public inbox for gcc-bugs@sourceware.org
help / color / mirror / Atom feed
From: "pinskia at gcc dot gnu.org" <gcc-bugzilla@gcc.gnu.org>
To: gcc-bugs@gcc.gnu.org
Subject: [Bug rtl-optimization/110701] [14 Regression] Wrong code at -O1/2/3/s on x86_64-linux-gnu
Date: Mon, 17 Jul 2023 17:51:08 +0000	[thread overview]
Message-ID: <bug-110701-4-9G5dh73lTk@http.gcc.gnu.org/bugzilla/> (raw)
In-Reply-To: <bug-110701-4@http.gcc.gnu.org/bugzilla/>

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=110701

Andrew Pinski <pinskia at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Keywords|                            |wrong-code
   Last reconfirmed|                            |2023-07-17
   Target Milestone|---                         |14.0
     Ever confirmed|0                           |1
             Status|UNCONFIRMED                 |NEW

--- Comment #2 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
Confirmed.

Trying 16 -> 17:
   16: {r94:HI=r92:DI#0^0x3;clobber flags:CC;}
      REG_DEAD r92:DI
      REG_UNUSED flags:CC
   17: r95:SI=sign_extend(r94:HI)
      REG_DEAD r94:HI
Successfully matched this instruction:
(set (reg:SI 95)
    (ior:SI (subreg:SI (reg:DI 92) 0)
        (const_int 3 [0x3])))
allowing combination of insns 16 and 17
original costs 4 + 4 = 8
replacement cost 4
deferring deletion of insn with uid = 16.
modifying insn i3    17: {r95:SI=r92:DI#0|0x3;clobber flags:CC;}
      REG_UNUSED flags:CC
      REG_DEAD r92:DI
deferring rescan insn with uid = 17.


BUT r92 is defined by
(insn 12 11 13 2 (parallel [
            (set (reg:DI 92)
                (ashiftrt:DI (reg:DI 93 [ bD.2759 ])
                    (const_int 63 [0x3f])))
            (clobber (reg:CC 17 flags))
        ]) "/app/example.cpp":6:53 901 {ashrdi3_cvt}
     (expr_list:REG_DEAD (reg:DI 93 [ bD.2759 ])
        (expr_list:REG_UNUSED (reg:CC 17 flags)
            (expr_list:REG_EQUAL (ashiftrt:DI (mem/c:DI (symbol_ref:DI ("b")
[flags 0x2]  <var_decl 0x7f6efc025d80 b>) [1 bD.2759+0 S8 A64])
                    (const_int 63 [0x3f]))
                (nil)))))
(insn 13 12 14 2 (set (subreg:HI (reg:DI 92) 0)
        (not:HI (subreg:HI (reg:DI 92) 0))) "/app/example.cpp":6:53 816
{*one_cmplhi2_1}
     (nil))
(insn 14 13 16 2 (parallel [
            (set (subreg:HI (reg:DI 92) 0)
                (and:HI (subreg:HI (reg:DI 92) 0)
                    (const_int 340 [0x154])))
            (clobber (reg:CC 17 flags))
        ]) "/app/example.cpp":6:53 596 {*andhi_1}
     (expr_list:REG_UNUSED (reg:CC 17 flags)
        (nil)))

So that should be ok, well no because SI != HI here ... and

  parent reply	other threads:[~2023-07-17 17:51 UTC|newest]

Thread overview: 10+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2023-07-17 10:21 [Bug c/110701] New: " shaohua.li at inf dot ethz.ch
2023-07-17 12:43 ` [Bug rtl-optimization/110701] [14 Regression] " xry111 at gcc dot gnu.org
2023-07-17 17:51 ` pinskia at gcc dot gnu.org [this message]
2023-07-17 17:55 ` pinskia at gcc dot gnu.org
2023-07-17 17:58 ` pinskia at gcc dot gnu.org
2023-07-18 15:38 ` roger at nextmovesoftware dot com
2023-07-18 17:17 ` roger at nextmovesoftware dot com
2023-07-27 18:23 ` roger at nextmovesoftware dot com
2023-10-04 16:12 ` cvs-commit at gcc dot gnu.org
2023-10-11  7:59 ` roger at nextmovesoftware 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-110701-4-9G5dh73lTk@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).