[Bug target/112411] ICE: SIGSEGV with --param=min-nondebug-insn-uid=2147483647 on powerpc64le-unknown-linux-gnu

["linkw at gcc dot gnu.org" <gcc-bugzilla@gcc.gnu.org>]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=112411

Kewen Lin <linkw at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
     Ever confirmed|0                           |1
             Status|UNCONFIRMED                 |NEW
   Last reconfirmed|                            |2023-11-07
                 CC|                            |linkw at gcc dot gnu.org,
                   |                            |rguenth at gcc dot gnu.org,
                   |                            |rsandifo at gcc dot gnu.org

--- Comment #2 from Kewen Lin <linkw at gcc dot gnu.org> ---
Confirmed, can be reproduced with native powerpc64 compiler.

It fails at the code:

 352│   if (insn_lengths_max_uid > INSN_UID (insn))
 353├───> return insn_lengths[INSN_UID (insn)];

p INSN_UID (insn)
$1 = (int &) @0x7ffff5830384: -2147483641

p insn_lengths_max_uid
$2 = 0

p insn_lengths
$3 = (int *) 0x0

The given --param=min-nondebug-insn-uid=2147483647 causes the assigned insn uid
becomes negative. At that time, insn_lengths isn't ready yet, it can workaround
with further checking insn_lengths first.

I noticed that those uid variables are of signed type, such as:

  x_cur_insn_uid @ emit_status
  x_cur_debug_insn_uid @ emit_status
  insn_uid @ rtx_def

shouldn't they be with unsigned type instead? or is it by design?