Re: [PATCH printk v2 22/26] printk: nbcon: Implement emergency sections

[Petr Mladek]

On Sun 2024-02-18 20:03:22, John Ogness wrote:
> From: Thomas Gleixner <tglx@xxxxxxxxxxxxx>
> 
> In emergency situations (something has gone wrong but the
> system continues to operate), usually important information
> (such as a backtrace) is generated via printk(). Each
> individual printk record has little meaning. It is the
> collection of printk messages that is most often needed by
> developers and users.
> 
> In order to help ensure that the collection of printk messages
> in an emergency situation are all stored to the ringbuffer as
> quickly as possible, disable console output for that CPU while
> it is in the emergency situation. When exiting the emergency
> situation, trigger the consoles to be flushed.
> 
> Add per-CPU emergency nesting tracking because an emergency
> can arise while in an emergency situation.
> 
> Add functions to mark the beginning and end of emergency
> sections where the urgent messages are generated.
> 
> Do not print if the current CPU is in an emergency state.
> 
> Trigger console flushing when exiting all emergency nesting.
> 
> Note that the emergency state is not system-wide. While one CPU
> is in an emergency state, another CPU may continue to print
> console messages.
> 
> --- a/kernel/printk/printk.c
> +++ b/kernel/printk/printk.c
> @@ -2404,16 +2404,29 @@ asmlinkage int vprintk_emit(int facility, int level,
>  		 * printing of all remaining records to all consoles so that
>  		 * this context can return as soon as possible. Hopefully
>  		 * another printk() caller will take over the printing.
> +		 *
> +		 * Also, nbcon_get_default_prio() requires migration disabled.
>  		 */
>  		preempt_disable();
> +
>  		/*
> -		 * Try to acquire and then immediately release the console
> -		 * semaphore. The release will print out buffers. With the
> -		 * spinning variant, this context tries to take over the
> -		 * printing from another printing context.
> +		 * Do not emit for EMERGENCY priority. The console will be
> +		 * explicitly flushed when exiting the emergency section.
>  		 */
> -		if (console_trylock_spinning())
> -			console_unlock();
> +		if (nbcon_get_default_prio() == NBCON_PRIO_EMERGENCY) {
> +			do_trylock_unlock = false;

This would cause calling defer_console_output() in this printk().
I think that we do not want it. It is done later by
nbcon_cpu_emergency_exit(). I think that we want something like:


		/*
		 * Try to acquire and then immediately release the
		 * console semaphore. The release will print out
		 * buffers. With the spinning variant, this context
		 * tries to take over the printing from another
		 * printing context.
		 *
		 * Skip it in EMERGENCY priority. The console will be
		 * explicitly flushed when exiting the emergency section.
		 */
		if (nbcon_get_default_prio() != NBCON_PRIO_EMERGENCY) {
			if (console_trylock_spinning())
				console_unlock();
		}


> +		} else {
> +			/*
> +			 * Try to acquire and then immediately release the
> +			 * console semaphore. The release will print out
> +			 * buffers. With the spinning variant, this context
> +			 * tries to take over the printing from another
> +			 * printing context.
> +			 */
> +			if (console_trylock_spinning())
> +				console_unlock();
> +		}
> +
>  		preempt_enable();
>  	}

Otherwise, it looks good.

Best Regards,
Petr