Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::3:6 as permitted sender) client-ip=2620:137:e000::3:6;
Message-ID: <ab1c4929-0d7d-45eb-ab70-7680dbebcdbb@ti.com>
Date:   Fri, 20 Oct 2023 11:43:04 -0500
MIME-Version: 1.0
User-Agent: Mozilla Thunderbird
Subject: Re: [PATCH v3] mux: mmio: use reg property when parent device is not
 a syscon
Content-Language: en-US
To:     Peter Rosin <peda@axentia.se>,
        Greg Kroah-Hartman <gregkh@linuxfoundation.org>,
        Rob Herring <robh+dt@kernel.org>,
        Krzysztof Kozlowski <krzysztof.kozlowski+dt@linaro.org>,
        Nishanth Menon <nm@ti.com>,
        Vignesh Raghavendra <vigneshr@ti.com>
CC:     <devicetree@vger.kernel.org>, <linux-kernel@vger.kernel.org>
References: <20230911151030.71100-1-afd@ti.com>
 <0cb645c7-f3c5-e4bb-7686-2a83d32274bb@axentia.se>
From:   Andrew Davis <afd@ti.com>
In-Reply-To: <0cb645c7-f3c5-e4bb-7686-2a83d32274bb@axentia.se>
Content-Type: text/plain; charset="UTF-8"; format=flowed
Content-Transfer-Encoding: 7bit
Precedence: bulk

On 10/20/23 9:28 AM, Peter Rosin wrote:
> Hi!
> 
> 2023-09-11 at 17:10, Andrew Davis wrote:
>> The DT binding for the reg-mux compatible states it can be used when the
>> "parent device of mux controller is not syscon device". It also allows
>> for a reg property. When the reg property is provided, use that to
>> identify the address space for this mux. If not provided fallback to
>> using the parent device as a regmap provider.
>>
>> Signed-off-by: Andrew Davis <afd@ti.com>
>> Reviewed-by: Nishanth Menon <nm@ti.com>
>> ---
>>
>> Changes from v2:
>>   - Rebased on v6.6-rc1
>>
>> Changes from v1:
>>   - Flip logic as suggested in v1[0]
>>
>> [0] https://lore.kernel.org/lkml/1c27d9d4-b1cc-c158-90f7-f7e47e02c424@ti.com/T/
>>
>>   drivers/mux/mmio.c | 9 ++++++---
>>   1 file changed, 6 insertions(+), 3 deletions(-)
>>
>> diff --git a/drivers/mux/mmio.c b/drivers/mux/mmio.c
>> index fd1d121a584ba..b6095b7853ed2 100644
>> --- a/drivers/mux/mmio.c
>> +++ b/drivers/mux/mmio.c
>> @@ -44,10 +44,13 @@ static int mux_mmio_probe(struct platform_device *pdev)
>>   	int ret;
>>   	int i;
>>   
>> -	if (of_device_is_compatible(np, "mmio-mux"))
>> +	if (of_device_is_compatible(np, "mmio-mux")) {
>>   		regmap = syscon_node_to_regmap(np->parent);
>> -	else
>> -		regmap = dev_get_regmap(dev->parent, NULL) ?: ERR_PTR(-ENODEV);
>> +	} else {
>> +		regmap = device_node_to_regmap(np);
> 
> I started digging in device_node_to_regmap() to try to find an error that
> could be used to trigger if the failover to dev_get_regmap() should be
> tried, instead of always doing the failover on error. I got lost fairly
> quickly, but it seems device_node_to_regmap() can return -EDEFER_PROBE.
> While I'm not certain that it is applicable, that case should probably
> not fall back to dev_get_regmap()...
> 
> Are there other error cases that should prevent the failover? I would
> guess that it's perhaps just a single error that should trigger trying
> the failover path? But I don't know, and which error if that's the case?
> 

Ideally the only error that will be returned is ENOMEM, which happens when
this node does not have a 'reg' property, and this is also the one case we
want to do the failover. So all should be well.

> How much badness can be caused if syscon_node_to_regmap() fails for some
> random obscure reason and the failover path is taken inadvertently? It
> certainly smells bad for -EDEFER_PROBE, but do you have any insight in
> other cases?
> 

If we take the failover inadvertently then we will check if the parent
node is a syscon, if it is then our offset will most likely be wrong
(parent will not match child 'reg').

> And after getting to approx that point a while back, I had other things
> to take care of, and this fell off the table. Sorry!
> 

No problem as long as we can find a way to get this in quickly (lot of
DT warning need cleaned up based on this patch).

Thanks
Andrew

> Cheers,
> Peter
> 
>> +		if (IS_ERR(regmap))
>> +			regmap = dev_get_regmap(dev->parent, NULL) ?: ERR_PTR(-ENODEV);
>> +	}
>>   	if (IS_ERR(regmap)) {
>>   		ret = PTR_ERR(regmap);
>>   		dev_err(dev, "failed to get regmap: %d\n", ret);