Re: [PATCH v6 0/3] i2c: designware: Improve support of multi-messages transfer

[Andy Shevchenko]

On Fri, Jan 30, 2026 at 04:52:28PM +0100, Benoît Monin wrote:
> Add support for the I2C_M_STOP flag to the .xfer() function of
> the designware driver. This allows grouping multiple accesses in a
> single call and changing the target address after a STOP flag. This is
> achieved by splitting i2c_dw_xfer_common() in two functions. The core
> logic handling the transaction is now in __i2c_dw_xfer_one_part(),
> while i2c_dw_xfer_common() loops over the messages to search for the
> I2C_M_STOP flag and calls __i2c_dw_xfer_one_part().
> 
> Handle controllers that lack the ability to emit a RESTART when two
> consecutive messages have the same address and direction by aborting
> transfers that contain such a sequence of messages. For those controllers,
> we also check that we do not get any unwanted STOP caused by a Tx FIFO
> underrun, as they lack the ability to hold the clock during a transaction.
> And we set the irq as non-threaded to prevent underrun on PREEMPT-RT
> kernel when filling the FIFO.
> 
> The I2C controllers found in the EyeQ6Lplus and EyeQ7H SoCs from Mobileye
> lack such capability, so a compatible string is added because this cannot
> be detected at runtime. The corresponding DT bindings has already been
> merged and is available in v6.19-rc2:
> 
>     d9b85d296f3a ("dt-bindings: i2c: dw: Add Mobileye I2C controllers")
> 
> This patch series also simplifies runtime PM handling in
> i2c_dw_xfer_common() and amd_i2c_dw_xfer_quirk() with PM_RUNTIME_* macros.
> 
> To test I2C_M_STOP and other message modifier flags, I posted a patch
> series for i2c-tools[1]. With it, i2cdetect shows the support of the
> protocol mangling and repeated start skipping functionalities, and
> i2ctransfer can set the modifier flags to the I2C messages.
> 
> The series is now based on i2c-host tree and only contains the last three
> patches that have not been merged yet. The first and last patches are
> functionally identical to the one posted in the previous revision. Only
> the second patch contains modified code to use PM_RUNTIME_* macros
> instead of ACQUIRE()/ACQUIRE_ERR(). Thus I let Mika's Acked-by on the
> first and third patch, and dropped it from the second patch. I hope it
> was the correct thing to do.

Looks good enough to me, and if anything, it may be fixed later on.
Reviewed-by: Andy Shevchenko <andriy.shevchenko@xxxxxxxxxxxxxxx>

-- 
With Best Regards,
Andy Shevchenko