Subject: Re: port-amd64/34203: kbc: cmd word write error
To: None <port-amd64-maintainer@netbsd.org, gnats-admin@netbsd.org,>
From: =?ISO-8859-1?Q?St=E9phane_Witzmann?= <stephane.witzmann@gmail.com>
List: netbsd-bugs
Date: 08/15/2006 12:40:02
The following reply was made to PR port-amd64/34203; it has been noted by GNATS.
From: "=?ISO-8859-1?Q?St=E9phane_Witzmann?=" <stephane.witzmann@gmail.com>
To: "Martin Husemann" <martin@duskware.de>
Cc: gnats-bugs@netbsd.org
Subject: Re: port-amd64/34203: kbc: cmd word write error
Date: Tue, 15 Aug 2006 14:39:01 +0200
On 8/15/06, Martin Husemann <martin@duskware.de> wrote:
> On Tue, Aug 15, 2006 at 11:35:01AM +0000, St=E9phane Witzmann wrote:
> > - If I add a panic() in main() before configure(), the keyboard works
> > in DDB. There are some pckbc functions called before the copyright is
> > printed, in consinit() I suppose, and none of them seems to fail.
>
> I'm not sure how delay works on amd64 at this early stage - the loop that
> fails does a KBD_DELAY, which is defined as
>
> #define KBD_DELAY DELAY(8)
>
> Can you try to increase this significantly (for testing only, of course),
> or change it to the lower case delay(...) variant?
Replaced 8 with 160, same problem. I also tried keeping the delay and
increasing the number of iterations in the loop (100000-> 10000000).
Same with delay().
> Maybe you could also try to compile this file (sys/dev/ic/pckbc.c) with -=
O0,
> to check for optimization problems.
Recompiled the whole kernel with -O0, no change.