public inbox for gcc-bugs@sourceware.org
help / color / mirror / Atom feed
From: "jakub at gcc dot gnu.org" <gcc-bugzilla@gcc.gnu.org>
To: gcc-bugs@gcc.gnu.org
Subject: [Bug target/57896] [4.8 Regression] ICE in expand_expr_real_2
Date: Thu, 20 Feb 2014 10:27:00 -0000 [thread overview]
Message-ID: <bug-57896-4-oKlTtx6diD@http.gcc.gnu.org/bugzilla/> (raw)
In-Reply-To: <bug-57896-4@http.gcc.gnu.org/bugzilla/>
http://gcc.gnu.org/bugzilla/show_bug.cgi?id=57896
Jakub Jelinek <jakub at gcc dot gnu.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
Component|middle-end |target
--- Comment #19 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
The bug is in the backend, gen_reg_rtx shouldn't be called when init_emit has
not been called.
If one looks e.g. at tree-ssa-loop-ivopts.c or even
ix86_vectorize_vec_perm_const_ok, it avoids using gen_reg_rtx and instead uses
gen_raw_REG:
d.target = gen_raw_REG (d.vmode, LAST_VIRTUAL_REGISTER + 1);
d.op1 = d.op0 = gen_raw_REG (d.vmode, LAST_VIRTUAL_REGISTER + 2);
if (!d.one_operand_p)
d.op1 = gen_raw_REG (d.vmode, LAST_VIRTUAL_REGISTER + 3);
So, we should avoid calling gen_reg_rtx if d->testing_p, we can create
gen_raw_REG instead if really needed. Let me skim through the routines.
next prev parent reply other threads:[~2014-02-20 10:27 UTC|newest]
Thread overview: 33+ messages / expand[flat|nested] mbox.gz Atom feed top
2013-07-15 6:44 [Bug c/57896] New: ICE in " zeccav at gmail dot com
2013-07-15 9:37 ` [Bug c/57896] " mikpe at it dot uu.se
2013-07-15 14:23 ` mikpe at it dot uu.se
2013-07-15 15:32 ` glisse at gcc dot gnu.org
2013-07-15 16:32 ` glisse at gcc dot gnu.org
2013-07-15 23:42 ` zeccav at gmail dot com
2013-07-16 6:19 ` zeccav at gmail dot com
2014-02-19 8:11 ` [Bug c/57896] [4.8 Regression] ICE " mpolacek at gcc dot gnu.org
2014-02-19 17:59 ` [Bug target/57896] " ubizjak at gmail dot com
2014-02-19 18:06 ` ubizjak at gmail dot com
2014-02-19 18:34 ` ubizjak at gmail dot com
2014-02-19 18:37 ` [Bug middle-end/57896] " ubizjak at gmail dot com
2014-02-19 19:19 ` mpolacek at gcc dot gnu.org
2014-02-20 9:29 ` ubizjak at gmail dot com
2014-02-20 9:58 ` ubizjak at gmail dot com
2014-02-20 10:15 ` ubizjak at gmail dot com
2014-02-20 10:27 ` jakub at gcc dot gnu.org [this message]
2014-02-20 10:41 ` [Bug target/57896] " jakub at gcc dot gnu.org
2014-02-20 10:45 ` ubizjak at gmail dot com
2014-02-20 10:57 ` jakub at gcc dot gnu.org
2014-02-20 11:01 ` ubizjak at gmail dot com
2014-02-20 11:07 ` [Bug target/57896] [4.7/4.8/4.9 " ubizjak at gmail dot com
2014-02-20 11:42 ` jakub at gcc dot gnu.org
2014-02-20 12:00 ` jakub at gcc dot gnu.org
2014-02-20 12:09 ` ubizjak at gmail dot com
2014-02-20 13:00 ` jakub at gcc dot gnu.org
2014-02-20 13:12 ` ubizjak at gmail dot com
2014-02-20 17:25 ` ubizjak at gmail dot com
2014-02-20 20:40 ` jakub at gcc dot gnu.org
2014-02-20 20:41 ` jakub at gcc dot gnu.org
2014-02-20 20:42 ` [Bug target/57896] [4.7 " jakub at gcc dot gnu.org
2014-02-21 7:56 ` uros at gcc dot gnu.org
2014-02-21 8:01 ` ubizjak at gmail dot com
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=bug-57896-4-oKlTtx6diD@http.gcc.gnu.org/bugzilla/ \
--to=gcc-bugzilla@gcc.gnu.org \
--cc=gcc-bugs@gcc.gnu.org \
/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).