ion: ion_page_pool: Check for fatal signals from OOM killer

The OOM killer sets tsk_is_oom_victim() which allows access to memory
reserves. These reserves are intended for use by processes which need to
allocate a small amount of memory for a short time, such that they can
free their caches/buffers/etc. Ion allocations are not used for this
usecase.

Exiting immediately allows handling the fatal signal to occur that much
sooner.

This patch does not attempt to fix a similar issue in handling of the
prefetch ioctl - Since prefetch is handled by a kernel thread, it will
never be killed by the OOM killer.

Change-Id: I9dca65920229026dfacd688cf10470504d1957ed
Signed-off-by: Patrick Daly <pdaly@codeaurora.org>