From: thomas.petazzoni@free-electrons.com (Thomas Petazzoni) Date: Wed, 9 Jul 2014 23:16:54 +0200 Subject: [PATCH RFC] Armada XP (mvebu) Hot-plug issues continued In-Reply-To: References: Message-ID: <20140709231654.20e28529@free-electrons.com> To: linux-arm-kernel@lists.infradead.org List-Id: linux-arm-kernel.lists.infradead.org Dear Matthew Minter, On Wed, 9 Jul 2014 18:08:07 +0100, Matthew Minter wrote: > This is something of a follow up to my previous thread: > Armada XP (mvebu) PCIe memory (BAR/window) re-allocation > > Thanks to the helpful patches provided to the mvebu PCIe subsystem, > hot-plug PCIe devices worked much better on this board, however I have > discovered one remaining issue regarding them, due to the fact the > legacy style IRQ is assigned during the fake "PCI BIOS" phase, devices > which are plugged after boot will not be given an IRQ and thus if they > do not support MSI/MSI-X will fail to initialize. > > This seems a fairly simple problem to squash and as such I have > written a small patch which helps solve this, however I am unsure if I > am doing this in a sensible way or if a different kind of rework is > needed here. Also, if this is the wrong mailing list to ask about this > please say so. > > Any comments regarding this patch or a better way to achieve this > would be greatly appreciated. Patch follows (sorry if the format is > off, still getting used to git format-patch). My knowledge of the PCI core is a bit too limited to judge whether this is the right solution or not. However, I would recommend you to send the patch and problem description to the linux-pci at vger.kernel.org mailing list, with Bjorn Helgaas in the recipients (he is the PCI subsystem maintainer). It would indeed be good if you could use git send-email, because your patch has been badly modified by your e-mail client :/ Best regards, Thomas -- Thomas Petazzoni, CTO, Free Electrons Embedded Linux, Kernel and Android engineering http://free-electrons.com