linux-kernel - Re: [PATCH 0/7] proc/stat: Maintain monotonicity of "intr" and "softirq"

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: <YT3In8SWc2eYZ/09@localhost.localdomain>
Date:   Sun, 12 Sep 2021 12:30:07 +0300
From:   Alexey Dobriyan <adobriyan@...il.com>
To:     Alexei Lozovsky <me@...mmy.net>
Cc:     Thomas Gleixner <tglx@...utronix.de>,
        Christoph Lameter <cl@...ux.com>,
        LKML <linux-kernel@...r.kernel.org>,
        linux-fsdevel@...r.kernel.org
Subject: Re: [PATCH 0/7] proc/stat: Maintain monotonicity of "intr" and
 "softirq"

On Sat, Sep 11, 2021 at 12:48:01PM +0900, Alexei Lozovsky wrote:
> Here's a patch set that makes /proc/stat report total interrupt counts
> as monotonically increasing values, just like individual counters for
> interrupt types and CPUs are.
> 
> This is as if the sum was a shared counter that all CPUs increment
> atomically together with their individual counters, with the sum
> correctly and expectedly wrapping around to zero once it reaches
> UINT_MAX.
> 
> I've also added some documentation bits to codify this behavior and make
> it explicit that wrap-arounds must be expected and handled if userspace
> wants to maintain accurate total interrupt count for whatever reasons.

How about making everything "unsigned long" or even "u64" like NIC
drivers do?