Re: [PATCH 1/2] serial: 8250: Initialize pointer to irq_info

[Alan Cox <alan@lxorguk.ukuu.org.uk>]

> > Use a newer compiler - modern ones get this right.
> 
> : static void serial_unlink_irq_chain(struct uart_8250_port *up)
> : {
> : 	struct irq_info *i;
> : 	struct hlist_node *n;
> : 	struct hlist_head *h;
> : 
> : 	mutex_lock(&hash_mutex);
> : 
> : 	h = &irq_lists[up->port.irq % NR_IRQ_HASH];
> : 
> : 	hlist_for_each(n, h) {
> : 		i = hlist_entry(n, struct irq_info, node);
> : 		if (i->irq == up->port.irq)
> : 			break;
> : 	}
> : 
> : 	BUG_ON(n == NULL);
> : 	BUG_ON(i->head == NULL);
> 
> How can any compiler possibly determine that the hlist_for_each() is
> never executed zero times?

We had this argument was it two years ago ?

Modern gcc is capable of at least working out it can't work it out and
doesn't emit a spurious warning.