On 4/13/22 17:55, Eric Blake wrote:
-g_assert(qemu_in_coroutine());
Why is this assert dropped? Is it because we've marked the function
with coroutine_fn? If so, should we drop it earlier in the series,
when you added the label?
The label doesn't guarantee much, but in this case it's pre
On Tue, Apr 12, 2022 at 09:42:01PM +0200, Paolo Bonzini wrote:
> The condition for waiting on the s->free_sema queue depends on
> both s->in_flight and s->state. The latter is currently using
> atomics, but this is quite dubious and probably wrong.
>
> Because s->state is written in the main thre
On 4/12/22 21:42, Paolo Bonzini wrote:
Because s->state is written in the main thread too, for example by
the reconnect timer callback, it cannot be protected by a CoMutex.
Actually by the yank callback, not the timer (which runs in the "right"
AioContext).
Paolo
The condition for waiting on the s->free_sema queue depends on
both s->in_flight and s->state. The latter is currently using
atomics, but this is quite dubious and probably wrong.
Because s->state is written in the main thread too, for example by
the reconnect timer callback, it cannot be protect
