Re: [PATCH v2] i2c-mpc: generate START condition after STOP caused by read i2c_msg

[Michael Lawnick <ml.lawnick@gmx.de>]

Ben Dooks said the following:
> On Thu, Dec 03, 2009 at 04:09:57PM +0100, Michael Lawnick wrote:
>> Ben Dooks said the following:
>> > On Tue, May 26, 2009 at 01:30:21PM +0200, Esben Haabendal wrote:
>> >> On Tue, May 19, 2009 at 7:22 AM, Esben Haabendal <eha@doredevelopment.dk> wrote:
>> >> > This fixes MAL (arbitration lost) bug caused by illegal use of
>> >> > RSTA (repeated START) after STOP condition generated after last byte
>> >> > of reads. With this patch, it is possible to do an i2c_transfer() with
>> >> > additional i2c_msg's following the I2C_M_RD messages.
>> >> >
>> >> > It still needs to be resolved if it is possible to fix this issue
>> >> > by removing the STOP condition after reads in a robust way.
>> >> >
>> >> > Signed-off-by: Esben Haabendal <eha@doredevelopment.dk>
>> >> > ---
>> >> > ?drivers/i2c/busses/i2c-mpc.c | ? ?9 +++++++--
>> >> > ?1 files changed, 7 insertions(+), 2 deletions(-)
>> >> 
>> >> Any blockers to get this accepted?
>> > 
>> > It would be nice to get an ack from someone who can actually test
>> > the driver before getting this merged.
>> >  
>> What is the state of this patch?
>> Shouldn't we attack the problem on a more general way by inventing a
>> Flag I2C_M_RESTART (or better I2C_M_NO_RESTART for backward compatibility)?
>> This way the client driver is able to decide what it needs. If we do the
>> choice within adapter, chance is about 50% to be wrong.
> 
> The documentation for 'struct i2c_msg' already says the STOP should only
> be generated for the last message of the transfer. If STOP is being
> generated for a message that isn't the last in the transfer than this
> is incorrect behaviour.

Ah, now I see, this is a mpc-only problem of implementation
(automatically generating stop in read function).
I couldn't find the above mentioned specification of STOP/RESTART (seems
I need new glasses) and was worried about whether my own implementation
for OCTEON is correct.

Thank you.
-- 
Michael