Re: Fwd: [PATCH net-next 1/2] lib: string: add strreplace_nonalnum

[Heiner Kallweit <hkallweit1@gmail.com>]

On 03.03.2019 18:55, Greg Kroah-Hartman wrote:
> On Sun, Mar 03, 2019 at 06:47:32PM +0100, Heiner Kallweit wrote:
>> I submitted this through the netdev tree, maybe relevant for you as well.
>> See also here: https://marc.info/?t=155103900100003&r=1&w=2
>>
>> -------- Forwarded Message --------
>> Subject: [PATCH net-next 1/2] lib: string: add strreplace_nonalnum
>> Date: Sun, 3 Mar 2019 18:20:50 +0100
>> From: Heiner Kallweit <hkallweit1@gmail.com>
>> To: Florian Fainelli <f.fainelli@gmail.com>, Andrew Lunn <andrew@lunn.ch>, David Miller <davem@davemloft.net>
>> CC: netdev@vger.kernel.org <netdev@vger.kernel.org>
>>
>> Add a new function strreplace_nonalnum that replaces all
>> non-alphanumeric characters. Such functionality is needed e.g. when a
>> string is supposed to be used in a sysfs file name. If '\0' is given
>> as new character then non-alphanumeric characters are cut. 
> 
> sysfs doesn't have any such requirements, it can use whatever you want
> to give it for a filename.
> 
Even a slash?
HWMON drivers is an example where such functionality occurs open-coded.

> So don't create a random kernel function for sysfs please.
> 
>>
>> Signed-off-by: Heiner Kallweit <hkallweit1@gmail.com>
>> ---
>>  include/linux/string.h |  1 +
>>  lib/string.c           | 27 +++++++++++++++++++++++++++
>>  2 files changed, 28 insertions(+)
>>
>> diff --git a/include/linux/string.h b/include/linux/string.h
>> index 7927b875f..d827b0b0f 100644
>> --- a/include/linux/string.h
>> +++ b/include/linux/string.h
>> @@ -169,6 +169,7 @@ static inline void memcpy_flushcache(void *dst, const void *src, size_t cnt)
>>  #endif
>>  void *memchr_inv(const void *s, int c, size_t n);
>>  char *strreplace(char *s, char old, char new);
>> +char *strreplace_nonalnum(char *s, char new);
>>  
>>  extern void kfree_const(const void *x);
>>  
>> diff --git a/lib/string.c b/lib/string.c
>> index 38e4ca08e..f2b1baf96 100644
>> --- a/lib/string.c
>> +++ b/lib/string.c
>> @@ -1047,6 +1047,33 @@ char *strreplace(char *s, char old, char new)
>>  }
>>  EXPORT_SYMBOL(strreplace);
>>  
>> +/**
>> + * strreplace_nonalnum - Replace all non-alphanumeric characters in a string.
>> + * @s: The string to operate on.
>> + * @new: The character non-alphanumeric characters are replaced with.
>> + *
>> + * If new is '\0' then non-alphanumeric characters are cut.
>> + *
>> + * Returns pointer to the nul byte at the end of the modified string.
> 
> Why do you need to point to the end of the string?
> 
I don't have to. I just tried to keep as close as possible to the original
strreplace().

> 
>> + */
>> +char *strreplace_nonalnum(char *s, char new)
>> +{
>> +	char *p = s;
>> +
>> +	for (; *s; ++s)
>> +		if (isalnum(*s)) {
>> +			if (p != s)
>> +				*p = *s;
>> +			++p;
>> +		} else if (new) {
>> +			*p++ = new;
>> +		}
>> +	*p = '\0';
> 
> No max length?  No error checking?  Surely we can do better, see the
> long thread on the kernel-hardnening list about string functions please.
> 
As for the review comment before, I basically used strreplace() as
template and extended it. What you say about hardening is right,
but it's open for quite a few functions, isn't it?
I'll have a look at the mail thread you mentioned.

> greg k-h
> 
Heiner