I've tested the following kernels:
unmodified - as shipped (recompile for control purposes) reverted - Linus's patch reverted (so using > TASK_STOPPED) removed - That check removed entirely.
The results (and what I expected) are as follows:
Kernel | Result | Expected unmodified | FAIL | FAIL reverted | PASS | FAIL removed | PASS | PASS
I'm rather surprised; it appears that reverting Linus's patch actually fixes the problem; I do not understand why yet.
I am recompiling again to make damned sure I got the test right, and didn't copy over the top of it, or something.
I've tested the following kernels:
unmodified - as shipped (recompile for control purposes)
reverted - Linus's patch reverted (so using > TASK_STOPPED)
removed - That check removed entirely.
The results (and what I expected) are as follows:
Kernel | Result | Expected
unmodified | FAIL | FAIL
reverted | PASS | FAIL
removed | PASS | PASS
I'm rather surprised; it appears that reverting Linus's patch actually fixes the problem; I do not understand why yet.
I am recompiling again to make damned sure I got the test right, and didn't copy over the top of it, or something.