From mboxrd@z Thu Jan 1 00:00:00 1970 From: Bjorn Helgaas Date: Tue, 13 Dec 2005 17:47:14 +0000 Subject: Re: [PATCH]: Prevent sn2 ptc code from executing on all ia64 subarches Message-Id: <200512131047.14264.bjorn.helgaas@hp.com> List-Id: References: <20051121180016.24224.2378.sendpatchset@prarit.boston.redhat.com> In-Reply-To: <20051121180016.24224.2378.sendpatchset@prarit.boston.redhat.com> MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit To: linux-ia64@vger.kernel.org On Tuesday 13 December 2005 10:24 am, Luck, Tony wrote: > > I think this is too complicated. Initcall ordering is fragile as it > > is. > ... > > > Can you use the existing machine vector infrastructure to solve this > > problem? For example, what if you added a "platform_device_setup()" > > hook that is a no-op on most platforms, but maps to sn_device_setup() > > in the SN2 machine vector? platform_device_setup() would itself be > > called as a device_initcall(). > > This solution would move all the sn2 initializaions into the > device_initcall sequence. Which might cause some ordering > problems. E.g. sn_pci_init() is a "subsys_initcall" ... and I > expect that it really needs to happen before any of the > device_initcalls. Well, sure. device_setup() was just an example. I expect they'd want to add more (subsys, etc). I doubt that all the levels would be needed, but even if they were, I'd prefer it over sprinkling "ia64_platform_is" everywhere.