Re: [Qemu-devel] [PATCH 2/5] qemu_coroutine_self should not be marked coroutine_fn as it cannot yield

[Charlie Shepherd <charlie@ctshepherd.com>]

On 07/08/2013 20:18, Stefan Hajnoczi wrote:
> On Mon, Aug 05, 2013 at 08:44:04PM +0200, Charlie Shepherd wrote:
>> From: Charlie Shepherd <cs648@cam.ac.uk>
>>
>> While it only really makes sense to call qemu_coroutine_self() in a coroutine
>> context, it cannot actually yield execution itself, so remove the coroutine_fn
>> annotation.
>> ---
>>   include/block/coroutine.h | 2 +-
>>   1 file changed, 1 insertion(+), 1 deletion(-)
> By removing coroutine_fn the rules have changed:
>
> It's now legal to call qemu_coroutine_self() outside a coroutine.
> Previously only callers that knew the internals of the coroutine
> implementation did that.

Yes, I agree that it's probably not helpful to loosen the restrictions 
on when its possible to call qemu_coroutine_self().

> coroutine_fn gives coroutine backend implementors more flexibility in
> how they choose to implement qemu_coroutine_self().  From an API
> perspective I prefer to keep qemu_coroutine_self() marked as a
> coroutine_fn.
>
> I guess the practical problem is that CPC will get
> upset that it's being called by the coroutine implementation from
> non-coroutine contexts.  This can be solved:
>
> 1. Keep the public qemu_coroutine_self() marked coroutine_fn.
>
> 2. Have it call an internal coroutine_self() function that is not
>     coroutine_fn.
>
> This way the API stays strict and the internal implementation doesn't
> violate the coroutine_fn calling rule.

This is a good solution, I'll add a patch implementing it to the series.


Charlie