Solution to various issues on resume resulting from WARN_ON in suspend_test_finish() ?

[TJ]

We've got various outstanding bug reports against Hardy, Intrepid and
Jaunty which are basically a false-positive WARN_ON() in
kernel/power/main.c::suspend_test_finish() which can cause bug noise in
the form of kernel Oops reports for hardware that is slow to resume -
usually hard disks - but is expected.

I wonder if we should consider slightly increasing TEST_SUSPEND_SECONDS
from its current 5 seconds?

kernel/power/main.c:147:#define TEST_SUSPEND_SECONDS    5
kernel/power/main.c:177:        WARN(msec > (TEST_SUSPEND_SECONDS * 1000), "Component: %s\n", label);
kernel/power/main.c:627:        rtc_time_to_tm(now + TEST_SUSPEND_SECONDS, &alm.time);