Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:18 as permitted sender) client-ip=2620:137:e000::1:18;
Message-ID: <795b9b2d-1d6b-359c-ef12-f25c6b3472be@roeck-us.net>
Date:   Wed, 30 Mar 2022 08:13:14 -0700
MIME-Version: 1.0
User-Agent: Mozilla/5.0 (X11; Linux x86_64; rv:91.0) Gecko/20100101
 Thunderbird/91.7.0
Content-Language: en-US
To:     David Laight <David.Laight@ACULAB.COM>,
        Michael Walle <michael@walle.cc>,
        Xu Yilun <yilun.xu@intel.com>, Tom Rix <trix@redhat.com>,
        Jean Delvare <jdelvare@suse.com>, Andrew Lunn <andrew@lunn.ch>,
        Heiner Kallweit <hkallweit1@gmail.com>,
        Russell King <linux@armlinux.org.uk>,
        "David S . Miller" <davem@davemloft.net>,
        Jakub Kicinski <kuba@kernel.org>,
        Paolo Abeni <pabeni@redhat.com>
Cc:     "linux-hwmon@vger.kernel.org" <linux-hwmon@vger.kernel.org>,
        "linux-kernel@vger.kernel.org" <linux-kernel@vger.kernel.org>,
        "netdev@vger.kernel.org" <netdev@vger.kernel.org>
References: <20220329160730.3265481-1-michael@walle.cc>
 <20220329160730.3265481-2-michael@walle.cc>
 <75093b82-4625-d806-a4ea-372b74e60c3b@roeck-us.net>
 <02545bf1c21b45f78eba5e8b37951748@AcuMS.aculab.com>
From:   Guenter Roeck <linux@roeck-us.net>
Subject: Re: [PATCH v2 1/5] hwmon: introduce hwmon_sanitize_name()
In-Reply-To: <02545bf1c21b45f78eba5e8b37951748@AcuMS.aculab.com>
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 7bit
Precedence: bulk

On 3/30/22 07:50, David Laight wrote:
> From: Guenter Roeck
>> Sent: 30 March 2022 15:23
>> On 3/29/22 09:07, Michael Walle wrote:
>>> More and more drivers will check for bad characters in the hwmon name
>>> and all are using the same code snippet. Consolidate that code by adding
>>> a new hwmon_sanitize_name() function.
>>>
>>> Signed-off-by: Michael Walle <michael@walle.cc>
>>> ---
>>>    Documentation/hwmon/hwmon-kernel-api.rst |  9 ++++-
>>>    drivers/hwmon/hwmon.c                    | 49 ++++++++++++++++++++++++
>>>    include/linux/hwmon.h                    |  3 ++
>>>    3 files changed, 60 insertions(+), 1 deletion(-)
>>>
>>> diff --git a/Documentation/hwmon/hwmon-kernel-api.rst b/Documentation/hwmon/hwmon-kernel-api.rst
>>> index c41eb6108103..12f4a9bcef04 100644
>>> --- a/Documentation/hwmon/hwmon-kernel-api.rst
>>> +++ b/Documentation/hwmon/hwmon-kernel-api.rst
>>> @@ -50,6 +50,10 @@ register/unregister functions::
>>>
>>>      void devm_hwmon_device_unregister(struct device *dev);
>>>
>>> +  char *hwmon_sanitize_name(const char *name);
>>> +
>>> +  char *devm_hwmon_sanitize_name(struct device *dev, const char *name);
>>> +
>>>    hwmon_device_register_with_groups registers a hardware monitoring device.
>>>    The first parameter of this function is a pointer to the parent device.
>>>    The name parameter is a pointer to the hwmon device name. The registration
>>> @@ -93,7 +97,10 @@ removal would be too late.
>>>
>>>    All supported hwmon device registration functions only accept valid device
>>>    names. Device names including invalid characters (whitespace, '*', or '-')
>>> -will be rejected. The 'name' parameter is mandatory.
>>> +will be rejected. The 'name' parameter is mandatory. Before calling a
>>> +register function you should either use hwmon_sanitize_name or
>>> +devm_hwmon_sanitize_name to replace any invalid characters with an
>>> +underscore.
>>
>> That needs more details and deserves its own paragraph. Calling one of
>> the functions is only necessary if the original name does or can include
>> unsupported characters; an unconditional "should" is therefore a bit
>> strong. Also, it is important to mention that the function duplicates
>> the name, and that it is the responsibility of the caller to release
>> the name if hwmon_sanitize_name() was called and the device is removed.
> 
> More worrying, and not documented, is that the buffer 'name' points
> to must persist.
> 

You mean the name argument passed to the hwmon registration functions ?
That has been the case forever, and I don't recall a single problem
with it. If it disturbs you, please feel free to submit a patch adding
more details to the documentation.

I would not want to change the code and always copy the name because in
almost all cases it _is_ a fixed string, and duplicating it would have
no value.

> ISTM that the kmalloc() in __hwmon_device_register() should include
> space for a copy of the name.
> It can then do what it will with whatever is passed in.
> 

Whatever is passed in is what the user wants. Registration functions
don't change the name. Providing a valid name is the responsibility
of the caller.

> Oh yes, it has my 'favourite construct':  if (!strlen(name)) ...
> (well str[strlen(str)] = 0 also happens!)
> 

Sorry, I don't understand what the problem is here.

Thanks,
Guenter