Hi Eugeniu, coming back to this topic, thanks for your patience everyone. > > > > You could wire up two R-Car I2C instances, set up one as an I2C slave > > handled by the I2C testunit and then use the other instance with > > SMBUS_BLOCK_PROC_CALL which also needs RECV_LEN. Check > > Documentation/i2c/slave-testunit-backend.rst for details. > > I am obviously not an SMBus expert, but I wonder if simply testing the > PCA9654 I/O Expander with SMBus support on the H3-Salvator-X target > could be acceptable as a test procedure? See some test results below. As long as the first read value is 8 (or lower than 32), it will work. But it is testing only this one value while my method above is more flexible and allows for arbitrary test patterns. However, your tests already showed that Andrew's patch seems to be not correct. > ############################################################ > #################### ANDREW'S V2 PATCH ##################### > ############################################################ > root@rcar-gen3:# i2cget -y 4 0x68 0 i 8 > 0x08 0xff 0x06 0xff 0x5f 0xff 0x11 0x08 > > root@rcar-gen3:# i2cget -y 4 0x68 0 s > 0x08 0xff 0x06 0xff 0x5f 0xff 0x11 0x08 This is wrong. The first byte is the length byte and should not be seen here. Check the i2c_smbus_read_block_data() implementation in i2c-tools. > ############################################################ > ##################### WOLFRAM'S PATCH ###################### > ############################################################ > root@rcar-gen3:# i2cget -y 4 0x68 0 i 8 > 0x08 0xff 0x06 0xff 0x5f 0xff 0x11 0x08 > > root@rcar-gen3:# i2cget -y 4 0x68 0 s > 0xff 0x06 0xff 0x5f 0xff 0x11 0x08 0x08 This is how it should look like IMO. Happy hacking, Wolfram
Attachment:
signature.asc
Description: PGP signature
