i2c: add bcm2835 driver
Subsystems (4)
PATCH v3 1/1 i2c: add bcm2835 driver
This implements a very basic I2C host driver for the BCM2835 SoC. Missing features so far are:
* 10-bit addressing.
* DMA.
---
v3:
* Return i2c_add_numbered_adapter() from probe() directly. v2:
* Implemented clock divider configuration based on desired bus rate.
* Make use of module_platform_driver().
* Removed use of devinit.
--- .../devicetree/bindings/i2c/brcm,bcm2835-i2c.txt | 20 ++ drivers/i2c/busses/Kconfig | 12 + drivers/i2c/busses/Makefile | 1 + drivers/i2c/busses/i2c-bcm2835.c | 340 ++++++++++++++++++++ 4 files changed, 373 insertions(+) create mode 100644 Documentation/devicetree/bindings/i2c/brcm,bcm2835-i2c.txt create mode 100644 drivers/i2c/busses/i2c-bcm2835.c
Documentation/devicetree/bindings/i2c/brcm,bcm2835-i2c.txt new file +20 -0
@@ -0,0 +1,20 @@ +Broadcom BCM2835 I2C controller + +Required properties: +- compatible : Should be "brcm,bcm2835-i2c". +- reg: Should contain register location and length. +- interrupts: Should contain interrupt. +- clocks : The clock feeding the I2C controller. + +Recommended properties: +- clock-frequency : desired I2C bus clock frequency in Hz. + +Example: + +i2c at 20205000 { + compatible = "brcm,bcm2835-i2c"; + reg = <0x7e205000 0x1000>; + interrupts = <2 21>; + clocks = <&clk_i2c>; + clock-frequency = <100000>; +};
drivers/i2c/busses/Kconfig +12 -0
@@ -319,6 +319,18 @@ config I2C_AU1550 This driver can also be built as a module. If so, the module will be called i2c-au1550. +config I2C_BCM2835 + tristate "Broadcom BCM2835 I2C controller" + depends on ARCH_BCM2835 + help + If you say yes to this option, support will be included for the + BCM2835 I2C controller. + + If you don't know what to do here, say N. + + This support is also available as a module. If so, the module + will be called i2c-bcm2835. + config I2C_BLACKFIN_TWI tristate "Blackfin TWI I2C support" depends on BLACKFIN
drivers/i2c/busses/Makefile +1 -0
@@ -30,6 +30,7 @@ obj-$(CONFIG_I2C_POWERMAC) += i2c-powermac.o # Embedded system I2C/SMBus host controller drivers obj-$(CONFIG_I2C_AT91) += i2c-at91.o obj-$(CONFIG_I2C_AU1550) += i2c-au1550.o +obj-$(CONFIG_I2C_BCM2835) += i2c-bcm2835.o obj-$(CONFIG_I2C_BLACKFIN_TWI) += i2c-bfin-twi.o obj-$(CONFIG_I2C_CBUS_GPIO) += i2c-cbus-gpio.o obj-$(CONFIG_I2C_CPM) += i2c-cpm.o
drivers/i2c/busses/i2c-bcm2835.c new file +340 -0
340 lines not on this page · raw message
Signature · 1 line
1.7.10.4
PATCH v3 1/1 i2c: add bcm2835 driver
Hi Stephen,
Quoted 9 lines · Stephen Warren
This implements a very basic I2C host driver for the BCM2835 SoC. Missing features so far are:
* 10-bit addressing.
* DMA.Reviewed-by Grant Likely <[email protected]>Signed-off-by Stephen Warren <[email protected]>---
...
Quoted 11 lines
drivers/i2c/busses/i2c-bcm2835.c new file +11
-0+struct bcm2835_i2c_dev { +struct device *dev; +void __iomem *regs; +struct clk *clk; +struct i2c_adapter adapter; +struct completion completion; +u32 msg_err; +u8 *msg_buf; +size_t msg_buf_remaining; +}; +
drivers/i2c/busses/i2c-bcm2835.c line 75
Quoted 10 lines
drivers/i2c/busses/i2c-bcm2835.c new file +10
-0+static inline void bcm2835_i2c_writel (struct bcm2835_i2c_dev *i2c_dev, +u32 reg,u32 val) +{ +writel (val, i2c_dev->regs + reg); +} + +static inline u32 bcm2835_i2c_readl (struct bcm2835_i2c_dev *i2c_dev,u32 reg) +{ +return readl (i2c_dev->regs + reg); +}
Hmm, not sure if those functions add readability, but no need to change.
Quoted 6 lines
drivers/i2c/busses/i2c-bcm2835.c new file +6
-0+ +static void bcm2835_fill_txfifo (struct bcm2835_i2c_dev *i2c_dev) +{ +u32 val; + +for (;;) {
drivers/i2c/busses/i2c-bcm2835.c line 91
Quoted 10 lines
drivers/i2c/busses/i2c-bcm2835.c new file +10
-0+if (!i2c_dev->msg_buf_remaining) +return ; + val =bcm2835_i2c_readl (i2c_dev,BCM2835_I2C_S ); +if (!(val &BCM2835_I2C_S_TXD )) +break ; +bcm2835_i2c_writel (i2c_dev,BCM2835_I2C_FIFO , + *i2c_dev->msg_buf); + i2c_dev->msg_buf++; + i2c_dev->msg_buf_remaining--; + }
while (i2c_dev->msg_buf_remaining) ...
?
Quoted 7 lines
drivers/i2c/busses/i2c-bcm2835.c new file +7
-0+} + +static void bcm2835_drain_rxfifo (struct bcm2835_i2c_dev *i2c_dev) +{ +u32 val; + +for (;;) {
drivers/i2c/busses/i2c-bcm2835.c line 108
Quoted 10 lines
drivers/i2c/busses/i2c-bcm2835.c new file +10
-0+if (!i2c_dev->msg_buf_remaining) +return ; + val =bcm2835_i2c_readl (i2c_dev,BCM2835_I2C_S ); +if (!(val &BCM2835_I2C_S_RXD )) +break ; + *i2c_dev->msg_buf =bcm2835_i2c_readl (i2c_dev, +BCM2835_I2C_FIFO ); + i2c_dev->msg_buf++; + i2c_dev->msg_buf_remaining--; + }
ditto?
...
drivers/i2c/busses/i2c-bcm2835.c line 179
Quoted 7 lines
drivers/i2c/busses/i2c-bcm2835.c new file +7
-0+ ret =wait_for_completion_timeout (&i2c_dev->completion, +BCM2835_I2C_TIMEOUT ); +bcm2835_i2c_writel (i2c_dev,BCM2835_I2C_C ,BCM2835_I2C_C_CLEAR ); +if (WARN_ON (ret ==0 )) { +dev_err (i2c_dev->dev,"i2c transfer timed out\n" ); +return -ETIMEDOUT ; + }
I'd suggest to skip the WARN_ON. Timeout is the expected case when you want to read from an EEPROM which is just in the process of erasing/writing data from the previous command.
...
drivers/i2c/busses/i2c-bcm2835.c line 300
Quoted 4 lines
drivers/i2c/busses/i2c-bcm2835.c new file +4
-0+ adap = &i2c_dev->adapter; +i2c_set_adapdata (adap, i2c_dev); + adap->owner =THIS_MODULE ; + adap->class =I2C_CLASS_HWMON ;
Do you really need the class?
drivers/i2c/busses/i2c-bcm2835.c line 304
Quoted 4 lines
drivers/i2c/busses/i2c-bcm2835.c new file +4
-0+strlcpy (adap->name,"bcm2835 I2C adapter" ,sizeof (adap->name)); + adap->algo = &bcm2835_i2c_algo; + adap->dev.parent = &pdev->dev; + adap->nr =-1 ;
If you are using '-1' anyhow, you could skip setting it and simply call i2c_add_adapter(). Or you can init it here to pdev->id. Whatever you prefer.
drivers/i2c/busses/i2c-bcm2835.c line 315
Quoted 11 lines
drivers/i2c/busses/i2c-bcm2835.c new file +11
-0+ +bcm2835_i2c_writel (i2c_dev,BCM2835_I2C_C ,0 ); + +return i2c_add_numbered_adapter (adap); +} + +static int bcm2835_i2c_remove (struct platform_device *pdev) +{ +struct bcm2835_i2c_dev *i2c_dev =platform_get_drvdata (pdev); + +i2c_del_adapter (&i2c_dev->adapter);
Is it guaranteed that interrupts cannot occur anymore? Otherwise OOPS might happen, since the adapter is cleared here already but devm will remove the interrupt only a little later.
Thanks,
Wolfram
PATCH v3 1/1 i2c: add bcm2835 driver
Quoted 8 lines · Wolfram Sang
Hi Stephen,
Quoted 5 lines · Stephen Warren
This implements a very basic I2C host driver for the BCM2835 SoC. Missing features so far are:
* 10-bit addressing.
* DMA.
...
Quoted 11 lines
drivers/i2c/busses/i2c-bcm2835.c line 179
Quoted 7 lines
drivers/i2c/busses/i2c-bcm2835.c new file +7
-0+ ret =wait_for_completion_timeout (&i2c_dev->completion, +BCM2835_I2C_TIMEOUT ); +bcm2835_i2c_writel (i2c_dev,BCM2835_I2C_C ,BCM2835_I2C_C_CLEAR ); +if (WARN_ON (ret ==0 )) { +dev_err (i2c_dev->dev,"i2c transfer timed out\n" ); +return -ETIMEDOUT ; + }I'd suggest to skip the
WARN_ON. Timeout is the expected case when you want to read from an EEPROM which is just in the process of erasing/writing data from the previous command.
I copied that from Tegra. Should that driver be changed too?
Quoted 6 lines
drivers/i2c/busses/i2c-bcm2835.c line 300
Quoted 4 lines
drivers/i2c/busses/i2c-bcm2835.c new file +4
-0+ adap = &i2c_dev->adapter; +i2c_set_adapdata (adap, i2c_dev); + adap->owner =THIS_MODULE ; + adap->class =I2C_CLASS_HWMON ;Do you really need the class?
If I grep for I2C_CLASS_HWMON, I see a ton of hits. I think I'll keep it unless you strongly object, since it seems typical.
I'll fix up the other points you mentioned.
PATCH v3 1/1 i2c: add bcm2835 driver
Quoted 13 lines
Quoted 12 lines
drivers/i2c/busses/i2c-bcm2835.c line 179
Quoted 8 lines
drivers/i2c/busses/i2c-bcm2835.c new file +7
-0+ ret =wait_for_completion_timeout (&i2c_dev->completion, +BCM2835_I2C_TIMEOUT ); +bcm2835_i2c_writel (i2c_dev,BCM2835_I2C_C ,BCM2835_I2C_C_CLEAR ); +if (WARN_ON (ret ==0 )) { +dev_err (i2c_dev->dev,"i2c transfer timed out\n" ); +return -ETIMEDOUT ; + }I'd suggest to skip the
WARN_ON. Timeout is the expected case when you want to read from an EEPROM which is just in the process of erasing/writing data from the previous command.
Quoted 1 line
I copied that from Tegra. Should that driver be changed too?
The suggestion applies to all drivers, yes :)