Re: [PATCH] synchronize_irq needs a barrier

[Benjamin Herrenschmidt]

> I have read this thread and I concluded few things:
> 
> 1) It is impossible to know that the card won't send more interrupts:
> Even if I do a read from the device, the IRQ can be pending in the bus/APIC
> It is even possible (and likely) that the IRQ line will be shared, thus the 
> handler can be called by non-relevant device.
> 
> 2) the synchronize_irq(); in .suspend is useless:
> an IRQ can happen immediately after this synchronize_irq();
> and interrupt even the .suspend()
> (At least theoretically)

It's not totally useless not. It guarantees that by the time your
are out of your suspend(), a simultaneous instance of the IRQ handler
is either finished, or it (or any subsequent occurence) have seen
your insuspend flag set to 1.

> Thus I now understand that .suspend() should do:
> 
> 	saa_writel(SAA7134_IRQ1, 0);
> 	saa_writel(SAA7134_IRQ2, 0);
> 	saa_writel(SAA7134_MAIN_CTRL, 0);
> 
> 	dev->insuspend = 1;
> 	smp_wmb();
> 
> 	/* at that point the _request to disable card's IRQs was issued, we don't know 
> 	   that there will be no irqs anymore.
> 	   the smp_mb(); guaranties that the IRQ handler will bail out in that case. */
> 	
> 	/* .......*/
> 
> 	pci_save_state(pci_dev);
> 	pci_set_power_state(pci_dev, pci_choose_state(pci_dev, state));
> 	return 0;

The above doesn't handle the case where the IRQ handle just passed the
if (insuspend) test. You may end up calling pci_set_power_state() while
in the middle of some further HW accesses in the handler.

You still need synchronize_irq() for that reason. Or use a spinlock.

> and the interrupt handler:
> 
> 	smp_rmb();
> 	if (dev->insuspend)
> 		goto out;
> 
> Am I right?

Not quite :-)

Also not that the work we are doing with synchronize_irq, if it gets
merged, renders it unnecessary for you to add those two memory barriers,
synchronize_irq will pretty much do the ordering for you.

Cheers,
Ben.

-
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/