Re: Silence gcc-8 warnings

[Franz Sirl <Franz dot Sirl-kernel at lauterbach dot com>]

Am 2018-04-24 um 14:07 schrieb Michael Matz:> Hi,
>
> On Tue, 24 Apr 2018, Alan Modra wrote:
>
>> No, it is not a bug.  We have internal and external structs, and a
>> field in the external struct that does not need to be NULL
>> terminated.
>
> Ah, that's the crucial point (plus the zero termination for 
convenience of
> the internal one).  Well, so the warning is correct still (according to
> its intent, which may or may not be ill advised), but of course doesn't
> take that into account.
>
>> What's more, we do want strncpy behaviour rather than memcpy, since for
>> defensive programming we don't want to copy possibly uninitialized data
>> past a string terminator in the internal representation, and do want to
>> fill the external representation with zeros.
>
> Yeah, understood.

Hi,

unfortunately it's too late for GCC-8, but for GCC-9 I intend to open an 
enhancement request to split -Wstringop-truncation (and a similar case 
for -Wsizeof-pointer-memaccess).
Something like -Wstringop-truncation will not warn for strncpy's 
documented (!) natural usage to copy to a fixed size record (non-null 
terminated char array), but -Wstringop-truncation-strict 
(-Wstringop-truncation-secure?) will warn.

Franz Sirl