[syslinux] CPU usage while at syslinux boot prompt

[H. Peter Anvin]

Ferenc Wagner wrote:
> "H. Peter Anvin" <hpa at zytor.com> writes:
> 
>> The *only* device that we have to talk to that we can guarantee
>> getting an interrupt for is the keyboard, the rest of the devices
>> *have* to be handled by polling, via an idle loop.
>> [...]
>> I guess I might put a HLT in *as an option* (because people *will*
>> complain that the serial port loses characters)
> 
> I get this very annoying dropout from the Xen hypervisor, and wonder:
> is this a problem of common serial hardware not supporting interrupt
> generation or a deeper theoretical one?  Can't you do reliable
> interrupt driver serial IO on common UARTs by principle?  Doesn't the
> Linux kernel do this?

It is extremely unsafe to use interrupts for serial ports while the 
hardware is still in BIOS mode.

	-hpa