[Cbe-oss-dev] [PATCH 2/2] [POWERPC] spufs: fix invalid scheduling of forgotten contexts

[Arnd Bergmann]

On Thursday 06 March 2008, Luke Browning wrote:

> This model is really just based on the way our library works.  Could we
> write a testcase where one thread calls spu_run and the spu stops
> waiting for some kind of input operation, while the second thread closes
> the file descriptor associated with the context.  Would this drive
> spu_forget while the controlling thread was in spu_run?

spu_run takes a reference on the file descriptor, so this test case should
result in spu_forget being delayed until we almost return from spu_run.

	Arnd <><