https://gcc.gnu.org/bugzilla/show_bug.cgi?id=93141
--- Comment #3 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
Untested fix, though this is just about double-word uaddv4, would be good to
handle double-word usubv4, addv4 and subv4 similarly.
--- gcc/config/i386/i386.md.jj 2020-01-03 11:10:43.839511446 +0100
+++ gcc/config/i386/i386.md 2020-01-03 17:05:51.435808734 +0100
@@ -1036,6 +1036,9 @@
(define_mode_iterator DWI [(DI "!TARGET_64BIT")
(TI "TARGET_64BIT")])
+;; SWI and DWI together.
+(define_mode_iterator SWIDWI [QI HI SI DI (TI "TARGET_64BIT")])
+
;; GET_MODE_SIZE for selected modes. As GET_MODE_SIZE is not
;; compile time constant, it is faster to use <MODE_SIZE> than
;; GET_MODE_SIZE (<MODE>mode). For XFmode which depends on
@@ -6121,12 +6124,12 @@
(define_expand "uaddv<mode>4"
[(parallel [(set (reg:CCC FLAGS_REG)
(compare:CCC
- (plus:SWI
- (match_operand:SWI 1 "nonimmediate_operand")
- (match_operand:SWI 2 "<general_operand>"))
+ (plus:SWIDWI
+ (match_operand:SWIDWI 1 "nonimmediate_operand")
+ (match_operand:SWIDWI 2 "<general_operand>"))
(match_dup 1)))
- (set (match_operand:SWI 0 "register_operand")
- (plus:SWI (match_dup 1) (match_dup 2)))])
+ (set (match_operand:SWIDWI 0 "register_operand")
+ (plus:SWIDWI (match_dup 1) (match_dup 2)))])
(set (pc) (if_then_else
(ltu (reg:CCC FLAGS_REG) (const_int 0))
(label_ref (match_operand 3))
@@ -6885,6 +6888,49 @@
[(set_attr "type" "alu")
(set_attr "mode" "SI")])
+(define_insn_and_split "*add<dwi>3_doubleword_cc_overflow_1"
+ [(set (reg:CCC FLAGS_REG)
+ (compare:CCC
+ (plus:<DWI>
+ (match_operand:<DWI> 1 "nonimmediate_operand" "%0,0")
+ (match_operand:<DWI> 2 "x86_64_hilo_general_operand" "r<di>,o"))
+ (match_dup 1)))
+ (set (match_operand:<DWI> 0 "nonimmediate_operand" "=ro,r")
+ (plus:<DWI> (match_dup 1) (match_dup 2)))]
+ "ix86_binary_operator_ok (PLUS, <DWI>mode, operands)"
+ "#"
+ "reload_completed"
+ [(parallel [(set (reg:CCC FLAGS_REG)
+ (compare:CCC
+ (plus:DWIH (match_dup 1) (match_dup 2))
+ (match_dup 1)))
+ (set (match_dup 0)
+ (plus:DWIH (match_dup 1) (match_dup 2)))])
+ (parallel [(set (reg:CCC FLAGS_REG)
+ (compare:CCC
+ (zero_extend:<DWI>
+ (plus:DWIH
+ (plus:DWIH
+ (ltu:DWIH (reg:CC FLAGS_REG) (const_int 0))
+ (match_dup 4))
+ (match_dup 5)))
+ (plus:<DWI>
+ (zero_extend:<DWI> (match_dup 5))
+ (ltu:<DWI> (reg:CC FLAGS_REG) (const_int 0)))))
+ (set (match_dup 3)
+ (plus:DWIH
+ (plus:DWIH (ltu:DWIH (reg:CC FLAGS_REG) (const_int 0))
+ (match_dup 4))
+ (match_dup 5)))])]
+{
+ split_double_mode (<DWI>mode, &operands[0], 3, &operands[0], &operands[3]);
+ if (operands[2] == const0_rtx)
+ {
+ emit_insn (gen_addcarry<mode>_0 (operands[3], operands[4],
operands[5]));
+ DONE;
+ }
+})
+
;; x == 0 with zero flag test can be done also as x < 1U with carry flag
;; test, where the latter is preferrable if we have some carry consuming
;; instruction.
