From mboxrd@z Thu Jan 1 00:00:00 1970 From: broonie@opensource.wolfsonmicro.com (Mark Brown) Date: Wed, 4 Jul 2012 16:33:08 +0100 Subject: [PATCH 1/4] mfd: support 88pm80x in 80x driver In-Reply-To: <201207041527.13954.arnd@arndb.de> References: <1341392115-9425-1-git-send-email-zhouqiao@marvell.com> <201207041137.49020.arnd@arndb.de> <201207041527.13954.arnd@arndb.de> Message-ID: <20120704153307.GE4111@opensource.wolfsonmicro.com> To: linux-arm-kernel@lists.infradead.org List-Id: linux-arm-kernel.lists.infradead.org On Wed, Jul 04, 2012 at 03:27:13PM +0000, Arnd Bergmann wrote: > On Wednesday 04 July 2012, Qiao Zhou wrote: > > >On the other hand, I think it probably makes sense to drop the irq_base > > >member in this struct and rely on irq domains to allocate them dynamically > > >as mentioned before. > > Do you mean that both regmap_add_irq_chip and mfd_add_devices api pass -1 as > > the irq_base, so that system can dynamically allocate the irq_base for it? > regmap_add_irq_chip should pass -1, mfd_add_devices should pass 0. > Mark can probably correct me if that's wrong. That's right. I do need to grovel through the irqdomain code and try to figure out if the stuff added recently for MFDs to pass an irqdomain about would also support doing the same mapping. Unfortunately all the irqdomain code I found that I didn't write was rather tied to DT which makes things more obscure, it's not clear what's for irqdomain and what's for DT. -------------- next part -------------- A non-text attachment was scrubbed... Name: signature.asc Type: application/pgp-signature Size: 836 bytes Desc: Digital signature URL: