New function, 'of_dev_get_platdata()'
 - provides unified handling of getting device platform data
 - supports DT and non-DT(legacy) cases
 - removes duplicated code from each driver
 - keeps driver specific code simple in each driver

Issues
------
On loading a driver, the driver tries to get the platform data.
* Check conditions prior to parsing the DT
  You can see various/not general way of checking the platform data.

  case 1) driver checks whether the platform data is null or not.

        foo_platform_data *pdata = dev_get_platdata(dev);
        if (!pdata) {
                pdata = foo_parse_dt();
                if (IS_ERR(pdata))
                        return PTR_ERR(pdata);
        }

  case 2) driver checks whether 'of_node' exists or not.

        if (dev.of_node) {
                pdata = foo_parse_dt();
                if (IS_ERR(pdata))
                        return PTR_ERR(pdata);
        }

  case 3) check both

        if (pdata) {
                copy pdata into driver private data
        } else if (dev.of_node) {
                pdata = foo_parse_dt();
        }

  Check conditions are very depend on the driver, but it can be unified.
  of_dev_get_platdata() provides unified handling. So the driver can reduce
  if-condition statements.

* Memory allocation in each parser
  In most cases, driver allocates the platform data inside its DT parser.

  static struct foo_platform_data *foo_parse_dt()
  {
        allocates memory for generating platform data.
        parsing DT properties and copy them into the platform data.
  }

  of_dev_get_platdata() allocates the platform data internally.
  And it calls back to the driver specific parser function. It removes
  memory allocation code in each driver.

* Two types of parser definition
  Many drivers implement DT parser in two cases, '#ifdef CONFIG_OF' and
  '#else'.

  #ifdef CONFIG_OF
  static struct foo_platform_data *foo_parse_dt()
  {
        (snip)
  }
  #else
  static struct foo_platform_data *foo_parse_dt()
  {
        return NULL;
  }
  #endif

  of_dev_get_platdata() supports both cases. It removes few lines of code
  in each driver.

What of_dev_get_platdata() does
-------------------------------
  if CONFIG_OF is not defined, return legacy code, 'dev_get_platdata()'.
  if platform data exists, just return it.
  if platform data is null, check the 'of node'.
    if of_node is null, then return NULL.
    Otherwise, allocates memory for platform data.
    Call back to the driver(caller) with allocated platform data and
    private data if needed.
    Then, driver will parse the DT internally and handle errors.

Examples
--------
  Following patches are examples.

  drivers/input/touchscreen/mms114.c
  drivers/mfd/tps65910.c
  drivers/usb/musb/ux500.c

Driver list
-----------
of_dev_get_platdata() can be applied into files below. You may find more
if you're interested in this. :)

  drivers/dma/ste_dma40.c
  drivers/gpio/gpio-rcar.c
  drivers/gpu/drm/exynos/exynos_dp_core.c
  drivers/gpu/drm/exynos/exynos_hdmi.c
  drivers/hwmon/ntc_thermistor.c
  drivers/i2c/busses/i2c-s3c2410.c
  drivers/iio/frequency/adf4350.c
  drivers/input/keyboard/matrix_keypad.c
  drivers/input/keyboard/samsung-keypad.c
  drivers/input/misc/drv260x.c
  drivers/input/misc/regulator-haptic.c
  drivers/input/misc/rotary_encoder.c
  drivers/input/touchscreen/atmel_mxt_ts.c
  drivers/input/touchscreen/auo-pixcir-ts.c
  drivers/input/touchscreen/bu21013_ts.c
  drivers/input/touchscreen/mms114.c
  drivers/input/touchscreen/pixcir_i2c_ts.c
  drivers/input/touchscreen/zforce_ts.c
  drivers/leds/leds-lp5521.c
  drivers/leds/leds-lp5523.c
  drivers/leds/leds-lp5562.c
  drivers/leds/leds-lp55xx-common.c
  drivers/leds/leds-lp8501.c
  drivers/leds/leds-mc13783.c
  drivers/leds/leds-pca963x.c
  drivers/leds/leds-tca6507.c
  drivers/mfd/max8997.c
  drivers/mfd/max8998.c
  drivers/mfd/sec-core.c
  drivers/mfd/tps6586x.c
  drivers/mfd/tps65910.c
  drivers/misc/lis3lv02d/lis3lv02d.c
  drivers/mmc/host/davinci_mmc.c
  drivers/mmc/host/dw_mmc.c
  drivers/mmc/host/sdhci-s3c.c
  drivers/mtd/devices/spear_smi.c
  drivers/mtd/nand/fsmc_nand.c
  drivers/mtd/nand/lpc32xx_mlc.c
  drivers/mtd/nand/lpc32xx_slc.c
  drivers/mtd/nand/sh_flctl.c
  drivers/net/can/sja1000/sja1000_platform.c
  drivers/net/ethernet/davicom/dm9000.c
  drivers/net/ethernet/marvell/mv643xx_eth.c
  drivers/net/ethernet/renesas/sh_eth.c
  drivers/net/wireless/brcm80211/brcmfmac/bcmsdh.c
  drivers/power/bq24735-charger.c
  drivers/power/gpio-charger.c
  drivers/power/sbs-battery.c
  drivers/power/tps65090-charger.c
  drivers/power/twl4030_charger.c
  drivers/pwm/pwm-lp3943.c
  drivers/pwm/pwm-samsung.c
  drivers/spi/spi-sh-msiof.c
  drivers/spi/spi/spi-s3c64xx.c
  drivers/thermal/db8500_thermal.c
  drivers/usb/misc/usb3503.c
  drivers/usb/musb/omap2430.c
  drivers/usb/musb/ux500.c
  drivers/usb/renesas_usbhs/common.c
  drivers/video/fbdev/simplefb.c
  drivers/video/backlight/lp855x_bl.c
  drivers/video/backlight/pwm_bl.c
  drivers/video/backlight/sky81452-backlight.c
  drivers/video/backlight/tps65217_bl.c

This is the RFC, so I would like to get some feedback prior to patching all
drivers. Any comments are welcome.

Cc: Dmitry Torokhov <[email protected]>
Cc: Felipe Balbi <[email protected]>
Cc: Grant Likely <[email protected]>
Cc: Greg Kroah-Hartman <[email protected]>
Cc: Lee Jones <[email protected]>
Cc: Rob Herring <[email protected]>
Cc: Samuel Ortiz <[email protected]>
Cc: Tony Lindgren <[email protected]>
Cc: [email protected]
Cc: [email protected]

Milo Kim (4):
  of: introduce of_dev_get_platdata()
  input: touchscree: mms114: use of_dev_get_platdata()
  mfd: tps65910: use of_dev_get_platdata()
  usb: musb: use of_dev_get_platdata()

 drivers/input/touchscreen/mms114.c | 34 ++++++++------------------
 drivers/mfd/tps65910.c             | 49 +++++++++++++-------------------------
 drivers/of/device.c                | 46 +++++++++++++++++++++++++++++++++++
 drivers/usb/musb/ux500.c           | 40 +++++++++++++------------------
 include/linux/of_device.h          | 12 ++++++++++
 5 files changed, 100 insertions(+), 81 deletions(-)

-- 
1.9.1

--
To unsubscribe from this list: send the line "unsubscribe devicetree" in
the body of a message to [email protected]
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Reply via email to