: <21f0aa0e-9df3-4b5a-957a-dcf68399cdce@cogentembedded.com>
Date: Fri, 20 Dec 2024 14:33:39 +0500
From: Nikita Yushchenko <nikita.yoush@...entembedded.com>
To: Andrew Lunn <andrew@...n.ch>
Cc: Yoshihiro Shimoda <yoshihiro.shimoda.uh@...esas.com>,
 Andrew Lunn <andrew+netdev@...n.ch>, "David S. Miller"
 <davem@...emloft.net>, Eric Dumazet <edumazet@...gle.com>,
 Jakub Kicinski <kuba@...nel.org>, Paolo Abeni <pabeni@...hat.com>,
 Geert Uytterhoeven <geert+renesas@...der.be>, netdev@...r.kernel.org,
 linux-renesas-soc@...r.kernel.org, linux-kernel@...r.kernel.org,
 Michael Dege <michael.dege@...esas.com>,
 Christian Mardmoeller <christian.mardmoeller@...esas.com>,
 Dennis Ostermann <dennis.ostermann@...esas.com>
Subject: Re: [PATCH net-next 1/2] net: renesas: rswitch: use per-port irq
 handlers

20.12.2024 14:19, Andrew Lunn wrote:
> On Fri, Dec 20, 2024 at 09:16:58AM +0500, Nikita Yushchenko wrote:
>> Instead of handling all possible data interrupts in the same handler,
>> switch to per-port handlers.
>>
>> This significantly simplifies handling: when the same interrupt is used
>> for several ports, system calls all handlers, and each handler only has
>> to check interrupts for one port's tx and rx queues.
>>
>> But it is not required to use the same interrupt for all ports - GWCA
>> provides 8 data interrupts and allows arbitrary per-queue assignment
>> of those. Support that by reading interrupt index for each port from
>> optional 'irq-index' device tree property.
> 
> It has been pointed out that adding this property breaks backwards
> compatibility with older DT blobs. 

It does not break backwards compatibility.

Current behavior is that everything is serviced by interrupt 0.
And in case of irq-index not defined, the fallback is exactly that.

(physically there is code that assigns interrupts per chain index, but in the current driver chains that 
get non-zero interrupts assigned are never used; anso currently multiple interrupts are just multiple 
entries to the exactly same handler that always services everything)

> I don't know this hardware...
> 
> How many ports are there? Less than 9? Can you just do a static
> allocation, port 0 gets interrupt 0, port 1 interrupt 1...

There are only 3 physical ports, however the use case I'm targeting is - virtual ports serving virtual 
machines (with offloading features making hardware directly L2-forward or L3-route most traffic between 
outside world and VM-owned virtual port frontends). In this setup, some of 8 GWCA irqs will be given to 
VMs and thus there are definitely not enough to per-consumer allocation.

Nikita