Instruction bundling on IA64 and the Intel Assembler
Richard Kenner
kenner@vlsi1.ultra.nyu.edu
Mon Dec 15 23:11:00 GMT 2003
Unfortunately I have no time for this.
This is a perfect example of why patches should not be approved until
they are adequately documented. It's *very* hard (often impossible)
to get that done later. Indeed it's usually recommended to write the
comments *before* the code.
As I wrote in my previous email I'll try to write more comments and
divide the function on several ones when the main line is in stage 1.
Please write the comments now. Dividing the function up can probably
safely be done now, but can indeed wait for stage 1.
Take a look at reorg.c. That's code of roughly equivalent complexity.
Try to match the commenting style in that code.
More information about the Gcc
mailing list