From mboxrd@z Thu Jan 1 00:00:00 1970 From: Stephen Warren Subject: Re: [PATCH 5/9] usb: phy: tegra: Register as an USB PHY. Date: Mon, 01 Jul 2013 15:53:16 -0600 Message-ID: <51D1FA4C.8000507@wwwdotorg.org> References: <1372455427-20898-1-git-send-email-ttynkkynen@nvidia.com> <1372455427-20898-6-git-send-email-ttynkkynen@nvidia.com> Mime-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit Return-path: In-Reply-To: <1372455427-20898-6-git-send-email-ttynkkynen@nvidia.com> Sender: linux-kernel-owner@vger.kernel.org To: Tuomas Tynkkynen Cc: balbi@ti.com, linux-tegra@vger.kernel.org, linux-kernel@vger.kernel.org, linux-usb@vger.kernel.org, gregkh@linuxfoundation.org List-Id: linux-tegra@vger.kernel.org On 06/28/2013 03:37 PM, Tuomas Tynkkynen wrote: > Register the Tegra PHY device instances with the PHY subsystem so that > the Tegra EHCI driver can locate a PHY via the standard APIs. > diff --git a/drivers/usb/phy/phy-tegra-usb.c b/drivers/usb/phy/phy-tegra-usb.c > + err = usb_add_phy_dev(&tegra_phy->u_phy); > + if (err < 0) { > + tegra_usb_phy_close(&tegra_phy->u_phy); > + return err; > + } Don't you want to do that a bit later? In particular ... > tegra_phy->u_phy.shutdown = tegra_usb_phy_close; > tegra_phy->u_phy.set_suspend = tegra_usb_phy_suspend; > > dev_set_drvdata(&pdev->dev, tegra_phy); ... that all happens *after* the call to usb_add_phy_dev() call, yet I think the USB core could legally call any of the PHY methods after the call to usb_add_phy_dev(), and that probably requires all those methods pointers and drvdata to be set up? I suggest moving the call to usb_add_phy_dev() to the very end of the function. > +static int tegra_usb_phy_remove(struct platform_device *pdev) > +{ > + struct tegra_usb_phy *tegra_phy = platform_get_drvdata(pdev); > + > + usb_remove_phy(&tegra_phy->u_phy); > + return 0; > +} Nit: I'd typically expect to see a blank line before the return statement.