Re: [PATCH 5/8] printk: Try to register each console as Braille first

[Chris Down]

Petr Mladek writes:
> @@ -3918,6 +3922,12 @@ static int try_enable_preferred_console(struct console *newcon,
> 		    newcon->match(newcon, pc->name, pc->index, pc->options) != 0) {
> 			/* default matching */
> 			BUILD_BUG_ON(sizeof(pc->name) != sizeof(newcon->name));
> +			/*
> +			 * Two entries might have the same pc->name when one was
> +			 * defined via "devname".
> +			 */
> +			if (try_only_braille && !is_braille_console_preferred(pc))
> +				continue;
> 			if (strcmp(pc->name, newcon->name) != 0)
> 				continue;
> 			if (newcon->index >= 0 &&
> @@ -3926,7 +3936,7 @@ static int try_enable_preferred_console(struct console *newcon,
> 			if (newcon->index < 0)
> 				newcon->index = pc->index;
>
> -			if (is_braille_console_preferred(pc))
> +			if (try_only_braille)
> 				return _braille_register_console(newcon, pc);
>
> 			err = console_call_setup(newcon, pc->options);

This doesn't look right to me. By putting the Braille dispatch inside the if 
block, you bypass it whenever newcon->match matches.

So that means if a console driver matches, the execution skips this 
default matching block completely, falls through, and silently sets CON_ENABLED 
without ever registering it as a Braille console or adding it to console_list, 
and the console is silently lost.

The try_only_braille and is_braille_console_preferred(pc) checks likely need to 
happen before or independently of the match() vs. default matching branch.