[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
[avr-libc-dev] Re: Iotn13.h
From: |
Theodore A. Roth |
Subject: |
[avr-libc-dev] Re: Iotn13.h |
Date: |
Fri, 21 May 2004 10:00:10 -0700 (PDT) |
On Thu, 20 May 2004, Bruce Graham wrote:
> Ted,
>
> Here is my suggested patch for the header iotn13.h I did not test it
> very well since I have been unable to run "avr-as" by itself
> successfully. I seem to be short the right document, which I probably
> have but can't find, or less likely don't have and can't find.
>
> Some things may be controversial:
> (1) When there is no External Memory -- should XRAMEND be 0? or should
> it be equal to RAMEND. Actually I could go either way and I can't see
> where the symbol is used -- probably in malloc.
> (2) Although PB5 is a handy shorthand, the tradition for bit names seems
> to be:
> (a) use the data sheet name if a name is given
> (b) concatenate numbers to the register name if a name is not given
> (3) I was not sure what SPEN referred to in register SPMCSR. I could
> not find it in the data sheet, nor could I find it in one or two other
> data sheets I checked. I hope it was just a mistake, but I could
> believe it came form an earlier version of the family.
Attached is what I committed (basically limited to comments only, but
with the fix for SPMEN).
A few notes about your original patch.
- Please don't add lots of white space at the end of comments.
- EEAR and EEARH don't need to be undef'd since they are handled by
avr/io.h.
- SP and SPL don't need to be undef'd either.
Now, I like the concept you introduced for setting _VECTORS_SIZE, but I
didn't include that change here since it should be done more globally. I
would rather see something like the following.
In avr/io.h:
/* The value defined in the device io header for _MAX_VECTOR_NUMBER
should be the highest number used in the SIG_* defines, while
the value for _VECTOR_INSN_SIZE should be the size of the vector
instruction in bytes. This will be either 2 (for devices that don't
support the JMP instruction) or 4 (for devices that can use the JMP
instruction). */
#define _VECTORS_SIZE ((_MAX_VECTOR_NUMBER)+1)*_VECTOR_INSN_SIZE
Then in each io*.h file (using iotn13.h as example):
#define _MAX_VECTOR_NUMBER 9
#define _VECTOR_INSN_SIZE 2
One last thing about your usage of CVS. I think you are making it more
difficult than it needs to be. It looks like you are trying to work with
a local repository with your own tree structure or a repo for each file
you are working on. This is making it a bit more difficult for you to
generate patches and for me to apply those patches.
Here's the steps you need to take to generate a patch (as unix shell
commands, but it shouldn't be too hard to figure what to do for
windows).
Check out the cvs tree:
$ cvs -z3 -d:ext:address@hidden:/cvsroot/avr-libc co avr-libc
Go into the tree:
$ cd avr-libc
Edit files to your hearts content:
$ vim include/avr/iotn13.h
When you are done editing, generate a patch:
$ cvs diff -u > my-changes.diff
Look the diff file over and if it looks right, post it to the
avr-libc-dev list.
Notice that I generated the diff file from the top of the tree. This
allows me or anyone else to apply the patch like this:
$ patch -p0 --dry-run < my-changes.diff
If the --dry-run doesn't spew a bunch of errors, I'll apply it:
$ patch -p0 < my-changes.diff
Hope that helps.
---
Ted Roth
PGP Key ID: 0x18F846E9
Jabber ID: address@hidden