Why use FIFO IO directly

[jokemanfire]

I find this issus from
https://github.com/containerd/containerd/issues/9568

The community shim links child procress through copyIO. The shim and containerd have separate pipe links, so the shim can close the pipe when the exec process exit. r-shim use the FIFO-IO directly.

[jokemanfire]

@abel-von @mxpv @Burning1020
problem find in ctr v1.7 "TestContainerDrainExecIOAfterExit"

[Burning1020]

As mentioned here: #24 (comment)
We think directly IO could be efficient and simple.

[jokemanfire]

but the exec process will occupy the fifo io , so the containerd in execsync function will loop forver ,in test "TestContainerDrainExecIOAfterExit " .how can we resolve this problem . that's why I pull this pr #235 . How to report errors at the underlying runtime , the runc can't trace the daemon like "sleep 365d &" , and
runc didn't return the error . I hope we can resolve this problem , if it is runc's problem .do you have any suggestions?