Received-SPF: pass (google.com: domain of linux-kernel+bounces-119697-linux.lists.archive=gmail.com@vger.kernel.org designates 2604:1380:45d1:ec00::1 as permitted sender) client-ip=2604:1380:45d1:ec00::1;
Message-ID: <a083fe56-19ba-4c12-8364-944d8bbcc043@ghiti.fr>
Date: Tue, 26 Mar 2024 19:43:35 +0100
Precedence: bulk
MIME-Version: 1.0
User-Agent: Mozilla Thunderbird
Subject: Re: [PATCH] riscv: Mark __se_sys_* functions __used
Content-Language: en-US
To: Sami Tolvanen <samitolvanen@google.com>,
 Paul Walmsley <paul.walmsley@sifive.com>, Palmer Dabbelt
 <palmer@dabbelt.com>, Albert Ou <aou@eecs.berkeley.edu>
Cc: Kees Cook <keescook@chromium.org>, linux-riscv@lists.infradead.org,
 llvm@lists.linux.dev, linux-kernel@vger.kernel.org,
 Linux Kernel Functional Testing <lkft@linaro.org>
References: <20240326153712.1839482-2-samitolvanen@google.com>
From: Alexandre Ghiti <alex@ghiti.fr>
In-Reply-To: <20240326153712.1839482-2-samitolvanen@google.com>
Content-Type: text/plain; charset=UTF-8; format=flowed
Content-Transfer-Encoding: 7bit

Hi Sami,

On 26/03/2024 16:37, Sami Tolvanen wrote:
> Clang doesn't think ___se_sys_* functions used even though they are
> aliased to __se_sys_*, resulting in -Wunused-function warnings when
> building rv32. For example:
>
>     mm/oom_kill.c:1195:1: warning: unused function '___se_sys_process_mrelease' [-Wunused-function]
>      1195 | SYSCALL_DEFINE2(process_mrelease, int, pidfd, unsigned int, flags)
>           | ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
>     include/linux/syscalls.h:221:36: note: expanded from macro 'SYSCALL_DEFINE2'
>       221 | #define SYSCALL_DEFINE2(name, ...) SYSCALL_DEFINEx(2, _##name, __VA_ARGS__)
>           |                                    ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
>     include/linux/syscalls.h:231:2: note: expanded from macro 'SYSCALL_DEFINEx'
>       231 |         __SYSCALL_DEFINEx(x, sname, __VA_ARGS__)
>           |         ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
>     arch/riscv/include/asm/syscall_wrapper.h:81:2: note: expanded from macro '__SYSCALL_DEFINEx'
>        81 |         __SYSCALL_SE_DEFINEx(x, sys, name, __VA_ARGS__)                         \
>           |         ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
>     arch/riscv/include/asm/syscall_wrapper.h:40:14: note: expanded from macro '__SYSCALL_SE_DEFINEx'
>        40 |         static long ___se_##prefix##name(__MAP(x,__SC_LONG,__VA_ARGS__))
>           |                     ^~~~~~~~~~~~~~~~~~~~
>     <scratch space>:30:1: note: expanded from here
>        30 | ___se_sys_process_mrelease
>           | ^~~~~~~~~~~~~~~~~~~~~~~~~~
>     1 warning generated.
>
> Mark the functions __used explicitly to fix the Clang warnings.
>
> Fixes: a9ad73295cc1 ("riscv: Fix syscall wrapper for >word-size arguments")
> Reported-by: Linux Kernel Functional Testing <lkft@linaro.org>


Do you have the link of the report? So we can add a Closes tag.


> Tested-by: Linux Kernel Functional Testing <lkft@linaro.org>
> Signed-off-by: Sami Tolvanen <samitolvanen@google.com>
> ---
>   arch/riscv/include/asm/syscall_wrapper.h | 3 ++-
>   1 file changed, 2 insertions(+), 1 deletion(-)
>
> diff --git a/arch/riscv/include/asm/syscall_wrapper.h b/arch/riscv/include/asm/syscall_wrapper.h
> index 980094c2e976..ac80216549ff 100644
> --- a/arch/riscv/include/asm/syscall_wrapper.h
> +++ b/arch/riscv/include/asm/syscall_wrapper.h
> @@ -36,7 +36,8 @@ asmlinkage long __riscv_sys_ni_syscall(const struct pt_regs *);
>   					ulong)						\
>   			__attribute__((alias(__stringify(___se_##prefix##name))));	\
>   	__diag_pop();									\
> -	static long noinline ___se_##prefix##name(__MAP(x,__SC_LONG,__VA_ARGS__));	\
> +	static long noinline ___se_##prefix##name(__MAP(x,__SC_LONG,__VA_ARGS__))	\
> +			__used;								\
>   	static long ___se_##prefix##name(__MAP(x,__SC_LONG,__VA_ARGS__))
>   
>   #define SC_RISCV_REGS_TO_ARGS(x, ...) \
>
> base-commit: 4cece764965020c22cff7665b18a012006359095


You can add:

Reviewed-by: Alexandre Ghiti <alexghiti@rivosinc.com>

As the commit fixed here was merged in 6.9, this should go into fixes.

Thanks,

Alex