Re: [PATCH v2 4/7] util: add qemu-co-timeout

[Hanna Reitz <hreitz@redhat.com>]

On 01.04.22 11:19, Vladimir Sementsov-Ogievskiy wrote:
> Add new API, to make a time limited call of the coroutine.
>
> Signed-off-by: Vladimir Sementsov-Ogievskiy <vsementsov@openvz.org>
> ---
>   include/qemu/coroutine.h | 13 ++++++
>   util/meson.build         |  1 +
>   util/qemu-co-timeout.c   | 89 ++++++++++++++++++++++++++++++++++++++++
>   3 files changed, 103 insertions(+)
>   create mode 100644 util/qemu-co-timeout.c

I don’t really understand what this does.  Intuitively, I would have 
assumed this makes the first yield of the respective coroutine not 
return when the timeout has elapsed, and instead, we switch back to 
qemu_co_timeout(), which returns to its callers.

But it looks like when this yield happens and we switch back to 
qemu_co_timeout(), the coroutine actually isn’t cancelled, and will just 
continue running, actually.  Is that right?  On first look, this looks 
like it’ll be quite difficult to think about what happens when this is 
used, because the coroutine in question is no longer run in sequence 
with its caller, but actually might run in parallel (even though it’s 
still a coroutine, so it’ll remain cooperative multitasking).