On Tue, 19 Jun 2012 20:47:04 +0200, Jean Delvare wrote:
> On Fri, 6 Jan 2012 18:58:21 +0800, Daniel Kurtz wrote:
> > @@ -879,8 +953,24 @@ static int __devinit i801_probe(struct pci_dev *dev,
> > i801_probe_optional_slaves(priv);
> >
> > pci_set_drvdata(dev, priv);
> > +
> > + if (priv->features & FEATURE_IRQ) {
> > + init_waitqueue_head(&priv->waitq);
> > + spin_lock_init(&priv->lock);
> > +
> > + err = request_irq(dev->irq, i801_isr, IRQF_SHARED,
> > + i801_driver.name, priv);
> > + if (err) {
> > + dev_err(&dev->dev, "Failed to allocate irq %d: %d",
>
> Missing "\n".
>
> > + dev->irq, err);
> > + goto exit_del_adapter;
> > + }
>
> I believe order is wrong, and interrupt handler should be installed
> _before_ registering the adapter. Otherwise you have a race condition
> where the handler could be called before the waitqueue and spinlock are
> initialized.
Oh, and it's not a theoretical thing. I can reproducibly crash the
kernel by removing and loading the i2c-i801 driver again. Swapping
request_irq() and i2c_add_adapter() solves it.
--
Jean Delvare
--
To unsubscribe from this list: send the line "unsubscribe linux-i2c" in
the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html
