Re: [PATCH] ACPI / memhotplug: Fix a stale pointer in error path

["Rafael J. Wysocki" <rjw@sisk.pl>]

On Friday, July 12, 2013 04:28:36 PM Toshi Kani wrote:
> On Fri, 2013-07-12 at 23:40 +0200, Rafael J. Wysocki wrote:
> > On Friday, July 12, 2013 03:12:24 PM Toshi Kani wrote:
> > > On Fri, 2013-07-12 at 23:13 +0200, Rafael J. Wysocki wrote:
> > > > On Friday, July 12, 2013 03:01:15 PM Toshi Kani wrote:
> > > > > On Fri, 2013-07-12 at 22:42 +0200, Rafael J. Wysocki wrote:
> > > > > > On Friday, July 12, 2013 08:51:29 AM Toshi Kani wrote:
> > > > > > > On Fri, 2013-07-12 at 09:24 +0900, Yasuaki Ishimatsu wrote:
> > > > > > > > (2013/07/11 1:47), Toshi Kani wrote:
> > > > > > > > > device->driver_data needs to be cleared when releasing its data,
> > > > > > > > > mem_device, in an error path of acpi_memory_device_add().
> > > > > > > > > 
> > > > > > > > > Signed-off-by: Toshi Kani <toshi.kani@hp.com>
> > > > > > > > > ---
> > > > > > > > 
> > > > > > > > Reviewed-by: Yasuaki Ishimatsu <isimatu.yasuaki@jp.fujitsu.com>
> > > > > > > 
> > > > > > > Thanks Yasuaki!
> > > > > > 
> > > > > > Queued up as a fix for 3.11.
> > > > > 
> > > > > Thanks!
> > > > > 
> > > > > > Do we need that in -stable as well?
> > > > > 
> > > > > Good point.  Yes, we need that in -stable as well.
> > > > 
> > > > What's the oldest mainline major release that fix is applicable to?
> > > 
> > > The fix is applicable all ways up to 2.6.32.
> > 
> > For -stable I'll need to say some more about what practical consequences of
> > the bug are.  Is it difficult to trigger?
> 
> The function evaluates _CRS of memory device objects, and fails when it
> gets an unexpected resource or cannot allocate a memory.

OK, so this is essentially about surviving unexpected external input, which
I suppose is serious enough.

> A kernel crash
> or data corruption may occur when the kernel accessed a stale pointer.
> That said, I am not sure how critical this issue is for old kernels
> since I do not think there are many platforms that support memory
> hotplug today.

Which doesn't matter.  People may want to run 3.10.y on future hardware too.

> After reading the recent -stable discussion on LKML, now
> I am not sure if this fix should be applied for -stable.

Well, I don't necessarily agree with some things being said there.  I guess
I'll need to say something in that thread. :-)

> I instrumented the kernel to generate an error for testing this change.

OK

Thanks,
Rafael


> > > > > > > > >   drivers/acpi/acpi_memhotplug.c |    1 +
> > > > > > > > >   1 file changed, 1 insertion(+)
> > > > > > > > > 
> > > > > > > > > diff --git a/drivers/acpi/acpi_memhotplug.c b/drivers/acpi/acpi_memhotplug.c
> > > > > > > > > index c711d11..999adb5 100644
> > > > > > > > > --- a/drivers/acpi/acpi_memhotplug.c
> > > > > > > > > +++ b/drivers/acpi/acpi_memhotplug.c
> > > > > > > > > @@ -323,6 +323,7 @@ static int acpi_memory_device_add(struct acpi_device *device,
> > > > > > > > >   	/* Get the range from the _CRS */
> > > > > > > > >   	result = acpi_memory_get_device_resources(mem_device);
> > > > > > > > >   	if (result) {
> > > > > > > > > +		device->driver_data = NULL;
> > > > > > > > >   		kfree(mem_device);
> > > > > > > > >   		return result;
> > > > > > > > >   	}
> > > > > > > > > --
> > > > > > > > > To unsubscribe from this list: send the line "unsubscribe linux-acpi" in
> > > > > > > > > the body of a message to majordomo@vger.kernel.org
> > > > > > > > > More majordomo info at  http://vger.kernel.org/majordomo-info.html
> > > > > > > > > 
> > > > > > > > 
> > > > > > > > 
> > > > > > > 
> > > > > > > 
> > > > > 
> > > > > 
> > > 
> > > 
> 
> 
> --
> To unsubscribe from this list: send the line "unsubscribe linux-acpi" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html
-- 
I speak only for myself.
Rafael J. Wysocki, Intel Open Source Technology Center.