From mboxrd@z Thu Jan 1 00:00:00 1970 From: Peter Ujfalusi Subject: Re: [PATCH] mfd: twl6040: Optional clk32k clock handling Date: Mon, 28 Apr 2014 13:46:50 +0300 Message-ID: <535E319A.9050104@ti.com> References: <1396522482-32036-1-git-send-email-peter.ujfalusi@ti.com> <20140428103629.GI21279@lee--X1> Mime-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: QUOTED-PRINTABLE Return-path: In-Reply-To: <20140428103629.GI21279@lee--X1> Sender: linux-kernel-owner@vger.kernel.org To: Lee Jones Cc: sameo@linux.intel.com, devicetree@vger.kernel.org, linux-doc@vger.kernel.org, linux-kernel@vger.kernel.org List-Id: devicetree@vger.kernel.org On 04/28/2014 01:36 PM, Lee Jones wrote: >> In certain boards the source for the clk32k clock can be gated. In t= hese >> boards the clk32k clock can be provided to the driver and it is goin= g to be >> enabled/disabled when it is needed. >> If the clk32k clock is not provided the driver will assume that it i= s always >> running. >> >> Signed-off-by: Peter Ujfalusi >> --- >> Documentation/devicetree/bindings/mfd/twl6040.txt | 2 ++ >> drivers/mfd/twl6040.c | 10 ++++++++++ >> include/linux/mfd/twl6040.h | 2 ++ >> 3 files changed, 14 insertions(+) >> >> diff --git a/Documentation/devicetree/bindings/mfd/twl6040.txt b/Doc= umentation/devicetree/bindings/mfd/twl6040.txt >> index 0f5dd709d752..a41157b5d930 100644 >> --- a/Documentation/devicetree/bindings/mfd/twl6040.txt >> +++ b/Documentation/devicetree/bindings/mfd/twl6040.txt >> @@ -19,6 +19,8 @@ Required properties: >> =20 >> Optional properties, nodes: >> - enable-active-high: To power on the twl6040 during boot. >> +- clocks: phandle to the clk32k clock provider >> +- clock-names: Must be "clk32k" >> =20 >> Vibra functionality >> Required properties: >> diff --git a/drivers/mfd/twl6040.c b/drivers/mfd/twl6040.c >> index 2e6504a8e1e3..12b314ea48dc 100644 >> --- a/drivers/mfd/twl6040.c >> +++ b/drivers/mfd/twl6040.c >> @@ -291,6 +291,8 @@ int twl6040_power(struct twl6040 *twl6040, int o= n) >> if (twl6040->power_count++) >> goto out; >> =20 >> + clk_prepare_enable(twl6040->clk32k); >> + >> /* Allow writes to the chip */ >> regcache_cache_only(twl6040->regmap, false); >> =20 >> @@ -346,6 +348,8 @@ int twl6040_power(struct twl6040 *twl6040, int o= n) >> =20 >> twl6040->sysclk =3D 0; >> twl6040->mclk =3D 0; >> + >> + clk_disable_unprepare(twl6040->clk32k); >> } >> =20 >> out: >> @@ -644,6 +648,12 @@ static int twl6040_probe(struct i2c_client *cli= ent, >> =20 >> i2c_set_clientdata(client, twl6040); >> =20 >> + twl6040->clk32k =3D devm_clk_get(&client->dev, "clk32k"); >> + if (IS_ERR(twl6040->clk32k)) { >> + dev_info(&client->dev, "clk32k is not handled\n"); >> + twl6040->clk32k =3D NULL; >=20 > So what happens if you pass a NULL clk reference to clk_prepare_enabl= e()? We have checks for !clk in all cases within the clk_* API. If we pass N= ULL, nothing will happen and from the calling driver point of view it will l= ook like a success. AFAIK this is the way to handle non esential/optional clocks without if= ()s in the driver code around the clock handling. --=20 P=C3=A9ter