Re: [PATCH 2/2] ACPI, PCI, ISA: Fix memory leak when there is no IRQ in the ACPI subsystem.

["Rafael J. Wysocki" <rjw@rjwysocki.net>]

On Tuesday, February 18, 2014 11:21:39 AM Tomasz Nowicki wrote:
> On 18.02.2014 02:02, Rafael J. Wysocki wrote:
> > On Monday, February 10, 2014 02:00:11 PM Tomasz Nowicki wrote:
> >> Whenever we register ISA interrupt or not, we need to free the IRQ routing
> >> table entry.
> >>
> >> Signed-off-by: Tomasz Nowicki <tomasz.nowicki@linaro.org>
> >> ---
> >>   drivers/acpi/pci_irq.c |    1 +
> >>   1 file changed, 1 insertion(+)
> >>
> >> diff --git a/drivers/acpi/pci_irq.c b/drivers/acpi/pci_irq.c
> >> index b0e31b6..6ec0f36 100644
> >> --- a/drivers/acpi/pci_irq.c
> >> +++ b/drivers/acpi/pci_irq.c
> >> @@ -432,6 +432,7 @@ int acpi_pci_irq_enable(struct pci_dev *dev)
> >>   #endif
> >>   		dev_warn(&dev->dev, "PCI INT %c: no GSI\n", pin_name(pin));
> >>
> >> +		kfree(entry);
> >
> > If I'm not mistaken, entry is always NULL here, isn't it?
>
> acpi_pci_link_allocate_irq() can return negative gsi even if entry != 
> NULL. For that case we'd have memory leak.

And that's what your changelog should be saying, isn't it?

Queued up for 3.14 and marked for -stable.  I had to rebase it, though, so
please check the result in linux-pm.git/linux-next.

Thanks!

-- 
I speak only for myself.
Rafael J. Wysocki, Intel Open Source Technology Center.