From: Jeff Law <law@redhat.com>
To: Jim Wilson <jim.wilson@linaro.org>,
"gcc-patches@gcc.gnu.org" <gcc-patches@gcc.gnu.org>
Subject: Re: [PATCH, ARM] stop changing signedness in PROMOTE_MODE
Date: Tue, 07 Jul 2015 15:07:00 -0000 [thread overview]
Message-ID: <559BEB2D.7040800@redhat.com> (raw)
In-Reply-To: <CABXYE2VXjy7=5Y=c1TCxLE8KuwLtwBYBhTB24xrWDvWAeiBwbQ@mail.gmail.com>
On 06/29/2015 07:15 PM, Jim Wilson wrote:
> This is my suggested fix for PR 65932, which is a linux kernel
> miscompile with gcc-5.1.
>
> The problem here is caused by a chain of events. The first is that
> the relatively new eipa_sra pass creates fake parameters that behave
> slightly differently than normal parameters. The second is that the
> optimizer creates phi nodes that copy local variables to fake
> parameters and/or vice versa. The third is that the ouf-of-ssa pass
> assumes that it can emit simple move instructions for these phi nodes.
> And the fourth is that the ARM port has a PROMOTE_MODE macro that
> forces QImode and HImode to unsigned, but a
> TARGET_PROMOTE_FUNCTION_MODE hook that does not. So signed char and
> short parameters have different in register representations than local
> variables, and require a conversion when copying between them, a
> conversion that the out-of-ssa pass can't easily emit.
So if these "copies" require a conversion, then isn't it fundamentally
wrong to have a PHI node which copies between them? That would seem to
implicate the eipa_sra pass as needing to be aware of these promotions
and avoid having these objects with different representations appearing
on the lhs/rhs of a PHI node.
Jeff
next prev parent reply other threads:[~2015-07-07 15:07 UTC|newest]
Thread overview: 32+ messages / expand[flat|nested] mbox.gz Atom feed top
2015-06-30 1:56 Jim Wilson
2015-07-02 9:07 ` Richard Earnshaw
2015-07-07 18:25 ` Jim Wilson
2015-07-07 15:07 ` Jeff Law [this message]
2015-07-07 16:29 ` Jim Wilson
2015-07-07 21:35 ` Richard Biener
2015-07-10 15:46 ` Jim Wilson
2015-07-13 8:19 ` Richard Biener
2015-07-13 15:29 ` Michael Matz
2015-07-13 15:35 ` H.J. Lu
2015-07-14 16:38 ` Richard Earnshaw
2015-07-14 16:49 ` Richard Biener
2015-07-14 17:07 ` Jim Wilson
2015-07-14 17:23 ` Richard Biener
2015-07-15 13:25 ` Michael Matz
2015-07-15 16:01 ` Jim Wilson
2015-07-16 9:40 ` Richard Earnshaw
2015-07-16 15:02 ` Michael Matz
2015-07-16 15:20 ` Richard Earnshaw
2015-07-15 13:04 ` Michael Matz
2015-07-08 22:54 ` Jeff Law
2015-07-10 15:35 ` Jim Wilson
2016-02-04 8:58 ` Ramana Radhakrishnan
2016-02-15 11:32 ` Kyrill Tkachov
2016-02-16 10:44 ` Ramana Radhakrishnan
2016-02-17 10:03 ` Christophe Lyon
2016-02-17 10:05 ` Kyrill Tkachov
2016-02-17 10:20 ` Christophe Lyon
2016-02-17 10:22 ` Kyrill Tkachov
2016-02-18 10:16 ` Christophe Lyon
2016-03-07 4:43 ` Ramana Radhakrishnan
2016-03-07 12:55 ` Christophe Lyon
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=559BEB2D.7040800@redhat.com \
--to=law@redhat.com \
--cc=gcc-patches@gcc.gnu.org \
--cc=jim.wilson@linaro.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).