Re: RTEMS Driver Interface for Microwindows --- Proposal

["Rosimildo DaSilva" <rdasilva@connecttel.com>]

From: Eric Norum <eric@cls.usask.ca>
To: Rosimildo DaSilva <rdasilva@connecttel.com>; rtems-snapshots@oarcorp.com
<rtems-snapshots@oarcorp.com>
Date: Monday, December 20, 1999 12:59 PM
Subject: Re: RTEMS Driver Interface for Microwindows --- Proposal


>Rosimildo DaSilva wrote:
>>
>> No problem. I'll wait for suggestions from Eric.
>
>I've posted my comments to the snapshot group.  Have you seen them?

No, I did not. Could you forward it to me ?
I do not seem to be registered with that list.

Joel, is it possible to get me up on that list ?


>> I have another question to him:
>>
>> 1. Why cfgetispeed(), cfsetispeed() are implemented differently of
>> cfgetospeed(), cfsetospeed() on RTEMS ?.
>> Looks like that  cfgetispeed(), cfsetispeed() are broken.
>> I have made them exactly as the output version, and got things working
ok.
>>
>
>I expect that the ispeed routines are just as they were when I lifted
>them from the FreeBSD distribution.  What is wrong with them?  I just
>had a look and could find nothing wrong.   I don't know if any serial
>port drivers actually use the CIBAUD bits, but the
>cfgetispeed/cfsetispeed routines seem correct.

Weird values are passed to the drivers when you try to set the baud rate
using them. If I select B1200, and i put a log in the driver, I see  the
9600 bps been
been set on the driver level. I do not know whether it did not take the
parameter
( B9600 is default ), or just got it wrong.
When I switched to ospeed routines, everything worked fine.

Looking at conSetAttr() routine at the driver level( /dev/console ),
I've got the impression that the ospeed() ones match with what the
drivers expects.

>
>--
>Eric Norum                                 eric@cls.usask.ca
>Canadian Light Source                      Phone: (306) 966-6308
>University of Saskatchewan                 FAX:   (306) 966-6058
>Saskatoon, Canada.
>