On Thu, May 27, 2021 at 6:37 PM Guenter Roeck [off-list ref] wrote:

On 5/27/21 7:58 AM, Robert Marko wrote:
[ ... ]

quoted

quoted

quoted

quoted

quoted

quoted

I tried applying the block support for mv64xx as well:
https://patchwork.ozlabs.org/project/linux-i2c/patch/20200118115820.9080-1-fuga@studiofuga.com/

That patch would be needed, but it looks buggy to me. This chunk:

+               drv_data->effective_length = data+1;
+               drv_data->bytes_left = data+1;
+               drv_data->msg->len = data+1;

should be:

+               drv_data->effective_length = data+1;
+               drv_data->bytes_left = data;
+               drv_data->msg->len = data+1;

It should also make sure that 'data' is not larger than I2C_SMBUS_BLOCK_MAX,
and bail out if it isn't.

Yeah, I did not check the contents, I just saw 2 reviews and tested it
since it can't hurt.

That patch doesn't work at all. Make the above change, and also change
the type of effective_length from u32 to int, and try again.

I was just looking and it, and doing the changes you recommended make
no difference at all.

Is the i2c controller compatible with marvell,mv78230-i2c ?
The block transfers would not work in that case. Let me know
and I'll send you a patch that might fix it.

Yes, the CPU is Armada 7040 and I can see that compatible is
marvell,mv78230-i2c indeed.

It would be great if you have the patch,
Regards
Robert

Guenter

-- 
Robert Marko
Staff Embedded Linux Engineer
Sartura Ltd.
Lendavska ulica 16a
10000 Zagreb, Croatia
Email: robert.marko@sartura.hr
Web: www.sartura.hr