[Cbe-oss-dev] [RFC 5/5] spufs: SPU-AES support (kernel side)

[Arnd Bergmann]

On Tuesday 12 June 2007, Brian Watt wrote:
> Got a question about this in general...
> So, let's assume that the SPU code issues a stop and signal to do 
> something 
> like printf or fopen, what happens since there is nothing in user space 
> that initiated the execution of the SPU code? How is that handled?

It's not handled at all. The KSPU code is part of the kernel and runs
with full access to all memory, so the author of a KSPU program
must be very careful about what the program can or cannot do.

stop-and-signal calls are not part of the KSPU API at this point,
because AES does not need that.

	Arnd <><