netdev - Re: [PATCH v5 net-next 1/2] net: Prevent RPS table overwrite for active flows

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: <CACLgkEYaJiHCc_XdbuFynujWcza6q5TDdX+mdCZPiCQV6Js8vg@mail.gmail.com>
Date: Wed, 23 Jul 2025 10:03:59 +0530
From: Krishna Kumar <krikku@...il.com>
To: Eric Dumazet <edumazet@...gle.com>
Cc: netdev@...r.kernel.org, davem@...emloft.net, tom@...bertland.com, 
	kuba@...nel.org, pabeni@...hat.com, horms@...nel.org, sdf@...ichev.me, 
	kuniyu@...gle.com, ahmed.zaki@...el.com, aleksander.lobakin@...el.com, 
	atenart@...nel.org, krishna.ku@...pkart.com
Subject: Re: [PATCH v5 net-next 1/2] net: Prevent RPS table overwrite for
 active flows

Hi Eric,

On Mon, Jul 21, 2025 at 1:44 PM Eric Dumazet <edumazet@...gle.com> wrote:
> >  struct rps_dev_flow {
> >         u16             cpu;
> >         u16             filter;
> >         unsigned int    last_qtail;
> > +       u32             hash;
>
> This is problematic, because adds an extra potential cache line miss in RPS.
>
> Some of us do not use CONFIG_RFS_ACCEL, make sure to not add extra
> costs for this configuration ?

I will send this fix today.

On your point on making the active check less obscure/costly,
my understanding is we need last_qtail for in-order processing
and add something like last_use for activity detection. If that’s
correct, I can send a patch for it after this.

Thanks,
- Krishna