From mboxrd@z Thu Jan 1 00:00:00 1970 From: lyz@rock-chips.com (Yunzhi Li) Date: Fri, 12 Dec 2014 11:43:25 +0800 Subject: [PATCH v6 1/5] phy: add a driver for the Rockchip SoC internal USB2.0 PHY In-Reply-To: References: <1418291722-25448-1-git-send-email-lyz@rock-chips.com> <1418291722-25448-2-git-send-email-lyz@rock-chips.com> Message-ID: <548A645D.8010608@rock-chips.com> To: linux-arm-kernel@lists.infradead.org List-Id: linux-arm-kernel.lists.infradead.org Hi Doug: On 2014/12/12 2:09, Doug Anderson wrote: > Yunzhi, > > On Thu, Dec 11, 2014 at 1:55 AM, Yunzhi Li wrote: >> + rk_phy->clk = of_clk_get(child, 0); >> + if (IS_ERR(rk_phy->clk)) { >> + dev_warn(dev, "failed to get clock\n"); >> + rk_phy->clk = NULL; >> + } > The device tree bindings don't specify a clock and the "dtsi" added to > rk3288 don't reference a clock. Take that code out and avoid a > warning in the logs at bootup. > > ...or should there be a clock? Actually, there is a clk gating control bit in CRU for each usb phy and I think we should manage these clocks by the usb phy driver, so I will add clock property to usb PHYs nodes in next version of patche set. > >> + rk_phy->phy = devm_phy_create(dev, NULL, &ops); > This has the wrong number of arguments. Even before the change that > added the 4th argument, this is still wrong because "ops" is supposed > to be the 2nd argument, not the 3rd. > > ...so I'm confused how this compiled for you. I think this ought to be: > > rk_phy->phy = devm_phy_create(dev, child, &ops, NULL); > > ...but please correct me if I'm mistaken! > > >