linux-kernel - Re: [PATCH] soc: amlogic: socinfo: Avoid soc_device_to

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] [day] [month] [year] [list]

Message-ID: <ee3a4c3b-0f99-cc79-a438-08254fe38694@baylibre.com>
Date:   Tue, 12 Nov 2019 10:06:35 +0100
From:   Neil Armstrong <narmstrong@...libre.com>
To:     Andreas Färber <afaerber@...e.de>,
        Greg Kroah-Hartman <gregkh@...uxfoundation.org>
Cc:     linux-amlogic@...ts.infradead.org,
        linux-arm-kernel@...ts.infradead.org, linux-kernel@...r.kernel.org,
        Martin Blumenstingl <martin.blumenstingl@...glemail.com>,
        Kevin Hilman <khilman@...libre.com>
Subject: Re: [PATCH] soc: amlogic: socinfo: Avoid soc_device_to_device()

On 12/11/2019 09:11, Andreas Färber wrote:
> Am 12.11.19 um 06:40 schrieb Greg Kroah-Hartman:
>> On Mon, Nov 11, 2019 at 11:15:21PM +0100, Andreas Färber wrote:
>>> The helper soc_device_to_device() is considered deprecated.
>>> For a driver __init function the predictable prefix text
>>> "soc soc0:" from dev_info() does not add real value, so use
>>> pr_info() to emit the info text without such prefix.
>>>
>>> While at it, normalize the casing of "detected" for GX.
>>>
>>> Cc: Greg Kroah-Hartman <gregkh@...uxfoundation.org>
>>> Cc: Martin Blumenstingl <martin.blumenstingl@...glemail.com>
>>> Cc: Neil Armstrong <narmstrong@...libre.com>
>>> Signed-off-by: Andreas Färber <afaerber@...e.de>
>>> ---
>>>  drivers/soc/amlogic/meson-gx-socinfo.c | 4 +---
>>>  drivers/soc/amlogic/meson-mx-socinfo.c | 4 ++--
>>>  2 files changed, 3 insertions(+), 5 deletions(-)
>>>
>>> diff --git a/drivers/soc/amlogic/meson-gx-socinfo.c b/drivers/soc/amlogic/meson-gx-socinfo.c
>>> index 01fc0d20a70d..105b819bbd5f 100644
>>> --- a/drivers/soc/amlogic/meson-gx-socinfo.c
>>> +++ b/drivers/soc/amlogic/meson-gx-socinfo.c
>>> @@ -129,7 +129,6 @@ static int __init meson_gx_socinfo_init(void)
>>>  	struct device_node *np;
>>>  	struct regmap *regmap;
>>>  	unsigned int socinfo;
>>> -	struct device *dev;
>>>  	int ret;
>>>  
>>>  	/* look up for chipid node */
>>> @@ -192,9 +191,8 @@ static int __init meson_gx_socinfo_init(void)
>>>  		kfree(soc_dev_attr);
>>>  		return PTR_ERR(soc_dev);
>>>  	}
>>> -	dev = soc_device_to_device(soc_dev);
>>>  
>>> -	dev_info(dev, "Amlogic Meson %s Revision %x:%x (%x:%x) Detected\n",
>>> +	pr_info("Amlogic Meson %s Revision %x:%x (%x:%x) detected\n",
>>
>> This should message should just be removed entirely.
>>
>>>  			soc_dev_attr->soc_id,
>>>  			socinfo_to_major(socinfo),
>>>  			socinfo_to_minor(socinfo),
>>> diff --git a/drivers/soc/amlogic/meson-mx-socinfo.c b/drivers/soc/amlogic/meson-mx-socinfo.c
>>> index 78f0f1aeca57..7db2c94a7130 100644
>>> --- a/drivers/soc/amlogic/meson-mx-socinfo.c
>>> +++ b/drivers/soc/amlogic/meson-mx-socinfo.c
>>> @@ -167,8 +167,8 @@ static int __init meson_mx_socinfo_init(void)
>>>  		return PTR_ERR(soc_dev);
>>>  	}
>>>  
>>> -	dev_info(soc_device_to_device(soc_dev), "Amlogic %s %s detected\n",
>>> -		 soc_dev_attr->soc_id, soc_dev_attr->revision);
>>> +	pr_info("Amlogic %s %s detected\n",
>>> +		soc_dev_attr->soc_id, soc_dev_attr->revision);
>>
>> Same here, no need to polute the kernel log for when all is going just
>> fine.
>>
>> That's why we created "common" driver init helpers, to prevent the
>> ability for this type of noise from even being able to be created at
>> all.
> 
> Let's have that discussion in the central thread please.
> 
> Fact here is that Amlogic GX's kernel output (and this code getting
> mirrored into U-Boot) made me aware of this driver in the first place.

As Uwe said in the common thread, this print doesn't polute the kernel log at all,
it's pretty useful for CI, like kernelCI, to check the test coverage on SoCs revisions.

Yes it's duplicated in U-boot, but a little fraction of the boards in kCi uses this
U-Boot version, they often use the vendor bootloader.

And BTW, by moving dev_info to pr_info actually changes the output format.

Can you be more precise about why "soc_device_to_device()" is now suddently deprecated ?

Neil

> 
> Regards,
> Andreas
>