Re: [PATCH for v5.5 2/2] Input: rmi_f54: read from FIFO in 32 byte blocks

[Dmitry Torokhov <dmitry.torokhov@gmail.com>]

Hi Hans,

On Wed, Jan 15, 2020 at 01:48:19PM +0100, Hans Verkuil wrote:
> The F54 Report Data is apparently read through a fifo and for
> the smbus protocol that means that between reading a block of 32
> bytes the rmiaddr shouldn't be incremented. However, changing
> that causes other non-fifo reads to fail and so that change was
> reverted.
> 
> This patch changes just the F54 function and it now reads 32 bytes
> at a time from the fifo, using the F54_FIFO_OFFSET to update the
> start address that is used when reading from the fifo.
> 
> This has only been tested with smbus, not with i2c or spi. But I
> suspect that the same is needed there since I think similar
> problems will occur there when reading more than 256 bytes.
> 
> Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
> Tested-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
> Reported-by: Timo Kaufmann <timokau@zoho.com>
> Fixes: a284e11c371e ("Input: synaptics-rmi4 - don't increment rmiaddr for SMBus transfers")
> Cc: stable@vger.kernel.org

As you mentioned this one is not urgent so I dropped the stable
designation (you may forward to stable once it cooks in 5.5 for a bit)
and also dropped fixes as it does not fixes this particular commit but
something that was done before.

Otherwise applied.

Thanks.

-- 
Dmitry