linux-kernel - Re: [PATCH v2 4/8] genirq: soft_moderation: implement adaptive moderation

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: <875xb77m0m.ffs@tglx>
Date: Tue, 18 Nov 2025 10:00:57 +0100
From: Thomas Gleixner <tglx@...utronix.de>
To: Luigi Rizzo <lrizzo@...gle.com>
Cc: Marc Zyngier <maz@...nel.org>, Luigi Rizzo <rizzo.unipi@...il.com>,
 Paolo Abeni <pabeni@...hat.com>, Andrew Morton
 <akpm@...ux-foundation.org>, Sean Christopherson <seanjc@...gle.com>,
 Jacob Pan <jacob.jun.pan@...ux.intel.com>, linux-kernel@...r.kernel.org,
 linux-arch@...r.kernel.org, Bjorn Helgaas <bhelgaas@...gle.com>, Willem de
 Bruijn <willemb@...gle.com>
Subject: Re: [PATCH v2 4/8] genirq: soft_moderation: implement adaptive
 moderation

On Mon, Nov 17 2025 at 22:34, Luigi Rizzo wrote:
> (First world problem, for sure: I have examples for AMD, Intel, Arm,
> all of them with 100+ CPUs per numa node, and 160-480 CPUs total)
> On some of the above platforms, MSIx interrupts cause heavy serialization
> of all other PCIe requests. As a result, when the total interrupt rate exceeds
> 1-2M intrs/s, I/O throughput degrades by up to 4x and more.

Is it actually the sum of all interrupts in the system or is it
segmented per root port?

Thanks,

        tglx