linux-kernel - Re: [PATCH v2 1/2] genirq: Retain depth for managed IRQs across CPU hotplug

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: <aFPLFKMnlvtBjD1P@hovoldconsulting.com>
Date: Thu, 19 Jun 2025 10:32:20 +0200
From: Johan Hovold <johan@...nel.org>
To: Brian Norris <briannorris@...omium.org>
Cc: Aleksandrs Vinarskis <alex.vinarskis@...il.com>,
	Thomas Gleixner <tglx@...utronix.de>,
	Tsai Sung-Fu <danielsftsai@...gle.com>,
	Douglas Anderson <dianders@...omium.org>,
	linux-kernel@...r.kernel.org
Subject: Re: [PATCH v2 1/2] genirq: Retain depth for managed IRQs across CPU
 hotplug

On Wed, Jun 18, 2025 at 10:10:14AM -0700, Brian Norris wrote:
> On Wed, Jun 18, 2025 at 12:17:17PM +0200, Johan Hovold wrote:
> > On Thu, Jun 12, 2025 at 11:40:38AM -0700, Brian Norris wrote:
> > > For the record, I was reminded that DWC/pcie-qcom does not, in fact,
> > > support irq_chip::irq_set_affinity(), which could perhaps be a unique
> > > factor in his systems' behavior.
> > 
> > No, we use the GIC ITS for the NVMe interrupts on X1E so that should
> > not be involved here.
> 
> Huh, interesting callout. I had looked at trying that previously on
> another DWC-based platform, for the same reasons noted in that
> switchover.
> 
> Anyway, all I can tell you is that Alex's logs clearly showed:
> 
>   set affinity failed(-22)
> 
> for what looked like the NVMe interrupts during CPU unplug / migration.

No, those warnings are for wakeup enabled GPIO interrupts (e.g. the lid
switch), which are currently partly disabled on this platform pending
some rework of the PDC driver:

	602cb14e310a ("pinctrl: qcom: x1e80100: Bypass PDC wakeup parent for now")

(And with the rework those warnings go away, while the IRQ suspend
regression is still there.)

Johan