Hi Andrew,

you are right, the FPGA code definitely needs some cleanup.

On Monday 12 November 2007 19:27:46 Andrew Dyer wrote:

A couple of "IMHO" comments about this code in no particular order:

Is there any reason to have spartan2.c, virtex2.c and spartan3.c? They are pretty much copies of each other with search and replace.

ack.

local PRINTF should be replaced with the global 'debug' macro.

ack.

this issue behind this patch should be addressed (relocating null pointers) http://thread.gmane.org/gmane.comp.boot-loaders.u-boot/22918

I posted a couple of patches a few minutes before your posting arrived. My 2nd patch addressed this issue (please apologize that I forgot to enumerate them as PATCH x/5. I forgot git-format-patch's -n option and noticed that to late.)

I believe the edict was given by wd@ that CFG_HZ was always to be 1000, and delays can be spec'd in units of ms rather than based off of CFG_HZ

Probably yes, but I am not sure if we might need sub-ms-timing.

could do with some whitespace cleanup & reformatting to 80 cols.

There are some other coding style issues like braces for single statements.

I am willing to implement some of your/our ideas. But first I am awaiting my five patches to get it in :-)

Matthias