Re: [patch V2 09/37] x86/microcode/intel: Remove pointless mutex

[Nikolay Borisov]

On 12.08.23 г. 22:58 ч., Thomas Gleixner wrote:
> From: Thomas Gleixner <tglx@xxxxxxxxxxxxx>
>
> There is no concurreny.
>
> Signed-off-by: Thomas Gleixner <tglx@xxxxxxxxxxxxx>
>
> ---
>   arch/x86/kernel/cpu/microcode/intel.c |   24 ++----------------------
>   1 file changed, 2 insertions(+), 22 deletions(-)
> ---
> --- a/arch/x86/kernel/cpu/microcode/intel.c
> +++ b/arch/x86/kernel/cpu/microcode/intel.c
> @@ -397,22 +397,6 @@ scan_microcode(void *data, size_t size,
>   	return patch;
>   }
>   
> -/*
> - * Save this microcode patch. It will be loaded early when a CPU is
> - * hot-added or resumes.
> - */
> -static void save_mc_for_early(struct ucode_cpu_info *uci, u8 *mc, unsigned int size)
> -{
> -	/* Synchronization during CPU hotplug. */
> -	static DEFINE_MUTEX(x86_cpu_microcode_mutex);
> -
> -	mutex_lock(&x86_cpu_microcode_mutex);
> -
> -	save_microcode_patch(uci, mc, size);
> -
> -	mutex_unlock(&x86_cpu_microcode_mutex);
> -}
> -


So this function is ultimately invoked from a write to the 'reload' 
sysfs file, what about 2 racing writes to this file ? Though if it's 
about cpu hotplug then in reload_store() this is handled via 
cpus_read_lock I guess so indeed this is correct.