linux-kernel - Re: [linux-sunxi] [PATCH 1/7] pinctrl: sunxi: add support for pin controllers without bus gate

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <7c3b4294-117a-c293-03d4-a35a823a796e@arm.com>
Date:   Thu, 11 Jan 2018 10:41:39 +0000
From:   Andre Przywara <andre.przywara@....com>
To:     Icenowy Zheng <icenowy@...c.io>, Rob Herring <robh+dt@...nel.org>,
        Maxime Ripard <maxime.ripard@...e-electrons.com>,
        Chen-Yu Tsai <wens@...e.org>,
        Linus Walleij <linus.walleij@...aro.org>
Cc:     linux-clk@...r.kernel.org, devicetree@...r.kernel.org,
        linux-arm-kernel@...ts.infradead.org, linux-kernel@...r.kernel.org,
        linux-gpio@...r.kernel.org, linux-sunxi@...glegroups.com
Subject: Re: [linux-sunxi] [PATCH 1/7] pinctrl: sunxi: add support for pin
 controllers without bus gate

Hi,

On 11/01/18 10:15, Icenowy Zheng wrote:
> 
> 
> 于 2018年1月11日 GMT+08:00 下午6:08:19, Andre Przywara <andre.przywara@....com> 写到:
>> Hi,
>>
>> On 06/01/18 04:23, Icenowy Zheng wrote:
>>> The Allwinner H6 pin controllers (both the main one and the CPUs one)
>>> have no bus gate clocks.
>>>
>>> Add support for this kind of pin controllers.
>>>
>>> Signed-off-by: Icenowy Zheng <icenowy@...c.io>
>>> ---
>>>  drivers/pinctrl/sunxi/pinctrl-sunxi.c | 30
>> ++++++++++++++++++++----------
>>>  drivers/pinctrl/sunxi/pinctrl-sunxi.h |  1 +
>>>  2 files changed, 21 insertions(+), 10 deletions(-)
>>>
>>> diff --git a/drivers/pinctrl/sunxi/pinctrl-sunxi.c
>> b/drivers/pinctrl/sunxi/pinctrl-sunxi.c
>>> index 4b6cb25bc796..68cd505679d9 100644
>>> --- a/drivers/pinctrl/sunxi/pinctrl-sunxi.c
>>> +++ b/drivers/pinctrl/sunxi/pinctrl-sunxi.c
>>> @@ -1182,7 +1182,12 @@ static int sunxi_pinctrl_setup_debounce(struct
>> sunxi_pinctrl *pctl,
>>>  	unsigned int hosc_div, losc_div;
>>>  	struct clk *hosc, *losc;
>>>  	u8 div, src;
>>> -	int i, ret;
>>> +	int i, ret, clk_count;
>>> +
>>> +	if (pctl->desc->without_bus_gate)
>>> +		clk_count = 2;
>>> +	else
>>> +		clk_count = 3;
>>>  
>>>  	/* Deal with old DTs that didn't have the oscillators */
>>>  	if (of_count_phandle_with_args(node, "clocks", "#clock-cells") !=
>> 3)

Just spotted: I guess you wanted to compare against that computed value
here?
But I wonder if we can get rid of this check at all? Don't we rely on
clock-names and input-debounce anyway? So we will bail out later anyway
if the DT does not have those?
Why do we need this check then?

>>> @@ -1360,15 +1365,19 @@ int sunxi_pinctrl_init_with_variant(struct
>> platform_device *pdev,
>>>  			goto gpiochip_error;
>>>  	}
>>>  
>>> -	clk = devm_clk_get(&pdev->dev, NULL);
>>> -	if (IS_ERR(clk)) {
>>> -		ret = PTR_ERR(clk);
>>> -		goto gpiochip_error;
>>> -	}
>>> +	if (!desc->without_bus_gate) {
>>
>> Do we really need explicit support for that case?
>> Can't we have something that works automatically?
> 
> It can be a sanity check. When a SoC comes with bus gate
> support but no apb is provided, there's something wrong.
> 
>>
>> if (node has clock-names property)		(A)
>> 	use clocks as enumerated and named there
>> else if (node has one clock reference)		(B)
>> 	use this as gate clock, no debounce support
>> else if (node has no clock property at all)	(C)
>> 	no gate clock needed, no debounce support
> 
> This should not happen in practice, as no gate clock is implemented
> after debounce.

But still you seem to somewhat support it with your changes above - by
bailing out if there aren't two clocks.
This kind of explicitly checking for a certain number of clocks sounds
not very robust and future proof to me, hence the suggestion to get rid
of it.

Cheers,
Andre.

>>
>> On top of that we should add the clock-names property to all DTs, even
>> for those with only a "apb" clock. Shouldn't hurt existing kernels.
>> Possibly even add debounce support for those on the way, if applicable.
>>
>> So we would just support case (B) and (C) for legacy reasons.
>>
>> Does that make sense?
>>
>> Cheers,
>> Andre.
>>
>>> +		clk = devm_clk_get(&pdev->dev, NULL);
>>> +		if (IS_ERR(clk)) {
>>> +			ret = PTR_ERR(clk);
>>> +			goto gpiochip_error;
>>> +		}
>>>  
>>> -	ret = clk_prepare_enable(clk);
>>> -	if (ret)
>>> -		goto gpiochip_error;
>>> +		ret = clk_prepare_enable(clk);
>>> +		if (ret)
>>> +			goto gpiochip_error;
>>> +	} else {
>>> +		clk = NULL;
>>> +	}
>>>  
>>>  	pctl->irq = devm_kcalloc(&pdev->dev,
>>>  				 pctl->desc->irq_banks,
>>> @@ -1425,7 +1434,8 @@ int sunxi_pinctrl_init_with_variant(struct
>> platform_device *pdev,
>>>  	return 0;
>>>  
>>>  clk_error:
>>> -	clk_disable_unprepare(clk);
>>> +	if (clk)
>>> +		clk_disable_unprepare(clk);
>>>  gpiochip_error:
>>>  	gpiochip_remove(pctl->chip);
>>>  	return ret;
>>> diff --git a/drivers/pinctrl/sunxi/pinctrl-sunxi.h
>> b/drivers/pinctrl/sunxi/pinctrl-sunxi.h
>>> index 11b128f54ed2..ccb6230f0bb5 100644
>>> --- a/drivers/pinctrl/sunxi/pinctrl-sunxi.h
>>> +++ b/drivers/pinctrl/sunxi/pinctrl-sunxi.h
>>> @@ -113,6 +113,7 @@ struct sunxi_pinctrl_desc {
>>>  	unsigned			irq_bank_base;
>>>  	bool				irq_read_needs_mux;
>>>  	bool				disable_strict_mode;
>>> +	bool				without_bus_gate;
>>>  };
>>>  
>>>  struct sunxi_pinctrl_function {
>>>