On Wednesday 10 October 2012 11:19:09 Ming Lei wrote:
> On Tue, Oct 9, 2012 at 4:47 PM, Oliver Neukum <[email protected]> wrote:
> >
> > Using GFP_KERNEL you preclude using those in resume() and error handling.
> > Please pass a gfp_t parameter.
>
> IMO, it is not a big deal because generally only several bytes are to be
> allocated inside these helpers.
Any allocation can do it, if the VM layer decides to block.
> If you still think the problem should be considered, another candidate fix
> is to take GFP_NOIO during system suspend/resume, and take GFP_KERNEL
> in other situations.
No, the problem is autoresume.
Suppose we have a device with two interface. Interface A be usbnet; interface B
something you page on. Now consider that you can only resume both interfaces
and this is (and needs to be) done synchronously.
Now we can have this code path:
autoresume of device -> resume() -> kmalloc(..., GFP_KERNEL) ->
VM layer decides to start paging out -> IO to interface B -> autoresume of
device
--> DEADLOCK
We need to use GFP_NOIO in situations the helper cannot know about.
Please add a gfp_t parameter. Then the caller will solve that.
Regards
Oliver
--
To unsubscribe from this list: send the line "unsubscribe linux-usb" in
the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html
