On 10/05/2017 10:13 AM, Pontus Andersson wrote:
On Thu, Oct 05, 2017 at 02:41:33PM +0200, Wolfram Sang wrote:
On Mon, Oct 02, 2017 at 02:45:19PM +0200, Pontus Andersson wrote:
Commit b6c159a9cb69 ("i2c: ismt: Don't duplicate the receive length for
block reads") broke I2C block reads. It aimed to fix normal SMBus block
read, but changed the correct behavior of I2C block read in the process.
According to Documentation/i2c/smbus-protocol, one vital difference
between normal SMBus block read and I2C block read is that there is no
byte count prefixed in the data sent on the wire:
SMBus Block Read: i2c_smbus_read_block_data()
S Addr Wr [A] Comm [A]
S Addr Rd [A] [Count] A [Data] A [Data] A ... A [Data] NA P
I2C Block Read: i2c_smbus_read_i2c_block_data()
S Addr Wr [A] Comm [A]
S Addr Rd [A] [Data] A [Data] A ... A [Data] NA P
Therefore the two transaction types need to be processed differently in
the driver by copying of the dma_buffer as done previously for the
I2C_SMBUS_I2C_BLOCK_DATA case.
Fixes: b6c159a9cb69 ("i2c: ismt: Don't duplicate the receive length for block reads")
Signed-off-by: Pontus Andersson <epontan@xxxxxxxxx>
Cc: stable@xxxxxxxxxxxxxxx
Applied to for-current, thanks!
Thank you! Is it safe to asume it will be backported to 4.4 stable
branch and upwards as the problematic commit did (once in mainline)?
Might have been good to CC the patch author of the problematic commit,
too.
Yes, of course! Stephen is CC now at least (kept the commit message
quoted at his convenience).
Doh. Looks like our regression test gave a false pass on the I2C access
case. I'd guess it's probably doing a series of byte reads instead of a
block read and missing the I2C_SMBUS_I2C_BLOCK_DATA case entirely.
Unfortunately I can't find the board that was modded with more I2C
devices on the ismt bus to verify this right now. I'll check with Dan
on Monday to see if he knows where that board is stashed.
Apologies for moving the bug instead of fixing it, and FWIW I'll have an
update Monday.
