diff options
author | Uros Bizjak <ubizjak@gmail.com> | 2022-01-03 20:58:16 +0100 |
---|---|---|
committer | Uros Bizjak <ubizjak@gmail.com> | 2022-01-03 20:59:19 +0100 |
commit | 1096ab1775636f35de9c6661f8f71f03299af998 (patch) | |
tree | 36ae2da3429df604850a758fff00d4cc23e04c1d /gcc/config | |
parent | 122a75488c98b898df1a6f2ed4e48836a5d72d7a (diff) | |
download | gcc-1096ab1775636f35de9c6661f8f71f03299af998.zip gcc-1096ab1775636f35de9c6661f8f71f03299af998.tar.gz gcc-1096ab1775636f35de9c6661f8f71f03299af998.tar.bz2 |
i386: Always enable mov<V_32:mode> patterns [PR103894]
Middle end tries to generate V4QImode moves to implement V2QImode inserts
and calls emit_move_multi_word when V4QImode moves are unavailable, as is
the case with 32-bit vector moves, constrainted with TARGET_SSE2.
However, this triggers
gcc_assert (mode_size >= UNITS_PER_WORD);
in emit_move_multi_word, since mode_size of V4QImode operand is less than
UNITS_PER_WORD of 64-bit targets.
The patch unconditionally enables 32-bit vector moves to match 16-bit
vector moves. This also enables implementation of 32-bit vector logic
operations with GPR in a follow-up patch.
2022-01-03 Uroš Bizjak <ubizjak@gmail.com>
gcc/ChangeLog:
PR target/103894
* config/i386/mmx.md (mov<V_32:mode>): Remove TARGET_SSE2 constraint.
(mov<V_32:mode>_internal): Ditto.
(*push<V_32:mode>_rex64): Ditto.
(movmisalign<V_32:mode>): Ditto.
(*push<V_32:mode>_rex64 splitter): Enable for
TARGET_64BIT && TARGET_SSE.
(*push<V_32:mode>2): Remove insn pattern.
gcc/testsuite/ChangeLog:
PR target/103894
* gcc.target/i386/pr103894.c: New test.
Diffstat (limited to 'gcc/config')
-rw-r--r-- | gcc/config/i386/mmx.md | 23 |
1 files changed, 6 insertions, 17 deletions
diff --git a/gcc/config/i386/mmx.md b/gcc/config/i386/mmx.md index 67b0266..5b33d3c 100644 --- a/gcc/config/i386/mmx.md +++ b/gcc/config/i386/mmx.md @@ -250,7 +250,7 @@ (define_expand "mov<mode>" [(set (match_operand:V_32 0 "nonimmediate_operand") (match_operand:V_32 1 "nonimmediate_operand"))] - "TARGET_SSE2" + "" { ix86_expand_vector_move (<MODE>mode, operands); DONE; @@ -261,8 +261,7 @@ "=r ,m ,v,v,v,m,r,v") (match_operand:V_32 1 "general_operand" "rmC,rC,C,v,m,v,v,r"))] - "TARGET_SSE2 - && !(MEM_P (operands[0]) && MEM_P (operands[1]))" + "!(MEM_P (operands[0]) && MEM_P (operands[1]))" { switch (get_attr_type (insn)) { @@ -321,29 +320,19 @@ (define_insn "*push<mode>2_rex64" [(set (match_operand:V_32 0 "push_operand" "=X,X") (match_operand:V_32 1 "nonmemory_no_elim_operand" "rC,*v"))] - "TARGET_SSE2 && TARGET_64BIT" + "TARGET_64BIT" "@ push{q}\t%q1 #" [(set_attr "type" "push,multi") (set_attr "mode" "DI")]) -(define_insn "*push<mode>2" - [(set (match_operand:V_32 0 "push_operand" "=<,<") - (match_operand:V_32 1 "general_no_elim_operand" "rC*m,*v"))] - "TARGET_SSE2 && !TARGET_64BIT" - "@ - push{l}\t%1 - #" - [(set_attr "type" "push,multi") - (set_attr "mode" "SI")]) - (define_split [(set (match_operand:V_32 0 "push_operand") (match_operand:V_32 1 "sse_reg_operand"))] - "TARGET_SSE2 && reload_completed" + "TARGET_64BIT && TARGET_SSE && reload_completed" [(set (reg:P SP_REG) (plus:P (reg:P SP_REG) (match_dup 2))) - (set (match_dup 0) (match_dup 1))] + (set (match_dup 0) (match_dup 1))] { operands[2] = GEN_INT (-PUSH_ROUNDING (GET_MODE_SIZE (<V_32:MODE>mode))); /* Preserve memory attributes. */ @@ -353,7 +342,7 @@ (define_expand "movmisalign<mode>" [(set (match_operand:V_32 0 "nonimmediate_operand") (match_operand:V_32 1 "nonimmediate_operand"))] - "TARGET_SSE2" + "" { ix86_expand_vector_move (<MODE>mode, operands); DONE; |