[Bug rtl-optimization/126967] New: [avr] missed branch optimization

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=126967

            Bug ID: 126967
           Summary: [avr] missed branch optimization
           Product: gcc
           Version: 17.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: rtl-optimization
          Assignee: unassigned at gcc dot gnu.org
          Reporter: gjl at gcc dot gnu.org
  Target Milestone: ---

char func (char x)
{
    if (x == 0x20) goto L1;
    if (x > 0x20)  goto L2;
    return 0;
L1: __asm ("nop ; 1");
L2: __asm ("nop ; 2");
    return x;
}

$ avr-gcc-17 x.c -S -Os

Generates the following asm:

func:
    cpi r24,lo8(32) ;  42   [c=4 l=1]  cmpqi3/0
    brne .L2        ;  43   [c=4 l=1]  branch
    nop ; 1
.L3:
    nop ; 2
    ret             ;  37   [c=0 l=1]  return
.L2:
    brlt .L5        ;  44   [c=4 l=1]  branch
    rjmp .L3        ;  52   [c=4 l=1]  jump
.L5:
    ldi r24,0       ;  45   [c=4 l=1]  movqi_insn/0
    ret             ;  48   [c=0 l=1]  return

Notice jump insns 44 and 52 that are doing

    if (REG_CC < 0)
      goto .L5
    goto .L3
   .L5:

which would better be:

    if (REG_CC >= 0)
      goto .L3
   .L5:

i.e.

    brge .L3 ; if (REG_CC >= 0) goto .L3
    .L5:

Target: avr
Configured with: ../../source/gcc-master/configure --target=avr --disable-nls
--with-dwarf2 --with-gnu-as --with-gnu-ld --with-long-double=64
--disable-libcc1 --disable-analyzer --enable-languages=c,c++
Thread model: single
Supported LTO compression algorithms: zlib
gcc version 17.0.0 20260817 (experimental) (GCC)