[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Simulavr-devel] Bitpos-patch

From: Theodore A. Roth
Subject: Re: [Simulavr-devel] Bitpos-patch
Date: Mon, 12 Jan 2004 00:22:03 -0800 (PST)

On Sun, 11 Jan 2004, Hermann Kraus wrote:

> Hello!
> I found some time in the christmas holydays to work on the patch for 
> simulavr. This one is reduced to the minimum requirements. It only changes 
> the eeprom files. The system has changed a bit since last time, when I 
> submitted the whole bitpos patch. Now I'm using a struct for the tables to 
> avoid that wrong assignement to the tables are done. I'd like to do the 
> same thing with the arrays left, but I can't figure out how to write a 
> function that produces bitmasks if the arrays are converted to structs.
> If the patch is ok, please commit it as soon as possible, so I can add the 
> other changes.

Hi Hermann,

I've looked over your patch. That's much easier for me to see what's going 
on now. Thanks for sticking with it.

I have a feeling that having a BitPos tab for every register of every device
is going to cause maintenance problems. It makes for a lot of work to add
support for a new divice.

Another thing I'm not too keen on is the fact that you have moved all the 
bit defs into a single structure. You've broken the encapsulation of the 
data for the IO registers. I would really like to keep all of the 
information for a given IO register in one place if possible.

My last concern is the overhead involved. What's the performance penalty? 
Although, this is a minor concern right now since I don't want to 
prematurely optimize. Better to get the code working than correctly first.

Can you look over the vdev/mem rewrite patch I just posted? It might be 
possible to install a bit position table for a given IO register into the 
memory cell or maybe passed into vdev_create() or vdev_AddAddr() functions 
I'm adding.

Sorry, but my patch is going to conflict with your work. Best to grab a 
clean cvs tree to work with it.

Ted Roth

reply via email to

[Prev in Thread] Current Thread [Next in Thread]