Forgive my previous misformatted email.
On 03-Apr-18 15:40, Jarkko Nikula wrote:
> On 03/26/2018 12:05 PM, Alexander Monakov wrote:
>> On Mon, 19 Mar 2018, Jarkko Nikula wrote:
>>> I see specification mentions that for older IP versions have to monitor two
>>> registers IC_STATUS and IC_RAW_INTR_STAT but don't say exactly how.
>>>
>>> Luis: Do you know can the availability of IC_ENABLE_STATUS register 0x9c be
>>> seen from DW_IC_COMP_VERSION and how to find the same enabled status from
>>> those two registers?
>>
>> Ping? The change introduced a regression for (at least) ARM Berlin and SPEAr.
>>
> Luis?
>
Sorry for the late reply. I couldn't find a solution for what you are asking
internally
>> Is there confidence that that was the right fix? Various sources quoting what
>> appears to be the DW_apb_i2c databook (e.g. alt_i2c.h in RTEMS) only say that
>> checking IC_ENABLE_STATUS is needed when disabling, not enabling.
>>
> Hard to say, at least it fixed the issue Ben saw while keeping the benefit of
> 2702ea7dbec5 ("i2c: designware: wait for disable/enable only if necessary") by
> not waiting for disable status at the end of transfer.
>
Luis
