Re: [PATCH 4/5] atmel_serial: Split the interrupt handler

[Haavard Skinnemoen]

On Wed, 19 Dec 2007 00:15:24 +0100
Jiri Slaby <jirislaby@xxxxxxxxx> wrote:

> On 12/18/2007 06:06 PM, Haavard Skinnemoen wrote:
> >  	port = &atmel_ports[pdev->id];
> >  	atmel_init_port(port, pdev);
> >  
> > +	data = kmalloc(ATMEL_SERIAL_RINGSIZE, GFP_KERNEL);
> > +	if (!data)
> 
> {
>         clk_disable(atmel_port->clk);
>         clk_put(atmel_port->clk);

Indeed, thanks.

Hmm...the existing error path gets this wrong too. It's actually a
bit risky to disable the clock here since we might end up losing the
console.

I wonder what we're really supposed to do if the console is initialized
successfully, but the corresponding device fails to probe...?

> > @@ -1022,6 +1144,7 @@ static int __devexit atmel_serial_remove(struct platform_device *pdev)
> >  
> >  	if (port) {
> >  		ret = uart_remove_one_port(&atmel_uart, port);
> 
> Don't you need tasklet_kill() here (or somewhere)?

Absolutely. Thanks again.

> > +		kfree(atmel_port->rx_ring.buf);
> >  		kfree(port);

Hmm...this actually looks like a bug too. "port" is allocated
statically, so it shouldn't be freed. I wonder if anyone ever use this
driver as a module?

Haavard
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/