linux-kernel - Re: [PATCH] softirq: Do not loop if running under a real-time task

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: <20240807153230.rDSKGnY9@linutronix.de>
Date: Wed, 7 Aug 2024 17:32:30 +0200
From: Sebastian Andrzej Siewior <bigeasy@...utronix.de>
To: Petr Malat <oss@...at.biz>
Cc: linux-kernel@...r.kernel.org, paulmck@...nel.org, tglx@...utronix.de,
	nsaenzju@...hat.com, frederic@...nel.org
Subject: Re: [PATCH] softirq: Do not loop if running under a real-time task

On 2024-07-25 08:44:02 [+0000], Petr Malat wrote:
> Hi Sebastian,
Hi Petr,

> what is the status of this softirq issue? By looking on the current
> upstream code, I think the problem is still there. I can resend my
> patch or rework it to use independent task struct flag to decide if
> softirq processing should be skipped in the current thread.

I was re-reading the thread. I don't see how the migration thread should
run softirqs. This should not happen or please show how if so.

Your example has a long running timer callback which probably gets
PI-boosted by a threaded-IRQ. The longterm plan is that long running
timer callbacks don't affect threaded interrupts as they do now.
The worked started with
	c5bcab7558220 ("locking/local_lock: Add local nested BH locking infrastructure.")
	https://git.kernel.org/torvalds/c/c5bcab7558220

and needs to continue.

> BR,
>   Petr

Sebastian