From mboxrd@z Thu Jan 1 00:00:00 1970 From: Russell King - ARM Linux Subject: Re: [PATCH 1/6] serial: 8250_dw: add support for clk api Date: Sat, 16 Mar 2013 00:29:52 +0000 Message-ID: <20130316002952.GA4977@n2100.arm.linux.org.uk> References: <1363377988-4966-1-git-send-email-maxime.ripard@free-electrons.com> <1363377988-4966-2-git-send-email-maxime.ripard@free-electrons.com> <20130315223917.GZ4977@n2100.arm.linux.org.uk> Mime-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable Return-path: Content-Disposition: inline In-Reply-To: List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Sender: "linux-arm-kernel" Errors-To: linux-arm-kernel-bounces+linux-arm-kernel=m.gmane.org@lists.infradead.org To: Emilio =?iso-8859-1?Q?L=F3pez?= Cc: Greg Kroah-Hartman , linux-kernel@vger.kernel.org, linux-serial@vger.kernel.org, sunny@allwinnertech.com, shuge@allwinnertech.com, Maxime Ripard , Jiri Slaby , kevin@allwinnertech.com, linux-arm-kernel@lists.infradead.org List-Id: linux-serial@vger.kernel.org On Fri, Mar 15, 2013 at 09:15:11PM -0300, Emilio L=F3pez wrote: > Hello Russell, > = > El 15/03/13 19:39, Russell King - ARM Linux escribi=F3: > > On Fri, Mar 15, 2013 at 09:06:23PM +0100, Maxime Ripard wrote: > >> + /* clock got configured through clk api, all done */ > >> + if (p->uartclk) > > = > > if (IS_ERR(p->uartclk)) > > = > = > Isn't IS_ERR for pointers? p->uartclk is an unsigned int Right, sorry, ignore that. > >> + return 0; > >> + > >> + /* try to find out clock frequency from DT as fallback */ > >> if (of_property_read_u32(np, "clock-frequency", &val)) { > >> - dev_err(p->dev, "no clock-frequency property set\n"); > >> + dev_err(p->dev, "clk or clock-frequency not defined\n"); > >> return -EINVAL; > >> } > >> p->uartclk =3D val; > >> @@ -294,9 +301,21 @@ static int dw8250_probe(struct platform_device *p= dev) > >> if (!uart.port.membase) > >> return -ENOMEM; > >> = > >> + data =3D devm_kzalloc(&pdev->dev, sizeof(*data), GFP_KERNEL); > >> + if (!data) > >> + return -ENOMEM; > >> + > >> + data->clk =3D devm_clk_get(&pdev->dev, NULL); > >> + if (IS_ERR(data->clk)) > >> + data->clk =3D NULL; > >> + else > >> + clk_prepare_enable(data->clk); > > = > > if (!IS_ERR(data->clk)) > > clk_prepare_enable(data->clk); > > = > = > See below > = > >> + > >> uart.port.iotype =3D UPIO_MEM; > >> uart.port.serial_in =3D dw8250_serial_in; > >> uart.port.serial_out =3D dw8250_serial_out; > >> + uart.port.private_data =3D data; > >> + uart.port.uartclk =3D clk_get_rate(data->clk); > > = > > What if data->clk is invalid? > > = > > if (!IS_ERR(data->clk) > > uart.port.uartclk =3D clk_get_rate(data->clk); > > = > = > I'm not sure if it is coincidental or the way it is supposed to be, but > when using the common clock framework, if you pass a NULL to > clk_get_rate, the function explicitly checks for it and returns 0. I > relied on that behaviour when implementing this; see the if..else block > above. Is this not always the case on other clock drivers? That's something that the common clock framework decided to do. It's not a defined part of the API though, so drivers shouldn't rely on this behaviour meaning anything special.