Re: [PATCH v2] clk: rs9: Fix I2C accessors

From: Marek Vasut <marex@denx.de>
To: Alexander Stein <alexander.stein@ew.tq-group.com>
Cc: linux-clk@vger.kernel.org,
	Michael Turquette <mturquette@baylibre.com>,
	Stephen Boyd <sboyd@kernel.org>
Subject: Re: [PATCH v2] clk: rs9: Fix I2C accessors
Date: Wed, 28 Sep 2022 14:53:12 +0200	[thread overview]
Message-ID: <7acfab7e-27d7-d2f6-22e8-00042193ef87@denx.de> (raw)
In-Reply-To: <13108964.uLZWGnKmhe@steina-w>

On 9/28/22 09:32, Alexander Stein wrote:
> Hi Marek,

Hi,

>>> Putting all together the regmap debug output is like this:
>>> $ cat /sys/kernel/debug/regmap/1-0068/registers
>>> 0: ff
>>> 1: 06
>>> 2: ff
>>> 3: 5f
>>> 4: 00
>>> 5: 01
>>> 6: 04
>>> 7: 01
>>
>> What about option 3 -- disable the cache altogether ?
> 
> Sure this works as well.

I sent a patch like that for now, let's revisit the cache topic in a 
separate patch.

>> I can imagine since the chip is configured with like 2-3 I2C writes on
>> boot and then never again written to, that might be the simplest approach.
> 
> I'm thinking about disabling unused clock outputs later on, so dynamic bit
> flipping at runtime would be required. For this case cache usage seems
> reasonable. For now disabling is ok, IMHO.
> 
>>> This is actually a 9FGV0441 using some queued patches on my side.
>>
>> Nice, do you plan to send a binding update for this one ?
> 
> Sure, but as the DIF bit definition is device specific, I want to settle on
> general access first.

ACK