Re: [PATCH printk v4 14/15] printk: extend console_lock for proper kthread support

[Petr Mladek <pmladek@suse.com>]

On Thu 2022-04-21 23:28:49, John Ogness wrote:
> Currently threaded console printers synchronize against each
> other using console_lock(). However, different console drivers
> are unrelated and do not require any synchronization between
> each other. Removing the synchronization between the threaded
> console printers will allow each console to print at its own
> speed.
> 
> But the threaded consoles printers do still need to synchronize
> against console_lock() callers. Introduce a per-console mutex
> and a new console flag CON_THD_BLOCKED to provide this
> synchronization.
> 
> console_lock() is modified so that it must acquire the mutex
> of each console in order to set the CON_THD_BLOCKED flag.
> Console printing threads will acquire their mutex while
> printing a record. If CON_THD_BLOCKED was set, the thread will
> go back to sleep instead of printing.
> 
> The reason for the CON_THD_BLOCKED flag is so that
> console_lock() callers do not need to acquire multiple console
> mutexes simultaneously, which would introduce unnecessary
> complexity due to nested mutex locking.
> 
> Threaded console printers also need to synchronize against
> console_trylock() callers. Since console_trylock() may be
> called from any context, the per-console mutex cannot be used
> for this synchronization. (mutex_trylock() cannot be called
> from atomic contexts.) Introduce a global atomic counter to
> identify if any threaded printers are active. The threaded
> printers will also check the atomic counter to identify if the
> console has been locked by another task via console_trylock().
> 
> Note that @console_sem is still used to provide synchronization
> between console_lock() and console_trylock() callers.
> 
> A locking overview for console_lock(), console_trylock(), and the
> threaded printers is as follows (pseudo code):
> 
> console_lock()
> {
>         down(&console_sem);
>         for_each_console(con) {
>                 mutex_lock(&con->lock);
>                 con->flags |= CON_THD_BLOCKED;
>                 mutex_unlock(&con->lock);
>         }
>         /* console_lock acquired */
> }
> 
> console_trylock()
> {
>         if (down_trylock(&console_sem) == 0) {
>                 if (atomic_cmpxchg(&console_kthreads_active, 0, -1) == 0) {
>                         /* console_lock acquired */
>                 }
>         }
> }
> 
> threaded_printer()
> {
>         mutex_lock(&con->lock);
>         if (!(con->flags & CON_THD_BLOCKED)) {
> 		/* console_lock() callers blocked */
> 
>                 if (atomic_inc_unless_negative(&console_kthreads_active)) {
>                         /* console_trylock() callers blocked */
> 
>                         con->write();
> 
>                         atomic_dec(&console_lock_count);
>                 }
>         }
>         mutex_unlock(&con->lock);
> }
> 
> The console owner and waiter logic now only applies between contexts
> that have taken the console_lock via console_trylock(). Threaded
> printers never take the console_lock, so they do not have a
> console_lock to handover. Tasks that have used console_lock() will
> block the threaded printers using a mutex and if the console_lock
> is handed over to an atomic context, it would be unable to unblock
> the threaded printers. However, the console_trylock() case is
> really the only scenario that is interesting for handovers anyway.
> 
> @panic_console_dropped must change to atomic_t since it is no longer
> protected exclusively by the console_lock.

I have finally understood why console_lock_single_hold() solved
the problem with con->flags. We should describe it in the commit
message as well. Something like:

    @con->flags must be updated WRITE_ONCE() under console_lock
    when the related kthread is running. The kthread printers
    read the flags only under con->mutex. They have to see
    the CON_THD_BLOCKED flag when the value might not
    be consistent.

Sigh, I agree that this approach is error prone and kind of ugly.
The approach with console_lock_single_hold() was not ideal either.
I still have to think about it.

Anyway, the approach with READ_ONCE()/WRITE_ONCE() looks good enough
for now.


> Since threaded printers remain asleep if they see that the console
> is locked, they now must be explicitly woken in __console_unlock().
> This means wake_up_klogd() calls following a console_unlock() are
> no longer necessary and are removed.
> 
> Also note that threaded printers no longer need to check
> @console_suspended. The check for the CON_THD_BLOCKED flag
> implicitly covers the suspended console case.
> 
> Signed-off-by: John Ogness <john.ogness@linutronix.de>

I do not see any further problem with this patch. So, with
the updated commit message and comment above the macros:

Reviewed-by: Petr Mladek <pmladek@suse.com>

Best Regards,
Petr