Re: [edk2-devel] [PATCH 06/13] OvmfPkg/Tcg2PhysicalPresenceLib: Use pcd for user response wait time

["Laszlo Ersek" <lersek@redhat.com>]

On 01/19/20 08:03, Gao, Zhichao wrote:
> Hi Laszlo,
> 
> Sorry for misunderstanding the logic in read key function. That makes the incorrect analysis.
> The read key function is waiting for "specific" key press not a key press. We should drop all the other pressing-key.
> That means we should create a hotkey service like BDS one. It is not simple. I think that is why the original use the busy loop.
> I don't think it makes sense to port the function from BDS for such a simple function. If there is a requirement for the hot key service to be public and common, it makes sense.
> So I prefer to keep the original busy loop instead of porting hot key service.

I don't understand why the fact that we're interested in a specific key
only, changes *how* we should wait for keys. In both approaches (wait
for event, vs. busy wait), we should consume and drop keys until we find
what we need. Yes, the wait-for-event approach needs another loop around
it, but that still doesn't make it a "busy" loop -- it only runs when
there's an event to process (keypress or timeout). This is a standard
event loop mechanism; block/sleep in the loop body until something
occurs, then process the event; lather, rinse, repeat.

Anyway, I don't feel strongly about this. If your original approach is
accepted for SecurityPkg, I can live with it in OvmfPkg too.

Thanks
Laszlo