From mboxrd@z Thu Jan 1 00:00:00 1970 From: Yoshihiro Shimoda Subject: Re: [PATCH v6 1/1] usb: host: xhci-plat: add support for the R-Car H2 and M2 xHCI controllers Date: Thu, 26 Jun 2014 09:18:19 +0900 Message-ID: <53AB66CB.7010401@renesas.com> References: <539AEC8F.4030003@renesas.com> <20140613142525.GE8319@saruman.home> Mime-Version: 1.0 Content-Type: text/plain; charset="UTF-8" Content-Transfer-Encoding: 7bit Return-path: In-Reply-To: Sender: linux-sh-owner@vger.kernel.org To: Magnus Damm , Felipe Balbi Cc: "mathias.nyman@intel.com" , Greg Kroah-Hartman , "linux-usb@vger.kernel.org" , SH-Linux , Geert Uytterhoeven , Grant Likely , Rob Herring , "devicetree@vger.kernel.org" List-Id: devicetree@vger.kernel.org Hi Felipe, (2014/06/18 14:15), Magnus Damm wrote: > Hi Felipe, > > On Fri, Jun 13, 2014 at 11:25 PM, Felipe Balbi wrote: >> Hi, >> >> On Fri, Jun 13, 2014 at 09:20:31PM +0900, Yoshihiro Shimoda wrote: < snip > >>> +/* This function needs to initialize a "phy" of usb before */ >> >> initializing a PHY looks like something that the PHY layer should do. >> Why don't you write a PHY driver and teach xhci-core about PHYs ? Then, >> more people would benefit. > > Could you please clarify what you would like Shimoda-san to do? > > Like Ben and Shimoda-san mentioned, there is already a PHY driver > developed for this SoC. The PHY driver is however shared in various > ways. For example, on one particular SoC the same PHY driver is > interfacing to a total of 3 different variants of USB controllers > where XHCI is one of them. And to make things even more complicated, > depending on SoC variant the XHCI hardware may or may not be available > - but the PHY portion is more or less the same. > > Putting the firmware loader in the XHCI driver like this at least > keeps it together with the rest of the XHCI stuff and also makes it > possible to easily access the XHCI I/O register window for firmware > loading. Moving the firmware loading to the PHY driver however > complicates the situation when it comes to Kconfig handling of XHCI > and PHY driver and also forces the PHY driver to access the XHCI > hardware I/O registers for firmware loading... Do you have any comment about Magnus's comment? Best regards, Yoshihiro Shimoda > Cheers, > > / magnus >