netdev - Re: [PATCH net-next 01/11] ice: Add support for classid based queue selection

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 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]

Date:   Fri, 29 Apr 2022 19:42:07 -0700
From:   Jakub Kicinski <kuba@...nel.org>
To:     "Nambiar, Amritha" <amritha.nambiar@...el.com>
Cc:     "Nguyen, Anthony L" <anthony.l.nguyen@...el.com>,
        "davem@...emloft.net" <davem@...emloft.net>,
        "pabeni@...hat.com" <pabeni@...hat.com>,
        "edumazet@...gle.com" <edumazet@...gle.com>,
        "netdev@...r.kernel.org" <netdev@...r.kernel.org>,
        "Mogilappagari, Sudheer" <sudheer.mogilappagari@...el.com>,
        "Samudrala, Sridhar" <sridhar.samudrala@...el.com>,
        "Sreenivas, Bharathi" <bharathi.sreenivas@...el.com>,
        Jamal Hadi Salim <jhs@...atatu.com>
Subject: Re: [PATCH net-next 01/11] ice: Add support for classid based queue
 selection

On Sat, 30 Apr 2022 02:00:05 +0000 Nambiar, Amritha wrote:
> IIUC, currently the action skbedit queue_mapping is for transmit queue selection,
> and the bound checking is w.r.to dev->real_num_tx_queues. Also, based on my
> discussion with Alex (https://www.spinics.net/lists/netdev/msg761581.html), it
> looks like this currently applies at the qdisc enqueue stage and not at the
> classifier level.

They both apply at enqueue stage, AFAIU. Setting classid on ingress
does exactly nothing, no? :)

Neither is perfect, at least skbedit seems more straightforward. 
I suspect modern DC operator may have little familiarity with classful
qdiscs and what classid is. Plus, again, you're assuming mqprio's
interpretation like it's a TC-wide thing.

skbedit OTOH is used with a clsact qdisc.

Also it would be good if what we did had some applicability to SW. 
Maybe extend skbedit with a way of calling skb_record_rx_queue()?