Re: [linux-fbdev] Re: console spin_lock

[Petr Vandrovec]

On 18 Jan 01 at 0:49, Andrew Morton wrote:

> Assumption:
> - Once the system is up and running, it's always safe to
>   call down() when in_interrupt() returns false - probably
>   not the case in parts of the exit path - tough.
> 
> Anyway, that's the thoughtware.  Sound sane?

Do not forget to handle printk() done by fbdev driver... It
may invoke printk() from user context, but with console_semaphore
already held... Something like reentrant_semaphore? Also, we
should declare which console/fbdev function can printk/can schedule 
and which must not, as using interrupts & schedule could yield
CPU to other tasks when (hardware assisted) operation is performed.
Some of them (clear, bmove) can take loong time to finish.
                                    Best regards,
                                            Petr Vandrovec
                                            [e-mail]