From mboxrd@z Thu Jan 1 00:00:00 1970 From: "Andrew F. Davis" Subject: Re: [PATCH v4 3/7] power: bq27xxx_battery: Platform initialization must declare a device Date: Mon, 21 Sep 2015 13:40:53 -0500 Message-ID: <56004F35.3040609@ti.com> References: <1442265973-32261-1-git-send-email-afd@ti.com> <1442265973-32261-4-git-send-email-afd@ti.com> <20150921084129.GH20113@pali> Mime-Version: 1.0 Content-Type: text/plain; charset=utf-8; format=flowed Content-Transfer-Encoding: QUOTED-PRINTABLE Return-path: Received: from bear.ext.ti.com ([192.94.94.41]:52195 "EHLO bear.ext.ti.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753072AbbIUSlV (ORCPT ); Mon, 21 Sep 2015 14:41:21 -0400 In-Reply-To: <20150921084129.GH20113@pali> Sender: linux-pm-owner@vger.kernel.org List-Id: linux-pm@vger.kernel.org To: =?UTF-8?Q?Pali_Roh=c3=a1r?= Cc: Sebastian Reichel , Dmitry Eremin-Solenikov , David Woodhouse , Dan Murphy , linux-pm@vger.kernel.org, linux-kernel@vger.kernel.org On 09/21/2015 03:41 AM, Pali Roh=C3=A1r wrote: > On Monday 14 September 2015 16:26:09 Andrew F. Davis wrote: >> When initialized as a platform device the initializer must now speci= fy >> a device. An empty device name is no longer valid. >> >> Signed-off-by: Andrew F. Davis >> --- >> drivers/power/bq27xxx_battery.c | 9 ++++++--- >> drivers/w1/slaves/w1_bq27000.c | 1 + >> include/linux/power/bq27xxx_battery.h | 6 +++++- >> 3 files changed, 12 insertions(+), 4 deletions(-) >> >> diff --git a/drivers/power/bq27xxx_battery.c b/drivers/power/bq27xxx= _battery.c >> index e72055c..0734413 100644 >> --- a/drivers/power/bq27xxx_battery.c >> +++ b/drivers/power/bq27xxx_battery.c >> @@ -91,8 +91,6 @@ struct bq27xxx_access_methods { >> int (*read)(struct bq27xxx_device_info *di, u8 reg, bool single); >> }; >> >> -enum bq27xxx_chip { BQ27000, BQ27500, BQ27425, BQ27742, BQ27510}; >> - >> struct bq27xxx_reg_cache { >> int temperature; >> int time_to_empty; >> @@ -1036,6 +1034,11 @@ static int bq27xxx_battery_platform_probe(str= uct platform_device *pdev) >> return -EINVAL; >> } >> >> + if (!pdata->chip) { >> + dev_err(&pdev->dev, "no device supplied\n"); >> + return -EINVAL; >> + } >> + >> di =3D devm_kzalloc(&pdev->dev, sizeof(*di), GFP_KERNEL); >> if (!di) >> return -ENOMEM; >> @@ -1043,7 +1046,7 @@ static int bq27xxx_battery_platform_probe(stru= ct platform_device *pdev) >> platform_set_drvdata(pdev, di); >> >> di->dev =3D &pdev->dev; >> - di->chip =3D BQ27000; >> + di->chip =3D pdata->chip; >> >> name =3D pdata->name ?: dev_name(&pdev->dev); >> di->bus.read =3D &bq27xxx_battery_platform_read; >> diff --git a/drivers/w1/slaves/w1_bq27000.c b/drivers/w1/slaves/w1_b= q27000.c >> index 8480531..9f4a86b 100644 >> --- a/drivers/w1/slaves/w1_bq27000.c >> +++ b/drivers/w1/slaves/w1_bq27000.c >> @@ -42,6 +42,7 @@ static int w1_bq27000_read(struct device *dev, uns= igned int reg) >> static struct bq27xxx_platform_data bq27000_battery_info =3D { >> .read =3D w1_bq27000_read, >> .name =3D "bq27000-battery", >> + .chip =3D BQ27000, >> }; >> >> static int w1_bq27000_add_slave(struct w1_slave *sl) >> diff --git a/include/linux/power/bq27xxx_battery.h b/include/linux/p= ower/bq27xxx_battery.h >> index e70a93a..a4efb10 100644 >> --- a/include/linux/power/bq27xxx_battery.h >> +++ b/include/linux/power/bq27xxx_battery.h >> @@ -3,7 +3,8 @@ >> >> /** >> * struct bq27xxx_plaform_data - Platform data for bq27xxx devices >> - * @name: Name of the battery. If NULL the driver will fallback to = "bq27000". >> + * @name: Name of the battery. >> + * @chip: Chip class number of this device. >> * @read: HDQ read callback. >> * This function should provide access to the HDQ bus the battery = is >> * connected to. >> @@ -11,8 +12,11 @@ >> * register to be read. The return value should either be the cont= ent of >> * the passed register or an error value. >> */ >> +enum bq27xxx_chip { BQ27000 =3D 1, BQ27500, BQ27425, BQ27742, BQ275= 10 }; >> + >> struct bq27xxx_platform_data { >> const char *name; >> + enum bq27xxx_chip chip; > > Do we still need name property if we have chip enum? And if yes who u= se > it and why? > The enum only relates to the family of chip, not its name, passing name= with platform data allows us to give unique names to the supply should we ne= ed. >> int (*read)(struct device *dev, unsigned int); >> }; >> >