| Message ID | 20140822131150.GA26724@amd |
|---|---|
| State | Superseded |
| Headers | show |
> _But_ this should probably be enabled using device tree entry, right? > Unfortunately, the driver is i2c driver, not platform one, so I don't > see how to do that easily... Trickle charging is usually depending on the hardware setup, so devicetree is actually a good place to put it. We'd just need generic bindings so that they could be applied to various RTC. Just guessing, resistor value would probably be one such property? We would not need putting plain register values into DT. There should be no difference regarding DT between i2c and platform drivers. You should add the devicetree ML for discussing such things. > Or would module parameter be acceptable? No, this is device dependent, not module dependent. > * Copyright (C) 2009 Semihalf. > + * Copyright (C) 2014 Pavel Machek <pavel@denx.de> IMO, such a small improvement does not justify claiming copyright on the driver.
On Fri 2014-08-22 09:28:52, Wolfram Sang wrote: > > > _But_ this should probably be enabled using device tree entry, right? > > Unfortunately, the driver is i2c driver, not platform one, so I don't > > see how to do that easily... > > Trickle charging is usually depending on the hardware setup, so > devicetree is actually a good place to put it. We'd just need generic > bindings so that they could be applied to various RTC. Just guessing, > resistor value would probably be one such property? We would not need > putting plain register values into DT. Well, for bq32000 there's selection between "no charging", "diode + resistor", "different resistor". > There should be no difference regarding DT between i2c and platform > drivers. Do you have example how to acces of_ from i2c driver? Platform drivers do: static int palmas_rtc_probe(struct platform_device *pdev) { struct palmas *palmas = dev_get_drvdata(pdev->dev.parent); struct palmas_rtc *palmas_rtc = NULL; int ret; bool enable_bb_charging = false; bool high_bb_charging; if (pdev->dev.of_node) { enable_bb_charging = of_property_read_bool(pdev->dev.of_node, "ti,backup-battery-chargeable"); high_bb_charging = of_property_read_bool(pdev->dev.of_node, "ti,backup-battery-charge-high-current"); Thanks, Pavel
> Do you have example how to acces of_ from i2c driver? > > Platform drivers do: > > if (pdev->dev.of_node) { As I said, this is basically the same. Just use the node from the client's device. Nothing special here.
diff --git a/drivers/rtc/rtc-bq32k.c b/drivers/rtc/rtc-bq32k.c index c74bf0d..b6a4dd0 100644 --- a/drivers/rtc/rtc-bq32k.c +++ b/drivers/rtc/rtc-bq32k.c @@ -2,10 +2,14 @@ * Driver for TI BQ32000 RTC. * * Copyright (C) 2009 Semihalf. + * Copyright (C) 2014 Pavel Machek <pavel@denx.de> * * This program is free software; you can redistribute it and/or modify * it under the terms of the GNU General Public License version 2 as * published by the Free Software Foundation. + * + * You can get hardware description at + * http://www.ti.com/lit/ds/symlink/bq32000.pdf */ #include <linux/module.h> @@ -27,6 +31,10 @@ #define BQ32K_CENT 0x40 /* Century flag */ #define BQ32K_CENT_EN 0x80 /* Century flag enable bit */ +#define BQ32K_CALIBRATION 0x07 /* CAL_CFG1, calibration and control */ +#define BQ32K_TCH2 0x08 /* Trickle charge enable */ +#define BQ32K_CFG2 0x09 /* Trickle charger control */ + struct bq32k_regs { uint8_t seconds; uint8_t minutes; @@ -153,6 +161,25 @@ static int bq32k_probe(struct i2c_client *client, if (error) return error; + if (0) { + /* + * TCHE[3:0] == 0x05, TCH2 == 1, TCFE == 0 (charging + * over diode and 940ohm resistor) + */ + + reg = 0x05; + error = bq32k_write(dev, ®, BQ32K_CFG2, 1); + if (error) + return error; + + reg = 0x20; + error = bq32k_write(dev, ®, BQ32K_TCH2, 1); + if (error) + return error; + + dev_info(dev, "Enabled trickle RTC battery charge.\n"); + } + rtc = devm_rtc_device_register(&client->dev, bq32k_driver.driver.name, &bq32k_rtc_ops, THIS_MODULE); if (IS_ERR(rtc))