pdt14 is miscompiled with -fipa-modref. This is triggered by handling fnspec, but it seems to only trigger latent problem. The only disambiguations are: ipa-modref: call stmt push_8 (&root, &C.4105); ipa-modref: call to push_8/6 does not clobber ref: __vtab_link_module_Pdtlink_8._deallocate alias sets: 12->11 ipa-modref: call stmt push_8 (&root, &C.4104); ipa-modref: call to push_8/6 does not clobber ref: __vtab_link_module_Pdtlink_8._deallocate alias sets: 12->11 ipa-modref: call stmt push_8 (&root, &C.4103); ipa-modref: call to push_8/6 does not clobber ref: __vtab_link_module_Pdtlink_8._deallocate alias sets: 12->11 ipa-modref: call stmt push_8 (&root, &C.4105); ipa-modref: call to push_8/6 does not clobber ref: __vtab_link_module_Pdtlink_8._deallocate alias sets: 12->11 ipa-modref: call stmt push_8 (&root, &C.4104); ipa-modref: call to push_8/6 does not clobber ref: __vtab_link_module_Pdtlink_8._deallocate alias sets: 12->11 ipa-modref: call stmt push_8 (&root, &C.4103); ipa-modref: call to push_8/6 does not clobber ref: __vtab_link_module_Pdtlink_8._deallocate alias sets: 12->11 these ought to be safe since __vtab_link_module_Pdtlink_8 is readonly in the testcase. With LTO we detect that variable as such (and the testcase stil work without modref and fails different with modref). fre3 does quite a lot of additional changes and I am not sure what gets wrong here: __attribute__((externally_visible)) main (integer(kind=4) argc, character(kind=1) * * argv) { + struct array01_unknown cdesc.10; + struct array01_unknown cdesc.9; + real(kind=8) res; + struct Pdtlink_8 * previous; + struct Pdtlink_8 * current; + real(kind=8) res; struct pdtlink_8 * root; static integer(kind=4) options.11[7] = {2150, 4095, 1, 1, 1, 0, 31}; - real(kind=8) _7; - integer(kind=4) _8; - real(kind=8) _9; - integer(kind=4) _10; - real(kind=8) _11; - integer(kind=4) _12; - real(kind=8) _13; - integer(kind=4) _14; + struct Pdtlink_8 * _15; + struct Pdtlink_8 * _17; + struct Pdtlink_8 * _21; + struct Pdtlink_8 * _22; + void (*<T5d4>) () _23; + struct Pdtlink_8 * _25; + void (*<T5d4>) () _26; <bb 2> [local count: 1073741824]: _gfortran_set_args (argc_2(D), argv_3(D)); @@ -1972,52 +2120,75 @@ push_8 (&root, &C.4103); push_8 (&root, &C.4104); push_8 (&root, &C.4105); - _7 = pop_8 (&root); - _8 = (integer(kind=4)) _7; - if (_8 != 3) - goto <bb 3>; [0.04%] + _15 = MEM[(struct Pdtlink_8 * &)&root]; + if (_15 != 0B) + goto <bb 3>; [70.00%] else - goto <bb 4>; [99.96%] + goto <bb 11>; [30.00%] - <bb 3> [local count: 429496]: - _gfortran_stop_numeric (1, 0); - - <bb 4> [local count: 1073312329]: - _9 = pop_8 (&root); - _10 = (integer(kind=4)) _9; - if (_10 != 2) - goto <bb 5>; [0.04%] + <bb 3> [local count: 75913541732]: + # current_16 = PHI <_15(2), _17(3)> + # previous_29 = PHI <_15(2), current_16(3)> + _17 = current_16->next; + if (_17 == 0B) + goto <bb 4>; [0.00%] else - goto <bb 6>; [99.96%] - - <bb 5> [local count: 429324]: - _gfortran_stop_numeric (2, 0); + goto <bb 3>; [100.00%] - <bb 6> [local count: 1072883005]: - _11 = pop_8 (&root); - _12 = (integer(kind=4)) _11; - if (_12 != 1) - goto <bb 7>; [0.04%] + <bb 4> [count: 0]: + res_19 = current_16->n; + _21 = previous_29->next; + if (_21 == 0B) + goto <bb 5>; [30.00%] else - goto <bb 8>; [99.96%] + goto <bb 8>; [70.00%] - <bb 7> [local count: 429152]: - _gfortran_stop_numeric (3, 0); + <bb 5> [count: 0]: + _22 = _15->next; + if (_22 != 0B) + goto <bb 6>; [70.00%] + else + goto <bb 7>; [30.00%] - <bb 8> [local count: 1072453853]: - _13 = pop_8 (&root); - _14 = (integer(kind=4)) _13; - if (_14 != 0) - goto <bb 9>; [0.04%] + <bb 6> [count: 0]: + MEM <c_char[8]> [(struct dtype_type *)&cdesc.9 + 24B] = {}; + cdesc.9.dtype.elem_len = 24; + cdesc.9.dtype.rank = 1; + cdesc.9.dtype.type = 11; + cdesc.9.dim[0].lbound = 1; + cdesc.9.dim[0].stride = 1; + cdesc.9.dim[0].ubound = 1; + cdesc.9.data = _22; + _23 = __vtab_link_module_Pdtlink_8._deallocate; + __builtin_unreachable (); + + <bb 7> [count: 0]: + __builtin_unreachable (); + + <bb 8> [count: 0]: + _25 = _21->next; + if (_25 != 0B) + goto <bb 9>; [70.00%] else - goto <bb 10>; [99.96%] + goto <bb 10>; [30.00%] + + <bb 9> [count: 0]: + MEM <c_char[8]> [(struct dtype_type *)&cdesc.10 + 24B] = {}; + cdesc.10.dtype.elem_len = 24; + cdesc.10.dtype.rank = 1; + cdesc.10.dtype.type = 11; + cdesc.10.dim[0].lbound = 1; + cdesc.10.dim[0].stride = 1; + cdesc.10.dim[0].ubound = 1; + cdesc.10.data = _25; + _26 = __vtab_link_module_Pdtlink_8._deallocate; + __builtin_unreachable (); - <bb 9> [local count: 428981]: - _gfortran_stop_numeric (4, 0); + <bb 10> [count: 0]: + __builtin_unreachable (); - <bb 10> [local count: 1072024872]: - root ={v} {CLOBBER}; - return 0; + <bb 11> [local count: 128815]: + _gfortran_stop_numeric (1, 0);
Actually there is another propagation happening in ipa-cp analysis: --- aa/pdt_14.f03.077i.cp 2020-10-31 09:00:52.809726530 +0100 +++ pdt_14.f03.077i.cp 2020-10-31 09:10:35.204755828 +0100 @@ -10,6 +10,8 @@ Starting walk at: push_8 (&root, &C.4104); instance pointer: &root Outer instance pointer: root offset: 0 (bits) vtbl reference: Function call may change dynamic type:push_8 (&root, &C.4103); +ipa-modref: call stmt push_8 (&root, &C.4103); +ipa-modref: call to push_8/6 does not clobber ref: root alias sets: 14->14 Determining dynamic type for call: push_8 (&root, &C.4104); Starting walk at: push_8 (&root, &C.4104); instance pointer: &C.4104 Outer instance pointer: C.4104 offset: 0 (bits) vtbl reference: @@ -19,6 +21,10 @@ instance pointer: &root Outer instance pointer: root offset: 0 (bits) vtbl reference: Function call may change dynamic type:push_8 (&root, &C.4104); Function call may change dynamic type:push_8 (&root, &C.4103); +ipa-modref: call stmt push_8 (&root, &C.4104); +ipa-modref: call to push_8/6 does not clobber ref: root alias sets: 14->14 +ipa-modref: call stmt push_8 (&root, &C.4103); +ipa-modref: call to push_8/6 does not clobber ref: root alias sets: 14->14 Determining dynamic type for call: push_8 (&root, &C.4105); Starting walk at: push_8 (&root, &C.4105); instance pointer: &C.4105 Outer instance pointer: C.4105 offset: 0 (bits) vtbl reference: @@ -30,6 +36,12 @@ Function call may change dynamic type:push_8 (&root, &C.4105); Function call may change dynamic type:push_8 (&root, &C.4104); Function call may change dynamic type:push_8 (&root, &C.4103); +ipa-modref: call stmt push_8 (&root, &C.4105); +ipa-modref: call to push_8/6 does not clobber ref: root alias sets: 14->14 +ipa-modref: call stmt push_8 (&root, &C.4104); +ipa-modref: call to push_8/6 does not clobber ref: root alias sets: 14->14 +ipa-modref: call stmt push_8 (&root, &C.4103); +ipa-modref: call to push_8/6 does not clobber ref: root alias sets: 14->14 Determining dynamic type for call: _3 = pop_8 (&root); Starting walk at: _3 = pop_8 (&root); instance pointer: &root Outer instance pointer: root offset: 0 (bits) vtbl reference: @@ -129,10 +141,14 @@ no arg info callsite ch2701/7 -> pop_8/5 : param 0: UNKNOWN + Aggregate passed by reference: + offset: 0, type: struct pdtlink_8 *, CONST: 0B value: 0x0, mask: 0xfffffffffffffff8 VR [1, -1] callsite ch2701/7 -> push_8/6 : param 0: UNKNOWN + Aggregate passed by reference: + offset: 0, type: struct pdtlink_8 *, CONST: 0B value: 0x0, mask: 0xfffffffffffffff8 VR [1, -1] param 1: CONST: &C.4105 -> 3.0e+0 @@ -140,6 +156,8 @@ Unknown VR callsite ch2701/7 -> push_8/6 : param 0: UNKNOWN + Aggregate passed by reference: + offset: 0, type: struct pdtlink_8 *, CONST: 0B value: 0x0, mask: 0xfffffffffffffff8 VR [1, -1] param 1: CONST: &C.4104 -> 2.0e+0 The jump function is not used for cloning, only triggers inline, but the conclusion seems wrong. push_8 can make root non-0. Root is of type pdtlink_8 so perhaps Frontend produces multiple copies of these. push_8 store is: - Analyzing store: *self_34(D) - Recording base_set=8 ref_set=8 parm=0 so indeed a different alias set than 14 used by ch2701
*** Bug 97672 has been marked as a duplicate of this bug. ***
*** Bug 97674 has been marked as a duplicate of this bug. ***
I can only see one __vtype_link_module_Pdtlink_8 and one Pdtlink_8 type (breaking at record_component_aliases).
Oh, btw: > /home/rguenther/obj-gcc2-g/gcc/testsuite/gfortran/../../gfortran -B/home/rguenther/obj-gcc2-g/gcc/testsuite/gfortran/../../ -B/home/rguenther/obj-gcc2-g/x86_64-pc-linux-gnu/./libgfortran/ /home/rguenther/src/gcc2/gcc/testsuite/gfortran.dg/pdt_14.f03 -fdiagnostics-plain-output -fdiagnostics-plain-output -O2 -pedantic-errors -B/home/rguenther/obj-gcc2-g/x86_64-pc-linux-gnu/./libgfortran/.libs -L/home/rguenther/obj-gcc2-g/x86_64-pc-linux-gnu/./libgfortran/.libs -L/home/rguenther/obj-gcc2-g/x86_64-pc-linux-gnu/./libgfortran/.libs -L/home/rguenther/obj-gcc2-g/x86_64-pc-linux-gnu/./libatomic/.libs -B/home/rguenther/obj-gcc2-g/x86_64-pc-linux-gnu/./libquadmath/.libs -L/home/rguenther/obj-gcc2-g/x86_64-pc-linux-gnu/./libquadmath/.libs -L/home/rguenther/obj-gcc2-g/x86_64-pc-linux-gnu/./libquadmath/.libs -lm -o ./pdt_14.exe -fno-ipa-cp -fno-ipa-vrp -fno-ipa-icf -fdump-ipa-all during IPA pass: inline dump file: ./pdt_14.f03.081i.inline /home/rguenther/src/gcc2/gcc/testsuite/gfortran.dg/pdt_14.f03:85:0: internal compiler error: Segmentation fault 0x14896d3 crash_signal /home/rguenther/src/gcc2/gcc/toplev.c:330 0xd26c14 cgraph_edge::redirect_call_stmt_to_callee(cgraph_edge*) /home/rguenther/src/gcc2/gcc/cgraph.c:1513 0x156e11a redirect_all_calls(copy_body_data*, basic_block_def*) /home/rguenther/src/gcc2/gcc/tree-inline.c:2963 0x156e906 copy_cfg_body /home/rguenther/src/gcc2/gcc/tree-inline.c:3118 0x156f0e5 copy_body /home/rguenther/src/gcc2/gcc/tree-inline.c:3294 0x1574028 expand_call_inline /home/rguenther/src/gcc2/gcc/tree-inline.c:5084 0x1574dca gimple_expand_calls_inline
And it looks like something (inlining) inserts __builtin_unreachable () calls that eventually make us optimize this to an endless loop. pdt_14.f03.081i.inline:Introduced new external node (__builtin_unreachable/19). pdt_14.f03.081i.inline:unreachable : 4 calls, 0.000000 freq, 0 count pdt_14.f03.081i.inline: __builtin_unreachable/19 unreachable ...
> And it looks like something (inlining) inserts __builtin_unreachable () calls > that eventually make us optimize this to an endless loop. > > pdt_14.f03.081i.inline:Introduced new external node (__builtin_unreachable/19). > pdt_14.f03.081i.inline:unreachable : > 4 calls, 0.000000 freq, 0 count > pdt_14.f03.081i.inline: __builtin_unreachable/19 unreachable > ... Aha, that will be the wrong jump function from ipa-cp I mentioned in earlier comment. We should not believe that root is not written to as we derive now from TBAA in push_8.
Ok, the unreachable () appears in place of indirect calls: _8 = __vtab_link_module_Pdtlink_8._deallocate; _8 (&cdesc.1); but -fno-devirtualize doesn't help.
(In reply to Jan Hubicka from comment #7) > > And it looks like something (inlining) inserts __builtin_unreachable () calls > > that eventually make us optimize this to an endless loop. > > > > pdt_14.f03.081i.inline:Introduced new external node (__builtin_unreachable/19). > > pdt_14.f03.081i.inline:unreachable : > > 4 calls, 0.000000 freq, 0 count > > pdt_14.f03.081i.inline: __builtin_unreachable/19 unreachable > > ... > Aha, that will be the wrong jump function from ipa-cp I mentioned in > earlier comment. We should not believe that root is not written to as > we derive now from TBAA in push_8. But -fno-ipa-cp doesn't help.
> > Aha, that will be the wrong jump function from ipa-cp I mentioned in > > earlier comment. We should not believe that root is not written to as > > we derive now from TBAA in push_8. > > But -fno-ipa-cp doesn't help. It is because jump functions are used both by inlining and ipa-cp. If inline predicates says that the path is impossible (which it will derriving from jump function) we redirect to unreachable. Honza
push_8 argument type is arguments <parm_decl 0x7ffff69cd600 self type <reference_type 0x7ffff69ec000 type <pointer_type 0x7ffff69e3000> unsigned DI size <integer_cst 0x7ffff680cbb8 64> unit-size <integer_cst 0x7ffff680cbd0 8> align:64 warn_if_not_align:0 symtab:0 alias-set -1 canonical-type 0x7ffff69ec000> readonly used unsigned DI passed-by-reference /home/rguenther/src/gcc2/gcc/testsuite/gfortran.dg/pdt_14.f03:47:0 size <integer_cst 0x7ffff680cbb8 64> unit-size <integer_cst 0x7ffff680cbd0 8> align:64 warn_if_not_align:0 context <function_decl 0x7ffff69e0400 pop_8> arg-type <reference_type 0x7ffff69ec000>> (gdb) p debug_tree (0x7ffff69e3000) <pointer_type 0x7ffff69e3000 type <record_type 0x7ffff69dcf18 Pdtlink_8 BLK size <integer_cst 0x7ffff682a078 constant 192> unit-size <integer_cst 0x7ffff682a048 constant 24> align:64 warn_if_not_align:0 symtab:0 alias-set -1 canonical-type 0x7ffff69dcf18 and in ch2701 the root variable is var_decl 0x7ffff69f6900 root type <pointer_type 0x7ffff69fb0a8> addressable used unsigned DI /home/rguenther/src/gcc2/gcc/testsuite/gfortran.dg/pdt_14.f03:78:48 size <integer_cst 0x7ffff680cbb8 64> unit-size <integer_cst 0x7ffff680cbd0 8> align:64 warn_if_not_align:0 context <function_decl 0x7ffff69e0b00 ch2701>> (gdb) p debug_tree (0x7ffff69fb0a8) <pointer_type 0x7ffff69fb0a8 type <record_type 0x7ffff69fb2a0 pdtlink_8 BLK size <integer_cst 0x7ffff682a078 constant 192> unit-size <integer_cst 0x7ffff682a048 constant 24> align:64 warn_if_not_align:0 symtab:0 alias-set -1 canonical-type 0x7ffff69fb2a0 which is a different type. The latter is created here: #0 0x0000000000c1d6f2 in gfc_get_derived_type (derived=0x39ebb70, codimen=0) at /home/rguenther/src/gcc2/gcc/fortran/trans-types.c:2647 #1 0x0000000000c18c21 in gfc_typenode_for_spec (spec=0x39eb090, codim=0) at /home/rguenther/src/gcc2/gcc/fortran/trans-types.c:1167 #2 0x0000000000c1c54b in gfc_sym_type (sym=0x39eb070) at /home/rguenther/src/gcc2/gcc/fortran/trans-types.c:2252 #3 0x0000000000b701d9 in gfc_get_symbol_decl (sym=0x39eb070) at /home/rguenther/src/gcc2/gcc/fortran/trans-decl.c:1790 #4 0x0000000000b7fe7f in generate_local_decl (sym=0x39eb070) at /home/rguenther/src/gcc2/gcc/fortran/trans-decl.c:5948 #5 0x0000000000b1843d in do_traverse_symtree (st=0x39ec880, st_func=0x0, sym_func=0xb7fdbb <generate_local_decl(gfc_symbol*)>) at /home/rguenther/src/gcc2/gcc/fortran/symbol.c:4170 #6 0x0000000000b184f8 in gfc_traverse_ns (ns=0x3a2f6d0, sym_func=0xb7fdbb <generate_local_decl(gfc_symbol*)>) at /home/rguenther/src/gcc2/gcc/fortran/symbol.c:4195 #7 0x0000000000b80677 in generate_local_vars (ns=0x3a2f6d0) at /home/rguenther/src/gcc2/gcc/fortran/trans-decl.c:6154 #8 0x0000000000b825af in gfc_generate_function_code (ns=0x3a2f6d0) at /home/rguenther/src/gcc2/gcc/fortran/trans-decl.c:6812 #9 0x0000000000b3e3fe in gfc_generate_code (ns=0x3a2f6d0) at /home/rguenther/src/gcc2/gcc/fortran/trans.c:2235 #10 0x0000000000abe7e0 in translate_all_program_units ( gfc_global_ns_list=0x39b18f0) while the former here: #0 0x0000000000c1d6f2 in gfc_get_derived_type (derived=0x39c6480, codimen=0) at /home/rguenther/src/gcc2/gcc/fortran/trans-types.c:2647 #1 0x0000000000c18c21 in gfc_typenode_for_spec (spec=0x39ce5a0, codim=0) at /home/rguenther/src/gcc2/gcc/fortran/trans-types.c:1167 #2 0x0000000000c1c54b in gfc_sym_type (sym=0x39ce580) at /home/rguenther/src/gcc2/gcc/fortran/trans-types.c:2252 #3 0x0000000000b701d9 in gfc_get_symbol_decl (sym=0x39ce580) at /home/rguenther/src/gcc2/gcc/fortran/trans-decl.c:1790 #4 0x0000000000b7dd72 in gfc_create_module_variable (sym=0x39ce580) at /home/rguenther/src/gcc2/gcc/fortran/trans-decl.c:5339 #5 0x0000000000b1843d in do_traverse_symtree (st=0x3951a30, st_func=0x0, sym_func=0xb7d65d <gfc_create_module_variable(gfc_symbol*)>) at /home/rguenther/src/gcc2/gcc/fortran/symbol.c:4170 #6 0x0000000000b184f8 in gfc_traverse_ns (ns=0x39b2f80, sym_func=0xb7d65d <gfc_create_module_variable(gfc_symbol*)>) at /home/rguenther/src/gcc2/gcc/fortran/symbol.c:4195 #7 0x0000000000b7fb05 in gfc_generate_module_vars (ns=0x39b2f80) at /home/rguenther/src/gcc2/gcc/fortran/trans-decl.c:5838 #8 0x0000000000b3e4f4 in gfc_generate_module_code (ns=0x39b2f80) at /home/rguenther/src/gcc2/gcc/fortran/trans.c:2259 #9 0x0000000000abe712 in translate_all_program_units ( gfc_global_ns_list=0x39b18f0) there is it seems code to deal with the type already been there but it seems this handling doesn't work. It does /* Store up the canonical type to be added to this one. */ if (got_canonical) { if (TYPE_CANONICAL (derived->backend_decl)) canonical = TYPE_CANONICAL (derived->backend_decl); else canonical = derived->backend_decl; derived->backend_decl = NULL_TREE; } /* derived->backend_decl != 0 means we saw it before, but its components' backend_decl may have not been built. */ if (derived->backend_decl) ... else { /* We see this derived type first time, so build the type node. */ typenode = make_node (RECORD_TYPE); TYPE_NAME (typenode) = get_identifier (derived->name); TYPE_PACKED (typenode) = flag_pack_derived; derived->backend_decl = typenode; } but we enver find a canonical type. Looks like a latent issue exposed now.
Note setting TYPE_TYPELESS_STORAGE on the aggregates doesn't help the testcase since the pointer types are the issue.
*** Bug 97686 has been marked as a duplicate of this bug. ***
The following works with and without -flto (fixes the endless loop without and the execute fail with) diff --git a/gcc/fortran/trans-types.c b/gcc/fortran/trans-types.c index b7129dcbe6d..4643fff243f 100644 --- a/gcc/fortran/trans-types.c +++ b/gcc/fortran/trans-types.c @@ -2647,6 +2647,8 @@ gfc_get_derived_type (gfc_symbol * derived, int codimen) typenode = make_node (RECORD_TYPE); TYPE_NAME (typenode) = get_identifier (derived->name); TYPE_PACKED (typenode) = flag_pack_derived; + if (!got_canonical) + SET_TYPE_STRUCTURAL_EQUALITY (typenode); derived->backend_decl = typenode; }
The master branch has been updated by Tobias Burnus <burnus@gcc.gnu.org>: https://gcc.gnu.org/g:c283a711c850efaab4fe3bca5ef7200eb854bba1 commit r11-4766-gc283a711c850efaab4fe3bca5ef7200eb854bba1 Author: Tobias Burnus <tobias@codesourcery.com> Date: Fri Nov 6 08:26:51 2020 +0100 Fortran: Fix type-decl for PDT / wrong-code pdt_14.f03 issue [PR97652] Parameterized derived types are handled in a special way and start with 'Pdt'. If the 'P' is not uppercase, gfc_get_derived_type (which calls gfc_get_module_backend_decl) does not find the existing declaration and builds a new type. The middle end then sees those types as being different and nonalising, creating an endless loop for pdt_14.f03. gcc/fortran/ChangeLog: PR fortran/97652 * module.c (mio_symbol): Fix symbol name for pdt_type.
Fixed.
The releases/gcc-10 branch has been updated by Tobias Burnus <burnus@gcc.gnu.org>: https://gcc.gnu.org/g:099857318ca92210e34cfb8975c5c58c00bf1587 commit r10-8989-g099857318ca92210e34cfb8975c5c58c00bf1587 Author: Tobias Burnus <tobias@codesourcery.com> Date: Fri Nov 6 08:26:51 2020 +0100 Fortran: Fix type-decl for PDT / wrong-code pdt_14.f03 issue [PR97652] Parameterized derived types are handled in a special way and start with 'Pdt'. If the 'P' is not uppercase, gfc_get_derived_type (which calls gfc_get_module_backend_decl) does not find the existing declaration and builds a new type. The middle end then sees those types as being different and nonalising, creating an endless loop for pdt_14.f03. gcc/fortran/ChangeLog: PR fortran/97652 * module.c (mio_symbol): Fix symbol name for pdt_type. (cherry picked from commit c283a711c850efaab4fe3bca5ef7200eb854bba1)
The releases/gcc-9 branch has been updated by Tobias Burnus <burnus@gcc.gnu.org>: https://gcc.gnu.org/g:542b564343fdb896ede9c9d5e32d45dcd96b2a00 commit r9-9028-g542b564343fdb896ede9c9d5e32d45dcd96b2a00 Author: Tobias Burnus <tobias@codesourcery.com> Date: Fri Nov 6 08:26:51 2020 +0100 Fortran: Fix type-decl for PDT / wrong-code pdt_14.f03 issue [PR97652] Parameterized derived types are handled in a special way and start with 'Pdt'. If the 'P' is not uppercase, gfc_get_derived_type (which calls gfc_get_module_backend_decl) does not find the existing declaration and builds a new type. The middle end then sees those types as being different and nonalising, creating an endless loop for pdt_14.f03. gcc/fortran/ChangeLog: PR fortran/97652 * module.c (mio_symbol): Fix symbol name for pdt_type. (cherry picked from commit c283a711c850efaab4fe3bca5ef7200eb854bba1)