netdev - Re: why all packets have same queue no when rps enabled?

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: <1298379718.2861.3.camel@edumazet-laptop>
Date:	Tue, 22 Feb 2011 14:01:58 +0100
From:	Eric Dumazet <eric.dumazet@...il.com>
To:	Ben Hutchings <bhutchings@...arflare.com>
Cc:	Jon Zhou <Jon.Zhou@...u.com>,
	"netdev@...r.kernel.org" <netdev@...r.kernel.org>
Subject: Re: why all packets have same queue no when rps enabled?

Le mardi 22 février 2011 à 12:45 +0000, Ben Hutchings a écrit :

> The queue number identifies a hardware queue.  RPS therefore does not
> update this number when queueing packets for processing on other CPUs.
> 
> If the hardware/driver provides a receive hash (probably Toeplitz) then
> this is used for RPS.  Otherwise a much cheaper hash is used.

The default is/should be : rxhash generated in network stack.

ethtool -k eth0 | grep hash
receive-hashing: off

To use the device/harwdare provided rxhash, you need to ask for it.

ethtool -K eth0 rxhash on



BTW, I am not sure what you mean by "much cheaper hash is used"...

I presume hardware provided hash is less expensive (for our cpu) than
computing our rxhash...



--
To unsubscribe from this list: send the line "unsubscribe netdev" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html