Re: [Qemu-devel] [PATCH] Port E9 hack, for debugging purposes

[Thomas Petazzoni <thomas.petazzoni@enix.org>]

[-- Attachment #1: Type: text/plain, Size: 1737 bytes --]

Hello,

James Mastros a écrit :
> I'm a bit of a newbie here, so perhaps this has already been covered, 
> though I haven't seen it in the archives -- but why port E9, and not 
> port 80?  80 seems to be the standard port for this sort of thing. 
> (Though output seems to normally be two-digit-hex, and not ASCII-based.)
> 
> (Port E9 seems to be largely undefined, but marked as "reserved" in most 
> lists I've seen (or just not listed), so it seems possible, but 
> unlikely, that things will think they know what should be connected 
> there.  OTOH, 80 is reserved for exactly what we're doing with it.  On 
> the third hand, the Linux kernel uses writes to port 0x80 to "slow down" 
> IO.  I'm not sure what the interactions would be with this.)

Mostly because this is the way Bochs implements it, and I thought it 
would be nice to have it working exactly the same way on Qemu. This 
feature is particularly interesting for people making their own toy OS. 
It allows to have simple debugging possibilities without taking care of 
writing a serial driver or a full-featured console driver.

The Ralf's Brown Port List says that 0x80 is when writing a 
"Manufacturing Diagnostics Port", but it doesn't say anything about reading.

Anyway, if Linux uses it to slow down I/O, this is not a good candidate 
in my opinion for the E9-style hack.

To sum up, I don't have any valid reason for choosing E9 instead of any 
other, except that it's implemented that way in Bochs.

Thomas
-- 
PETAZZONI Thomas - thomas.petazzoni@enix.org
http://thomas.enix.org - Jabber: thomas.petazzoni@jabber.dk
http://kos.enix.org, http://sos.enix.org
Fingerprint : 0BE1 4CF3 CEA4 AC9D CC6E  1624 F653 CB30 98D3 F7A7


[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 256 bytes --]