[Bug target/65489] New: [5.0 Regression] Testcase for PR 65427 ICEs at -Os on arm-none-eabi in hoist pass

[Bug target/65489] New: [5.0 Regression] Testcase for PR 65427 ICEs at -Os on arm-none-eabi in hoist pass

[ktkachov at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=65489

            Bug ID: 65489
           Summary: [5.0 Regression] Testcase for PR 65427 ICEs at -Os on
                    arm-none-eabi in hoist pass
           Product: gcc
           Version: 5.0
            Status: UNCONFIRMED
          Keywords: ice-on-valid-code
          Severity: normal
          Priority: P3
         Component: target
          Assignee: unassigned at gcc dot gnu.org
          Reporter: ktkachov at gcc dot gnu.org
            Target: arm*

As reported at:
https://gcc.gnu.org/ml/gcc-patches/2015-03/msg00949.html

The testcase gcc.c-torture/execute/pr65427.c ICEs on arm at -Os with:
0x7e6f11 process_insert_insn
        $TOP/gcc/gcse.c:2174
0x7e8a80 insert_insn_end_basic_block
        $TOP/gcc/gcse.c:2196
0x7eab01 hoist_code
        $TOP/gcc/gcse.c:3492
0x7eab01 one_code_hoisting_pass
        $TOP/gcc/gcse.c:3722
0x7eab01 execute_rtl_hoist
        $TOP/gcc/gcse.c:4212
0x7eab01 execute
        $TOP/gcc/gcse.c:4293

The ICE is unrelated to that PR and is just exposed by the new testcase.
I can reproduce it with on arm-none-eabi with -Os -mfloat-abi=hard
-mcpu=cortex-a9 -mfpu=neon -mfloat-abi=softfp

[Bug target/65489] [5.0 Regression] Testcase for PR 65427 ICEs at -Os on arm-none-eabi in hoist pass

[ktkachov at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=65489

ktkachov at gcc dot gnu.org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |ASSIGNED
   Last reconfirmed|                            |2015-03-20
      Known to work|                            |4.9.3
           Assignee|unassigned at gcc dot gnu.org      |ktkachov at gcc dot gnu.org
   Target Milestone|---                         |5.0
     Ever confirmed|0                           |1
      Known to fail|                            |5.0

--- Comment #1 from ktkachov at gcc dot gnu.org ---
I'm working on a fix. I think the arm backend implementation of
TARGET_LEGITIMATE_CONSTANT_P is at fault

[Bug target/65489] [5 Regression] Testcase for PR 65427 ICEs at -Os on arm-none-eabi in hoist pass

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=65489

Richard Biener <rguenth at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Priority|P3                          |P1

[Bug target/65489] [5 Regression] Testcase for PR 65427 ICEs at -Os on arm-none-eabi in hoist pass

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=65489

--- Comment #2 from Richard Biener <rguenth at gcc dot gnu.org> ---
Ping?  If TARGET_LEGITIMATE_CONSTANT_P is at fault isn't this a latent issue
also present on branches?

[Bug target/65489] [5 Regression] Testcase for PR 65427 ICEs at -Os on arm-none-eabi in hoist pass

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=65489

--- Comment #3 from Richard Biener <rguenth at gcc dot gnu.org> ---
A posted patch was approved but not yet checked in.

[Bug target/65489] [5 Regression] Testcase for PR 65427 ICEs at -Os on arm-none-eabi in hoist pass

[ktkachov at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=65489

--- Comment #4 from ktkachov at gcc dot gnu.org ---
Author: ktkachov
Date: Tue Apr  7 10:24:19 2015
New Revision: 221892

URL: https://gcc.gnu.org/viewcvs?rev=221892&root=gcc&view=rev
Log:
[ARM] PR 65489: Accept VSTRUCT constants in arm_legitimate_constant_p

    PR target/65489
    * config/arm/arm.c (arm_legitimate_constant_p_1): Remove restriction
    on constants for NEON VSTRUCT modes.

Modified:
    trunk/gcc/ChangeLog
    trunk/gcc/config/arm/arm.c

[Bug target/65489] [5 Regression] Testcase for PR 65427 ICEs at -Os on arm-none-eabi in hoist pass

[ktkachov at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=65489

ktkachov at gcc dot gnu.org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|ASSIGNED                    |RESOLVED
         Resolution|---                         |FIXED

--- Comment #5 from ktkachov at gcc dot gnu.org ---
Fixed on trunk.
I couldn't get any testcase to fail from this issue on the other branches, so I
can't confirm that it's present there.
If someone manages to reproduce it there, please reopen this or a new PR