gnupic: Thread: gpsim CVS update


[<<] [<] Page 1 of 1 [>] [>>]
Subject: gpsim CVS update
From: Scott Dattalo ####@####.####
Date: 15 Mar 2000 13:52:05 -0000
Message-Id: <Pine.LNX.4.21.0003150711370.4589-100000@tempest2.blackhat.net>

FYI,

After taking a week off from gpsim, I jumped back in and added support for USART
in the 14-bit core. It's not fully functional, but here's what's supported:

asynchronous transfers
9-bit mode
Interrupt on rx and tx
mapping to I/O pins


Right now, only the 40-pin versions of the 14-bit core devices have the
uart. This includes the 16c65 (not the c64 - it doesn't support the uart in the
real device), 16c74, 16f874, and 16f877.

As a simple test, I wrote a routine called usart_14.asm (in cvs
examples/14bit) that reads and writes to and from the uart. If you create a node
that loops tx back to rx, you can actually receive the bytes that are
transmitted - just like a real rs-232 loop back!

There were some fixes to the 10-bit a/d converter in the 16f877 as well. Now,
the ADFM bit properly selects the format (the way the 10-bit value is mapped
into two 8-bit registers). In addition, the analog scaling works properly
now. Consequently, analog stimuli (which are expressed in terms of voltages) are
scaled such that if the A/D converter is 8 or 10 bits the digital value is
correct.


I had to make some significant changes to the stimulus infrastructure too.
Partially because it has become somewhat broken due to neglect and partially
because the analog portion was just plain broken.


Ralf has just started the program memory browser that will you to modify the
program memory much in the same way you can modify the register memory. 

Scott


Subject: gpsim CVS update
From: Scott Dattalo ####@####.####
Date: 27 Mar 2000 00:35:57 -0000
Message-Id: <Pine.LNX.4.21.0003261814370.12539-100000@tempest2.blackhat.net>

gpsim CVS:
https://sourceforge.net/cvs/?group_id=2341

There's not a whole lot to report...

The usart rx sampling was completely wrong. So I fixed this and added usart rx
interrupts as well. While I haven't tested every baud rate, I do believe that
the asynchonous mode of the usart is completed. 

I added a new file 

examples/14bit/usart.stc

It looks like so:


load s usart_14.cod

# Stimuli:
# All we do is tie rx and tx together to create a loop back.
# This will cause whatever is transmitted to be received...

node loop_back
attach loop_back portc6 portc7

-----------

The thing to notice is that you can tie two pins together using nodes. I use
this to create a simple rs232 loop back.

I then modified examples/14bit/usart_14.asm to repeatedly transmit the string
"0123456789ABCDEF". An interrupt routine in usart_14.asm will respond to
receiver interrupts and copy the received bytes into a buffer. This code
illustrates most of the usart behavior and (as far as I can tell) works.

Scott

Subject: gpsim CVS update
From: Scott Dattalo ####@####.####
Date: 3 Apr 2000 12:45:04 -0000
Message-Id: <Pine.LNX.4.21.0004030726340.30368-100000@tempest2.blackhat.net>

I'm getting ready to release version 19.0. I'll create the tarballs and RPMs so
you won't have to check everything out of CVS.

New:

gpsim has a new developer: Daniel Schudel. Daniel has added support for the
16x5x processors (and in particular the 16c54). Welcome aboard Daniel!

Thc CLI-only mode has been fixed. Now you can build gpsim without the gui
(again). This must have broke in version 16 or 17.

Scott

Subject: gpsim CVS update
From: Scott Dattalo ####@####.####
Date: 27 May 2001 04:49:41 -0000
Message-Id: <Pine.LNX.4.21.0105262326350.15080-100000@tempest2.blackhat.net>

Added module attributes

A few days ago I added the "set" option to module command
with the idea of supporting attributes. This support has
now been added. A new file called "src/attribute.cc" has
been added. There's a new class called Attribute that serves
as the base to FloatAttribute, IntAttribute, and StringAttribute 
classes. From the names you probably can guess that integer,
float, and string type attributes are supported. 

The attributes are supported in a very generic/flexible way
(from a programming point of view). For example, attributes
can be dynamically created and attached to modules. While I
haven't tried it, you could create a user defined (via a 
program) attribute and attach it to a simulated PIC.


The pullup resistor is the only module that supports attributes
right now. Check out examples/modules/mod_test/ in CVS for an
example of the usage. Here's a brief summary:

A command file with these contents:


load s mod_test.cod
module library libgpsim_modules.so
module load pullup   R1
node n1
attach n1 portb0 R1

(whole command file:
http://cvs.sourceforge.net/cgi-bin/viewcvs.cgi/gpsim/examples/modules/mod_test/mod_test.stc
)

Then while in gpsim you can do this:

gpsim> module set R1
resistance = 2000
gpsim> module set R1 resistance 1000
gpsim> module set R1
resistance = 1000

The pullup resistor only has one attribute: its resistance.
The default value is 2k (there's an adhocness with this).
This is shown by "set"ting R1, but not specifying an attribute.
In other words, all attributes (only one in this case) are
displayed when no attributes are specified with the set
command.

Then the resistance is changed to 1000 and displayed.

---------

The next module is going to be a USART. This one will make
extensive use of the attribute command.

Scott

Subject: Re: gpsim CVS update
From: Alex Holden ####@####.####
Date: 27 May 2001 18:11:17 -0000
Message-Id: <Pine.LNX.4.04.10105271905220.615-100000@hyperspace.linuxhacker.org>

On Sat, 26 May 2001, Scott Dattalo wrote:
> The next module is going to be a USART. This one will make
> extensive use of the attribute command.

A switch (both toggle and momentary) module with a GUI interface to it
might be useful. Perhaps it could even simulate switch bounce in order to 
test software debouncing.

-- 
------- Alex Holden -------
http://www.linuxhacker.org/
 http://www.robogeeks.org/

Subject: Re: gpsim CVS update
From: Scott Dattalo ####@####.####
Date: 27 May 2001 23:44:57 -0000
Message-Id: <Pine.LNX.4.21.0105271837500.20780-100000@tempest2.blackhat.net>

Does any one have any nice graphics for switches?

Scott


On Sun, 27 May 2001, Alex Holden wrote:

> On Sat, 26 May 2001, Scott Dattalo wrote:
> > The next module is going to be a USART. This one will make
> > extensive use of the attribute command.
> 
> A switch (both toggle and momentary) module with a GUI interface to it
> might be useful. Perhaps it could even simulate switch bounce in order to 
> test software debouncing.
> 
> -- 
> ------- Alex Holden -------
> http://www.linuxhacker.org/
>  http://www.robogeeks.org/
> 

[<<] [<] Page 1 of 1 [>] [>>]


Powered by ezmlm-browse 0.20.