[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <Pine.WNT.4.64.1001080023050.15320@ppwaskie-MOBL2.amr.corp.intel.com>
Date: Fri, 8 Jan 2010 00:25:26 -0800 (Pacific Standard Time)
From: "Waskiewicz Jr, Peter P" <peter.p.waskiewicz.jr@...el.com>
To: David Miller <davem@...emloft.net>
cc: "Kirsher, Jeffrey T" <jeffrey.t.kirsher@...el.com>,
"netdev@...r.kernel.org" <netdev@...r.kernel.org>,
"gospo@...hat.com" <gospo@...hat.com>,
"Waskiewicz Jr, Peter P" <peter.p.waskiewicz.jr@...el.com>
Subject: Re: [net-next-2.6 PATCH 2/5] ixgbe: Make descriptor ring allocations
NUMA-aware
On Fri, 8 Jan 2010, David Miller wrote:
> From: Jeff Kirsher <jeffrey.t.kirsher@...el.com>
> Date: Wed, 06 Jan 2010 20:48:46 -0800
>
> > @@ -147,7 +147,7 @@ struct ixgbe_ring {
> >
> > #ifdef CONFIG_IXGBE_DCA
> > /* cpu for tx queue */
> > - int cpu;
> > + u8 cpu;
> > #endif
> >
> > u16 work_limit; /* max work per interrupt */
>
> Is truncating cpu and node numbers to 8-bits ok? I really don't
> see how it can be fine, even for DCA.
In our hardware (82598 and 82599), the CPU field in the DCA registers is 8
bits.
The reason I truncated this was to fit these values into the
first cacheline of the struct. I also didn't figure we'd see a system
that would have 255 NUMA nodes before 10 GbE was something that was on the
shelf collecting dust. :-)
> This is especially the case since dca3_get_tag() and the
> DCA ->get_tag() callback explicitly take an 'int' argument
> too.
Cheers,
-PJ
--
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
Powered by blists - more mailing lists