[PATCH v3] drivers: ipmi: Support raw i2c packet in IPMB

[Corey Minyard]

On Wed, Nov 13, 2019 at 09:28:10PM +0000, Vijay Khemka wrote:
> 
> 
> On 11/13/19, 12:33 PM, "Asmaa Mnebhi" <Asmaa at mellanox.com> wrote:
> 
>     Inline response:
>     
>     -----Original Message-----
>     From: Vijay Khemka <vijaykhemka at fb.com> 
>     Sent: Wednesday, November 13, 2019 2:23 PM
>     To: Corey Minyard <minyard at acm.org>; Arnd Bergmann <arnd at arndb.de>; Greg Kroah-Hartman <gregkh at linuxfoundation.org>; openipmi-developer at lists.sourceforge.net; linux-kernel at vger.kernel.org
>     Cc: vijaykhemka at fb.com; cminyard at mvista.com; Asmaa Mnebhi <Asmaa at mellanox.com>; joel at jms.id.au; linux-aspeed at lists.ozlabs.org; sdasari at fb.com
>     Subject: [PATCH v3] drivers: ipmi: Support raw i2c packet in IPMB
>     
>     Many IPMB devices doesn't support smbus protocol and current driver support only smbus devices. Added support for raw i2c packets.
>     
>     User can define use-i2c-block in device tree to use i2c raw transfer.
>     
>     Asmaa>> Fix the description: "The ipmb_dev_int driver only supports the smbus protocol at the moment. Add support for the i2c protocol as well. There will be a variable passed by though the device tree or ACPI table which sets the configures the protocol as either i2c or smbus."
>     
>     Signed-off-by: Vijay Khemka <vijaykhemka at fb.com>
>     ---
>      drivers/char/ipmi/ipmb_dev_int.c | 48 ++++++++++++++++++++++++++++++++
>      1 file changed, 48 insertions(+)
>     
>     diff --git a/drivers/char/ipmi/ipmb_dev_int.c b/drivers/char/ipmi/ipmb_dev_int.c
>     index ae3bfba27526..16d5d4b636a9 100644
>     --- a/drivers/char/ipmi/ipmb_dev_int.c
>     +++ b/drivers/char/ipmi/ipmb_dev_int.c
>     @@ -63,6 +63,7 @@ struct ipmb_dev {
>      	spinlock_t lock;
>      	wait_queue_head_t wait_queue;
>      	struct mutex file_mutex;
>     +	bool use_i2c;
>      };
>      
>     Asmaa>> rename this variable : is_i2c_protocol
> Done.
>     
>      static inline struct ipmb_dev *to_ipmb_dev(struct file *file) @@ -112,6 +113,39 @@ static ssize_t ipmb_read(struct file *file, char __user *buf, size_t count,
>      	return ret < 0 ? ret : count;
>      }
>      
>     +static int ipmb_i2c_write(struct i2c_client *client, u8 *msg) {
>     +	unsigned char *i2c_buf;
>     +	struct i2c_msg i2c_msg;
>     +	ssize_t ret;
>     +	u8 msg_len;
>     +
>     +	/*
>     +	 * subtract 1 byte (rq_sa) from the length of the msg passed to
>     +	 * raw i2c_transfer
>     +	 */
>     +	msg_len = msg[IPMB_MSG_LEN_IDX] - 1;
>     +
>     +	i2c_buf = kzalloc(msg_len, GFP_KERNEL);
>     
>     Asmaa >> We do not want to use kzalloc every time you execute this write function. It would create so much fragmentation.
>     You don't really need to use kzalloc anyways. 
> We need to allocate memory to pass to i2c_transfer. That's what being done in i2c_smbus_xfer function as well.
>     
>     Also, this code chunk is short, so you can call it directly from the write function. I don't think you need a separate function for it.
> I wanted to keep this change as clean as possible.

I'd agree.  Fragmentation is not a big deal here.

However, why not just pass in msg + 2?  That would be cleaner, faster,
and less wasteful.

>     
>     +	if (!i2c_buf)
>     +		return -EFAULT;

-ENOMEM?  (Assuming you keep the malloc)

>     +
>     +	/* Copy message to buffer except first 2 bytes (length and address) */
>     +	memcpy(i2c_buf, msg+2, msg_len);
>     +
>     +	i2c_msg.addr = GET_7BIT_ADDR(msg[RQ_SA_8BIT_IDX]);
>     +	i2c_msg.flags = client->flags &
>     +			(I2C_M_TEN | I2C_CLIENT_PEC | I2C_CLIENT_SCCB);
>     Asmaa>> I don't think ipmb supports 10 bit addresses. The max number of bits in the IPMB address field is 8.
> Done.
>     
>     +	i2c_msg.len = msg_len;
>     +	i2c_msg.buf = i2c_buf;
>     +
>     +	ret = i2c_transfer(client->adapter, &i2c_msg, 1);
>     +	kfree(i2c_buf);
>     +
>     +	return ret;
>     +
>     +}
>     +
>      static ssize_t ipmb_write(struct file *file, const char __user *buf,
>      			size_t count, loff_t *ppos)
>      {
>     @@ -133,6 +167,12 @@ static ssize_t ipmb_write(struct file *file, const char __user *buf,
>      	rq_sa = GET_7BIT_ADDR(msg[RQ_SA_8BIT_IDX]);
>      	netf_rq_lun = msg[NETFN_LUN_IDX];
>      
>     +	/* Check i2c block transfer vs smbus */
>     +	if (ipmb_dev->use_i2c) {
>     +		ret = ipmb_i2c_write(ipmb_dev->client, msg);
>     +		return (ret == 1) ? count : ret;
>     +	}
>     +
>      	/*
>      	 * subtract rq_sa and netf_rq_lun from the length of the msg passed to
>      	 * i2c_smbus_xfer
>     @@ -277,6 +317,7 @@ static int ipmb_probe(struct i2c_client *client,
>      			const struct i2c_device_id *id)
>      {
>      	struct ipmb_dev *ipmb_dev;
>     +	struct device_node *np;
>      	int ret;
>      
>      	ipmb_dev = devm_kzalloc(&client->dev, sizeof(*ipmb_dev), @@ -302,6 +343,13 @@ static int ipmb_probe(struct i2c_client *client,
>      	if (ret)
>      		return ret;
>      
>     +	/* Check if i2c block xmit needs to use instead of smbus */
>     +	np = client->dev.of_node;
>     +	if (np && of_get_property(np, "use-i2c-block", NULL))
>     Asmaa>> Rename this variable i2c-protocol. And also, apply this to ACPI as well.
> Done.

I don't think ACPI is that important at the moment.  Rename is good.

>     +		ipmb_dev->use_i2c = true;
>     +	else
>     +		ipmb_dev->use_i2c = false;

The above two lines are unnecessary.

-corey

>     +
>      	ipmb_dev->client = client;
>      	i2c_set_clientdata(client, ipmb_dev);
>      	ret = i2c_slave_register(client, ipmb_slave_cb);
>     --
>     2.17.1
>     
>     
>