From mboxrd@z Thu Jan 1 00:00:00 1970 From: Eran Duchan Subject: Re: [PATCH] i2c: Adding the i2c-bit-platform bus Date: Sat, 7 May 2011 18:47:45 +0300 Message-ID: References: Mime-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: QUOTED-PRINTABLE Return-path: In-Reply-To: Sender: linux-i2c-owner-u79uwXL29TY76Z2rM5mHXA@public.gmane.org To: =?UTF-8?Q?H=C3=A5vard_Skinnemoen?= , linux-i2c-u79uwXL29TY76Z2rM5mHXA@public.gmane.org, Grant Likely List-Id: linux-i2c@vger.kernel.org Argh, the patch won't pass checkpatch because I brilliantly submitted it via gmail. Once the 2 issues (initializing SDA/SCL and platform structure) are discussed and agreed I'll resubmit the patch via git. Thanks Eran On Sat, May 7, 2011 at 2:53 PM, Eran Duchan wrote: > Hi guys, > > Following the discussion on the matter of i2c-gpio performance, I'm > submitting this very simple patch which adds i2c-bit-platform. Using > this driver, a platform can register callbacks called by i2c-bit-algo > so that it can modify GPIO pins directly or do pretty much anything > else. Besides style, I see 2 points for discussion: > > 1) Should the i2c-bit-platform call setscl/setsda passing state to 1 > (so that SDA/SCL are explicitly pulled high) on probe or should this > be implemented in the platforms which mandate this? In the patch I > left it up to the platform. > > 2) Which platform_data structure is passed to the driver? There are > three options: > 2.a) Currently in the patch - just expect a i2c_algo_bit_data > structure. No new structures introduced, but this does expose some > underlying bit-algo detail to the platform and does not elegantly > support passing any future i2c-bit-platform specific fields, should > they be required. > 2.b) Declare a new platform structure which has a i2c_algo_bit_data > member (for supporting future i2c-bit-platform fields). Solves the > future field issue but still exposes so bit-algo stuff to the platfor= m > 2.c) Declare a new platform structure with members similar to > i2c_algo_bit_data, but only those who should be exposed to the > platform. The only downside to this is that any future change to > i2c_algo_bit_data may require change to this new platform structure. > > Tested on an MPC875 running @ 50MHz to achieve near perfect 100kHz I2= C. > > Eran > > Signed-off-by: Eran Duchan > --- > =C2=A0drivers/i2c/busses/Kconfig =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2= =A0| =C2=A0 10 +++ > =C2=A0drivers/i2c/busses/Makefile =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 = | =C2=A0 =C2=A01 + > =C2=A0drivers/i2c/busses/i2c-bit-platform.c | =C2=A0105 +++++++++++++= ++++++++++++++++++++ > =C2=A03 files changed, 116 insertions(+), 0 deletions(-) > =C2=A0create mode 100644 drivers/i2c/busses/i2c-bit-platform.c > > diff --git a/drivers/i2c/busses/Kconfig b/drivers/i2c/busses/Kconfig > index 3a6321c..f6c6b8c 100644 > --- a/drivers/i2c/busses/Kconfig > +++ b/drivers/i2c/busses/Kconfig > @@ -365,6 +365,16 @@ config I2C_GPIO > =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0This is a very simple bitbanging I2= C driver utilizing the > =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0arch-neutral GPIO API to control th= e SCL and SDA lines. > > +config I2C_BIT_PLATFORM > + =C2=A0 =C2=A0 =C2=A0 tristate "I2c-bit-algo as platform device" > + =C2=A0 =C2=A0 =C2=A0 select I2C_ALGOBIT > + =C2=A0 =C2=A0 =C2=A0 help > + =C2=A0 =C2=A0 =C2=A0 =C2=A0 An I2C bus adapter which delegates the = bit-algo callback registration > + =C2=A0 =C2=A0 =C2=A0 =C2=A0 to the platform. Useful in cases where = existing adapters such as > + =C2=A0 =C2=A0 =C2=A0 =C2=A0 i2c-gpio are too slow. Keep in mind tha= t no resource locking of any kind > + =C2=A0 =C2=A0 =C2=A0 =C2=A0 is performed by the adapter or algo, so= any such contention must be > + =C2=A0 =C2=A0 =C2=A0 =C2=A0 handled by the platform. > + > =C2=A0config I2C_HIGHLANDER > =C2=A0 =C2=A0 =C2=A0 =C2=A0tristate "Highlander FPGA SMBus interface" > =C2=A0 =C2=A0 =C2=A0 =C2=A0depends on SH_HIGHLANDER > diff --git a/drivers/i2c/busses/Makefile b/drivers/i2c/busses/Makefil= e > index 84cb16a..65bced4 100644 > --- a/drivers/i2c/busses/Makefile > +++ b/drivers/i2c/busses/Makefile > @@ -35,6 +35,7 @@ obj-$(CONFIG_I2C_CPM) =C2=A0 =C2=A0 =C2=A0 =C2=A0 += =3D i2c-cpm.o > =C2=A0obj-$(CONFIG_I2C_DAVINCI) =C2=A0 =C2=A0 =C2=A0+=3D i2c-davinci.= o > =C2=A0obj-$(CONFIG_I2C_DESIGNWARE) =C2=A0 +=3D i2c-designware.o > =C2=A0obj-$(CONFIG_I2C_GPIO) =C2=A0 =C2=A0 =C2=A0 =C2=A0 +=3D i2c-gpi= o.o > +obj-$(CONFIG_I2C_BIT_PLATFORM) +=3D i2c-bit-platform.o > =C2=A0obj-$(CONFIG_I2C_HIGHLANDER) =C2=A0 +=3D i2c-highlander.o > =C2=A0obj-$(CONFIG_I2C_IBM_IIC) =C2=A0 =C2=A0 =C2=A0+=3D i2c-ibm_iic.= o > =C2=A0obj-$(CONFIG_I2C_IMX) =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0+=3D i2= c-imx.o > diff --git a/drivers/i2c/busses/i2c-bit-platform.c > b/drivers/i2c/busses/i2c-bit-platform.c > new file mode 100644 > index 0000000..cb98006 > --- /dev/null > +++ b/drivers/i2c/busses/i2c-bit-platform.c > @@ -0,0 +1,105 @@ > +/* > + =C2=A0 =C2=A0i2c-bit-algo as platform device > + =C2=A0 =C2=A0Delegates the bit-algo callback registration to the pl= atform > + > + =C2=A0 =C2=A0Copyright (C) 2011 Eran Duchan > + > + =C2=A0 =C2=A0This program is free software; you can redistribute it= and/or modify > + =C2=A0 =C2=A0it under the terms of the GNU General Public License a= s published by > + =C2=A0 =C2=A0the Free Software Foundation; either version 2 of the = License, or > + =C2=A0 =C2=A0(at your option) any later version. > +*/ > + > +#include > +#include > +#include > + > +static int __devinit i2c_bit_platform_probe(struct platform_device *= pdev) > +{ > + =C2=A0 =C2=A0 =C2=A0 struct i2c_algo_bit_data *bit_data; > + =C2=A0 =C2=A0 =C2=A0 struct i2c_adapter *adap; > + =C2=A0 =C2=A0 =C2=A0 int ret; > + > + =C2=A0 =C2=A0 =C2=A0 if (pdev->dev.platform_data =3D=3D NULL) > + =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 ret =3D -ENXIO; > + > + =C2=A0 =C2=A0 =C2=A0 bit_data =3D pdev->dev.platform_data; > + > + =C2=A0 =C2=A0 =C2=A0 ret =3D -ENOMEM; > + =C2=A0 =C2=A0 =C2=A0 adap =3D kzalloc(sizeof(struct i2c_adapter), G= =46P_KERNEL); > + =C2=A0 =C2=A0 =C2=A0 if (adap =3D=3D NULL) > + =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 goto err_alloc_ada= p; > + > + =C2=A0 =C2=A0 =C2=A0 adap->owner =3D THIS_MODULE; > + =C2=A0 =C2=A0 =C2=A0 adap->algo_data =3D bit_data; > + =C2=A0 =C2=A0 =C2=A0 adap->class =3D I2C_CLASS_HWMON | I2C_CLASS_SP= D; > + =C2=A0 =C2=A0 =C2=A0 adap->dev.parent =3D &pdev->dev; > + > + =C2=A0 =C2=A0 =C2=A0 snprintf(adap->name, sizeof(adap->name), "i2c-= bit-platform%d", > + =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0= =C2=A0 =C2=A0pdev->id); > + =C2=A0 =C2=A0 =C2=A0 adap->name[sizeof(adap->name) - 1] =3D '\0'; > + > + =C2=A0 =C2=A0 =C2=A0 /* > + =C2=A0 =C2=A0 =C2=A0 =C2=A0* If "dev->id" is negative we consider i= t as zero. > + =C2=A0 =C2=A0 =C2=A0 =C2=A0* The reason to do so is to avoid sysfs = names that only make > + =C2=A0 =C2=A0 =C2=A0 =C2=A0* sense when there are multiple adapters= =2E > + =C2=A0 =C2=A0 =C2=A0 =C2=A0*/ > + =C2=A0 =C2=A0 =C2=A0 adap->nr =3D (pdev->id !=3D -1) ? pdev->id : 0= ; > + =C2=A0 =C2=A0 =C2=A0 ret =3D i2c_bit_add_numbered_bus(adap); > + =C2=A0 =C2=A0 =C2=A0 if (ret) > + =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 goto err_add_bus; > + > + =C2=A0 =C2=A0 =C2=A0 platform_set_drvdata(pdev, adap); > + > + =C2=A0 =C2=A0 =C2=A0 return 0; > + > +err_add_bus: > +err_alloc_adap: > + > + =C2=A0 =C2=A0 =C2=A0 kfree(adap); > + > + =C2=A0 =C2=A0 =C2=A0 return ret; > +} > + > +static int __devexit i2c_bit_platform_remove(struct platform_device = *pdev) > +{ > + =C2=A0 =C2=A0 =C2=A0 struct i2c_adapter *adap; > + > + =C2=A0 =C2=A0 =C2=A0 adap =3D platform_get_drvdata(pdev); > + > + =C2=A0 =C2=A0 =C2=A0 i2c_del_adapter(adap); > + =C2=A0 =C2=A0 =C2=A0 kfree(adap); > + > + =C2=A0 =C2=A0 =C2=A0 return 0; > +} > + > +static struct platform_driver i2c_bit_platform_driver =3D { > + =C2=A0 =C2=A0 =C2=A0 .driver =C2=A0 =C2=A0 =C2=A0 =C2=A0 =3D { > + =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 .name =C2=A0 =3D "= i2c-bit-platform", > + =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 .owner =C2=A0=3D T= HIS_MODULE, > + =C2=A0 =C2=A0 =C2=A0 }, > + =C2=A0 =C2=A0 =C2=A0 .probe =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0=3D i= 2c_bit_platform_probe, > + =C2=A0 =C2=A0 =C2=A0 .remove =C2=A0 =C2=A0 =C2=A0 =C2=A0 =3D __deve= xit_p(i2c_bit_platform_remove), > +}; > + > +static int __init i2c_bit_platform_init(void) > +{ > + =C2=A0 =C2=A0 =C2=A0 int ret; > + > + =C2=A0 =C2=A0 =C2=A0 ret =3D platform_driver_register(&i2c_bit_plat= form_driver); > + =C2=A0 =C2=A0 =C2=A0 if (ret) > + =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 =C2=A0 printk(KERN_ERR "i= 2c-bit-platform: probe failed: %d\n", ret); > + > + =C2=A0 =C2=A0 =C2=A0 return ret; > +} > +subsys_initcall(i2c_bit_platform_init); > + > +static void __exit i2c_bit_platform_exit(void) > +{ > + =C2=A0 =C2=A0 =C2=A0 platform_driver_unregister(&i2c_bit_platform_d= river); > +} > +module_exit(i2c_bit_platform_exit); > + > +MODULE_AUTHOR("Eran Duchan "); > +MODULE_DESCRIPTION("Thin platform adapter for i2c-bit-algo"); > +MODULE_LICENSE("GPL"); > -- > 1.7.0.4 >