From mboxrd@z Thu Jan 1 00:00:00 1970 From: Greg KH Date: Thu, 16 Mar 2006 23:49:06 +0000 Subject: Re: [PATCH] (-mm) drivers/pci/msi: explicit declaration of msi_register Message-Id: <20060316234906.GA24675@suse.de> List-Id: References: <44172F0E.6070708@ce.jp.nec.com> <20060314134535.72eb7243.akpm@osdl.org> <44176502.9050109@ce.jp.nec.com> <20060315235544.GA6504@suse.de> <44198210.6090109@ce.jp.nec.com> <20060316181934.GM13666@sgi.com> <20060316234118.GB9746@esmail.cup.hp.com> In-Reply-To: <20060316234118.GB9746@esmail.cup.hp.com> MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: Grant Grundler Cc: Mark Maule , Jun'ichi Nomura , Andrew Morton , linux-kernel@vger.kernel.org, linux-ia64@vger.kernel.org, shaohua.li@intel.com On Thu, Mar 16, 2006 at 03:41:18PM -0800, Grant Grundler wrote: > On Thu, Mar 16, 2006 at 12:19:34PM -0600, Mark Maule wrote: > > If there's objectins to having struct msi_ops declared in pci.h, then I guess > > we need to come up with another solution. > > There are other transaction based interrupt subsystems that are typically > arch specific (e.g. GSC device interrupts on PA-RISC). So far, MSI is the > only generic one and that is clearly part of the PCI spec. Yes, that's fine. But the core pci msi structures do not need to be exported for the whole kernel to see, right? That's my only point here. pci.h is cluttered enough as it is :) thanks, greg k-h