Re: [PATCH V2] [tty] Fix possible race in n_tty_read()

[Stanislav Kozina <skozina@redhat.com>]

I believe that BUG_ON is different issue.

I'm not sure what's the correct locking scheme for tty->read_buf. I 
guess that it's allocated as part of tty initialization, and only 
removed during n_tty_close(), correct? So we should not need any lock there.
Perhaps we should change that BUG_ON(!tty->read_buf); to just return 
-EAGAIN, as e.g. n_tty_receive_buf does?

         if (!tty->read_buf)
                 return;

I will submit a separate patch changing the BUG_ON to return, I believe 
it's still better to drop a n_tty_read() rather than panic whole system.
But I still don't understand how this can happen - that we have a tty 
operational without a buffer allocated. There still must be some other race.

-Stanislav

> Should this also be removing the BUG_ON check you noted in the other
> email was not valid now ?
>
> Alan