This is the mail archive of the gcc@gcc.gnu.org mailing list for the GCC project.

Index Nav:	[Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav:	[Date Prev] [Date Next]	[Thread Prev] [Thread Next]
Other format:	[Raw text]

Re: tree-eh.c vs. CFGs

From: Stuart Hastings <stuart at apple dot com>
To: Richard Henderson <rth at redhat dot com>
Cc: gcc at gcc dot gnu dot org
Date: Thu, 17 Jun 2004 16:04:47 -0700
Subject: Re: tree-eh.c vs. CFGs
References: <22A40AFA-C099-11D8-B6A0-000A95A83B3C@apple.com> <20040617203451.GA7584@redhat.com>

On Jun 17, 2004, at 1:34 PM, Richard Henderson wrote:

On Thu, Jun 17, 2004 at 01:01:35PM -0700, Stuart Hastings wrote:
I believe the Right Thing to do is revise tree-eh.c to work on CFGs.
No. You can't build a correct CFG without running pass_lower_eh.

Moot point.

The inliner doesn't really care if the CFG is "correct" or not; it's mostly concerned with finding the calls and splicing the blocks. Inlining before lowering EH nodes would be easier for the inliner (but nobody else seems to be concerned about that ;-).

All those edges are just a nuisance... :-)

If you want to do inlining in the presence of CFGs, then you have
to run *everything* up to pass_build_cfg.  And then you have to
make the inliner be aware of the EH data on the side, and merge it.

O.K.

stuart hastings
Apple Computer

References:
- tree-eh.c vs. CFGs
  - From: Stuart Hastings
- Re: tree-eh.c vs. CFGs
  - From: Richard Henderson

Index Nav:	[Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav:	[Date Prev] [Date Next]	[Thread Prev] [Thread Next]