Re: [PATCH RT 1/2 v2] x86: Do not disable preemption in int3 on32bit

[Steven Rostedt]

On Thu, 2012-01-26 at 22:59 +0900, Masami Hiramatsu wrote:

> Better. If the functions are only for signaling, how
> about using the "signal" in name? :)
> 
> conditional_sti/cli_for_signal()

I don't think they are only for signals. The conditional_sti/cli is for
all callers of do_trap(). But those callers that run on an IST stack use
the preempt_conditional_sti/cli() code. Perhaps we should call it:

conditional_sti/cli_ist() ?

Actually, I think this is the best name. The "preempt_" one, didn't give
any annotation about why it was disabling preemption. It was done
because the stack is on the IST and we can not schedule out. Thus, if we
add the "_ist()" to it, not only does it annotate why this call is
special, but also allows x86_32, which does not have an IST, not to have
to disable preemption.

-- Steve


--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/