Date:   Sun, 5 Aug 2018 11:23:00 +0200
From:   Miguel Ojeda <miguel.ojeda.sandonis@...il.com>
To:     Joe Perches <joe@...ches.com>
Cc:     Ingo Molnar <mingo@...nel.org>,
        Andrew Morton <akpm@...ux-foundation.org>,
        Josh Poimboeuf <jpoimboe@...hat.com>,
        Kees Cook <keescook@...omium.org>,
        Geert Uytterhoeven <geert@...ux-m68k.org>,
        Will Deacon <will.deacon@....com>,
        Greg Kroah-Hartman <gregkh@...uxfoundation.org>,
        David Rientjes <rientjes@...gle.com>,
        Martin Sebor <msebor@...il.com>, Arnd Bergmann <arnd@...db.de>,
        linux-kernel <linux-kernel@...r.kernel.org>
Subject: Re: [PATCH v2 1/2] Support the nonstring variable attribute (gcc >= 8)

On Sun, Aug 5, 2018 at 11:14 AM, Miguel Ojeda
<miguel.ojeda.sandonis@...il.com> wrote:
> On Sun, Aug 5, 2018 at 3:38 AM, Joe Perches <joe@...ches.com> wrote:
>> On Wed, 2018-08-01 at 19:54 +0200, Miguel Ojeda wrote:
>>> From the GCC manual:
>>>
>>> The nonstring variable attribute specifies that an object or member
>>> declaration with type array of char or pointer to char is intended to
>>> store character arrays that do not necessarily contain a terminating NUL
>>> character. This is useful in detecting uses of such arrays or pointers
>>> with functions that expect NUL-terminated strings, and to avoid warnings
>>> when such an array or pointer is used as an argument to a bounded string
>>> manipulation function such as strncpy.
>>>
>>>   https://gcc.gnu.org/onlinedocs/gcc/Common-Variable-Attributes.html
>>>
>>> Some reports are already coming to the LKML regarding these
>>> warnings. When they are false positives, like this one
>>>
>>>   https://lkml.org/lkml/2018/1/16/135
>>>
>>> we can use __nonstring to let gcc know a NUL character is not required.
>>>
>>> Cc: Ingo Molnar <mingo@...nel.org>
>>> Cc: Josh Poimboeuf <jpoimboe@...hat.com>
>>> Cc: Kees Cook <keescook@...omium.org>
>>> Cc: Andrew Morton <akpm@...ux-foundation.org>
>>> Cc: Geert Uytterhoeven <geert@...ux-m68k.org>
>>> Cc: Will Deacon <will.deacon@....com>
>>> Cc: Greg Kroah-Hartman <gregkh@...uxfoundation.org>
>>> Cc: David Rientjes <rientjes@...gle.com>
>>> Cc: Martin Sebor <msebor@...il.com>
>>> Cc: Arnd Bergmann <arnd@...db.de>
>>> Signed-off-by: Miguel Ojeda <miguel.ojeda.sandonis@...il.com>
>>> ---
>>> Re-sending this since a few months have passed, Martin has improved
>>> GCC's feature and warnings are appearing in Geert's build bot.
>>> Added an example in the second patch as requested by David.
>>>
>>>  include/linux/compiler-gcc.h   | 14 ++++++++++++++
>>>  include/linux/compiler_types.h |  4 ++++
>>>  2 files changed, 18 insertions(+)
>>>
>>> diff --git a/include/linux/compiler-gcc.h b/include/linux/compiler-gcc.h
>>> index 573f5a7d42d4..fab4e904f1fe 100644
>>> --- a/include/linux/compiler-gcc.h
>>> +++ b/include/linux/compiler-gcc.h
>>> @@ -343,6 +343,20 @@
>>>  #define __designated_init __attribute__((designated_init))
>>>  #endif
>>>
>>> +#if GCC_VERSION >= 80000
>>> +/*
>>> + * The nonstring variable attribute specifies that an object or member
>>> + * declaration with type array of char or pointer to char is intended
>>> + * to store character arrays that do not necessarily contain a terminating
>>> + * NUL character. This is useful in detecting uses of such arrays or pointers
>>> + * with functions that expect NUL-terminated strings, and to avoid warnings
>>> + * when such an array or pointer is used as an argument to a bounded string
>>> + * manipulation function such as strncpy.
>>> + * https://gcc.gnu.org/onlinedocs/gcc/Common-Variable-Attributes.html
>>> + */
>>
>> Please move this down to the already existing test
>> for GCC_VERSION >= 80000 near the bottom of the
>> file so that version number tests are always in
>> increasing order in the file.
>
> Ah, good catch! The test was added in the v1->v2 meantime. Will do, thanks!
>

On the other hand, it may get messy given that this has an "else" section:

    #if GCC_VERSION >= 80000
    #define __diag_GCC_8(s) __diag(s)
    #else
    #define __diag_GCC_8(s)
    #endif

Doing it feature-by-feature seems more readable. e.g. doing sorted
single tests for versions would imply splitting the __diag feature.

For the moment I will move __nonstring to the bottom in v3, which
looks better nevertheless, and I will think about how to do this.

Cheers,
Miguel

> By the way, the file is a mess... Some other tests go inside the big
> "#if GCC_VERSION >= 40000", others go at the end. We should clean it
> up and sort it. I might just do it...
>
> Cheers,
> Miguel
>
>>
>>> +#define __nonstring __attribute__((nonstring))
>>> +#endif
>>> +
>>>  #endif       /* gcc version >= 40000 specific checks */
>>>
>>>  #if !defined(__noclone)
>>> diff --git a/include/linux/compiler_types.h b/include/linux/compiler_types.h
>>> index a8ba6b04152c..9c07be36e86a 100644
>>> --- a/include/linux/compiler_types.h
>>> +++ b/include/linux/compiler_types.h
>>> @@ -289,4 +289,8 @@ struct ftrace_likely_data {
>>>  #define __diag_error(compiler, version, option, comment) \
>>>       __diag_ ## compiler(version, error, option)
>>>
>>> +#ifndef __nonstring
>>> +# define __nonstring
>>> +#endif
>>> +
>>>  #endif /* __LINUX_COMPILER_TYPES_H */

Powered by blists - more mailing lists