From mboxrd@z Thu Jan 1 00:00:00 1970 From: maxime.ripard@free-electrons.com (Maxime Ripard) Date: Fri, 18 Sep 2015 15:55:38 +0200 Subject: [PATCH 0/5] ARM: sunxi: Introduce CHIP support In-Reply-To: <13485899.XsimukeSAF@wuerfel> References: <1442566130-3497-1-git-send-email-maxime.ripard@free-electrons.com> <2353279.QTDzFUHVnF@wuerfel> <20150918121905.GO4684@lukather> <13485899.XsimukeSAF@wuerfel> Message-ID: <20150918135538.GS4684@lukather> To: linux-arm-kernel@lists.infradead.org List-Id: linux-arm-kernel.lists.infradead.org On Fri, Sep 18, 2015 at 03:18:57PM +0200, Arnd Bergmann wrote: > On Friday 18 September 2015 14:19:05 Maxime Ripard wrote: > > > > There's actually two issues that are pretty orthogonal: > > > > * the fact that the regulator to power the wifi chip needs to be > > enabled before it enumerates and the driver is probed. The MMC > > pwrseq stuff seems to fix that, even if it doesn't really fit in > > the usual DT model. > > Ok. If there is anything missing in the pwrseq implementation or > binding at the moment, you'll have to work with Ulf to get it to > drive your hardware then. Yep. > > * The fact that that regulator on this board is actually two > > regulators in parallel that needs to be controlled in parallel > > too. Here, it will probably require a new regulator driver to > > combine two regulators into one. > > I don't yet understand what it means to control them in parallel. > Do you mean the hardware uses two outputs to support a higher > current than you could do with one? Yes. > How critical is the timing? Would it be enough if the mmc > code knew how to set more than one regulator and do them > one at a time? I don't think it's critical as in both should happen exactly at the same time (but I've not really tried it either, so I'm not 100% sure). And I don't think it's something the MMC core should deal with. The device itself has a single pin to provide its power, it's the board that is wired that way. The way I was seeing it was more that I needed to create a regulator driver that would take two inputs, that the mmc core would use. Last time I discussed it with Mark (Brown), it semes like he was open to the idea, so I just need to find some time to do just that. Maxime -- Maxime Ripard, Free Electrons Embedded Linux, Kernel and Android engineering http://free-electrons.com -------------- next part -------------- A non-text attachment was scrubbed... Name: signature.asc Type: application/pgp-signature Size: 819 bytes Desc: Digital signature URL: