https://gcc.gnu.org/bugzilla/show_bug.cgi?id=82524
--- Comment #5 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
If I diff the combine dumps before/after this commit, the difference is:
-(insn 32 31 33 2 (parallel [
- (set (reg:SI 160)
- (ashiftrt:SI (reg:SI 159)
- (const_int 8 [0x8])))
- (clobber (reg:CC 17 flags))
- ]) "pr82545.c":16 542 {*ashrsi3_1}
-(insn 36 35 37 2 (parallel [
- (set (reg:SI 164)
- (ashiftrt:SI (reg:SI 163)
- (const_int 8 [0x8])))
- (clobber (reg:CC 17 flags))
- ]) "pr82545.c":16 542 {*ashrsi3_1}
-(insn 37 36 38 2 (parallel [
- (set (reg:QI 165)
- (plus:QI (subreg:QI (reg:SI 160) 0)
- (subreg:QI (reg:SI 164) 0)))
- (clobber (reg:CC 17 flags))
- ]) "pr82545.c":27 214 {*addqi_1}
-(insn 38 37 39 2 (set (zero_extract:SI (reg/v:SI 135 [ result ])
- (const_int 8 [0x8])
- (const_int 8 [0x8]))
- (subreg:SI (reg:QI 165) 0)) "pr82545.c":27 109 {insvsi_1}
vs.
+(note 32 31 33 2 NOTE_INSN_DELETED)
+(note 36 35 37 2 NOTE_INSN_DELETED)
+(note 37 36 38 2 NOTE_INSN_DELETED)
+(insn 38 37 39 2 (parallel [
+ (set (zero_extract:SI (reg/v:SI 135 [ result ])
+ (const_int 8 [0x8])
+ (const_int 8 [0x8]))
+ (subreg:SI (plus:QI (subreg:QI (zero_extract:SI (reg:SI 159)
+ (const_int 8 [0x8])
+ (const_int 8 [0x8])) 0)
+ (subreg:QI (zero_extract:SI (reg:SI 163)
+ (const_int 8 [0x8])
+ (const_int 8 [0x8])) 0)) 0))
+ (clobber (reg:CC 17 flags))
+ ]) "pr82545.c":27 235 {*addqi_ext_2}
which looks fine to me. So there is likely just some later bug.
