netdev - Re: [PATCH 3/4] ethtool: can't set combined and tx/rx channel counts at the same time

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: <1454968816.25441.20.camel@intel.com>
Date:	Mon, 8 Feb 2016 22:00:16 +0000
From:	"Keller, Jacob E" <jacob.e.keller@...el.com>
To:	"moorray3@...pl" <moorray3@...pl>
CC:	"netdev@...r.kernel.org" <netdev@...r.kernel.org>,
	"davem@...emloft.net" <davem@...emloft.net>
Subject: Re: [PATCH 3/4] ethtool: can't set combined and tx/rx channel
 counts at the same time

On Mon, 2016-02-08 at 20:52 +0000, Jakub Kicinski wrote:
> On Mon,  8 Feb 2016 12:06:04 -0800, Jacob Keller wrote:
> > +	/* can't set combined and separate channels at the same
> > time */
> > +	if ((channels.combined_count &&
> > +	     (channels.rx_count || channels.tx_count))
> > +	    return -EINVAL;
> > +
> >  	/* ensure the new Rx count fits within the configured Rx
> > flow
> >  	 * indirection table settings */
> >  	if (netif_is_rxfh_configured(dev) &&
> 
> My understanding is that unsymmetrical rx/tx queue configuration with
> combined IRQ vectors should be expressed as:
> 
> combined = min(tx, rx);
> num_tx = tx - combined;
> num_rx = rx - combined;
> 
> Please see:
> 
> https://www.mail-archive.com/netdev@vger.kernel.org/msg93977.html
> 
> Patches which I cooked up for this are rotting in my queue blocked by
> other things, unfortunately.

That's definitely not how any driver currently uses it today, that I
can see... At least drivers/net/ethernet/broadcom/bnxt/bnxt_ethtool.c
is currently assuming you can't set combined and asymmetric queues at
the same time.

I can ok dropping this change, and looking at implementing what you
suggest, as that makes more sense to me.

Regards,
Jake