[Bug rtl-optimization/111971] New: ICE: maximum number of generated reload insns per insn achieved (90)

[Bug rtl-optimization/111971] New: ICE: maximum number of generated reload insns per insn achieved (90)

[guojiufu at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=111971

            Bug ID: 111971
           Summary: ICE: maximum number of generated reload insns per insn
                    achieved (90)
           Product: gcc
           Version: unknown
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: rtl-optimization
          Assignee: unassigned at gcc dot gnu.org
          Reporter: guojiufu at gcc dot gnu.org
  Target Milestone: ---

For the below code, an ICE occurs when built with "-m32 -O2".
```
void
foo (unsigned long long *a)
{
  register long long d asm ("r0") = 0x24;
  long long n;
  asm ("mr %0, %1" : "=r"(n) : "r"(d));
  *a++ = n;
}

```
---------------
    8 | }
      | ^
0x2000007c4ca3 __libc_start_call_main
        ../sysdeps/nptl/libc_start_call_main.h:58
0x2000007c4f07 generic_start_main
        ../csu/libc-start.c:360
0x2000007c4f07 __libc_start_main_impl
        ../sysdeps/unix/sysv/linux/powerpc/libc-start.c:109
Please submit a full bug report, with preprocessed source (by using
-freport-bug).
Please include the complete backtrace with any bug report.
See <file:///usr/share/doc/gcc-13/README.Bugs> for instructions.

----------------

It is ok if it is "register long d asm ("r0") = 0x24;".

The 'd' is 'long long' DImode(64bits), but with -m32, "r0" is not 64bits
without -mpowerpc64.  So, it would say this code would be invalid in some
aspects.

[Bug rtl-optimization/111971] ICE: maximum number of generated reload insns per insn achieved (90)

[guojiufu at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=111971

--- Comment #1 from Jiu Fu Guo <guojiufu at gcc dot gnu.org> ---
This issue can be reproduced on 'ppc64' BE machine with -m32.

[Bug rtl-optimization/111971] ICE: maximum number of generated reload insns per insn achieved (90)

[guojiufu at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=111971

--- Comment #2 from Jiu Fu Guo <guojiufu at gcc dot gnu.org> ---
It seems gcc11 is ok.

[Bug rtl-optimization/111971] ICE: maximum number of generated reload insns per insn achieved (90)

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=111971

--- Comment #3 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
D should be a pair of registers, r0 and r1.

[Bug rtl-optimization/111971] ICE: maximum number of generated reload insns per insn achieved (90)

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=111971

--- Comment #4 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
But r1 is the argument register.

[Bug rtl-optimization/111971] ICE: maximum number of generated reload insns per insn achieved (90)

[guojiufu at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=111971

--- Comment #5 from Jiu Fu Guo <guojiufu at gcc dot gnu.org> ---
With a bisect, the result shows "85419ac59724b7ce710ebb4acf03dbd747edeea3 is
the first bad commit".

[Bug rtl-optimization/111971] [12/13/14 regression] ICE: maximum number of generated reload insns per insn achieved (90) since r12-6803-g85419ac59724b7

[vmakarov at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=111971

--- Comment #6 from Vladimir Makarov <vmakarov at gcc dot gnu.org> ---
(In reply to Andrew Pinski from comment #4)
> But r1 is the argument register.

It is even worse, r1 is a stack pointer.  Still the compilation should not
finish by LRA failure.

I've just started to work on this problem. I hope a patch fixing this will be
committed on this week or at the beginning of the next week.

[Bug rtl-optimization/111971] [12/13/14 regression] ICE: maximum number of generated reload insns per insn achieved (90) since r12-6803-g85419ac59724b7

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=111971

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Target|                            |powerpc
            Version|unknown                     |14.0
           Keywords|                            |ra
   Target Milestone|---                         |12.4

[Bug rtl-optimization/111971] [12/13/14 regression] ICE: maximum number of generated reload insns per insn achieved (90) since r12-6803-g85419ac59724b7

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=111971

--- Comment #7 from CVS Commits <cvs-commit at gcc dot gnu.org> ---
The trunk branch has been updated by Vladimir Makarov <vmakarov@gcc.gnu.org>:

https://gcc.gnu.org/g:9119b008b4195e06012a485ec01a8bb0e43266be

commit r14-5037-g9119b008b4195e06012a485ec01a8bb0e43266be
Author: Vladimir N. Makarov <vmakarov@redhat.com>
Date:   Tue Oct 31 10:54:43 2023 -0400

    [RA]: Fixing LRA cycling for multi-reg variable containing a fixed reg

    PR111971 test case uses a multi-reg variable containing a fixed reg.  LRA
    rejects such multi-reg because of this when matching the constraint for
    an asm insn.  The rejection results in LRA cycling.  The patch fixes this
issue.

    gcc/ChangeLog:

            PR rtl-optimization/111971
            * lra-constraints.cc: (process_alt_operands): Don't check start
            hard regs for regs originated from register variables.

    gcc/testsuite/ChangeLog:

            PR rtl-optimization/111971
            * gcc.target/powerpc/pr111971.c: New test.

[Bug rtl-optimization/111971] [12/13/14 regression] ICE: maximum number of generated reload insns per insn achieved (90) since r12-6803-g85419ac59724b7

[guojiufu at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=111971

Jiu Fu Guo <guojiufu at gcc dot gnu.org> changed:

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

--- Comment #8 from Jiu Fu Guo <guojiufu at gcc dot gnu.org> ---
Fixed in the trunk.