This is the mail archive of the
mailing list for the GCC project.
Re: [PATCH] Make ir_type differentiate between cfgrtl and cfglayout mode
- From: Roger Sayle <roger at eyesopen dot com>
- To: Steven Bosscher <stevenb dot gcc at gmail dot com>
- Cc: gcc-patches at gcc dot gnu dot org
- Date: Wed, 1 Nov 2006 11:24:02 -0700 (MST)
- Subject: Re: [PATCH] Make ir_type differentiate between cfgrtl and cfglayout mode
On Wed, 1 Nov 2006, Steven Bosscher wrote:
> 2006-11-01 Steven Bosscher <firstname.lastname@example.org>
> * tree-dump.c (dump_enable_all): Rename local variable
> ir_type to avoid name conflicts.
> * cfgloopmanip.c (lv_adjust_loop_entry_edge): Check for IR_GIMPLE
> instead of using ir_type().
> * profile.c (tree_register_profile_hooks): Likewise.
> * value-prof.c (tree_register_value_prof_hooks): Likewise.
> * basic-block.h (struct edge_def): Likewise.
> * config/arm/arm.c (legitimize_pic_address): Likewise.
> * coretypes.h (ir_type): New enum of all intermediate languages
> used in GCC.
> * cfghooks.c (ir_type): Rename to...
> (current_ir_type): ...this. Distinguish between cfgrtl and
> cfglayout mode when the current IR is RTL. Return enum ir_type.
> * cfghooks.h (ir_type): Replace with current_ir_type prototype.
This is OK for mainline. Thanks. I agree that the ARM backend change
is trivial; but I'm sure someone will let us know if it isn't.
Hopefully, this is a short-term issue until all the RTL optimizers
can be standardized. However, if any one decides to add more IR
states, we should probably switch to a bit-based approch where we
can easily test for "ir_type & RTL", rather than "!= GIMPLE".
However, I'm happy to let that someone else face this clean-up
when they come to it, as the current approach seems reasonable for