On Fri, 3 Mar 2017, Jakub Jelinek wrote:
> On Fri, Mar 03, 2017 at 09:21:55AM +0100, Richard Biener wrote:
> > > --- gcc/gimple-fold.c.jj 2017-02-07 16:40:45.000000000 +0100
> > > +++ gcc/gimple-fold.c 2017-03-02 16:04:51.304850077 +0100
> > > @@ -3533,6 +3533,8 @@ fold_builtin_atomic_compare_exchange (gi
> > > tree itype = TREE_VALUE (TREE_CHAIN (TREE_CHAIN (parmt)));
> > > tree ctype = build_complex_type (itype);
> > > tree expected = TREE_OPERAND (gimple_call_arg (stmt, 1), 0);
> > > + bool may_throw = false;
> > > + edge e = NULL;
> > > gimple *g = gimple_build_assign (make_ssa_name (TREE_TYPE (expected)),
> > > expected);
> > > gsi_insert_before (gsi, g, GSI_SAME_STMT);
> > > @@ -3558,19 +3560,43 @@ fold_builtin_atomic_compare_exchange (gi
> > > gimple_set_vdef (g, gimple_vdef (stmt));
> > > gimple_set_vuse (g, gimple_vuse (stmt));
> > > SSA_NAME_DEF_STMT (gimple_vdef (g)) = g;
> > > - if (gimple_call_lhs (stmt))
> > > + tree oldlhs = gimple_call_lhs (stmt);
> > > + if (flag_non_call_exceptions && stmt_ends_bb_p (stmt))
> >
> > I think a more appropriate check is stmt_can_throw_internal (stmt)
>
> Ok. I suppose without "flag_non_call_exceptions && ".
Yeah.
> > > {
> > > - gsi_insert_before (gsi, g, GSI_SAME_STMT);
> > > + may_throw = true;
> >
> > and 'throws' rather than 'may_throw'.
>
> Ack.
>
> > > + gimple_call_set_lhs (stmt, NULL_TREE);
> > > + gsi_replace (gsi, g, true);
> > > + e = find_fallthru_edge (gsi_bb (*gsi)->succs);
> > > + }
> > > + gimple_call_set_nothrow (as_a <gcall *> (g), flag_non_call_exceptions
> > > == 0);
> >
> > it should copy nothrow state from the source(s) of the folding if you
> > consider flag_non_call_exceptions == true in this fn but working on
> > inlined code from a flag_non_call_exceptions == false function. At least
> > for calls in the source(s) that works, for other stmts I think we do
> > not have an explicit nothrow on the stmt.
>
> So just
> gimple_call_set_nothrow (as_a <gcall *> (g), !throws);
> then?
Hmm, no. That will result in stmt_can_throw_external to return false.
Simply
gimple_call_set_nothrow (g, gimple_call_nothrow (orig-call))
> > > + if (oldlhs)
> > > + {
> > > + if (!may_throw)
> > > + gsi_insert_before (gsi, g, GSI_SAME_STMT);
> > > g = gimple_build_assign (make_ssa_name (itype), IMAGPART_EXPR,
> > > build1 (IMAGPART_EXPR, itype, lhs));
> > > - gsi_insert_before (gsi, g, GSI_SAME_STMT);
> > > - g = gimple_build_assign (gimple_call_lhs (stmt), NOP_EXPR,
> > > - gimple_assign_lhs (g));
> > > + if (may_throw)
> > > + {
> > > + gsi_insert_on_edge_immediate (e, g);
> > > + *gsi = gsi_for_stmt (g);
> > > + }
> > > + else
> > > + gsi_insert_before (gsi, g, GSI_SAME_STMT);
> > > + g = gimple_build_assign (oldlhs, NOP_EXPR, gimple_assign_lhs (g));
> > > + if (may_throw)
> > > + gsi_insert_after (gsi, g, GSI_NEW_STMT);
> > > }
> > > - gsi_replace (gsi, g, true);
> > > + if (!may_throw)
> > > + gsi_replace (gsi, g, true);
> > > g = gimple_build_assign (make_ssa_name (itype), REALPART_EXPR,
> > > build1 (REALPART_EXPR, itype, lhs));
> > > - gsi_insert_after (gsi, g, GSI_NEW_STMT);
> > > + if (may_throw && oldlhs == NULL_TREE)
> > > + {
> > > + gsi_insert_on_edge_immediate (e, g);
> > > + *gsi = gsi_for_stmt (g);
> > > + }
> > > + else
> > > + gsi_insert_after (gsi, g, GSI_NEW_STMT);
> >
> > rather than these dances with if (may_throw) isn't it easier to
> > compute an insert gsi after finding the fallthru edge, splitting
> > the edge if it is a critical one? (basically doing a less
> > fancy gimple_find_edge_insert_loc)
>
> I can't see how that could simplify the code, gimple_find_edge_insert_loc
> is lots of code, not exported and can return ins_after/!ins_after which
> afterwards needs to be honored.
>
> I've simplified the code a little bit by doing gsi_replace early, then
> the difference is pretty much just:
> - gsi_insert_after (gsi, g, GSI_NEW_STMT);
> + if (throws)
> + {
> + gsi_insert_on_edge_immediate (e, g);
> + *gsi = gsi_for_stmt (g);
> + }
> + else
> + gsi_insert_after (gsi, g, GSI_NEW_STMT);
> that has to be done on the first insertion after the call (which
> unfortunately means two locations). Another option I see is to
> gimple_seq_add_stmt_without_update all the stmts following the call
> into a seq and then conditionally insert that seq on the edge (immediate)
> or insert the seq after the call. It would make sense only for
> gimple-fold.c, in tree-ssa-ccp.c there is just one stmt added this way,
> the debug stmt needs extra care.
As you prefer.
Looks good apart from the set_nothrow issue above.
Thanks,
Richard.
> 2017-03-03 Jakub Jelinek <ja...@redhat.com>
>
> PR middle-end/79805
> * internal-fn.def (ATOMIC_BIT_TEST_AND_SET, ATOMIC_BIT_TEST_AND_RESET,
> ATOMIC_BIT_TEST_AND_COMPLEMENT, ATOMIC_COMPARE_EXCHANGE): Remove
> ECF_NOTHROW.
> * gimple-fold.c (fold_builtin_atomic_compare_exchange): Set
> gimple_call_nothrow_p flag based on whether original builtin can throw.
> If it can, emit following stmts on the fallthrough edge.
> * tree-ssa-ccp.c (optimize_atomic_bit_test_and): Similarly, except
> don't create new bb if inserting just debug stmts on the edge, try to
> insert them on the fallthru bb or just reset debug stmts.
>
> * g++.dg/opt/pr79805.C: New test.
>
> --- gcc/internal-fn.def.jj 2017-03-02 16:41:12.723121784 +0100
> +++ gcc/internal-fn.def 2017-03-03 11:48:58.225451010 +0100
> @@ -205,11 +205,13 @@ DEF_INTERNAL_FN (GOACC_TILE, ECF_NOTHROW
> current target. */
> DEF_INTERNAL_FN (SET_EDOM, ECF_LEAF | ECF_NOTHROW, NULL)
>
> -/* Atomic functions. */
> -DEF_INTERNAL_FN (ATOMIC_BIT_TEST_AND_SET, ECF_LEAF | ECF_NOTHROW, NULL)
> -DEF_INTERNAL_FN (ATOMIC_BIT_TEST_AND_COMPLEMENT, ECF_LEAF | ECF_NOTHROW,
> NULL)
> -DEF_INTERNAL_FN (ATOMIC_BIT_TEST_AND_RESET, ECF_LEAF | ECF_NOTHROW, NULL)
> -DEF_INTERNAL_FN (ATOMIC_COMPARE_EXCHANGE, ECF_LEAF | ECF_NOTHROW, NULL)
> +/* Atomic functions. These don't have ECF_NOTHROW because for
> + -fnon-call-exceptions they can throw, otherwise we set
> + gimple_call_nothrow_p on it. */
> +DEF_INTERNAL_FN (ATOMIC_BIT_TEST_AND_SET, ECF_LEAF, NULL)
> +DEF_INTERNAL_FN (ATOMIC_BIT_TEST_AND_COMPLEMENT, ECF_LEAF, NULL)
> +DEF_INTERNAL_FN (ATOMIC_BIT_TEST_AND_RESET, ECF_LEAF, NULL)
> +DEF_INTERNAL_FN (ATOMIC_COMPARE_EXCHANGE, ECF_LEAF, NULL)
>
> /* To implement [[fallthrough]]. */
> DEF_INTERNAL_FN (FALLTHROUGH, ECF_LEAF | ECF_NOTHROW, NULL)
> --- gcc/gimple-fold.c.jj 2017-03-02 16:41:12.745121494 +0100
> +++ gcc/gimple-fold.c 2017-03-03 11:52:59.679219649 +0100
> @@ -3533,6 +3533,8 @@ fold_builtin_atomic_compare_exchange (gi
> tree itype = TREE_VALUE (TREE_CHAIN (TREE_CHAIN (parmt)));
> tree ctype = build_complex_type (itype);
> tree expected = TREE_OPERAND (gimple_call_arg (stmt, 1), 0);
> + bool throws = false;
> + edge e = NULL;
> gimple *g = gimple_build_assign (make_ssa_name (TREE_TYPE (expected)),
> expected);
> gsi_insert_before (gsi, g, GSI_SAME_STMT);
> @@ -3558,19 +3560,38 @@ fold_builtin_atomic_compare_exchange (gi
> gimple_set_vdef (g, gimple_vdef (stmt));
> gimple_set_vuse (g, gimple_vuse (stmt));
> SSA_NAME_DEF_STMT (gimple_vdef (g)) = g;
> - if (gimple_call_lhs (stmt))
> + tree oldlhs = gimple_call_lhs (stmt);
> + if (stmt_can_throw_internal (stmt))
> + {
> + throws = true;
> + e = find_fallthru_edge (gsi_bb (*gsi)->succs);
> + }
> + gimple_call_set_nothrow (as_a <gcall *> (g), !throws);
> + gimple_call_set_lhs (stmt, NULL_TREE);
> + gsi_replace (gsi, g, true);
> + if (oldlhs)
> {
> - gsi_insert_before (gsi, g, GSI_SAME_STMT);
> g = gimple_build_assign (make_ssa_name (itype), IMAGPART_EXPR,
> build1 (IMAGPART_EXPR, itype, lhs));
> - gsi_insert_before (gsi, g, GSI_SAME_STMT);
> - g = gimple_build_assign (gimple_call_lhs (stmt), NOP_EXPR,
> - gimple_assign_lhs (g));
> + if (throws)
> + {
> + gsi_insert_on_edge_immediate (e, g);
> + *gsi = gsi_for_stmt (g);
> + }
> + else
> + gsi_insert_after (gsi, g, GSI_NEW_STMT);
> + g = gimple_build_assign (oldlhs, NOP_EXPR, gimple_assign_lhs (g));
> + gsi_insert_after (gsi, g, GSI_NEW_STMT);
> }
> - gsi_replace (gsi, g, true);
> g = gimple_build_assign (make_ssa_name (itype), REALPART_EXPR,
> build1 (REALPART_EXPR, itype, lhs));
> - gsi_insert_after (gsi, g, GSI_NEW_STMT);
> + if (throws && oldlhs == NULL_TREE)
> + {
> + gsi_insert_on_edge_immediate (e, g);
> + *gsi = gsi_for_stmt (g);
> + }
> + else
> + gsi_insert_after (gsi, g, GSI_NEW_STMT);
> if (!useless_type_conversion_p (TREE_TYPE (expected), itype))
> {
> g = gimple_build_assign (make_ssa_name (TREE_TYPE (expected)),
> --- gcc/tree-ssa-ccp.c.jj 2017-03-02 16:41:12.686122271 +0100
> +++ gcc/tree-ssa-ccp.c 2017-03-03 11:54:07.707309233 +0100
> @@ -2890,9 +2890,18 @@ optimize_atomic_bit_test_and (gimple_stm
> gimple_set_location (g, gimple_location (call));
> gimple_set_vuse (g, gimple_vuse (call));
> gimple_set_vdef (g, gimple_vdef (call));
> + bool throws = stmt_can_throw_internal (call);
> + gimple_call_set_nothrow (as_a <gcall *> (g), !throws);
> SSA_NAME_DEF_STMT (gimple_vdef (call)) = g;
> gimple_stmt_iterator gsi = *gsip;
> gsi_insert_after (&gsi, g, GSI_NEW_STMT);
> + edge e = NULL;
> + if (throws)
> + {
> + maybe_clean_or_replace_eh_stmt (call, g);
> + if (after || (use_bool && has_debug_uses))
> + e = find_fallthru_edge (gsi_bb (gsi)->succs);
> + }
> if (after)
> {
> /* The internal function returns the value of the specified bit
> @@ -2905,23 +2914,42 @@ optimize_atomic_bit_test_and (gimple_stm
> use_bool ? build_int_cst (TREE_TYPE (lhs), 1)
> : mask);
> new_lhs = gimple_assign_lhs (g);
> - gsi_insert_after (&gsi, g, GSI_NEW_STMT);
> + if (throws)
> + {
> + gsi_insert_on_edge_immediate (e, g);
> + gsi = gsi_for_stmt (g);
> + }
> + else
> + gsi_insert_after (&gsi, g, GSI_NEW_STMT);
> }
> if (use_bool && has_debug_uses)
> {
> - tree temp = make_node (DEBUG_EXPR_DECL);
> - DECL_ARTIFICIAL (temp) = 1;
> - TREE_TYPE (temp) = TREE_TYPE (lhs);
> - SET_DECL_MODE (temp, TYPE_MODE (TREE_TYPE (lhs)));
> - tree t = build2 (LSHIFT_EXPR, TREE_TYPE (lhs), new_lhs, bit);
> - g = gimple_build_debug_bind (temp, t, g);
> - gsi_insert_after (&gsi, g, GSI_NEW_STMT);
> + tree temp = NULL_TREE;
> + if (!throws || after || single_pred_p (e->dest))
> + {
> + temp = make_node (DEBUG_EXPR_DECL);
> + DECL_ARTIFICIAL (temp) = 1;
> + TREE_TYPE (temp) = TREE_TYPE (lhs);
> + SET_DECL_MODE (temp, TYPE_MODE (TREE_TYPE (lhs)));
> + tree t = build2 (LSHIFT_EXPR, TREE_TYPE (lhs), new_lhs, bit);
> + g = gimple_build_debug_bind (temp, t, g);
> + if (throws && !after)
> + {
> + gsi = gsi_after_labels (e->dest);
> + gsi_insert_before (&gsi, g, GSI_SAME_STMT);
> + }
> + else
> + gsi_insert_after (&gsi, g, GSI_NEW_STMT);
> + }
> FOR_EACH_IMM_USE_STMT (g, iter, use_lhs)
> if (is_gimple_debug (g))
> {
> use_operand_p use_p;
> - FOR_EACH_IMM_USE_ON_STMT (use_p, iter)
> - SET_USE (use_p, temp);
> + if (temp == NULL_TREE)
> + gimple_debug_bind_reset_value (g);
> + else
> + FOR_EACH_IMM_USE_ON_STMT (use_p, iter)
> + SET_USE (use_p, temp);
> update_stmt (g);
> }
> }
> --- gcc/testsuite/g++.dg/opt/pr79805.C.jj 2017-03-03 11:48:58.228450970
> +0100
> +++ gcc/testsuite/g++.dg/opt/pr79805.C 2017-03-03 11:48:58.228450970
> +0100
> @@ -0,0 +1,219 @@
> +// PR middle-end/79805
> +// { dg-do compile }
> +// { dg-options "-O2 -fnon-call-exceptions" }
> +
> +struct A { A (); ~A (); };
> +
> +void bar (void);
> +
> +int
> +f0 (int *d, int f)
> +{
> + A z;
> + int e = __atomic_compare_exchange_n (d, &f, 1, 1, __ATOMIC_RELAXED,
> __ATOMIC_RELAXED);
> + return e;
> +}
> +
> +int
> +f1 (int *a, int bit)
> +{
> + A z;
> + unsigned int mask = (1u << bit);
> + return (__sync_fetch_and_or (a, mask) & mask) != 0;
> +}
> +
> +int
> +f2 (int *a, int bit)
> +{
> + A z;
> + unsigned int mask = (1u << bit);
> + unsigned int t1 = __atomic_fetch_or (a, mask, __ATOMIC_RELAXED);
> + unsigned int t2 = t1 & mask;
> + return t2 != 0;
> +}
> +
> +long int
> +f3 (long int *a, int bit)
> +{
> + A z;
> + unsigned long int mask = (1ul << bit);
> + return (__atomic_fetch_or (a, mask, __ATOMIC_SEQ_CST) & mask) == 0;
> +}
> +
> +int
> +f4 (int *a)
> +{
> + A z;
> + unsigned int mask = (1u << 7);
> + return (__sync_fetch_and_or (a, mask) & mask) != 0;
> +}
> +
> +int
> +f5 (int *a)
> +{
> + A z;
> + unsigned int mask = (1u << 13);
> + return (__atomic_fetch_or (a, mask, __ATOMIC_RELAXED) & mask) != 0;
> +}
> +
> +int
> +f6 (int *a)
> +{
> + A z;
> + unsigned int mask = (1u << 0);
> + return (__atomic_fetch_or (a, mask, __ATOMIC_SEQ_CST) & mask) != 0;
> +}
> +
> +void
> +f7 (int *a, int bit)
> +{
> + A z;
> + unsigned int mask = (1u << bit);
> + if ((__sync_fetch_and_xor (a, mask) & mask) != 0)
> + bar ();
> +}
> +
> +void
> +f8 (int *a, int bit)
> +{
> + A z;
> + unsigned int mask = (1u << bit);
> + if ((__atomic_fetch_xor (a, mask, __ATOMIC_RELAXED) & mask) == 0)
> + bar ();
> +}
> +
> +int
> +f9 (int *a, int bit)
> +{
> + A z;
> + unsigned int mask = (1u << bit);
> + return (__atomic_fetch_xor (a, mask, __ATOMIC_SEQ_CST) & mask) != 0;
> +}
> +
> +int
> +f10 (int *a)
> +{
> + A z;
> + unsigned int mask = (1u << 7);
> + return (__sync_fetch_and_xor (a, mask) & mask) != 0;
> +}
> +
> +int
> +f11 (int *a)
> +{
> + A z;
> + unsigned int mask = (1u << 13);
> + return (__atomic_fetch_xor (a, mask, __ATOMIC_RELAXED) & mask) != 0;
> +}
> +
> +int
> +f12 (int *a)
> +{
> + A z;
> + unsigned int mask = (1u << 0);
> + return (__atomic_fetch_xor (a, mask, __ATOMIC_SEQ_CST) & mask) != 0;
> +}
> +
> +int
> +f13 (int *a, int bit)
> +{
> + A z;
> + unsigned int mask = (1u << bit);
> + return (__sync_fetch_and_and (a, ~mask) & mask) != 0;
> +}
> +
> +int
> +f14 (int *a, int bit)
> +{
> + A z;
> + unsigned int mask = (1u << bit);
> + return (__atomic_fetch_and (a, ~mask, __ATOMIC_RELAXED) & mask) != 0;
> +}
> +
> +int
> +f15 (int *a, int bit)
> +{
> + A z;
> + unsigned int mask = (1u << bit);
> + return (__atomic_fetch_and (a, ~mask, __ATOMIC_SEQ_CST) & mask) != 0;
> +}
> +
> +int
> +f16 (int *a)
> +{
> + A z;
> + unsigned int mask = (1u << 7);
> + return (__sync_fetch_and_and (a, ~mask) & mask) != 0;
> +}
> +
> +int
> +f17 (int *a)
> +{
> + A z;
> + unsigned int mask = (1u << 13);
> + return (__atomic_fetch_and (a, ~mask, __ATOMIC_RELAXED) & mask) != 0;
> +}
> +
> +int
> +f18 (int *a)
> +{
> + A z;
> + unsigned int mask = (1u << 0);
> + return (__atomic_fetch_and (a, ~mask, __ATOMIC_SEQ_CST) & mask) != 0;
> +}
> +
> +unsigned long int
> +f19 (unsigned long int *a, int bit)
> +{
> + A z;
> + unsigned long int mask = (1ul << bit);
> + return (__atomic_xor_fetch (a, mask, __ATOMIC_SEQ_CST) & mask) != 0;
> +}
> +
> +unsigned long int
> +f20 (unsigned long int *a)
> +{
> + A z;
> + unsigned long int mask = (1ul << 7);
> + return (__atomic_xor_fetch (a, mask, __ATOMIC_SEQ_CST) & mask) == 0;
> +}
> +
> +int
> +f21 (int *a, int bit)
> +{
> + A z;
> + unsigned int mask = (1u << bit);
> + return (__sync_fetch_and_or (a, mask) & mask);
> +}
> +
> +unsigned long int
> +f22 (unsigned long int *a)
> +{
> + A z;
> + unsigned long int mask = (1ul << 7);
> + return (__atomic_xor_fetch (a, mask, __ATOMIC_SEQ_CST) & mask);
> +}
> +
> +unsigned long int
> +f23 (unsigned long int *a)
> +{
> + A z;
> + unsigned long int mask = (1ul << 7);
> + return (__atomic_fetch_xor (a, mask, __ATOMIC_SEQ_CST) & mask);
> +}
> +
> +unsigned short int
> +f24 (unsigned short int *a)
> +{
> + A z;
> + unsigned short int mask = (1u << 7);
> + return (__sync_fetch_and_or (a, mask) & mask) != 0;
> +}
> +
> +unsigned short int
> +f25 (unsigned short int *a)
> +{
> + A z;
> + unsigned short int mask = (1u << 7);
> + return (__atomic_fetch_or (a, mask, __ATOMIC_SEQ_CST) & mask) != 0;
> +}
>
>
> Jakub
>
>
--
Richard Biener <rguent...@suse.de>
SUSE LINUX GmbH, GF: Felix Imendoerffer, Jane Smithard, Graham Norton, HRB
21284 (AG Nuernberg)
