From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1750779AbWHVFtl (ORCPT ); Tue, 22 Aug 2006 01:49:41 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1750775AbWHVFtl (ORCPT ); Tue, 22 Aug 2006 01:49:41 -0400 Received: from gate.crashing.org ([63.228.1.57]:1960 "EHLO gate.crashing.org") by vger.kernel.org with ESMTP id S1750785AbWHVFtk (ORCPT ); Tue, 22 Aug 2006 01:49:40 -0400 Subject: struct msix_entry bogosity From: Benjamin Herrenschmidt To: linux-pci maillist Cc: Greg KH , Linux Kernel list Content-Type: text/plain Date: Tue, 22 Aug 2006 15:49:22 +1000 Message-Id: <1156225762.21752.82.camel@localhost.localdomain> Mime-Version: 1.0 X-Mailer: Evolution 2.6.1 Content-Transfer-Encoding: 7bit Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org Any objection to something like that ? (And of course fixing the few drivers for the name change) struct msix_entry { - u16 vector; /* kernel uses to write allocated vector */ + unsigned int irq; /* kernel uses to write allocated interrupt */ u16 entry; /* driver uses to specify entry, OS writes */ }; Maybe x86 has a 1:1 hw vector <-> linux irq numbers (does it btw ?) but other archs certainly don't. pci_enable_msix API should return a linux IRQ number for the driver to pass to request_irq(), certainly not a hw vector number, and the normal type for an irq is unsigned int :) If it's ok, I'll do a patch changing that and fixing all in-tree users. Ben.