linux-kernel - Re: [PATCH v5 1/1] x86/cpu: Make sure flag_is_changeable

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: <106B0D31-F58E-49EA-9FEC-5573B684ACC7@zytor.com>
Date: Fri, 08 Nov 2024 18:34:15 +0100
From: "H. Peter Anvin" <hpa@...or.com>
To: Nathan Chancellor <nathan@...nel.org>
CC: Andy Shevchenko <andriy.shevchenko@...ux.intel.com>,
        Thomas Gleixner <tglx@...utronix.de>, linux-kernel@...r.kernel.org,
        llvm@...ts.linux.dev, Ingo Molnar <mingo@...hat.com>,
        Borislav Petkov <bp@...en8.de>,
        Dave Hansen <dave.hansen@...ux.intel.com>, x86@...nel.org,
        Nick Desaulniers <ndesaulniers@...gle.com>,
        Bill Wendling <morbo@...gle.com>,
        Justin Stitt <justinstitt@...gle.com>,
        Dave Hansen <dave.hansen@...el.com>
Subject: Re: [PATCH v5 1/1] x86/cpu: Make sure flag_is_changeable_p() is always being used

On November 8, 2024 6:29:20 PM GMT+01:00, Nathan Chancellor <nathan@...nel.org> wrote:
>On Fri, Nov 08, 2024 at 04:48:16PM +0100, H. Peter Anvin wrote:
>> On November 8, 2024 4:41:16 PM GMT+01:00, Andy Shevchenko <andriy.shevchenko@...ux.intel.com> wrote:
>> >On Fri, Nov 08, 2024 at 04:35:17PM +0100, H. Peter Anvin wrote:
>> >> On November 8, 2024 4:30:10 PM GMT+01:00, Andy Shevchenko <andriy.shevchenko@...ux.intel.com> wrote:
>> >
>> >> >See also commit 6863f5643dd7 ("kbuild: allow Clang to find unused static
>> >> >inline functions for W=1 build").
>> >
>> >^^^ (1)
>> >
>> >> Looks good to me:
>> >> 
>> >> Reviewed-by: H. Peter Anvin (Intel) <hpa@...or.com>
>> >
>> >Thank you!
>> >
>> >> But another question: why the hell does clang complain about an unused static inline function?!
>> >
>> >Does (1) shed a bit of light to this?
>> >
>> 
>> How on earth is that supposed to work?! We have static inline functions in headers all over the place that are only used in certain circumstances. 
>> 
>> Is this a good thing, really? Or is it noise?
>
>Did you read the commit message of 6863f5643dd7 or just the title?
>
>The difference between Clang and GCC is only around static inline
>function in .c files, not .h files.
>
>  $ cat test.h
>  static inline void unused_inline_in_h(void) {}
>
>  $ cat test.c
>  #include "test.h"
>
>  static inline void unused_inline_in_c(void) {}
>
>  static void unused_in_c(void) {}
>
>  $ gcc -Wall -c -o /dev/null test.c
>  test.c:5:13: warning: ‘unused_in_c’ defined but not used [-Wunused-function]
>      5 | static void unused_in_c(void) {}
>        |             ^~~~~~~~~~~
>
>  $ clang -fsyntax-only -Wall test.c
>  test.c:3:20: warning: unused function 'unused_inline_in_c' [-Wunused-function]
>      3 | static inline void unused_inline_in_c(void) {}
>        |                    ^~~~~~~~~~~~~~~~~~
>  test.c:5:13: warning: unused function 'unused_in_c' [-Wunused-function]
>      5 | static void unused_in_c(void) {}
>        |             ^~~~~~~~~~~
>  2 warnings generated.
>
>I do not think there are too many instances of unused static inline
>functions in .c files but Andy might be able to speak more around how
>many instances he has had to fix across the tree. I can see how this
>difference can be useful for catching dead code and maybe even making
>code cleaner but if it proves to be too much of an annoyance for the
>wider community, we could potentially discuss reverting 6863f5643dd7.
>
>Cheers,
>Nathan

I'm on the road traveling and have limited ability to look things up at the moment. 

However, in .c files the value becomes very very small: in .h files an unused inline becomes a drag on compile time because it effects a number of compilation units, but for a .c file it is just one.