On Tue, Aug 13, 2024 at 03:12:01PM -0700, Jason Merrill wrote: > On 8/12/24 7:21 PM, Marek Polacek wrote: > > On Fri, Aug 09, 2024 at 05:15:05PM -0400, Jason Merrill wrote: > > > On 8/9/24 4:21 PM, Marek Polacek wrote: > > > > On Fri, Aug 09, 2024 at 12:58:34PM -0400, Jason Merrill wrote: > > > > > On 8/8/24 1:37 PM, Marek Polacek wrote: > > > > > > Bootstrapped/regtested on x86_64-pc-linux-gnu, ok for trunk? > > > > > > > > > > > > -- >8 -- > > > > > > The problem in this PR is that we ended up with > > > > > > > > > > > > {.rows=(&<PLACEHOLDER_EXPR struct Widget>)->n, > > > > > > .outer_stride=(&<PLACEHOLDER_EXPR struct MatrixLayout>)->rows} > > > > > > > > > > > > that is, two PLACEHOLDER_EXPRs for different types on the same level > > > > > > in one { }. That should not happen; we may, for instance, neglect > > > > > > to > > > > > > replace a PLACEHOLDER_EXPR due to CONSTRUCTOR_PLACEHOLDER_BOUNDARY > > > > > > on > > > > > > the constructor. > > > > > > > > > > > > The same problem happened in PR100252, which I fixed by introducing > > > > > > replace_placeholders_for_class_temp_r. That didn't work here, > > > > > > though, > > > > > > because r_p_for_c_t_r only works for non-eliding TARGET_EXPRs: > > > > > > replacing > > > > > > a PLACEHOLDER_EXPR with a temporary that is going to be elided will > > > > > > result in a crash in gimplify_var_or_parm_decl when it encounters > > > > > > such > > > > > > a loose decl. > > > > > > > > > > > > But leaving the PLACEHOLDER_EXPRs in is also bad because then we end > > > > > > up with this PR. > > > > > > > > > > > > TARGET_EXPRs for function arguments are elided in gimplify_arg. The > > > > > > argument will get a real temporary only in get_formal_tmp_var. One > > > > > > idea was to use the temporary that is going to be elided anyway, and > > > > > > then replace_decl it with the real object once we get it. But that > > > > > > didn't work out: one problem is that we elide the TARGET_EXPR for an > > > > > > argument before we create the real temporary for the argument, and > > > > > > when we get it, the context that this was a TARGET_EXPR for an > > > > > > argument > > > > > > has been lost. We're also in the middle end territory now, even > > > > > > though > > > > > > this is a C++-specific problem. > > > > > > > > > > How complex! > > > > > > > > > > > I figured that since the to-be-elided temporary is going to stay > > > > > > around > > > > > > until gimplification, the front end is free to use it. Once we're > > > > > > done > > > > > > with things like store_init_value, which replaces PLACEHOLDER_EXPRs > > > > > > with > > > > > > the decl it is initializing, we can turn those to-be-elided > > > > > > temporaries > > > > > > into PLACEHOLDER_EXPRs again, so that cp_gimplify_init_expr can > > > > > > replace > > > > > > them with the real object once available. The context is not lost > > > > > > so we > > > > > > do not need an extra flag for these makeshift temporaries. > > > > > > > > > > Clever, that makes a lot of sense. But I wonder if we can avoid the > > > > > problem > > > > > more simply than working around it? > > > > > > > > > > I see that the get_formal_tmp_var happens directly from gimplify_arg, > > > > > so it > > > > > strips the TARGET_EXPR to avoid a temporary...and then immediately > > > > > turns > > > > > around and creates a new temporary. > > > > > > > > > > Would it work to stop stripping the TARGET_EXPR in gimplify_arg and > > > > > therefore stop setting TARGET_EXPR_ELIDING_P in > > > > > convert_for_arg_passing? > > > > > > > > Well, it does fix the ICE. But then a number of testcases fail :(. > > > > For instance, pr23372.C. .gimple diff w/ and w/o stripping the > > > > TARGET_EXPR: > > > > > > > > @@ -1,6 +1,9 @@ > > > > void g (struct A * a) > > > > { > > > > - f (MEM[(const struct A &)a]); > > > > + struct A D.2829; > > > > + > > > > + D.2829 = MEM[(const struct A &)a]; > > > > + f (D.2829); > > > > } > > > > > > > > The extra copy is there even in .optimized with -O2. > > > > > > > > > > > > It's always sad when we have to add complicated code just to work around > > > > a corner case, but the above pessimization looks pretty important :(. > > > > > > Ah, good point. In that case, the stripping avoids the copy because the > > > TARGET_EXPR_INITIAL is already (adjustable into) a suitable lvalue. The > > > current code already fails to avoid the redundant copy when _INITIAL is a > > > CONSTRUCTOR: > > > > > > void g (struct A * a) > > > { > > > struct A D.2805; > > > > > > D.2805 = {}; // boo > > > f (D.2805); > > > } > > > > > > I'm failing to find the PR about this issue. > > > > I also haven't found it (that doesn't mean it doesn't exist :)). I can file > > one if you'd like... > > > > Note that if we do fix that, we may be facing this problem again. > > Please do. The way I would expect the bug to get fixed would be to assign > the temporary the location of the argument slot, and then recognize that the > copy is a no-op.
https://gcc.gnu.org/PR116375 > > > How about the change I mentioned only in the CONSTRUCTOR case? > > > > That works! > > > > Bootstrapped/regtested on x86_64-pc-linux-gnu, ok for trunk? > > > > -- >8 -- > > The problem in this PR is that we ended up with > > > > {.rows=(&<PLACEHOLDER_EXPR struct Widget>)->n, > > .outer_stride=(&<PLACEHOLDER_EXPR struct MatrixLayout>)->rows} > > > > that is, two PLACEHOLDER_EXPRs for different types on the same level > > in one { }. That should not happen; we may, for instance, neglect to > > replace a PLACEHOLDER_EXPR due to CONSTRUCTOR_PLACEHOLDER_BOUNDARY on > > the constructor. > > > > The same problem happened in PR100252, which I fixed by introducing > > replace_placeholders_for_class_temp_r. That didn't work here, though, > > because r_p_for_c_t_r only works for non-eliding TARGET_EXPRs: replacing > > a PLACEHOLDER_EXPR with a temporary that is going to be elided will > > result in a crash in gimplify_var_or_parm_decl when it encounters such > > a loose decl. > > > > But leaving the PLACEHOLDER_EXPRs in is also bad because then we end > > up with this PR. > > > > TARGET_EXPRs for function arguments are elided in gimplify_arg. The > > argument will get a real temporary only in get_formal_tmp_var. One > > idea was to use the temporary that is going to be elided anyway, and > > then replace_decl it with the real object once we get it. But that > > didn't work out: one problem is that we elide the TARGET_EXPR for an > > argument before we create the real temporary for the argument, and > > when we get it, the context that this was a TARGET_EXPR for an argument > > has been lost. We're also in the middle end territory now, even though > > this is a C++-specific problem. > > > > A solution is to simply stop eliding TARGET_EXPRs whose initializer is > > a CONSTRUCTOR. Such copies can't be (at the moment) elided anyway. But > > not eliding all TARGET_EXPRs would be a pessimization. > > > > PR c++/116015 > > > > gcc/cp/ChangeLog: > > > > * call.cc (convert_for_arg_passing): Don't set_target_expr_eliding > > when the TARGET_EXPR initializer is a CONSTRUCTOR. > > > > gcc/ChangeLog: > > > > * gimplify.cc (gimplify_arg): Do not strip a TARGET_EXPR whose > > initializer is a CONSTRUCTOR. > > > > gcc/testsuite/ChangeLog: > > > > * g++.dg/cpp1y/nsdmi-aggr23.C: New test. > > --- > > gcc/cp/call.cc | 7 ++++-- > > gcc/gimplify.cc | 14 +++++++++++- > > gcc/testsuite/g++.dg/cpp1y/nsdmi-aggr23.C | 26 +++++++++++++++++++++++ > > 3 files changed, 44 insertions(+), 3 deletions(-) > > create mode 100644 gcc/testsuite/g++.dg/cpp1y/nsdmi-aggr23.C > > > > diff --git a/gcc/cp/call.cc b/gcc/cp/call.cc > > index 94015db4e65..0fe679aae9f 100644 > > --- a/gcc/cp/call.cc > > +++ b/gcc/cp/call.cc > > @@ -9468,8 +9468,11 @@ convert_for_arg_passing (tree type, tree val, > > tsubst_flags_t complain) > > if (complain & tf_warning) > > warn_for_address_of_packed_member (type, val); > > - /* gimplify_arg elides TARGET_EXPRs that initialize a function argument. > > */ > > - if (SIMPLE_TARGET_EXPR_P (val)) > > + /* gimplify_arg elides TARGET_EXPRs that initialize a function argument, > > + unless the initializer is a CONSTRUCTOR. In that case, we fail to > > + elide the copy anyway. See that function for more information. */ > > + if (SIMPLE_TARGET_EXPR_P (val) > > + && TREE_CODE (TARGET_EXPR_INITIAL (val)) != CONSTRUCTOR) > > set_target_expr_eliding (val); > > return val; > > diff --git a/gcc/gimplify.cc b/gcc/gimplify.cc > > index 71cc6c38d80..37d0d3a59f5 100644 > > --- a/gcc/gimplify.cc > > +++ b/gcc/gimplify.cc > > @@ -3760,7 +3760,19 @@ gimplify_arg (tree *arg_p, gimple_seq *pre_p, > > location_t call_location, > > { > > tree init = TARGET_EXPR_INITIAL (*arg_p); > > if (init > > - && !VOID_TYPE_P (TREE_TYPE (init))) > > + && !VOID_TYPE_P (TREE_TYPE (init)) > > + /* Currently, due to c++/116015, it is not desirable to > > + strip a TARGET_EXPR whose initializer is a {}. The > > + problem is that if we do elide it, we also have to > > + replace all the occurrences of the slot temporary in the > > + initializer with the temporary created for the argument. > > + But we do not have that temporary yet so the replacement > > + would be quite awkward and it might be needed to resort > > + back to a PLACEHOLDER_EXPR. > > + > > + See convert_for_arg_passing for the C++ code that marks > > + the TARGET_EXPR as eliding or not. */ > > Please also mention that stripping the TARGET_EXPR wouldn't help anyway in > this case, as gimplify_expr will immediately allocate a temporary to store > the CONSTRUCTOR into. And reference the new PR discussed above. Done. > OK with that change. Here is what I'm pushing. Thanks! -- >8 -- The problem in this PR is that we ended up with {.rows=(&<PLACEHOLDER_EXPR struct Widget>)->n, .outer_stride=(&<PLACEHOLDER_EXPR struct MatrixLayout>)->rows} that is, two PLACEHOLDER_EXPRs for different types on the same level in one { }. That should not happen; we may, for instance, neglect to replace a PLACEHOLDER_EXPR due to CONSTRUCTOR_PLACEHOLDER_BOUNDARY on the constructor. The same problem happened in PR100252, which I fixed by introducing replace_placeholders_for_class_temp_r. That didn't work here, though, because r_p_for_c_t_r only works for non-eliding TARGET_EXPRs: replacing a PLACEHOLDER_EXPR with a temporary that is going to be elided will result in a crash in gimplify_var_or_parm_decl when it encounters such a loose decl. But leaving the PLACEHOLDER_EXPRs in is also bad because then we end up with this PR. TARGET_EXPRs for function arguments are elided in gimplify_arg. The argument will get a real temporary only in get_formal_tmp_var. One idea was to use the temporary that is going to be elided anyway, and then replace_decl it with the real object once we get it. But that didn't work out: one problem is that we elide the TARGET_EXPR for an argument before we create the real temporary for the argument, and when we get it, the context that this was a TARGET_EXPR for an argument has been lost. We're also in the middle end territory now, even though this is a C++-specific problem. A solution is to simply stop eliding TARGET_EXPRs whose initializer is a CONSTRUCTOR. Such copies can't be (at the moment) elided anyway. But not eliding all TARGET_EXPRs would be a pessimization. PR c++/116015 gcc/cp/ChangeLog: * call.cc (convert_for_arg_passing): Don't set_target_expr_eliding when the TARGET_EXPR initializer is a CONSTRUCTOR. gcc/ChangeLog: * gimplify.cc (gimplify_arg): Do not strip a TARGET_EXPR whose initializer is a CONSTRUCTOR. gcc/testsuite/ChangeLog: * g++.dg/cpp1y/nsdmi-aggr23.C: New test. --- gcc/cp/call.cc | 7 ++++-- gcc/gimplify.cc | 17 ++++++++++++++- gcc/testsuite/g++.dg/cpp1y/nsdmi-aggr23.C | 26 +++++++++++++++++++++++ 3 files changed, 47 insertions(+), 3 deletions(-) create mode 100644 gcc/testsuite/g++.dg/cpp1y/nsdmi-aggr23.C diff --git a/gcc/cp/call.cc b/gcc/cp/call.cc index 94015db4e65..0fe679aae9f 100644 --- a/gcc/cp/call.cc +++ b/gcc/cp/call.cc @@ -9468,8 +9468,11 @@ convert_for_arg_passing (tree type, tree val, tsubst_flags_t complain) if (complain & tf_warning) warn_for_address_of_packed_member (type, val); - /* gimplify_arg elides TARGET_EXPRs that initialize a function argument. */ - if (SIMPLE_TARGET_EXPR_P (val)) + /* gimplify_arg elides TARGET_EXPRs that initialize a function argument, + unless the initializer is a CONSTRUCTOR. In that case, we fail to + elide the copy anyway. See that function for more information. */ + if (SIMPLE_TARGET_EXPR_P (val) + && TREE_CODE (TARGET_EXPR_INITIAL (val)) != CONSTRUCTOR) set_target_expr_eliding (val); return val; diff --git a/gcc/gimplify.cc b/gcc/gimplify.cc index 71cc6c38d80..26a216e151d 100644 --- a/gcc/gimplify.cc +++ b/gcc/gimplify.cc @@ -3760,7 +3760,22 @@ gimplify_arg (tree *arg_p, gimple_seq *pre_p, location_t call_location, { tree init = TARGET_EXPR_INITIAL (*arg_p); if (init - && !VOID_TYPE_P (TREE_TYPE (init))) + && !VOID_TYPE_P (TREE_TYPE (init)) + /* Currently, due to c++/116015, it is not desirable to + strip a TARGET_EXPR whose initializer is a {}. The + problem is that if we do elide it, we also have to + replace all the occurrences of the slot temporary in the + initializer with the temporary created for the argument. + But we do not have that temporary yet so the replacement + would be quite awkward and it might be needed to resort + back to a PLACEHOLDER_EXPR. Note that stripping the + TARGET_EXPR wouldn't help anyway, as gimplify_expr would + just allocate a temporary to store the CONSTRUCTOR into. + (FIXME PR116375.) + + See convert_for_arg_passing for the C++ code that marks + the TARGET_EXPR as eliding or not. */ + && TREE_CODE (init) != CONSTRUCTOR) *arg_p = init; } } diff --git a/gcc/testsuite/g++.dg/cpp1y/nsdmi-aggr23.C b/gcc/testsuite/g++.dg/cpp1y/nsdmi-aggr23.C new file mode 100644 index 00000000000..2f5b8ca97bf --- /dev/null +++ b/gcc/testsuite/g++.dg/cpp1y/nsdmi-aggr23.C @@ -0,0 +1,26 @@ +// PR c++/116015 +// { dg-do compile { target c++14 } } +// { dg-additional-options "-Wno-c++20-extensions" } + +struct MatrixLayout { + int rows = 0; + int outer_stride = rows; +}; +struct Matrix { + Matrix(MatrixLayout m) {} +}; +struct Widget { + int n = 5; + Matrix A0{{}}; + Matrix A1{{n}}; + Matrix A1_{{.rows = n}}; + Matrix A2{{n, n}}; +}; + +int +main () +{ + Widget w{}; + Widget w1; + Widget w2 = {}; +} base-commit: a247088adaf122116919235f4a40189506139495 -- 2.46.0