linux-kernel - Re: [PATCH] arm64: move smp_send

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: <aFVE3uuNCbHF7f5z@arm.com>
Date: Fri, 20 Jun 2025 12:24:14 +0100
From: Catalin Marinas <catalin.marinas@....com>
To: Arnd Bergmann <arnd@...nel.org>
Cc: Will Deacon <will@...nel.org>, Arnd Bergmann <arnd@...db.de>,
	Jonathan Cameron <Jonathan.Cameron@...wei.com>,
	Gavin Shan <gshan@...hat.com>,
	Jean-Philippe Brucker <jean-philippe@...aro.org>,
	James Morse <james.morse@....com>,
	Douglas Anderson <dianders@...omium.org>,
	linux-arm-kernel@...ts.infradead.org, linux-kernel@...r.kernel.org
Subject: Re: [PATCH] arm64: move smp_send_stop() cpu mask off stack

On Fri, Jun 20, 2025 at 01:10:41PM +0200, Arnd Bergmann wrote:
> From: Arnd Bergmann <arnd@...db.de>
> 
> For really large values of CONFIG_NR_CPUS, a CPU mask value should
> not be put on the stack:
> 
> arch/arm64/kernel/smp.c:1188:1: error: the frame size of 8544 bytes is larger than 1536 bytes [-Werror=frame-larger-than=]
> 
> This could be achieved using alloc_cpumask_var(), which makes it
> depend on CONFIG_CPUMASK_OFFSTACK, but as this function is already
> serialized and can only run on one CPU, making the variable 'static'
> is easier.
> 
> Signed-off-by: Arnd Bergmann <arnd@...db.de>
> ---
>  arch/arm64/kernel/smp.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/arch/arm64/kernel/smp.c b/arch/arm64/kernel/smp.c
> index 15987100c0cf..5c605dc7f5be 100644
> --- a/arch/arm64/kernel/smp.c
> +++ b/arch/arm64/kernel/smp.c
> @@ -1107,7 +1107,7 @@ static inline unsigned int num_other_online_cpus(void)
>  void smp_send_stop(void)
>  {
>  	static unsigned long stop_in_progress;
> -	cpumask_t mask;
> +	static cpumask_t mask;
>  	unsigned long timeout;

This would work, there's a stop_in_progress check and only one CPU would
modify the mask.

Reviewed-by: Catalin Marinas <catalin.marinas@....com>