linux-kernel - Re: [RFC V1 RESEND 2/6] PCI/MSI: Dynamic allocation of MSI-X vectors by group

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: <48a44ffc-4b5b-5eef-73de-020f1710c41e@arm.com>
Date:   Wed, 7 Aug 2019 15:18:59 +0100
From:   Marc Zyngier <marc.zyngier@....com>
To:     Thomas Gleixner <tglx@...utronix.de>,
        Megha Dey <megha.dey@...el.com>
Cc:     bhelgaas@...gle.com, linux-pci@...r.kernel.org,
        linux-kernel@...r.kernel.org, ashok.raj@...el.com,
        jacob.jun.pan@...ux.intel.com
Subject: Re: [RFC V1 RESEND 2/6] PCI/MSI: Dynamic allocation of MSI-X vectors
 by group

On 07/08/2019 14:56, Thomas Gleixner wrote:
> Megha,
> 
> On Tue, 6 Aug 2019, Megha Dey wrote:
>> On Sat, 2019-06-29 at 09:59 +0200, Thomas Gleixner wrote:
>>> On Fri, 21 Jun 2019, Megha Dey wrote:
>>
>> Totally agreed. The request to add a dynamic MSI-X infrastructure came
>> from some driver teams internally and currently they do not have
>> bandwidth to come up with relevant test cases. <sigh>
> 
> Hahahaha.
> 
>> But we hope that this patch set could serve as a precursor to the
>> interrupt message store (IMS) patch set, and we can use this patch set
>> as the baseline for the IMS patches.
> 
> If IMS needs the same functionality, then we need to think about it
> slightly differently because IMS is not necessarily tied to PCI.
>  
> IMS has some similarity to the ARM GIC ITS stuff IIRC, which already
> provides these things outside of PCI. Marc?

Indeed. We have MSI-like functionality almost everywhere, and make heavy
use of the generic MSI framework. Platform-MSI is probably the most
generic example we have (it's the Far West transposed to MSIs).

> We probably need some generic infrastructure for this so PCI and everything
> else can use it.

Indeed. Overall, I'd like the concept of MSI on whatever bus to have one
single behaviour across the board, as long as it makes sense for that
bus (nobody needs another PCI MultiMSI, for example).

Thanks,

	M.
-- 
Jazz is not dead. It just smells funny...