From: Chris Lattner <sabre@nondot.org>
To: Zack Weinberg <zack@codesourcery.com>
Cc: Richard.Earnshaw@arm.com, Jim Wilson <wilson@specifixinc.com>,
<gcc-patches@gcc.gnu.org>
Subject: Re: avoid unnecessary register saves for setjmp
Date: Thu, 27 Nov 2003 09:19:00 -0000 [thread overview]
Message-ID: <Pine.LNX.4.44.0311270301180.28662-100000@nondot.org> (raw)
Zack Weinberg said:
> I'm going to take this opportunity to throw in a suggestion I've made
> before, which is that setjmp/longjmp should - always, not just when
> special variants are used - be converted by the front end into
> invocations of the exception-unwinding facility, which gives us tons
> more control.
This is exactly what we do in LLVM, which makes a lot of things much
nicer. For example, because it uses the normal EH facilities in LLVM,
inlining a longjmp'ing function into the setjmp'ing function automatically
turns the stack unwinding operation into a direct branch, same as normal
EH.
This allows us to do aggressive optimizations on both the setjmp'ing and
longjmp'ing functions: there are no special cases needed or optimizations
disabled. The only negative of doing this is that it adds a lot of edges
to the CFG's of some functions, but you have to represent the possible
control flow SOMEHOW. :) Also, the normal LLVM EH pruner automatically
deletes the edges from function calls that provably cannot unwind, so it
gets rid of a lot of the problem.
Just food for thought,
-Chris
--
http://llvm.cs.uiuc.edu/
http://www.nondot.org/~sabre/Projects/
next reply other threads:[~2003-11-27 9:06 UTC|newest]
Thread overview: 33+ messages / expand[flat|nested] mbox.gz Atom feed top
2003-11-27 9:19 Chris Lattner [this message]
2003-11-27 9:21 ` Andrew Haley
2003-11-27 9:22 ` Chris Lattner
2003-11-27 9:39 ` Andrew Haley
2003-11-27 9:43 ` Chris Lattner
2003-11-27 10:14 ` Zack Weinberg
2003-11-27 10:15 ` Chris Lattner
2003-11-27 11:01 ` Zack Weinberg
2003-11-27 20:28 ` Chris Lattner
2003-11-27 20:51 ` Gabriel Dos Reis
2003-11-27 15:41 ` Jan Vroonhof
2003-11-27 16:23 ` Jan Vroonhof
2003-11-27 10:31 ` Andrew Haley
2003-11-27 10:53 ` Zack Weinberg
[not found] <5cad8ef043da68f2a3332f00bd6a186a3fc6195b@mail.esmertec.com>
2003-11-27 20:44 ` Chris Lattner
-- strict thread matches above, loose matches on Subject: below --
2003-11-22 4:13 Richard Kenner
2003-12-01 3:29 ` Jim Wilson
2003-11-21 5:54 Jim Wilson
2003-11-21 18:39 ` Daniel Jacobowitz
2003-11-22 0:22 ` Jim Wilson
2003-11-22 3:57 ` Daniel Jacobowitz
2003-11-22 9:06 ` Jim Wilson
2003-11-22 14:37 ` Richard Earnshaw
2003-11-22 16:33 ` Daniel Jacobowitz
2003-11-27 9:11 ` Zack Weinberg
2003-11-27 17:11 ` Andrew Pinski
2003-11-27 18:23 ` Zack Weinberg
2003-11-28 5:15 ` Geert Bosch
2003-11-28 11:14 ` Richard Earnshaw
2003-11-28 14:02 ` Richard Earnshaw
2003-11-21 20:14 ` Geoff Keating
2003-11-22 0:20 ` Jim Wilson
2003-11-30 9:23 ` Jim Wilson
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=Pine.LNX.4.44.0311270301180.28662-100000@nondot.org \
--to=sabre@nondot.org \
--cc=Richard.Earnshaw@arm.com \
--cc=gcc-patches@gcc.gnu.org \
--cc=wilson@specifixinc.com \
--cc=zack@codesourcery.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).