Re: [PATCH] user_intr: a lock to protect main_intr_queue

[Junling Ma]

In the next patch, we disallow deliver_user_intr from touching the queue, so we 
do not need to lock it. In fact, by doing that, we prevent interrupts from 
pulling the rug under our feet, so there is no need to do splhigh/splx in 
irq_acknowledge and intr_thread.

Junling

> On Aug 4, 2020, at 4:58 PM, Samuel Thibault <samuel.thibault@gnu.org> wrote:
> 
> Oh, btw, in queue_intr you also need to take the simple_lock which
> modifying the entry values.
> 
> Samuel
>