From mboxrd@z Thu Jan 1 00:00:00 1970 From: David Miller Subject: Re: [PATCH 2/2]: softirq: Add support for triggering softirq work on softirqs. Date: Sat, 20 Sep 2008 03:35:45 -0700 (PDT) Message-ID: <20080920.033545.171821293.davem@davemloft.net> References: <20080919.234832.127229997.davem@davemloft.net> <20080920004608.a3123052.akpm@linux-foundation.org> Mime-Version: 1.0 Content-Type: Text/Plain; charset=us-ascii Content-Transfer-Encoding: 7bit Cc: linux-kernel@vger.kernel.org, netdev@vger.kernel.org, jens.axboe@oracle.com, steffen.klassert@secunet.com To: akpm@linux-foundation.org Return-path: Received: from 74-93-104-97-Washington.hfc.comcastbusiness.net ([74.93.104.97]:39066 "EHLO sunset.davemloft.net" rhost-flags-OK-FAIL-OK-OK) by vger.kernel.org with ESMTP id S1751382AbYITKf6 (ORCPT ); Sat, 20 Sep 2008 06:35:58 -0400 In-Reply-To: <20080920004608.a3123052.akpm@linux-foundation.org> Sender: netdev-owner@vger.kernel.org List-ID: From: Andrew Morton Date: Sat, 20 Sep 2008 00:46:08 -0700 > Seems from the above `>> 16' you're stuffing the softirq ID into the > upper 16 bits of call_single_data.flags. Is it needed? Another u32 > can be added to call_single_data for free on 64-bit, or it could be > split into two u16's. call_single_data is already too large, simply adding it to struct sk_buff had negative performance impacts that are forcing me to trim the size of the existing structure a bit. I think just documenting the usage of the bits is better, and I'll do that in my next rev.