Re: [PATCH] arm64: ftrace: Don't forbid CALL_OPS+CC_OPTIMIZE_FOR_SIZE with Clang

From: Nathan Chancellor <nathan@kernel.org>
To: Stephen Boyd <swboyd@chromium.org>
Cc: Catalin Marinas <catalin.marinas@arm.com>,
	Will Deacon <will@kernel.org>,
	linux-kernel@vger.kernel.org, patches@lists.linux.dev,
	linux-arm-kernel@lists.infradead.org,
	Jason Ling <jasonling@chromium.org>,
	Florian Fainelli <f.fainelli@gmail.com>,
	Mark Rutland <mark.rutland@arm.com>,
	Nick Desaulniers <ndesaulniers@google.com>,
	Bill Wendling <morbo@google.com>,
	Justin Stitt <justinstitt@google.com>,
	llvm@lists.linux.dev
Subject: Re: [PATCH] arm64: ftrace: Don't forbid CALL_OPS+CC_OPTIMIZE_FOR_SIZE with Clang
Date: Sun, 25 Feb 2024 12:54:39 -0700	[thread overview]
Message-ID: <20240225195439.GA2972471@dev-arch.thelio-3990X> (raw)
In-Reply-To: <20240223064032.3463229-1-swboyd@chromium.org>

On Thu, Feb 22, 2024 at 10:40:29PM -0800, Stephen Boyd wrote:
> Per commit b3f11af9b2ce ("arm64: ftrace: forbid CALL_OPS with
> CC_OPTIMIZE_FOR_SIZE"), GCC is silently ignoring `-falign-functions=N`
> when passed `-Os`, causing functions to be improperly aligned. This
> doesn't seem to be a problem with Clang though, where enabling CALL_OPS
> with CC_OPTIMIZE_FOR_SIZE doesn't spit out any warnings at boot about
> misaligned patch-sites. Only forbid CALL_OPS if GCC is used and we're
> optimizing for size so that CALL_OPS can be used with clang optimizing
> for size.
> 
> Cc: Jason Ling <jasonling@chromium.org>
> Cc: Florian Fainelli <f.fainelli@gmail.com>
> Cc: Mark Rutland <mark.rutland@arm.com>
> Cc: Nathan Chancellor <nathan@kernel.org>
> Cc: Nick Desaulniers <ndesaulniers@google.com>
> Cc: Bill Wendling <morbo@google.com>
> Cc: Justin Stitt <justinstitt@google.com>
> Cc: <llvm@lists.linux.dev>
> Fixes: b3f11af9b2ce ("arm64: ftrace: forbid CALL_OPS with CC_OPTIMIZE_FOR_SIZE")
> Signed-off-by: Stephen Boyd <swboyd@chromium.org>

Seems reasonable to me.

Reviewed-by: Nathan Chancellor <nathan@kernel.org>

I wonder if that GCC issue is resolved with '-fmin-function-alignment'
that is being introduced in GCC 14 [1]. If it is, I suspect this
conditional could actually be CONFIG_CC_HAS_SANE_FUNCTION_ALIGNMENT
instead of CONFIG_CC_IS_CLANG, but that can obviously happen as a clean
up later down the road (more just food for thought for the arm64 folks,
not sure how much '-Os' is used in general)

[1]: https://git.kernel.org/masahiroy/linux-kbuild/c/5270316c9fec8cc99aa0e0a258509c5c7f789d12

> ---
>  arch/arm64/Kconfig | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/arch/arm64/Kconfig b/arch/arm64/Kconfig
> index aa7c1d435139..6b96d75a3a3d 100644
> --- a/arch/arm64/Kconfig
> +++ b/arch/arm64/Kconfig
> @@ -198,7 +198,7 @@ config ARM64
>  		if DYNAMIC_FTRACE_WITH_ARGS && DYNAMIC_FTRACE_WITH_CALL_OPS
>  	select HAVE_DYNAMIC_FTRACE_WITH_CALL_OPS \
>  		if (DYNAMIC_FTRACE_WITH_ARGS && !CFI_CLANG && \
> -		    !CC_OPTIMIZE_FOR_SIZE)
> +		    (CC_IS_CLANG || !CC_OPTIMIZE_FOR_SIZE))
>  	select FTRACE_MCOUNT_USE_PATCHABLE_FUNCTION_ENTRY \
>  		if DYNAMIC_FTRACE_WITH_ARGS
>  	select HAVE_SAMPLE_FTRACE_DIRECT
> 
> base-commit: 6613476e225e090cc9aad49be7fa504e290dd33d
> -- 
> https://chromeos.dev
> 

_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel