[Bug rtl-optimization/99830] [11 Regression] ICE: in lra_eliminate_regs_1, at lra-eliminations.c:659 with -O2 -fno-expensive-optimizations -fno-split-wide-types -g

2021-04-08 Thread jakub at gcc dot gnu.org via Gcc-bugs 
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=99830

Jakub Jelinek  changed:

   What|Removed |Added

 CC||jakub at gcc dot gnu.org

--- Comment #3 from Jakub Jelinek  ---
In normal insns such clobbers would be rejected by recog, but for DEBUG_INSNs
we don't have strict validity tests, but guess we need to throw away at least
the worst garbage.

[Bug rtl-optimization/99830] [11 Regression] ICE: in lra_eliminate_regs_1, at lra-eliminations.c:659 with -O2 -fno-expensive-optimizations -fno-split-wide-types -g

2021-04-08 Thread rguenth at gcc dot gnu.org via Gcc-bugs 
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=99830

Richard Biener  changed:

   What|Removed |Added

   Priority|P3  |P1
   Keywords||ra
 CC||segher at gcc dot gnu.org

--- Comment #2 from Richard Biener  ---
#10 0x011550f2 in process_insn_for_elimination (insn=0x76577680, 
final_p=true, first_p=false) at ../../src/trunk/gcc/lra-eliminations.c:1333
1333  eliminate_regs_in_insn (insn, final_p, first_p, 0);
(gdb) p debug_rtx (insn)
(debug_insn 18 4 19 2 (var_location:HI u16_1 (subreg:HI (ashiftrt:SI
(sign_extend:SI (subreg:HI (reg/v:SI 100 [ u16_1 ]) 0))
(zero_extend:SI (subreg:QI (ior:TI (and:TI (mem/c:TI (reg/f:DI 65
ap) [1 u128_1+0 S16 A128])
(const_int -16711681 [0xff00]))
(ashift:TI (zero_extend:TI (clobber:TI (const_int 0
[0])))
(const_int 16 [0x10]))) 0))) 0)) "t.c":4:9 -1
 (nil))

and it chokes on the (zero_extend:TI (clobber ...)) which combine introduces.
Before combine it was

(debug_insn 18 17 19 2 (var_location:HI u16_1 (subreg:HI (ashiftrt:SI
(sign_extend:SI (subreg:HI (reg/v:SI 100 [ u16_1 ]) 0))
(zero_extend:SI (subreg:QI (reg/v:TI 103 [ u128_1 ]) 0))) 0))
"t.c":4:9 -1
 (nil))

not sure what goes wrong but I suppose if sth doesn't work out combine should
reset the debug_insn rather than leaving garbage.  Possibly latent of course.

[Bug rtl-optimization/99830] [11 Regression] ICE: in lra_eliminate_regs_1, at lra-eliminations.c:659 with -O2 -fno-expensive-optimizations -fno-split-wide-types -g

2021-03-31 Thread rguenth at gcc dot gnu.org via Gcc-bugs 
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=99830

Richard Biener  changed:

   What|Removed |Added

   Target Milestone|--- |11.0

[Bug rtl-optimization/99830] [11 Regression] ICE: in lra_eliminate_regs_1, at lra-eliminations.c:659 with -O2 -fno-expensive-optimizations -fno-split-wide-types -g

2021-03-30 Thread acoplan at gcc dot gnu.org via Gcc-bugs 
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=99830

Alex Coplan  changed:

   What|Removed |Added

   Last reconfirmed||2021-03-30
 CC||acoplan at gcc dot gnu.org
 Status|UNCONFIRMED |NEW
 Ever confirmed|0   |1

--- Comment #1 from Alex Coplan  ---
Confirmed. Started with r11-5185-gd0d8b5d83614d8f0d0e40c0520d4f40ffa01f8d9 so
must be latent. I'll see if I can recover a testcase that ICEs prior to that
rev.