Re: [PATCH v2] i2c: designware: Add driver support for AMD NAVI GPU

[Jarkko Nikula]

Hi

On 3/22/21 6:59 PM, Sanket Goswami wrote:
> The Latest AMD NAVI GPU card has an integrated Type-C controller and
> Designware I2C with PCI Interface. The Type-C controller can be
> accessed over I2C. The client driver is part of the USB Type-C UCSI
> driver.
>
> Also, there exists a couple of notable IP limitations that are dealt as
> workarounds:
> - I2C transaction work on a polling mode as IP does not generate
> interrupt.
> - I2C read command sent twice to address the IP issues.
>
> Reviewed-by: Shyam Sundar S K <Shyam-sundar.S-k@xxxxxxx>
> Co-developed-by: Nehal Bakulchandra Shah <Nehal-Bakulchandra.shah@xxxxxxx>
> Signed-off-by: Nehal Bakulchandra Shah <Nehal-Bakulchandra.shah@xxxxxxx>
> Signed-off-by: Sanket Goswami <Sanket.Goswami@xxxxxxx>
> ---
> Changes in v2:
> - Utilized existing functionality of i2c_dw_xfer_init to configure I2C
>    bus.
> - Removed i2c_dw_populate_client and rewrrient navi_amd_register_client
>    to deduplicate from existing drivers.
> - Addressed review comments from Andy.
>   
> drivers/i2c/busses/i2c-designware-common.c |   3 +
>   drivers/i2c/busses/i2c-designware-core.h   |   3 +
>   drivers/i2c/busses/i2c-designware-master.c | 136 +++++++++++++++++++++
>   drivers/i2c/busses/i2c-designware-pcidrv.c |  57 +++++++++
>   4 files changed, 199 insertions(+)
>
> diff --git a/drivers/i2c/busses/i2c-designware-common.c b/drivers/i2c/busses/i2c-designware-common.c
> index 3c19aada4b30..50883a70b482 100644
> --- a/drivers/i2c/busses/i2c-designware-common.c
> +++ b/drivers/i2c/busses/i2c-designware-common.c
> @@ -150,6 +150,9 @@ int i2c_dw_init_regmap(struct dw_i2c_dev *dev)
>   	reg = readl(dev->base + DW_IC_COMP_TYPE);
>   	i2c_dw_release_lock(dev);
>   
> +	if (dev->flags & AMD_NON_INTR_MODE)
> +		map_cfg.max_register = AMD_UCSI_INTR_REG;
> +
>   	if (reg == swab32(DW_IC_COMP_TYPE_VALUE)) {
>   		map_cfg.reg_read = dw_reg_read_swab;
>   		map_cfg.reg_write = dw_reg_write_swab;
> diff --git a/drivers/i2c/busses/i2c-designware-core.h b/drivers/i2c/busses/i2c-designware-core.h
> index 5392b82f68a4..f29923c452ac 100644
> --- a/drivers/i2c/busses/i2c-designware-core.h
> +++ b/drivers/i2c/busses/i2c-designware-core.h
> @@ -293,9 +293,12 @@ struct dw_i2c_dev {
>   #define ACCESS_INTR_MASK	BIT(0)
>   #define ACCESS_NO_IRQ_SUSPEND	BIT(1)
>   
> +#define AMD_NON_INTR_MODE	BIT(2)
>   #define MODEL_MSCC_OCELOT	BIT(8)
>   #define MODEL_BAIKAL_BT1	BIT(9)
>   #define MODEL_MASK		GENMASK(11, 8)
> +#define AMD_UCSI_INTR_EN	0xd
> +#define AMD_UCSI_INTR_REG	0x474
>   
>   int i2c_dw_init_regmap(struct dw_i2c_dev *dev);
>   u32 i2c_dw_scl_hcnt(u32 ic_clk, u32 tSYMBOL, u32 tf, int cond, int offset);
> diff --git a/drivers/i2c/busses/i2c-designware-master.c b/drivers/i2c/busses/i2c-designware-master.c
> index dd27b9dbe931..a76e1f992850 100644
> --- a/drivers/i2c/busses/i2c-designware-master.c
> +++ b/drivers/i2c/busses/i2c-designware-master.c
> @@ -23,6 +23,10 @@
>   
>   #include "i2c-designware-core.h"
>   
> +#define AMD_TIMEOUT_MIN_MSEC	10000
> +#define AMD_TIMEOUT_MAX_MSEC	11000
> +#define AMD_MASTERCFG_MASK	GENMASK(15, 0)
> +
>   static void i2c_dw_configure_fifo_master(struct dw_i2c_dev *dev)
>   {
>   	/* Configure Tx/Rx FIFO threshold levels */
> @@ -208,6 +212,13 @@ static int i2c_dw_init_master(struct dw_i2c_dev *dev)
>   	if (dev->sda_hold_time)
>   		regmap_write(dev->map, DW_IC_SDA_HOLD, dev->sda_hold_time);
>   
> +	/*
> +	 * In order to enable the interrupt for UCSI i.e. AMD NAVI GPU card,
> +	 * it is mandatory to set the right value in specific register
> +	 * (offset:0x474) as per the hardware IP specification.
> +	 */
> +	if (dev->flags & AMD_NON_INTR_MODE)
> +		regmap_write(dev->map, AMD_UCSI_INTR_REG, AMD_UCSI_INTR_EN);

This confuses me - this patch is about adding support for DesignWare IP 
that does not generate interrupts but here code is enabling an 
interrupt. I guess it's for UCSI but should above code then go to a 
driver handling that IP?

> +static int i2c_dw_check_stopbit(struct dw_i2c_dev *i2cd)
...
> +static int i2c_dw_status(struct dw_i2c_dev *i2cd)
...
> +static int amd_i2c_dw_master_xfer(struct i2c_adapter *adap, struct i2c_msg *msgs, int num_msgs)
...
> +{
> +	struct dw_i2c_dev *i2cd = i2c_get_adapdata(adap);

For uniformity I'd use struct dw_i2c_dev *dev name instead of *i2cd 
since that what driver uses currently in other places.

> @@ -461,6 +587,13 @@ i2c_dw_xfer(struct i2c_adapter *adap, struct i2c_msg msgs[], int num)
>   	dev_dbg(dev->dev, "%s: msgs: %d\n", __func__, num);
>   
>   	pm_runtime_get_sync(dev->dev);
> +	/*
> +	 * Initiate I2C message transfer when AMD NAVI GPU card is enabled,
> +	 * As it is polling based transfer mechanism, which does not support
> +	 * interrupt based functionalities of existing DesignWare driver.
> +	 */
> +	if (dev->flags & AMD_NON_INTR_MODE)
> +		return amd_i2c_dw_master_xfer(adap, msgs, num);
>   
Does runtime PM go out of sync here?

Jarkko