Netbooks & BSD

[David Brodbeck]

On Wed, Oct 20, 2010 at 4:19 PM, Chip Camden
<sterling at camdensoftware.com> wrote:
> Quoth David Brodbeck on Wednesday, 20 October 2010:
>
>> Now, the USB keyboard protocol...ugh, they really dropped the ball on
>> that one.  It's standardized, which is good, but it's a polling
>> interface and tends to occasionally lose events under high CPU load,
>> which is bad.  Especially if it's a key-up event that gets lost.
>
> Ugh.
>
> The silver lining: that explains a lot of omissions I was beginning to blame on
> senility.

I first noticed it on an underpowered Linux box, where scrolling an
xterm took basically 100% CPU.  When using a USB keyboard, holding
down ENTER in the xterm would result in an endless stream of ENTERs
because the CPU would become too loaded to register the key being
released.

I don't see it as much now -- my computers are faster -- but I still
see it occasionally.

The IPMI serial-over-LAN console interface seems to have the same
problem in spades.  I have to really limit my typing speed when I'm
controlling a system that way or nearly every other character
disappears.