[Bug rtl-optimization/113510] New: [ARM Thumb] ICE in extract_constrain_insn with CPU cortex-m23

[Bug rtl-optimization/113510] New: [ARM Thumb] ICE in extract_constrain_insn with CPU cortex-m23

[thiago.bauermann at linaro dot org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=113510

            Bug ID: 113510
           Summary: [ARM Thumb] ICE in extract_constrain_insn with CPU
                    cortex-m23
           Product: gcc
           Version: 14.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: rtl-optimization
          Assignee: unassigned at gcc dot gnu.org
          Reporter: thiago.bauermann at linaro dot org
  Target Milestone: ---

Created attachment 57167
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=57167&action=edit
Preprocessed version of nested-3.c

After commit a729b6e002fe ("[PR112918][LRA]: Fixing IRA ICE on m68k"), the
following failures started appearing on --target=arm-none-eabi
--with-mode=thumb --with-cpu=cortex-m23:

        === gcc tests ===

Running gcc:gcc.c-torture/compile/compile.exp ...
FAIL: gcc.c-torture/compile/nested-3.c -O2  (internal compiler error: in
extract_constrain_insn, at recog.cc:2713)
FAIL: gcc.c-torture/compile/nested-3.c -O2  (test for excess errors)
FAIL: gcc.c-torture/compile/nested-3.c -O2 -flto -fno-use-linker-plugin
-flto-partition=none  (internal compiler error: in extract_constrain_insn, at
recog.cc:2713)
FAIL: gcc.c-torture/compile/nested-3.c -O2 -flto -fno-use-linker-plugin
-flto-partition=none  (test for excess errors)
FAIL: gcc.c-torture/compile/nested-3.c -O3 -g  (internal compiler error: in
extract_constrain_insn, at recog.cc:2713)
FAIL: gcc.c-torture/compile/nested-3.c -O3 -g  (test for excess errors)
FAIL: gcc.c-torture/compile/nested-3.c -Os  (internal compiler error: in
extract_constrain_insn, at recog.cc:2713)
FAIL: gcc.c-torture/compile/nested-3.c -Os  (test for excess errors)

Running gcc:gcc.c-torture/execute/execute.exp ...
FAIL: gcc.c-torture/execute/pr71494.c -O2  (internal compiler error: in
extract_constrain_insn, at recog.cc:2713)
FAIL: gcc.c-torture/execute/pr71494.c -O2  (test for excess errors)
UNRESOLVED: gcc.c-torture/execute/pr71494.c -O2  compilation failed to produce
executable
FAIL: gcc.c-torture/execute/pr71494.c -O2 -flto -fno-use-linker-plugin
-flto-partition=none  (internal compiler error: in extract_constrain_insn, at
recog.cc:2713)
FAIL: gcc.c-torture/execute/pr71494.c -O2 -flto -fno-use-linker-plugin
-flto-partition=none  (test for excess errors)
UNRESOLVED: gcc.c-torture/execute/pr71494.c -O2 -flto -fno-use-linker-plugin
-flto-partition=none  compilation failed to produce executable
FAIL: gcc.c-torture/execute/pr71494.c -O2 -flto -fuse-linker-plugin
-fno-fat-lto-objects  (internal compiler error: in extract_constrain_insn, at
recog.cc:2713)
FAIL: gcc.c-torture/execute/pr71494.c -O2 -flto -fuse-linker-plugin
-fno-fat-lto-objects  (test for excess errors)
UNRESOLVED: gcc.c-torture/execute/pr71494.c -O2 -flto -fuse-linker-plugin
-fno-fat-lto-objects  compilation failed to produce executable
FAIL: gcc.c-torture/execute/pr71494.c -O3 -g  (internal compiler error: in
extract_constrain_insn, at recog.cc:2713)
FAIL: gcc.c-torture/execute/pr71494.c -O3 -g  (test for excess errors)
UNRESOLVED: gcc.c-torture/execute/pr71494.c -O3 -g  compilation failed to
produce executable
FAIL: gcc.c-torture/execute/pr71494.c -Os  (internal compiler error: in
extract_constrain_insn, at recog.cc:2713)
FAIL: gcc.c-torture/execute/pr71494.c -Os  (test for excess errors)
UNRESOLVED: gcc.c-torture/execute/pr71494.c -Os  compilation failed to produce
executable

I tested on today's trunk (commit 07b392550f37) and the failures are still
present. I'm attaching the gcc.{sum,log} files, and also the .i file that
causes the ICE.

Here's how to reproduce on an x86_64-linux machine:

1. First, create a combined tree:

$ mkdir /tmp/combined-tree-src /tmp/combined-tree-build
$ cd ~/src/newlib-cygwin && find . -print | cpio -pdlm /tmp/combined-tree-src
$ cd ~/src/binutils-gdb && find . -print | cpio -pdlmu /tmp/combined-tree-src
$ cd ~/src/gcc && find . -print | cpio -pdlmu /tmp/combined-tree-src

2. Then build and install the cross GCC:

$ cd /tmp/combined-tree-build
$ /tmp/combined-tree-src/configure \
    SHELL=/bin/bash \
    --with-gnu-as \
    --with-gnu-ld \
    --disable-libmudflap \
    --enable-lto \
    --enable-shared \
    --without-included-gettext \
    --enable-nls \
    --with-system-zlib \
    --disable-sjlj-exceptions \
    --enable-gnu-unique-object \
    --enable-linker-build-id \
    --disable-libstdcxx-pch \
    --enable-c99 \
    --enable-clocale=gnu \
    --enable-libstdcxx-debug \
    --enable-long-long \
    --with-cloog=no \
    --with-ppl=no \
    --with-isl=no \
    --enable-threads=no \
    --disable-multiarch \
    --disable-multilib \
    --with-mode=thumb \
    --with-cpu=cortex-m23 \
    --with-float=soft \
    --with-newlib \
    --with-headers=yes \
    --with-native-system-header-dir=/include \
    --enable-checking=yes \
    --disable-bootstrap \
    --enable-languages=c,lto \
    --prefix=/tmp/arm-none-eabi \
    --build=x86_64-pc-linux-gnu \
    --host=x86_64-pc-linux-gnu \
    --disable-source-highlight \
    --target=arm-none-eabi \
&& exec make \
        SHELL=/bin/bash \
        -w \
        -j $(nproc) \
        LDFLAGS_FOR_TARGET="--specs=rdimon.specs" \
        CFLAGS_FOR_BUILD="-pipe -g -O2" \
        CXXFLAGS_FOR_BUILD="-pipe -g -O2" \
        LDFLAGS_FOR_BUILD="-static-libgcc" \
        MAKEINFOFLAGS=--force \
        BUILD_INFO="" \
        MAKEINFO=echo \
&& make install

3. Finally, use it to compile the problematic .c files:

$ /tmp/arm-none-eabi/bin/arm-none-eabi-gcc -fdiagnostics-plain-output -O2 -w -c
-o nested-3.o
/tmp/combined-tree-src/gcc/testsuite/gcc.c-torture/compile/nested-3.c
/tmp/combined-tree-src/gcc/testsuite/gcc.c-torture/compile/nested-3.c: In
function ‘f’:
/tmp/combined-tree-src/gcc/testsuite/gcc.c-torture/compile/nested-3.c:18:1:
error: insn does not satisfy its constraints:
(insn 39 13 16 2 (set (reg:SI 12 ip [127])
        (plus:SI (reg/f:SI 13 sp)
            (const_int 8 [0x8])))
"/tmp/combined-tree-src/gcc/testsuite/gcc.c-torture/compile/nested-3.c":17:10
935 {*thumb1_addsi3}
     (nil))
during RTL pass: cprop_hardreg
/tmp/combined-tree-src/gcc/testsuite/gcc.c-torture/compile/nested-3.c:18:1:
internal compiler error: in extract_constrain_insn, at recog.cc:2713
0x744632 _fatal_insn(char const*, rtx_def const*, char const*, int, char
const*)
        ../../combined-tree/gcc/rtl-error.cc:108
0x74465e _fatal_insn_not_found(rtx_def const*, char const*, int, char const*)
        ../../combined-tree/gcc/rtl-error.cc:118
0x7430fb extract_constrain_insn(rtx_insn*)
        ../../combined-tree/gcc/recog.cc:2713
0xef7636 copyprop_hardreg_forward_1
        ../../combined-tree/gcc/regcprop.cc:836
0xef8784 execute
        ../../combined-tree/gcc/regcprop.cc:1423
Please submit a full bug report, with preprocessed source (by using
-freport-bug).
Please include the complete backtrace with any bug report.
See <https://gcc.gnu.org/bugs/> for instructions.

$ /tmp/arm-none-eabi/bin/arm-none-eabi-gcc
/tmp/combined-tree-src/gcc/testsuite/gcc.c-torture/execute/pr71494.c
-fdiagnostics-plain-output -O2 -w
-B/tmp/combined-tree-build/arm-none-eabi/./libgloss/arm/
-L/tmp/combined-tree-build/arm-none-eabi/./libgloss/arm
-L/tmp/combined-tree-src/libgloss/arm -L/tmp/combined-tree-build/ld
-B/tmp/combined-tree-build/arm-none-eabi/./newlib/
-L/tmp/combined-tree-build/arm-none-eabi/./newlib -specs=rdimon.specs -lm -o
./pr71494.exe
/tmp/combined-tree-src/gcc/testsuite/gcc.c-torture/execute/pr71494.c: In
function ‘main’:
/tmp/combined-tree-src/gcc/testsuite/gcc.c-torture/execute/pr71494.c:23:1:
error: insn does not satisfy its constraints:
(insn 64 9 12 2 (set (reg:SI 12 ip [123])
        (plus:SI (reg/f:SI 13 sp)
            (const_int 4 [0x4])))
"/tmp/combined-tree-src/gcc/testsuite/gcc.c-torture/execute/pr71494.c":19:3 935
{*thumb1_addsi3}
     (nil))
during RTL pass: cprop_hardreg
/tmp/combined-tree-src/gcc/testsuite/gcc.c-torture/execute/pr71494.c:23:1:
internal compiler error: in extract_constrain_insn, at recog.cc:2713
0x744632 _fatal_insn(char const*, rtx_def const*, char const*, int, char
const*)
        ../../combined-tree/gcc/rtl-error.cc:108
0x74465e _fatal_insn_not_found(rtx_def const*, char const*, int, char const*)
        ../../combined-tree/gcc/rtl-error.cc:118
0x7430fb extract_constrain_insn(rtx_insn*)
        ../../combined-tree/gcc/recog.cc:2713
0xef7636 copyprop_hardreg_forward_1
        ../../combined-tree/gcc/regcprop.cc:836
0xef8784 execute
        ../../combined-tree/gcc/regcprop.cc:1423
Please submit a full bug report, with preprocessed source (by using
-freport-bug).
Please include the complete backtrace with any bug report.
See <https://gcc.gnu.org/bugs/> for instructions.

[Bug rtl-optimization/113510] [ARM Thumb] ICE in extract_constrain_insn with CPU cortex-m23

[thiago.bauermann at linaro dot org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=113510

--- Comment #1 from Thiago Jung Bauermann <thiago.bauermann at linaro dot org> ---
Created attachment 57168
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=57168&action=edit
Preprocessed version of pr71494.c

[Bug rtl-optimization/113510] [ARM Thumb] ICE in extract_constrain_insn with CPU cortex-m23

[thiago.bauermann at linaro dot org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=113510

Thiago Jung Bauermann <thiago.bauermann at linaro dot org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |thiago.bauermann at linaro dot org

--- Comment #2 from Thiago Jung Bauermann <thiago.bauermann at linaro dot org> ---
Created attachment 57169
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=57169&action=edit
Testsuite log from today's trunk with both testcases failing.

I said I was going to send gcc.sum but on second thought, I think gcc.log is
sufficient.

[Bug target/113510] [14 Regression ARM Thumb] ICE in extract_constrain_insn with CPU cortex-m23

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=113510

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
          Component|rtl-optimization            |target
   Target Milestone|---                         |14.0
           Keywords|                            |ice-on-valid-code, ra

--- Comment #3 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
g:a729b6e002fe

[Bug target/113510] [14 Regression] [ARM Thumb] ICE in extract_constrain_insn with CPU cortex-m23

[vmakarov at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=113510

Vladimir Makarov <vmakarov at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |vmakarov at gcc dot gnu.org

--- Comment #4 from Vladimir Makarov <vmakarov at gcc dot gnu.org> ---
This is not a RA bug.  I believe it is a bug in peephole optimization.

Right after LRA and before peephole2 we have:

(insn 28 13 15 2 (set (reg:SI 12 ip [127])
        (const_int 8 [0x8]))
"../../gcc/gcc/testsuite/gcc.c-torture/compile/nested-3.c":17:10 959
{*thumb1_movsi_insn}
     (nil))
(insn 15 28 16 2 (set (reg:SI 12 ip [127])
        (plus:SI (reg:SI 12 ip [127])
            (reg/f:SI 13 sp)))
"../../gcc/gcc/testsuite/gcc.c-torture/compile/nested-3.c":17:10 935
{*thumb1_addsi3}
     (nil))

and peephole2 combines these two insns into

(insn 39 13 16 2 (set (reg:SI 12 ip [127])
        (plus:SI (reg/f:SI 13 sp)
            (const_int 8 [0x8])))
"../../gcc/gcc/testsuite/gcc.c-torture/compile/nested-3.c":17:10 -1
     (nil))

which is wrong as output and the 1st operand of *thumb1_addsi3 should be a low
register but r12 is not a low register.  If peephole2 took this into account,
it would have not combined the 2 insns and we would have not this PR.

[Bug target/113510] [14 Regression] [ARM Thumb] ICE in extract_constrain_insn with CPU cortex-m23

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=113510

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
     Ever confirmed|0                           |1
   Last reconfirmed|                            |2024-01-24
             Status|UNCONFIRMED                 |NEW

--- Comment #5 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
Yes the peephole2 in thumb1.md looks wrong:
```
;; Reloading and elimination of the frame pointer can
;; sometimes cause this optimization to be missed.
(define_peephole2
  [(set (match_operand:SI 0 "arm_general_register_operand" "")
        (match_operand:SI 1 "const_int_operand" ""))
   (set (match_dup 0)
        (plus:SI (match_dup 0) (reg:SI SP_REGNUM)))]
  "TARGET_THUMB1
   && UINTVAL (operands[1]) < 1024
   && (UINTVAL (operands[1]) & 3) == 0"
  [(set (match_dup 0) (plus:SI (reg:SI SP_REGNUM) (match_dup 1)))]
  ""
)
```

Confirmed.

[Bug target/113510] [14 Regression] [ARM Thumb] ICE in extract_constrain_insn with CPU cortex-m23

[rearnsha at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=113510

--- Comment #6 from Richard Earnshaw <rearnsha at gcc dot gnu.org> ---
(In reply to Andrew Pinski from comment #5)
> Yes the peephole2 in thumb1.md looks wrong:
> ```
> ;; Reloading and elimination of the frame pointer can
> ;; sometimes cause this optimization to be missed.
> (define_peephole2
>   [(set (match_operand:SI 0 "arm_general_register_operand" "")
>         (match_operand:SI 1 "const_int_operand" ""))
>    (set (match_dup 0)
>         (plus:SI (match_dup 0) (reg:SI SP_REGNUM)))]
>   "TARGET_THUMB1
>    && UINTVAL (operands[1]) < 1024
>    && (UINTVAL (operands[1]) & 3) == 0"
>   [(set (match_dup 0) (plus:SI (reg:SI SP_REGNUM) (match_dup 1)))]
>   ""
> )
> ```
> 
> Confirmed.

Since this is a peephole and we're dealing with hard regs, we can just use
"low_register_operand" as the predicate for operand 0.

[Bug target/113510] [14 Regression] [ARM Thumb] ICE in extract_constrain_insn with CPU cortex-m23

[rearnsha at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=113510

Richard Earnshaw <rearnsha at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Assignee|unassigned at gcc dot gnu.org      |rearnsha at gcc dot gnu.org
             Status|NEW                         |ASSIGNED

--- Comment #7 from Richard Earnshaw <rearnsha at gcc dot gnu.org> ---
mine

[Bug target/113510] [14 Regression] [ARM Thumb] ICE in extract_constrain_insn with CPU cortex-m23

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=113510

--- Comment #8 from GCC Commits <cvs-commit at gcc dot gnu.org> ---
The master branch has been updated by Richard Earnshaw <rearnsha@gcc.gnu.org>:

https://gcc.gnu.org/g:067a012bde15bfb62d9af309d9d524ebfe91b705

commit r14-9322-g067a012bde15bfb62d9af309d9d524ebfe91b705
Author: Richard Earnshaw <rearnsha@arm.com>
Date:   Tue Mar 5 17:21:43 2024 +0000

    arm: check for low register before applying peephole [PR113510]

    For thumb1, when using a peephole to fuse

            mov reg, #const
            add reg, reg, SP

    into

            add reg, SP, #const

    we must first check that reg is a low register, otherwise we will ICE
    when trying to recognize the resulting insn.

    gcc/ChangeLog:

            PR target/113510
            * config/arm/thumb1.md (peephole2 to fuse mov imm/add SP): Use
            low_register_operand.

[Bug target/113510] [14 Regression] [ARM Thumb] ICE in extract_constrain_insn with CPU cortex-m23

[rearnsha at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=113510

Richard Earnshaw <rearnsha at gcc dot gnu.org> changed:

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

--- Comment #9 from Richard Earnshaw <rearnsha at gcc dot gnu.org> ---
Fixed