Re: [PATCH v2 2/2] i2c: aspeed: added documentation for Aspeed I2C driver
From: Joel Stanley <joel@jms.id.au>
Date: 2016-09-23 06:02:29
Also in:
linux-i2c, lkml, openbmc
From: Joel Stanley <joel@jms.id.au>
Date: 2016-09-23 06:02:29
Also in:
linux-i2c, lkml, openbmc
Hi Brendan, On Wed, Sep 21, 2016 at 3:35 AM, Brendan Higgins [off-list ref] wrote:
First off, someone pointed out to me that the mapping that I used between addresses and bus numbers is not actually valid for busses 8-14. This could be fixed by checking the offset, but I am wondering if that is the right way to do it. It seems like this is not completely trivial so maybe this should be specified in the device tree? If that is the case, should I do this as another reg entry or go back to the way I was doing it before?
I don't see an alternative way to derive these numbers. As you mention, the block of SRAM in the middle of the IP screwing up the linear mapping. I suggest going back to what you have in v2, with a mention in the commit message as to why the bus property is necessary. Cheers, Joel