| Message ID | nycvar.YFH.7.76.2009231013440.11491@elmra |
|---|---|
| State | New |
| Headers | show |
| Series | tree-optimization/97151 - improve PTA for C++ operator delete | expand |
On 9/23/20 4:14 AM, Richard Biener wrote: > C++ operator delete, when DECL_IS_REPLACEABLE_OPERATOR_DELETE_P, > does not cause the deleted object to be escaped. It also has no > other interesting side-effects for PTA so skip it like we do > for BUILT_IN_FREE. Hmm, this is true of the default implementation, but since the function is replaceable, we don't know what a user definition might do with the pointer. > Bootstrapped and tested on x86_64-unknown-linux-gnu, pushed. > > Richard. > > 2020-09-23 Richard Biener <rguenther@suse.de> > > PR tree-optimization/97151 > * tree-ssa-structalias.c (find_func_aliases_for_call): > DECL_IS_REPLACEABLE_OPERATOR_DELETE_P has no effect on > arguments. > > * g++.dg/cpp1y/new1.C: Adjust for two more handled transforms. > --- > gcc/testsuite/g++.dg/cpp1y/new1.C | 4 ++-- > gcc/tree-ssa-structalias.c | 2 ++ > 2 files changed, 4 insertions(+), 2 deletions(-) > > diff --git a/gcc/testsuite/g++.dg/cpp1y/new1.C b/gcc/testsuite/g++.dg/cpp1y/new1.C > index aa5f647d535..fec0088cb40 100644 > --- a/gcc/testsuite/g++.dg/cpp1y/new1.C > +++ b/gcc/testsuite/g++.dg/cpp1y/new1.C > @@ -69,5 +69,5 @@ test_unused() { > delete p; > } > > -/* { dg-final { scan-tree-dump-times "Deleting : operator delete" 5 "cddce1"} } */ > -/* { dg-final { scan-tree-dump-times "Deleting : _\\d+ = operator new" 7 "cddce1"} } */ > +/* { dg-final { scan-tree-dump-times "Deleting : operator delete" 6 "cddce1"} } */ > +/* { dg-final { scan-tree-dump-times "Deleting : _\\d+ = operator new" 8 "cddce1"} } */ > diff --git a/gcc/tree-ssa-structalias.c b/gcc/tree-ssa-structalias.c > index 44fe52e0f65..f676bf91e95 100644 > --- a/gcc/tree-ssa-structalias.c > +++ b/gcc/tree-ssa-structalias.c > @@ -4857,6 +4857,8 @@ find_func_aliases_for_call (struct function *fn, gcall *t) > point for reachable memory of their arguments. */ > else if (flags & (ECF_PURE|ECF_LOOPING_CONST_OR_PURE)) > handle_pure_call (t, &rhsc); > + else if (fndecl && DECL_IS_REPLACEABLE_OPERATOR_DELETE_P (fndecl)) > + ; > else > handle_rhs_call (t, &rhsc); > if (gimple_call_lhs (t)) >
On September 23, 2020 7:53:18 PM GMT+02:00, Jason Merrill <jason@redhat.com> wrote: >On 9/23/20 4:14 AM, Richard Biener wrote: >> C++ operator delete, when DECL_IS_REPLACEABLE_OPERATOR_DELETE_P, >> does not cause the deleted object to be escaped. It also has no >> other interesting side-effects for PTA so skip it like we do >> for BUILT_IN_FREE. > >Hmm, this is true of the default implementation, but since the function > >is replaceable, we don't know what a user definition might do with the >pointer. But can the object still be 'used' after delete? Can delete fail / throw? What guarantee does the predicate give us? >> Bootstrapped and tested on x86_64-unknown-linux-gnu, pushed. >> >> Richard. >> >> 2020-09-23 Richard Biener <rguenther@suse.de> >> >> PR tree-optimization/97151 >> * tree-ssa-structalias.c (find_func_aliases_for_call): >> DECL_IS_REPLACEABLE_OPERATOR_DELETE_P has no effect on >> arguments. >> >> * g++.dg/cpp1y/new1.C: Adjust for two more handled transforms. >> --- >> gcc/testsuite/g++.dg/cpp1y/new1.C | 4 ++-- >> gcc/tree-ssa-structalias.c | 2 ++ >> 2 files changed, 4 insertions(+), 2 deletions(-) >> >> diff --git a/gcc/testsuite/g++.dg/cpp1y/new1.C >b/gcc/testsuite/g++.dg/cpp1y/new1.C >> index aa5f647d535..fec0088cb40 100644 >> --- a/gcc/testsuite/g++.dg/cpp1y/new1.C >> +++ b/gcc/testsuite/g++.dg/cpp1y/new1.C >> @@ -69,5 +69,5 @@ test_unused() { >> delete p; >> } >> >> -/* { dg-final { scan-tree-dump-times "Deleting : operator delete" 5 >"cddce1"} } */ >> -/* { dg-final { scan-tree-dump-times "Deleting : _\\d+ = operator >new" 7 "cddce1"} } */ >> +/* { dg-final { scan-tree-dump-times "Deleting : operator delete" 6 >"cddce1"} } */ >> +/* { dg-final { scan-tree-dump-times "Deleting : _\\d+ = operator >new" 8 "cddce1"} } */ >> diff --git a/gcc/tree-ssa-structalias.c b/gcc/tree-ssa-structalias.c >> index 44fe52e0f65..f676bf91e95 100644 >> --- a/gcc/tree-ssa-structalias.c >> +++ b/gcc/tree-ssa-structalias.c >> @@ -4857,6 +4857,8 @@ find_func_aliases_for_call (struct function >*fn, gcall *t) >> point for reachable memory of their arguments. */ >> else if (flags & (ECF_PURE|ECF_LOOPING_CONST_OR_PURE)) >> handle_pure_call (t, &rhsc); >> + else if (fndecl && DECL_IS_REPLACEABLE_OPERATOR_DELETE_P >(fndecl)) >> + ; >> else >> handle_rhs_call (t, &rhsc); >> if (gimple_call_lhs (t)) >>
On 9/23/20 2:42 PM, Richard Biener wrote: > On September 23, 2020 7:53:18 PM GMT+02:00, Jason Merrill <jason@redhat.com> wrote: >> On 9/23/20 4:14 AM, Richard Biener wrote: >>> C++ operator delete, when DECL_IS_REPLACEABLE_OPERATOR_DELETE_P, >>> does not cause the deleted object to be escaped. It also has no >>> other interesting side-effects for PTA so skip it like we do >>> for BUILT_IN_FREE. >> >> Hmm, this is true of the default implementation, but since the function >> >> is replaceable, we don't know what a user definition might do with the >> pointer. > > But can the object still be 'used' after delete? Can delete fail / throw? > > What guarantee does the predicate give us? The deallocation function is called as part of a delete expression in order to release the storage for an object, ending its lifetime (if it was not ended by a destructor), so no, the object can't be used afterward. A deallocation function that throws has undefined behavior. >>> Bootstrapped and tested on x86_64-unknown-linux-gnu, pushed. >>> >>> Richard. >>> >>> 2020-09-23 Richard Biener <rguenther@suse.de> >>> >>> PR tree-optimization/97151 >>> * tree-ssa-structalias.c (find_func_aliases_for_call): >>> DECL_IS_REPLACEABLE_OPERATOR_DELETE_P has no effect on >>> arguments. >>> >>> * g++.dg/cpp1y/new1.C: Adjust for two more handled transforms. >>> --- >>> gcc/testsuite/g++.dg/cpp1y/new1.C | 4 ++-- >>> gcc/tree-ssa-structalias.c | 2 ++ >>> 2 files changed, 4 insertions(+), 2 deletions(-) >>> >>> diff --git a/gcc/testsuite/g++.dg/cpp1y/new1.C >> b/gcc/testsuite/g++.dg/cpp1y/new1.C >>> index aa5f647d535..fec0088cb40 100644 >>> --- a/gcc/testsuite/g++.dg/cpp1y/new1.C >>> +++ b/gcc/testsuite/g++.dg/cpp1y/new1.C >>> @@ -69,5 +69,5 @@ test_unused() { >>> delete p; >>> } >>> >>> -/* { dg-final { scan-tree-dump-times "Deleting : operator delete" 5 >> "cddce1"} } */ >>> -/* { dg-final { scan-tree-dump-times "Deleting : _\\d+ = operator >> new" 7 "cddce1"} } */ >>> +/* { dg-final { scan-tree-dump-times "Deleting : operator delete" 6 >> "cddce1"} } */ >>> +/* { dg-final { scan-tree-dump-times "Deleting : _\\d+ = operator >> new" 8 "cddce1"} } */ >>> diff --git a/gcc/tree-ssa-structalias.c b/gcc/tree-ssa-structalias.c >>> index 44fe52e0f65..f676bf91e95 100644 >>> --- a/gcc/tree-ssa-structalias.c >>> +++ b/gcc/tree-ssa-structalias.c >>> @@ -4857,6 +4857,8 @@ find_func_aliases_for_call (struct function >> *fn, gcall *t) >>> point for reachable memory of their arguments. */ >>> else if (flags & (ECF_PURE|ECF_LOOPING_CONST_OR_PURE)) >>> handle_pure_call (t, &rhsc); >>> + else if (fndecl && DECL_IS_REPLACEABLE_OPERATOR_DELETE_P >> (fndecl)) >>> + ; >>> else >>> handle_rhs_call (t, &rhsc); >>> if (gimple_call_lhs (t)) >>> >
On Wed, 23 Sep 2020, Jason Merrill wrote: > On 9/23/20 2:42 PM, Richard Biener wrote: > > On September 23, 2020 7:53:18 PM GMT+02:00, Jason Merrill <jason@redhat.com> > > wrote: > >> On 9/23/20 4:14 AM, Richard Biener wrote: > >>> C++ operator delete, when DECL_IS_REPLACEABLE_OPERATOR_DELETE_P, > >>> does not cause the deleted object to be escaped. It also has no > >>> other interesting side-effects for PTA so skip it like we do > >>> for BUILT_IN_FREE. > >> > >> Hmm, this is true of the default implementation, but since the function > >> > >> is replaceable, we don't know what a user definition might do with the > >> pointer. > > > > But can the object still be 'used' after delete? Can delete fail / throw? > > > > What guarantee does the predicate give us? > > The deallocation function is called as part of a delete expression in order to > release the storage for an object, ending its lifetime (if it was not ended by > a destructor), so no, the object can't be used afterward. OK, but the delete operator can access the object contents if there wasn't a destructor ... > A deallocation function that throws has undefined behavior. OK, so it seems the 'replaceable' operators are the global ones (for user-defined/class-specific placement variants I see arbitrary extra arguments that we'd possibly need to handle). I'm happy to revert but I'd like to have a testcase that FAILs with the patch ;) Now, the following aborts: struct X { static struct X saved; int *p; X() { __builtin_memcpy (this, &saved, sizeof (X)); } }; void operator delete (void *p) { __builtin_memcpy (&X::saved, p, sizeof (X)); } int main() { int y = 1; X *p = new X; p->p = &y; delete p; X *q = new X; *(q->p) = 2; if (y != 2) __builtin_abort (); } and I could fix this by not making *p but what *p points to escape. The testcase is of course maximally awkward, but hey ... ;) Now this would all be moot if operator delete may not access the object (or if the object contents are undefined at that point). Oh, and the testcase segfaults when compiled with GCC 10 because there we elide the new X / delete p pair ... which is invalid then? Hmm, we emit MEM[(struct X *)_8] ={v} {CLOBBER}; operator delete (_8, 8); so the object contents are undefined _before_ calling delete even when I do not have a DTOR? That is, the above, w/o -fno-lifetime-dse, makes the PTA patch OK for the testcase. Richard. > >>> Bootstrapped and tested on x86_64-unknown-linux-gnu, pushed. > >>> > >>> Richard. > >>> > >>> 2020-09-23 Richard Biener <rguenther@suse.de> > >>> > >>> PR tree-optimization/97151 > >>> * tree-ssa-structalias.c (find_func_aliases_for_call): > >>> DECL_IS_REPLACEABLE_OPERATOR_DELETE_P has no effect on > >>> arguments. > >>> > >>> * g++.dg/cpp1y/new1.C: Adjust for two more handled transforms. > >>> --- > >>> gcc/testsuite/g++.dg/cpp1y/new1.C | 4 ++-- > >>> gcc/tree-ssa-structalias.c | 2 ++ > >>> 2 files changed, 4 insertions(+), 2 deletions(-) > >>> > >>> diff --git a/gcc/testsuite/g++.dg/cpp1y/new1.C > >> b/gcc/testsuite/g++.dg/cpp1y/new1.C > >>> index aa5f647d535..fec0088cb40 100644 > >>> --- a/gcc/testsuite/g++.dg/cpp1y/new1.C > >>> +++ b/gcc/testsuite/g++.dg/cpp1y/new1.C > >>> @@ -69,5 +69,5 @@ test_unused() { > >>> delete p; > >>> } > >>> > >>> -/* { dg-final { scan-tree-dump-times "Deleting : operator delete" 5 > >> "cddce1"} } */ > >>> -/* { dg-final { scan-tree-dump-times "Deleting : _\\d+ = operator > >> new" 7 "cddce1"} } */ > >>> +/* { dg-final { scan-tree-dump-times "Deleting : operator delete" 6 > >> "cddce1"} } */ > >>> +/* { dg-final { scan-tree-dump-times "Deleting : _\\d+ = operator > >> new" 8 "cddce1"} } */ > >>> diff --git a/gcc/tree-ssa-structalias.c b/gcc/tree-ssa-structalias.c > >>> index 44fe52e0f65..f676bf91e95 100644 > >>> --- a/gcc/tree-ssa-structalias.c > >>> +++ b/gcc/tree-ssa-structalias.c > >>> @@ -4857,6 +4857,8 @@ find_func_aliases_for_call (struct function > >> *fn, gcall *t) > >>> point for reachable memory of their arguments. */ > >>> else if (flags & (ECF_PURE|ECF_LOOPING_CONST_OR_PURE)) > >>> handle_pure_call (t, &rhsc); > >>> + else if (fndecl && DECL_IS_REPLACEABLE_OPERATOR_DELETE_P > >> (fndecl)) > >>> + ; > >>> else > >>> handle_rhs_call (t, &rhsc); > >>> if (gimple_call_lhs (t)) > >>> > > > >
On 9/24/20 3:43 AM, Richard Biener wrote: > On Wed, 23 Sep 2020, Jason Merrill wrote: > >> On 9/23/20 2:42 PM, Richard Biener wrote: >>> On September 23, 2020 7:53:18 PM GMT+02:00, Jason Merrill <jason@redhat.com> >>> wrote: >>>> On 9/23/20 4:14 AM, Richard Biener wrote: >>>>> C++ operator delete, when DECL_IS_REPLACEABLE_OPERATOR_DELETE_P, >>>>> does not cause the deleted object to be escaped. It also has no >>>>> other interesting side-effects for PTA so skip it like we do >>>>> for BUILT_IN_FREE. >>>> >>>> Hmm, this is true of the default implementation, but since the function >>>> >>>> is replaceable, we don't know what a user definition might do with the >>>> pointer. >>> >>> But can the object still be 'used' after delete? Can delete fail / throw? >>> >>> What guarantee does the predicate give us? >> >> The deallocation function is called as part of a delete expression in order to >> release the storage for an object, ending its lifetime (if it was not ended by >> a destructor), so no, the object can't be used afterward. > > OK, but the delete operator can access the object contents if there > wasn't a destructor ... >> A deallocation function that throws has undefined behavior. > > OK, so it seems the 'replaceable' operators are the global ones > (for user-defined/class-specific placement variants I see arbitrary > extra arguments that we'd possibly need to handle). > > I'm happy to revert but I'd like to have a testcase that FAILs > with the patch ;) > > Now, the following aborts: > > struct X { > static struct X saved; > int *p; > X() { __builtin_memcpy (this, &saved, sizeof (X)); } > }; > void operator delete (void *p) > { > __builtin_memcpy (&X::saved, p, sizeof (X)); > } > int main() > { > int y = 1; > X *p = new X; > p->p = &y; > delete p; > X *q = new X; > *(q->p) = 2; > if (y != 2) > __builtin_abort (); > } > > and I could fix this by not making *p but what *p points to escape. > The testcase is of course maximally awkward, but hey ... ;) > > Now this would all be moot if operator delete may not access > the object (or if the object contents are undefined at that point). > > Oh, and the testcase segfaults when compiled with GCC 10 because > there we elide the new X / delete p pair ... which is invalid then? > Hmm, we emit > > MEM[(struct X *)_8] ={v} {CLOBBER}; > operator delete (_8, 8); > > so the object contents are undefined _before_ calling delete > even when I do not have a DTOR? That is, the above, > w/o -fno-lifetime-dse, makes the PTA patch OK for the testcase. Yes, all classes have a destructor, even if it's trivial, so the object's lifetime definitely ends before the call to operator delete. This is less clear for scalar objects, but treating them similarly would be consistent with other recent changes, so I think it's fine for us to assume that scalar objects are also invalidated before the call to operator delete. But of course this doesn't apply to explicit calls to operator delete outside of a delete expression. > Richard. > >>>>> Bootstrapped and tested on x86_64-unknown-linux-gnu, pushed. >>>>> >>>>> Richard. >>>>> >>>>> 2020-09-23 Richard Biener <rguenther@suse.de> >>>>> >>>>> PR tree-optimization/97151 >>>>> * tree-ssa-structalias.c (find_func_aliases_for_call): >>>>> DECL_IS_REPLACEABLE_OPERATOR_DELETE_P has no effect on >>>>> arguments. >>>>> >>>>> * g++.dg/cpp1y/new1.C: Adjust for two more handled transforms. >>>>> --- >>>>> gcc/testsuite/g++.dg/cpp1y/new1.C | 4 ++-- >>>>> gcc/tree-ssa-structalias.c | 2 ++ >>>>> 2 files changed, 4 insertions(+), 2 deletions(-) >>>>> >>>>> diff --git a/gcc/testsuite/g++.dg/cpp1y/new1.C >>>> b/gcc/testsuite/g++.dg/cpp1y/new1.C >>>>> index aa5f647d535..fec0088cb40 100644 >>>>> --- a/gcc/testsuite/g++.dg/cpp1y/new1.C >>>>> +++ b/gcc/testsuite/g++.dg/cpp1y/new1.C >>>>> @@ -69,5 +69,5 @@ test_unused() { >>>>> delete p; >>>>> } >>>>> >>>>> -/* { dg-final { scan-tree-dump-times "Deleting : operator delete" 5 >>>> "cddce1"} } */ >>>>> -/* { dg-final { scan-tree-dump-times "Deleting : _\\d+ = operator >>>> new" 7 "cddce1"} } */ >>>>> +/* { dg-final { scan-tree-dump-times "Deleting : operator delete" 6 >>>> "cddce1"} } */ >>>>> +/* { dg-final { scan-tree-dump-times "Deleting : _\\d+ = operator >>>> new" 8 "cddce1"} } */ >>>>> diff --git a/gcc/tree-ssa-structalias.c b/gcc/tree-ssa-structalias.c >>>>> index 44fe52e0f65..f676bf91e95 100644 >>>>> --- a/gcc/tree-ssa-structalias.c >>>>> +++ b/gcc/tree-ssa-structalias.c >>>>> @@ -4857,6 +4857,8 @@ find_func_aliases_for_call (struct function >>>> *fn, gcall *t) >>>>> point for reachable memory of their arguments. */ >>>>> else if (flags & (ECF_PURE|ECF_LOOPING_CONST_OR_PURE)) >>>>> handle_pure_call (t, &rhsc); >>>>> + else if (fndecl && DECL_IS_REPLACEABLE_OPERATOR_DELETE_P >>>> (fndecl)) >>>>> + ; >>>>> else >>>>> handle_rhs_call (t, &rhsc); >>>>> if (gimple_call_lhs (t)) >>>>> >>> >> >> >
On Thu, 24 Sep 2020, Jason Merrill wrote: > On 9/24/20 3:43 AM, Richard Biener wrote: > > On Wed, 23 Sep 2020, Jason Merrill wrote: > > > >> On 9/23/20 2:42 PM, Richard Biener wrote: > >>> On September 23, 2020 7:53:18 PM GMT+02:00, Jason Merrill > >>> <jason@redhat.com> > >>> wrote: > >>>> On 9/23/20 4:14 AM, Richard Biener wrote: > >>>>> C++ operator delete, when DECL_IS_REPLACEABLE_OPERATOR_DELETE_P, > >>>>> does not cause the deleted object to be escaped. It also has no > >>>>> other interesting side-effects for PTA so skip it like we do > >>>>> for BUILT_IN_FREE. > >>>> > >>>> Hmm, this is true of the default implementation, but since the function > >>>> > >>>> is replaceable, we don't know what a user definition might do with the > >>>> pointer. > >>> > >>> But can the object still be 'used' after delete? Can delete fail / throw? > >>> > >>> What guarantee does the predicate give us? > >> > >> The deallocation function is called as part of a delete expression in order > >> to > >> release the storage for an object, ending its lifetime (if it was not ended > >> by > >> a destructor), so no, the object can't be used afterward. > > > > OK, but the delete operator can access the object contents if there > > wasn't a destructor ... > > >> A deallocation function that throws has undefined behavior. > > > > OK, so it seems the 'replaceable' operators are the global ones > > (for user-defined/class-specific placement variants I see arbitrary > > extra arguments that we'd possibly need to handle). > > > > I'm happy to revert but I'd like to have a testcase that FAILs > > with the patch ;) > > > > Now, the following aborts: > > > > struct X { > > static struct X saved; > > int *p; > > X() { __builtin_memcpy (this, &saved, sizeof (X)); } > > }; > > void operator delete (void *p) > > { > > __builtin_memcpy (&X::saved, p, sizeof (X)); > > } > > int main() > > { > > int y = 1; > > X *p = new X; > > p->p = &y; > > delete p; > > X *q = new X; > > *(q->p) = 2; > > if (y != 2) > > __builtin_abort (); > > } > > > > and I could fix this by not making *p but what *p points to escape. > > The testcase is of course maximally awkward, but hey ... ;) > > > > Now this would all be moot if operator delete may not access > > the object (or if the object contents are undefined at that point). > > > > Oh, and the testcase segfaults when compiled with GCC 10 because > > there we elide the new X / delete p pair ... which is invalid then? > > Hmm, we emit > > > > MEM[(struct X *)_8] ={v} {CLOBBER}; > > operator delete (_8, 8); > > > > so the object contents are undefined _before_ calling delete > > even when I do not have a DTOR? That is, the above, > > w/o -fno-lifetime-dse, makes the PTA patch OK for the testcase. > > Yes, all classes have a destructor, even if it's trivial, so the object's > lifetime definitely ends before the call to operator delete. This is less > clear for scalar objects, but treating them similarly would be consistent with > other recent changes, so I think it's fine for us to assume that scalar > objects are also invalidated before the call to operator delete. But of > course this doesn't apply to explicit calls to operator delete outside of a > delete expression. OK, so change the testcase main slightly to int main() { int y = 1; X *p = new X; p->p = &y; ::operator delete(p); X *q = new X; *(q->p) = 2; if (y != 2) __builtin_abort (); } in this case the lifetime of *p does not end before calling ::operator delete() and delete can stash the object contents somewhere before ending its lifetime. For the very same reason we may not elide a new/delete pair like in int main() { int *p = new int; *p = 1; ::operator delete (p); } which we before the change did not do only because calling operator delete made p escape. Unfortunately points-to analysis cannot really reconstruct whether delete was called as part of a delete expression or directly (and thus whether object lifetime ended already), neither can DCE. So I guess we need to mark the operator delete call in some way to make those transforms safe. At least currently any operator delete call makes the alias guarantee of a operator new call moot by forcing the object to be aliased with all global and escaped memory ... Looks like there are some unallocated flags for CALL_EXPR we could pick but I wonder if we can recycle protected_flag which is CALL_FROM_THUNK_P and CALL_ALLOCA_FOR_VAR_P in CALL_EXPR for calls to DECL_IS_OPERATOR_{NEW,DELETE}_P, thus whether we have CALL_FROM_THUNK_P for those operators. Guess picking a new flag is safer. But, does it seem correct that we need to distinguish delete expressions from plain calls to operator delete? In this context I also wonder about non-replaceable operator delete, specifically operator delete in classes - are there any semantic differences between those or why did we choose to only mark the replaceable ones? Thanks, Richard. > > Richard. > > > >>>>> Bootstrapped and tested on x86_64-unknown-linux-gnu, pushed. > >>>>> > >>>>> Richard. > >>>>> > >>>>> 2020-09-23 Richard Biener <rguenther@suse.de> > >>>>> > >>>>> PR tree-optimization/97151 > >>>>> * tree-ssa-structalias.c (find_func_aliases_for_call): > >>>>> DECL_IS_REPLACEABLE_OPERATOR_DELETE_P has no effect on > >>>>> arguments. > >>>>> > >>>>> * g++.dg/cpp1y/new1.C: Adjust for two more handled transforms. > >>>>> --- > >>>>> gcc/testsuite/g++.dg/cpp1y/new1.C | 4 ++-- > >>>>> gcc/tree-ssa-structalias.c | 2 ++ > >>>>> 2 files changed, 4 insertions(+), 2 deletions(-) > >>>>> > >>>>> diff --git a/gcc/testsuite/g++.dg/cpp1y/new1.C > >>>> b/gcc/testsuite/g++.dg/cpp1y/new1.C > >>>>> index aa5f647d535..fec0088cb40 100644 > >>>>> --- a/gcc/testsuite/g++.dg/cpp1y/new1.C > >>>>> +++ b/gcc/testsuite/g++.dg/cpp1y/new1.C > >>>>> @@ -69,5 +69,5 @@ test_unused() { > >>>>> delete p; > >>>>> } > >>>>> > >>>>> -/* { dg-final { scan-tree-dump-times "Deleting : operator delete" 5 > >>>> "cddce1"} } */ > >>>>> -/* { dg-final { scan-tree-dump-times "Deleting : _\\d+ = operator > >>>> new" 7 "cddce1"} } */ > >>>>> +/* { dg-final { scan-tree-dump-times "Deleting : operator delete" 6 > >>>> "cddce1"} } */ > >>>>> +/* { dg-final { scan-tree-dump-times "Deleting : _\\d+ = operator > >>>> new" 8 "cddce1"} } */ > >>>>> diff --git a/gcc/tree-ssa-structalias.c b/gcc/tree-ssa-structalias.c > >>>>> index 44fe52e0f65..f676bf91e95 100644 > >>>>> --- a/gcc/tree-ssa-structalias.c > >>>>> +++ b/gcc/tree-ssa-structalias.c > >>>>> @@ -4857,6 +4857,8 @@ find_func_aliases_for_call (struct function > >>>> *fn, gcall *t) > >>>>> point for reachable memory of their arguments. */ > >>>>> else if (flags & (ECF_PURE|ECF_LOOPING_CONST_OR_PURE)) > >>>>> handle_pure_call (t, &rhsc); > >>>>> + else if (fndecl && DECL_IS_REPLACEABLE_OPERATOR_DELETE_P > >>>> (fndecl)) > >>>>> + ; > >>>>> else > >>>>> handle_rhs_call (t, &rhsc); > >>>>> if (gimple_call_lhs (t)) > >>>>> > >>> > >> > >> > > > > >
On 9/25/20 2:30 AM, Richard Biener wrote: > On Thu, 24 Sep 2020, Jason Merrill wrote: > >> On 9/24/20 3:43 AM, Richard Biener wrote: >>> On Wed, 23 Sep 2020, Jason Merrill wrote: >>> >>>> On 9/23/20 2:42 PM, Richard Biener wrote: >>>>> On September 23, 2020 7:53:18 PM GMT+02:00, Jason Merrill >>>>> <jason@redhat.com> >>>>> wrote: >>>>>> On 9/23/20 4:14 AM, Richard Biener wrote: >>>>>>> C++ operator delete, when DECL_IS_REPLACEABLE_OPERATOR_DELETE_P, >>>>>>> does not cause the deleted object to be escaped. It also has no >>>>>>> other interesting side-effects for PTA so skip it like we do >>>>>>> for BUILT_IN_FREE. >>>>>> >>>>>> Hmm, this is true of the default implementation, but since the function >>>>>> >>>>>> is replaceable, we don't know what a user definition might do with the >>>>>> pointer. >>>>> >>>>> But can the object still be 'used' after delete? Can delete fail / throw? >>>>> >>>>> What guarantee does the predicate give us? >>>> >>>> The deallocation function is called as part of a delete expression in order >>>> to >>>> release the storage for an object, ending its lifetime (if it was not ended >>>> by >>>> a destructor), so no, the object can't be used afterward. >>> >>> OK, but the delete operator can access the object contents if there >>> wasn't a destructor ... >> >>>> A deallocation function that throws has undefined behavior. >>> >>> OK, so it seems the 'replaceable' operators are the global ones >>> (for user-defined/class-specific placement variants I see arbitrary >>> extra arguments that we'd possibly need to handle). >>> >>> I'm happy to revert but I'd like to have a testcase that FAILs >>> with the patch ;) >>> >>> Now, the following aborts: >>> >>> struct X { >>> static struct X saved; >>> int *p; >>> X() { __builtin_memcpy (this, &saved, sizeof (X)); } >>> }; >>> void operator delete (void *p) >>> { >>> __builtin_memcpy (&X::saved, p, sizeof (X)); >>> } >>> int main() >>> { >>> int y = 1; >>> X *p = new X; >>> p->p = &y; >>> delete p; >>> X *q = new X; >>> *(q->p) = 2; >>> if (y != 2) >>> __builtin_abort (); >>> } >>> >>> and I could fix this by not making *p but what *p points to escape. >>> The testcase is of course maximally awkward, but hey ... ;) >>> >>> Now this would all be moot if operator delete may not access >>> the object (or if the object contents are undefined at that point). >>> >>> Oh, and the testcase segfaults when compiled with GCC 10 because >>> there we elide the new X / delete p pair ... which is invalid then? >>> Hmm, we emit >>> >>> MEM[(struct X *)_8] ={v} {CLOBBER}; >>> operator delete (_8, 8); >>> >>> so the object contents are undefined _before_ calling delete >>> even when I do not have a DTOR? That is, the above, >>> w/o -fno-lifetime-dse, makes the PTA patch OK for the testcase. >> >> Yes, all classes have a destructor, even if it's trivial, so the object's >> lifetime definitely ends before the call to operator delete. This is less >> clear for scalar objects, but treating them similarly would be consistent with >> other recent changes, so I think it's fine for us to assume that scalar >> objects are also invalidated before the call to operator delete. But of >> course this doesn't apply to explicit calls to operator delete outside of a >> delete expression. > > OK, so change the testcase main slightly to > > int main() > { > int y = 1; > X *p = new X; > p->p = &y; > ::operator delete(p); > X *q = new X; > *(q->p) = 2; > if (y != 2) > __builtin_abort (); > } > > in this case the lifetime of *p does not end before calling > ::operator delete() and delete can stash the object contents > somewhere before ending its lifetime. For the very same reason > we may not elide a new/delete pair like in > > int main() > { > int *p = new int; > *p = 1; > ::operator delete (p); > } Correct; the permission to elide new/delete pairs are for the expressions, not the functions. > which we before the change did not do only because calling > operator delete made p escape. Unfortunately points-to analysis > cannot really reconstruct whether delete was called as part of > a delete expression or directly (and thus whether object lifetime > ended already), neither can DCE. So I guess we need to mark > the operator delete call in some way to make those transforms > safe. At least currently any operator delete call makes the > alias guarantee of a operator new call moot by forcing the object > to be aliased with all global and escaped memory ... > > Looks like there are some unallocated flags for CALL_EXPR we could > pick but I wonder if we can recycle protected_flag which is > > CALL_FROM_THUNK_P and > CALL_ALLOCA_FOR_VAR_P in > CALL_EXPR > > for calls to DECL_IS_OPERATOR_{NEW,DELETE}_P, thus whether > we have CALL_FROM_THUNK_P for those operators. Guess picking > a new flag is safer. We won't ever call those operators from a thunk, so it should be OK to reuse it. > But, does it seem correct that we need to distinguish > delete expressions from plain calls to operator delete? A reason for that distinction came up in the context of omitting new/delete pairs: we want to consider the operator first called by the new or delete expression, not a call from that first operator to another operator new/delete and exposed by inlining. https://gcc.gnu.org/pipermail/gcc-patches/2020-April/543404.html > In this context I also wonder about non-replaceable operator delete, > specifically operator delete in classes - are there any semantic > differences between those or why did we choose to only mark > the replaceable ones? The standard says that for omitting a 'new' allocation, the operator new has to be a replaceable one, but does not say the same about 'delete'; it just says that if the allocation was omitted, the delete-expression does not call a deallocation function. It may not be necessary to make this distinction for delete. And this distinction could be local to the front end. In the front end, we currently have cxx_replaceable_global_alloc_fn that already ignores the replaceability of operator delete. And we have CALL_FROM_NEW_OR_DELETE_P, that would just need to move into the middle end. And perhaps get renamed to CALL_OMITTABLE_NEW_OR_DELETE_P, and not get set for calls to non-replaceable operator new. Jason
On Fri, 25 Sep 2020, Jason Merrill wrote: > On 9/25/20 2:30 AM, Richard Biener wrote: > > On Thu, 24 Sep 2020, Jason Merrill wrote: > > > >> On 9/24/20 3:43 AM, Richard Biener wrote: > >>> On Wed, 23 Sep 2020, Jason Merrill wrote: > >>> > >>>> On 9/23/20 2:42 PM, Richard Biener wrote: > >>>>> On September 23, 2020 7:53:18 PM GMT+02:00, Jason Merrill > >>>>> <jason@redhat.com> > >>>>> wrote: > >>>>>> On 9/23/20 4:14 AM, Richard Biener wrote: > >>>>>>> C++ operator delete, when DECL_IS_REPLACEABLE_OPERATOR_DELETE_P, > >>>>>>> does not cause the deleted object to be escaped. It also has no > >>>>>>> other interesting side-effects for PTA so skip it like we do > >>>>>>> for BUILT_IN_FREE. > >>>>>> > >>>>>> Hmm, this is true of the default implementation, but since the function > >>>>>> > >>>>>> is replaceable, we don't know what a user definition might do with the > >>>>>> pointer. > >>>>> > >>>>> But can the object still be 'used' after delete? Can delete fail / > >>>>> throw? > >>>>> > >>>>> What guarantee does the predicate give us? > >>>> > >>>> The deallocation function is called as part of a delete expression in > >>>> order > >>>> to > >>>> release the storage for an object, ending its lifetime (if it was not > >>>> ended > >>>> by > >>>> a destructor), so no, the object can't be used afterward. > >>> > >>> OK, but the delete operator can access the object contents if there > >>> wasn't a destructor ... > >> > >>>> A deallocation function that throws has undefined behavior. > >>> > >>> OK, so it seems the 'replaceable' operators are the global ones > >>> (for user-defined/class-specific placement variants I see arbitrary > >>> extra arguments that we'd possibly need to handle). > >>> > >>> I'm happy to revert but I'd like to have a testcase that FAILs > >>> with the patch ;) > >>> > >>> Now, the following aborts: > >>> > >>> struct X { > >>> static struct X saved; > >>> int *p; > >>> X() { __builtin_memcpy (this, &saved, sizeof (X)); } > >>> }; > >>> void operator delete (void *p) > >>> { > >>> __builtin_memcpy (&X::saved, p, sizeof (X)); > >>> } > >>> int main() > >>> { > >>> int y = 1; > >>> X *p = new X; > >>> p->p = &y; > >>> delete p; > >>> X *q = new X; > >>> *(q->p) = 2; > >>> if (y != 2) > >>> __builtin_abort (); > >>> } > >>> > >>> and I could fix this by not making *p but what *p points to escape. > >>> The testcase is of course maximally awkward, but hey ... ;) > >>> > >>> Now this would all be moot if operator delete may not access > >>> the object (or if the object contents are undefined at that point). > >>> > >>> Oh, and the testcase segfaults when compiled with GCC 10 because > >>> there we elide the new X / delete p pair ... which is invalid then? > >>> Hmm, we emit > >>> > >>> MEM[(struct X *)_8] ={v} {CLOBBER}; > >>> operator delete (_8, 8); > >>> > >>> so the object contents are undefined _before_ calling delete > >>> even when I do not have a DTOR? That is, the above, > >>> w/o -fno-lifetime-dse, makes the PTA patch OK for the testcase. > >> > >> Yes, all classes have a destructor, even if it's trivial, so the object's > >> lifetime definitely ends before the call to operator delete. This is less > >> clear for scalar objects, but treating them similarly would be consistent > >> with > >> other recent changes, so I think it's fine for us to assume that scalar > >> objects are also invalidated before the call to operator delete. But of > >> course this doesn't apply to explicit calls to operator delete outside of a > >> delete expression. > > > > OK, so change the testcase main slightly to > > > > int main() > > { > > int y = 1; > > X *p = new X; > > p->p = &y; > > ::operator delete(p); > > X *q = new X; > > *(q->p) = 2; > > if (y != 2) > > __builtin_abort (); > > } > > > > in this case the lifetime of *p does not end before calling > > ::operator delete() and delete can stash the object contents > > somewhere before ending its lifetime. For the very same reason > > we may not elide a new/delete pair like in > > > > int main() > > { > > int *p = new int; > > *p = 1; > > ::operator delete (p); > > } > > Correct; the permission to elide new/delete pairs are for the expressions, not > the functions. > > > which we before the change did not do only because calling > > operator delete made p escape. Unfortunately points-to analysis > > cannot really reconstruct whether delete was called as part of > > a delete expression or directly (and thus whether object lifetime > > ended already), neither can DCE. So I guess we need to mark > > the operator delete call in some way to make those transforms > > safe. At least currently any operator delete call makes the > > alias guarantee of a operator new call moot by forcing the object > > to be aliased with all global and escaped memory ... > > > > Looks like there are some unallocated flags for CALL_EXPR we could > > pick but I wonder if we can recycle protected_flag which is > > > > CALL_FROM_THUNK_P and > > CALL_ALLOCA_FOR_VAR_P in > > CALL_EXPR > > > > for calls to DECL_IS_OPERATOR_{NEW,DELETE}_P, thus whether > > we have CALL_FROM_THUNK_P for those operators. Guess picking > > a new flag is safer. > > We won't ever call those operators from a thunk, so it should be OK to reuse > it. > > > But, does it seem correct that we need to distinguish > > delete expressions from plain calls to operator delete? > > A reason for that distinction came up in the context of omitting new/delete > pairs: we want to consider the operator first called by the new or delete > expression, not a call from that first operator to another operator new/delete > and exposed by inlining. > > https://gcc.gnu.org/pipermail/gcc-patches/2020-April/543404.html > > > In this context I also wonder about non-replaceable operator delete, > > specifically operator delete in classes - are there any semantic > > differences between those or why did we choose to only mark > > the replaceable ones? > > The standard says that for omitting a 'new' allocation, the operator new has > to be a replaceable one, but does not say the same about 'delete'; it just > says that if the allocation was omitted, the delete-expression does not call a > deallocation function. It may not be necessary to make this distinction for > delete. And this distinction could be local to the front end. > > In the front end, we currently have cxx_replaceable_global_alloc_fn that > already ignores the replaceability of operator delete. And we have > CALL_FROM_NEW_OR_DELETE_P, that would just need to move into the middle end. > And perhaps get renamed to CALL_OMITTABLE_NEW_OR_DELETE_P, and not get set for > calls to non-replaceable operator new. CALL_FROM_NEW_OR_DELETE_P indeed looks like the best fit - it's only evaluated when cxx_replaceable_global_alloc_fn matches in the C++ FE so could be made to cover only replaceable variants. CALL_REPLACEABLE_NEW_OR_DELETE_P () maybe, since we already use REPLACEABLE for the fndecl flags? OMITTABLE is too specific for the PTA case where it really matters whether the object lifetime ends before the delete call, not whether it can be omitted (hmm, guess that's not 100% overlap then either...). Mind doing the C++ side of things recycling protected_flag as suggested? Thanks, Richard.
On 9/28/20 3:56 AM, Richard Biener wrote: > On Fri, 25 Sep 2020, Jason Merrill wrote: > >> On 9/25/20 2:30 AM, Richard Biener wrote: >>> On Thu, 24 Sep 2020, Jason Merrill wrote: >>> >>>> On 9/24/20 3:43 AM, Richard Biener wrote: >>>>> On Wed, 23 Sep 2020, Jason Merrill wrote: >>>>> >>>>>> On 9/23/20 2:42 PM, Richard Biener wrote: >>>>>>> On September 23, 2020 7:53:18 PM GMT+02:00, Jason Merrill >>>>>>> <jason@redhat.com> >>>>>>> wrote: >>>>>>>> On 9/23/20 4:14 AM, Richard Biener wrote: >>>>>>>>> C++ operator delete, when DECL_IS_REPLACEABLE_OPERATOR_DELETE_P, >>>>>>>>> does not cause the deleted object to be escaped. It also has no >>>>>>>>> other interesting side-effects for PTA so skip it like we do >>>>>>>>> for BUILT_IN_FREE. >>>>>>>> >>>>>>>> Hmm, this is true of the default implementation, but since the function >>>>>>>> >>>>>>>> is replaceable, we don't know what a user definition might do with the >>>>>>>> pointer. >>>>>>> >>>>>>> But can the object still be 'used' after delete? Can delete fail / >>>>>>> throw? >>>>>>> >>>>>>> What guarantee does the predicate give us? >>>>>> >>>>>> The deallocation function is called as part of a delete expression in >>>>>> order >>>>>> to >>>>>> release the storage for an object, ending its lifetime (if it was not >>>>>> ended >>>>>> by >>>>>> a destructor), so no, the object can't be used afterward. >>>>> >>>>> OK, but the delete operator can access the object contents if there >>>>> wasn't a destructor ... >>>> >>>>>> A deallocation function that throws has undefined behavior. >>>>> >>>>> OK, so it seems the 'replaceable' operators are the global ones >>>>> (for user-defined/class-specific placement variants I see arbitrary >>>>> extra arguments that we'd possibly need to handle). >>>>> >>>>> I'm happy to revert but I'd like to have a testcase that FAILs >>>>> with the patch ;) >>>>> >>>>> Now, the following aborts: >>>>> >>>>> struct X { >>>>> static struct X saved; >>>>> int *p; >>>>> X() { __builtin_memcpy (this, &saved, sizeof (X)); } >>>>> }; >>>>> void operator delete (void *p) >>>>> { >>>>> __builtin_memcpy (&X::saved, p, sizeof (X)); >>>>> } >>>>> int main() >>>>> { >>>>> int y = 1; >>>>> X *p = new X; >>>>> p->p = &y; >>>>> delete p; >>>>> X *q = new X; >>>>> *(q->p) = 2; >>>>> if (y != 2) >>>>> __builtin_abort (); >>>>> } >>>>> >>>>> and I could fix this by not making *p but what *p points to escape. >>>>> The testcase is of course maximally awkward, but hey ... ;) >>>>> >>>>> Now this would all be moot if operator delete may not access >>>>> the object (or if the object contents are undefined at that point). >>>>> >>>>> Oh, and the testcase segfaults when compiled with GCC 10 because >>>>> there we elide the new X / delete p pair ... which is invalid then? >>>>> Hmm, we emit >>>>> >>>>> MEM[(struct X *)_8] ={v} {CLOBBER}; >>>>> operator delete (_8, 8); >>>>> >>>>> so the object contents are undefined _before_ calling delete >>>>> even when I do not have a DTOR? That is, the above, >>>>> w/o -fno-lifetime-dse, makes the PTA patch OK for the testcase. >>>> >>>> Yes, all classes have a destructor, even if it's trivial, so the object's >>>> lifetime definitely ends before the call to operator delete. This is less >>>> clear for scalar objects, but treating them similarly would be consistent >>>> with >>>> other recent changes, so I think it's fine for us to assume that scalar >>>> objects are also invalidated before the call to operator delete. But of >>>> course this doesn't apply to explicit calls to operator delete outside of a >>>> delete expression. >>> >>> OK, so change the testcase main slightly to >>> >>> int main() >>> { >>> int y = 1; >>> X *p = new X; >>> p->p = &y; >>> ::operator delete(p); >>> X *q = new X; >>> *(q->p) = 2; >>> if (y != 2) >>> __builtin_abort (); >>> } >>> >>> in this case the lifetime of *p does not end before calling >>> ::operator delete() and delete can stash the object contents >>> somewhere before ending its lifetime. For the very same reason >>> we may not elide a new/delete pair like in >>> >>> int main() >>> { >>> int *p = new int; >>> *p = 1; >>> ::operator delete (p); >>> } >> >> Correct; the permission to elide new/delete pairs are for the expressions, not >> the functions. >> >>> which we before the change did not do only because calling >>> operator delete made p escape. Unfortunately points-to analysis >>> cannot really reconstruct whether delete was called as part of >>> a delete expression or directly (and thus whether object lifetime >>> ended already), neither can DCE. So I guess we need to mark >>> the operator delete call in some way to make those transforms >>> safe. At least currently any operator delete call makes the >>> alias guarantee of a operator new call moot by forcing the object >>> to be aliased with all global and escaped memory ... >>> >>> Looks like there are some unallocated flags for CALL_EXPR we could >>> pick but I wonder if we can recycle protected_flag which is >>> >>> CALL_FROM_THUNK_P and >>> CALL_ALLOCA_FOR_VAR_P in >>> CALL_EXPR >>> >>> for calls to DECL_IS_OPERATOR_{NEW,DELETE}_P, thus whether >>> we have CALL_FROM_THUNK_P for those operators. Guess picking >>> a new flag is safer. >> >> We won't ever call those operators from a thunk, so it should be OK to reuse >> it. >> >>> But, does it seem correct that we need to distinguish >>> delete expressions from plain calls to operator delete? >> >> A reason for that distinction came up in the context of omitting new/delete >> pairs: we want to consider the operator first called by the new or delete >> expression, not a call from that first operator to another operator new/delete >> and exposed by inlining. >> >> https://gcc.gnu.org/pipermail/gcc-patches/2020-April/543404.html >> >>> In this context I also wonder about non-replaceable operator delete, >>> specifically operator delete in classes - are there any semantic >>> differences between those or why did we choose to only mark >>> the replaceable ones? >> >> The standard says that for omitting a 'new' allocation, the operator new has >> to be a replaceable one, but does not say the same about 'delete'; it just >> says that if the allocation was omitted, the delete-expression does not call a >> deallocation function. It may not be necessary to make this distinction for >> delete. And this distinction could be local to the front end. >> >> In the front end, we currently have cxx_replaceable_global_alloc_fn that >> already ignores the replaceability of operator delete. And we have >> CALL_FROM_NEW_OR_DELETE_P, that would just need to move into the middle end. >> And perhaps get renamed to CALL_OMITTABLE_NEW_OR_DELETE_P, and not get set for >> calls to non-replaceable operator new. > > CALL_FROM_NEW_OR_DELETE_P indeed looks like the best fit - it's > only evaluated when cxx_replaceable_global_alloc_fn matches in the C++ > FE so could be made to cover only replaceable variants. > > CALL_REPLACEABLE_NEW_OR_DELETE_P () maybe, since we already use > REPLACEABLE for the fndecl flags? OMITTABLE is too specific > for the PTA case where it really matters whether the object > lifetime ends before the delete call, not whether it can be > omitted (hmm, guess that's not 100% overlap then either...). That seems like good overlap to me, if we agree that object lifetime ends before any delete call from a delete-expression, whether or not the operator delete is replaceable. > Mind doing the C++ side of things recycling protected_flag as suggested? OK. Jason
On 9/28/20 3:09 PM, Jason Merrill wrote: > On 9/28/20 3:56 AM, Richard Biener wrote: >> On Fri, 25 Sep 2020, Jason Merrill wrote: >> >>> On 9/25/20 2:30 AM, Richard Biener wrote: >>>> On Thu, 24 Sep 2020, Jason Merrill wrote: >>>> >>>>> On 9/24/20 3:43 AM, Richard Biener wrote: >>>>>> On Wed, 23 Sep 2020, Jason Merrill wrote: >>>>>> >>>>>>> On 9/23/20 2:42 PM, Richard Biener wrote: >>>>>>>> On September 23, 2020 7:53:18 PM GMT+02:00, Jason Merrill >>>>>>>> <jason@redhat.com> >>>>>>>> wrote: >>>>>>>>> On 9/23/20 4:14 AM, Richard Biener wrote: >>>>>>>>>> C++ operator delete, when DECL_IS_REPLACEABLE_OPERATOR_DELETE_P, >>>>>>>>>> does not cause the deleted object to be escaped. It also has no >>>>>>>>>> other interesting side-effects for PTA so skip it like we do >>>>>>>>>> for BUILT_IN_FREE. >>>>>>>>> >>>>>>>>> Hmm, this is true of the default implementation, but since the >>>>>>>>> function >>>>>>>>> >>>>>>>>> is replaceable, we don't know what a user definition might do >>>>>>>>> with the >>>>>>>>> pointer. >>>>>>>> >>>>>>>> But can the object still be 'used' after delete? Can delete fail / >>>>>>>> throw? >>>>>>>> >>>>>>>> What guarantee does the predicate give us? >>>>>>> >>>>>>> The deallocation function is called as part of a delete >>>>>>> expression in >>>>>>> order >>>>>>> to >>>>>>> release the storage for an object, ending its lifetime (if it was >>>>>>> not >>>>>>> ended >>>>>>> by >>>>>>> a destructor), so no, the object can't be used afterward. >>>>>> >>>>>> OK, but the delete operator can access the object contents if there >>>>>> wasn't a destructor ... >>>>> >>>>>>> A deallocation function that throws has undefined behavior. >>>>>> >>>>>> OK, so it seems the 'replaceable' operators are the global ones >>>>>> (for user-defined/class-specific placement variants I see arbitrary >>>>>> extra arguments that we'd possibly need to handle). >>>>>> >>>>>> I'm happy to revert but I'd like to have a testcase that FAILs >>>>>> with the patch ;) >>>>>> >>>>>> Now, the following aborts: >>>>>> >>>>>> struct X { >>>>>>     static struct X saved; >>>>>>     int *p; >>>>>>     X() { __builtin_memcpy (this, &saved, sizeof (X)); } >>>>>> }; >>>>>> void operator delete (void *p) >>>>>> { >>>>>>     __builtin_memcpy (&X::saved, p, sizeof (X)); >>>>>> } >>>>>> int main() >>>>>> { >>>>>>     int y = 1; >>>>>>     X *p = new X; >>>>>>     p->p = &y; >>>>>>     delete p; >>>>>>     X *q = new X; >>>>>>     *(q->p) = 2; >>>>>>     if (y != 2) >>>>>>       __builtin_abort (); >>>>>> } >>>>>> >>>>>> and I could fix this by not making *p but what *p points to escape. >>>>>> The testcase is of course maximally awkward, but hey ... ;) >>>>>> >>>>>> Now this would all be moot if operator delete may not access >>>>>> the object (or if the object contents are undefined at that point). >>>>>> >>>>>> Oh, and the testcase segfaults when compiled with GCC 10 because >>>>>> there we elide the new X / delete p pair ... which is invalid then? >>>>>> Hmm, we emit >>>>>> >>>>>>     MEM[(struct X *)_8] ={v} {CLOBBER}; >>>>>>     operator delete (_8, 8); >>>>>> >>>>>> so the object contents are undefined _before_ calling delete >>>>>> even when I do not have a DTOR? That is, the above, >>>>>> w/o -fno-lifetime-dse, makes the PTA patch OK for the testcase. >>>>> >>>>> Yes, all classes have a destructor, even if it's trivial, so the >>>>> object's >>>>> lifetime definitely ends before the call to operator delete. This >>>>> is less >>>>> clear for scalar objects, but treating them similarly would be >>>>> consistent >>>>> with >>>>> other recent changes, so I think it's fine for us to assume that >>>>> scalar >>>>> objects are also invalidated before the call to operator delete. >>>>> But of >>>>> course this doesn't apply to explicit calls to operator delete >>>>> outside of a >>>>> delete expression. >>>> >>>> OK, so change the testcase main slightly to >>>> >>>> int main() >>>> { >>>>    int y = 1; >>>>    X *p = new X; >>>>    p->p = &y; >>>>    ::operator delete(p); >>>>    X *q = new X; >>>>    *(q->p) = 2; >>>>    if (y != 2) >>>>      __builtin_abort (); >>>> } >>>> >>>> in this case the lifetime of *p does not end before calling >>>> ::operator delete() and delete can stash the object contents >>>> somewhere before ending its lifetime. For the very same reason >>>> we may not elide a new/delete pair like in >>>> >>>> int main() >>>> { >>>>    int *p = new int; >>>>    *p = 1; >>>>    ::operator delete (p); >>>> } >>> >>> Correct; the permission to elide new/delete pairs are for the >>> expressions, not >>> the functions. >>> >>>> which we before the change did not do only because calling >>>> operator delete made p escape. Unfortunately points-to analysis >>>> cannot really reconstruct whether delete was called as part of >>>> a delete expression or directly (and thus whether object lifetime >>>> ended already), neither can DCE. So I guess we need to mark >>>> the operator delete call in some way to make those transforms >>>> safe. At least currently any operator delete call makes the >>>> alias guarantee of a operator new call moot by forcing the object >>>> to be aliased with all global and escaped memory ... >>>> >>>> Looks like there are some unallocated flags for CALL_EXPR we could >>>> pick but I wonder if we can recycle protected_flag which is >>>> >>>>         CALL_FROM_THUNK_P and >>>>         CALL_ALLOCA_FOR_VAR_P in >>>>             CALL_EXPR >>>> >>>> for calls to DECL_IS_OPERATOR_{NEW,DELETE}_P, thus whether >>>> we have CALL_FROM_THUNK_P for those operators. Guess picking >>>> a new flag is safer. >>> >>> We won't ever call those operators from a thunk, so it should be OK >>> to reuse >>> it. >>> >>>> But, does it seem correct that we need to distinguish >>>> delete expressions from plain calls to operator delete? >>> >>> A reason for that distinction came up in the context of omitting >>> new/delete >>> pairs: we want to consider the operator first called by the new or >>> delete >>> expression, not a call from that first operator to another operator >>> new/delete >>> and exposed by inlining. >>> >>> https://gcc.gnu.org/pipermail/gcc-patches/2020-April/543404.html >>> >>>> In this context I also wonder about non-replaceable operator delete, >>>> specifically operator delete in classes - are there any semantic >>>> differences between those or why did we choose to only mark >>>> the replaceable ones? >>> >>> The standard says that for omitting a 'new' allocation, the operator >>> new has >>> to be a replaceable one, but does not say the same about 'delete'; it >>> just >>> says that if the allocation was omitted, the delete-expression does >>> not call a >>> deallocation function. It may not be necessary to make this >>> distinction for >>> delete. And this distinction could be local to the front end. >>> >>> In the front end, we currently have cxx_replaceable_global_alloc_fn that >>> already ignores the replaceability of operator delete. And we have >>> CALL_FROM_NEW_OR_DELETE_P, that would just need to move into the >>> middle end. >>> And perhaps get renamed to CALL_OMITTABLE_NEW_OR_DELETE_P, and not >>> get set for >>> calls to non-replaceable operator new. >> >> CALL_FROM_NEW_OR_DELETE_P indeed looks like the best fit - it's >> only evaluated when cxx_replaceable_global_alloc_fn matches in the C++ >> FE so could be made to cover only replaceable variants. >> >> CALL_REPLACEABLE_NEW_OR_DELETE_P () maybe, since we already use >> REPLACEABLE for the fndecl flags? OMITTABLE is too specific >> for the PTA case where it really matters whether the object >> lifetime ends before the delete call, not whether it can be >> omitted (hmm, guess that's not 100% overlap then either...). > > That seems like good overlap to me, if we agree that object lifetime > ends before any delete call from a delete-expression, whether or not the > operator delete is replaceable. > >> Mind doing the C++ side of things recycling protected_flag as suggested? > > OK.
On Wed, 30 Sep 2020, Jason Merrill wrote: > On 9/28/20 3:09 PM, Jason Merrill wrote: > > On 9/28/20 3:56 AM, Richard Biener wrote: > >> On Fri, 25 Sep 2020, Jason Merrill wrote: > >> > >>> On 9/25/20 2:30 AM, Richard Biener wrote: > >>>> On Thu, 24 Sep 2020, Jason Merrill wrote: > >>>> > >>>>> On 9/24/20 3:43 AM, Richard Biener wrote: > >>>>>> On Wed, 23 Sep 2020, Jason Merrill wrote: > >>>>>> > >>>>>>> On 9/23/20 2:42 PM, Richard Biener wrote: > >>>>>>>> On September 23, 2020 7:53:18 PM GMT+02:00, Jason Merrill > >>>>>>>> <jason@redhat.com> > >>>>>>>> wrote: > >>>>>>>>> On 9/23/20 4:14 AM, Richard Biener wrote: > >>>>>>>>>> C++ operator delete, when DECL_IS_REPLACEABLE_OPERATOR_DELETE_P, > >>>>>>>>>> does not cause the deleted object to be escaped.? It also has no > >>>>>>>>>> other interesting side-effects for PTA so skip it like we do > >>>>>>>>>> for BUILT_IN_FREE. > >>>>>>>>> > >>>>>>>>> Hmm, this is true of the default implementation, but since the > >>>>>>>>> function > >>>>>>>>> > >>>>>>>>> is replaceable, we don't know what a user definition might do with > >>>>>>>>> the > >>>>>>>>> pointer. > >>>>>>>> > >>>>>>>> But can the object still be 'used' after delete? Can delete fail / > >>>>>>>> throw? > >>>>>>>> > >>>>>>>> What guarantee does the predicate give us? > >>>>>>> > >>>>>>> The deallocation function is called as part of a delete expression in > >>>>>>> order > >>>>>>> to > >>>>>>> release the storage for an object, ending its lifetime (if it was not > >>>>>>> ended > >>>>>>> by > >>>>>>> a destructor), so no, the object can't be used afterward. > >>>>>> > >>>>>> OK, but the delete operator can access the object contents if there > >>>>>> wasn't a destructor ... > >>>>> > >>>>>>> A deallocation function that throws has undefined behavior. > >>>>>> > >>>>>> OK, so it seems the 'replaceable' operators are the global ones > >>>>>> (for user-defined/class-specific placement variants I see arbitrary > >>>>>> extra arguments that we'd possibly need to handle). > >>>>>> > >>>>>> I'm happy to revert but I'd like to have a testcase that FAILs > >>>>>> with the patch ;) > >>>>>> > >>>>>> Now, the following aborts: > >>>>>> > >>>>>> struct X { > >>>>>> ???? static struct X saved; > >>>>>> ???? int *p; > >>>>>> ???? X() { __builtin_memcpy (this, &saved, sizeof (X)); } > >>>>>> }; > >>>>>> void operator delete (void *p) > >>>>>> { > >>>>>> ???? __builtin_memcpy (&X::saved, p, sizeof (X)); > >>>>>> } > >>>>>> int main() > >>>>>> { > >>>>>> ???? int y = 1; > >>>>>> ???? X *p = new X; > >>>>>> ???? p->p = &y; > >>>>>> ???? delete p; > >>>>>> ???? X *q = new X; > >>>>>> ???? *(q->p) = 2; > >>>>>> ???? if (y != 2) > >>>>>> ?????? __builtin_abort (); > >>>>>> } > >>>>>> > >>>>>> and I could fix this by not making *p but what *p points to escape. > >>>>>> The testcase is of course maximally awkward, but hey ... ;) > >>>>>> > >>>>>> Now this would all be moot if operator delete may not access > >>>>>> the object (or if the object contents are undefined at that point). > >>>>>> > >>>>>> Oh, and the testcase segfaults when compiled with GCC 10 because > >>>>>> there we elide the new X / delete p pair ... which is invalid then? > >>>>>> Hmm, we emit > >>>>>> > >>>>>> ???? MEM[(struct X *)_8] ={v} {CLOBBER}; > >>>>>> ???? operator delete (_8, 8); > >>>>>> > >>>>>> so the object contents are undefined _before_ calling delete > >>>>>> even when I do not have a DTOR?? That is, the above, > >>>>>> w/o -fno-lifetime-dse, makes the PTA patch OK for the testcase. > >>>>> > >>>>> Yes, all classes have a destructor, even if it's trivial, so the > >>>>> object's > >>>>> lifetime definitely ends before the call to operator delete. This is > >>>>> less > >>>>> clear for scalar objects, but treating them similarly would be > >>>>> consistent > >>>>> with > >>>>> other recent changes, so I think it's fine for us to assume that scalar > >>>>> objects are also invalidated before the call to operator delete. But of > >>>>> course this doesn't apply to explicit calls to operator delete outside > >>>>> of a > >>>>> delete expression. > >>>> > >>>> OK, so change the testcase main slightly to > >>>> > >>>> int main() > >>>> { > >>>> ??? int y = 1; > >>>> ??? X *p = new X; > >>>> ??? p->p = &y; > >>>> ??? ::operator delete(p); > >>>> ??? X *q = new X; > >>>> ??? *(q->p) = 2; > >>>> ??? if (y != 2) > >>>> ????? __builtin_abort (); > >>>> } > >>>> > >>>> in this case the lifetime of *p does not end before calling > >>>> ::operator delete() and delete can stash the object contents > >>>> somewhere before ending its lifetime.? For the very same reason > >>>> we may not elide a new/delete pair like in > >>>> > >>>> int main() > >>>> { > >>>> ??? int *p = new int; > >>>> ??? *p = 1; > >>>> ??? ::operator delete (p); > >>>> } > >>> > >>> Correct; the permission to elide new/delete pairs are for the expressions, > >>> not > >>> the functions. > >>> > >>>> which we before the change did not do only because calling > >>>> operator delete made p escape.? Unfortunately points-to analysis > >>>> cannot really reconstruct whether delete was called as part of > >>>> a delete expression or directly (and thus whether object lifetime > >>>> ended already), neither can DCE.? So I guess we need to mark > >>>> the operator delete call in some way to make those transforms > >>>> safe.? At least currently any operator delete call makes the > >>>> alias guarantee of a operator new call moot by forcing the object > >>>> to be aliased with all global and escaped memory ... > >>>> > >>>> Looks like there are some unallocated flags for CALL_EXPR we could > >>>> pick but I wonder if we can recycle protected_flag which is > >>>> > >>>> ???????? CALL_FROM_THUNK_P and > >>>> ???????? CALL_ALLOCA_FOR_VAR_P in > >>>> ???????????? CALL_EXPR > >>>> > >>>> for calls to DECL_IS_OPERATOR_{NEW,DELETE}_P, thus whether > >>>> we have CALL_FROM_THUNK_P for those operators.? Guess picking > >>>> a new flag is safer. > >>> > >>> We won't ever call those operators from a thunk, so it should be OK to > >>> reuse > >>> it. > >>> > >>>> But, does it seem correct that we need to distinguish > >>>> delete expressions from plain calls to operator delete? > >>> > >>> A reason for that distinction came up in the context of omitting > >>> new/delete > >>> pairs: we want to consider the operator first called by the new or delete > >>> expression, not a call from that first operator to another operator > >>> new/delete > >>> and exposed by inlining. > >>> > >>> https://gcc.gnu.org/pipermail/gcc-patches/2020-April/543404.html > >>> > >>>> In this context I also wonder about non-replaceable operator delete, > >>>> specifically operator delete in classes - are there any semantic > >>>> differences between those or why did we choose to only mark > >>>> the replaceable ones? > >>> > >>> The standard says that for omitting a 'new' allocation, the operator new > >>> has > >>> to be a replaceable one, but does not say the same about 'delete'; it just > >>> says that if the allocation was omitted, the delete-expression does not > >>> call a > >>> deallocation function.? It may not be necessary to make this distinction > >>> for > >>> delete.? And this distinction could be local to the front end. > >>> > >>> In the front end, we currently have cxx_replaceable_global_alloc_fn that > >>> already ignores the replaceability of operator delete.? And we have > >>> CALL_FROM_NEW_OR_DELETE_P, that would just need to move into the middle > >>> end. > >>> And perhaps get renamed to CALL_OMITTABLE_NEW_OR_DELETE_P, and not get set > >>> for > >>> calls to non-replaceable operator new. > >> > >> CALL_FROM_NEW_OR_DELETE_P indeed looks like the best fit - it's > >> only evaluated when cxx_replaceable_global_alloc_fn matches in the C++ > >> FE so could be made to cover only replaceable variants. > >> > >> CALL_REPLACEABLE_NEW_OR_DELETE_P () maybe, since we already use > >> REPLACEABLE for the fndecl flags?? OMITTABLE is too specific > >> for the PTA case where it really matters whether the object > >> lifetime ends before the delete call, not whether it can be > >> omitted (hmm, guess that's not 100% overlap then either...). > > > > That seems like good overlap to me, if we agree that object lifetime ends > > before any delete call from a delete-expression, whether or not the operator > > delete is replaceable. > > > >> Mind doing the C++ side of things recycling protected_flag as suggested? > > > > OK. Find attached a patch series, your patch plus the GIMPLE side of the fix. This shows that the C++ FE side is possibly incomplete with for example g++.dg/pr94314-2.C now FAILing. There we have A *a = new A (argc); delete a; being expanded to <<cleanup_point <<< Unknown tree: expr_stmt (void) (a = TARGET_EXPR <D.2357, operator new (1)>;, try { A::A ((struct A *) D.2357, argc); } catch { operator delete (D.2357); }, (struct A *) D.2357;) >>>>>; <<cleanup_point <<< Unknown tree: expr_stmt if (SAVE_EXPR <a> != 0B) { try { *SAVE_EXPR <a> = {CLOBBER}; } finally { operator delete ((void *) SAVE_EXPR <a>); } } else { <<< Unknown tree: void_cst >>> } >>>>>; return <retval> = 0; where the operator delete call in the catch {} expression is not marked as CALL_FROM_NEW_OR_DELETE_P, possibly because this call is compiler generated. So it's probably technically true that CALL_FROM_NEW_OR_DELETE_P is false but we still expect the same guarantees to hold here, in particular we expect to elide the new/delete pair? Thanks, Richard. From 602853335c5f01adb8594b6b1dc49c365c5cd731 Mon Sep 17 00:00:00 2001 From: Jason Merrill <jason@redhat.com> Date: Thu, 1 Oct 2020 10:08:58 +0200 Subject: [PATCH 1/2] c++: Move CALL_FROM_NEW_OR_DELETE_P to tree.h To: gcc-patches@gcc.gnu.org As discussed with richi, we should be able to use TREE_PROTECTED for this flag, since CALL_FROM_THUNK_P will never be set on a call to an operator new or delete. 2020-10-01 Jason Merril <jason@redhat.com> gcc/cp/ChangeLog: * lambda.c (call_from_lambda_thunk_p): New. * cp-gimplify.c (cp_genericize_r): Use it. * pt.c (tsubst_copy_and_build): Use it. * typeck.c (check_return_expr): Use it. * cp-tree.h: Declare it. (CALL_FROM_NEW_OR_DELETE_P): Move to gcc/tree.h. gcc/ChangeLog: * tree.h (CALL_FROM_NEW_OR_DELETE_P): Move from cp-tree.h. --- gcc/cp/cp-gimplify.c | 2 +- gcc/cp/cp-tree.h | 7 +------ gcc/cp/lambda.c | 7 +++++++ gcc/cp/pt.c | 2 +- gcc/cp/typeck.c | 2 +- gcc/tree-core.h | 3 ++- gcc/tree.h | 9 ++++++++- 7 files changed, 21 insertions(+), 11 deletions(-) diff --git a/gcc/cp/cp-gimplify.c b/gcc/cp/cp-gimplify.c index bc8a03c7b41..07549828dc9 100644 --- a/gcc/cp/cp-gimplify.c +++ b/gcc/cp/cp-gimplify.c @@ -962,7 +962,7 @@ cp_genericize_r (tree *stmt_p, int *walk_subtrees, void *data) omp_cxx_notice_variable (wtd->omp_ctx, stmt); /* Don't dereference parms in a thunk, pass the references through. */ - if ((TREE_CODE (stmt) == CALL_EXPR && CALL_FROM_THUNK_P (stmt)) + if ((TREE_CODE (stmt) == CALL_EXPR && call_from_lambda_thunk_p (stmt)) || (TREE_CODE (stmt) == AGGR_INIT_EXPR && AGGR_INIT_FROM_THUNK_P (stmt))) { *walk_subtrees = 0; diff --git a/gcc/cp/cp-tree.h b/gcc/cp/cp-tree.h index 48a4074b370..cf6e02c3bc5 100644 --- a/gcc/cp/cp-tree.h +++ b/gcc/cp/cp-tree.h @@ -464,7 +464,6 @@ extern GTY(()) tree cp_global_trees[CPTI_MAX]; SWITCH_STMT_NO_BREAK_P (in SWITCH_STMT) LAMBDA_EXPR_CAPTURE_OPTIMIZED (in LAMBDA_EXPR) IMPLICIT_CONV_EXPR_BRACED_INIT (in IMPLICIT_CONV_EXPR) - CALL_FROM_NEW_OR_DELETE_P (in CALL_EXPR) 3: IMPLICIT_RVALUE_P (in NON_LVALUE_EXPR or STATIC_CAST_EXPR) ICS_BAD_FLAG (in _CONV) FN_TRY_BLOCK_P (in TRY_BLOCK) @@ -3839,11 +3838,6 @@ struct GTY(()) lang_decl { should be performed at instantiation time. */ #define KOENIG_LOOKUP_P(NODE) TREE_LANG_FLAG_0 (CALL_EXPR_CHECK (NODE)) -/* In a CALL_EXPR, true for allocator calls from new or delete - expressions. */ -#define CALL_FROM_NEW_OR_DELETE_P(NODE) \ - TREE_LANG_FLAG_2 (CALL_EXPR_CHECK (NODE)) - /* True if the arguments to NODE should be evaluated in left-to-right order regardless of PUSH_ARGS_REVERSED. */ #define CALL_EXPR_ORDERED_ARGS(NODE) \ @@ -7279,6 +7273,7 @@ extern bool lambda_fn_in_template_p (tree); extern void maybe_add_lambda_conv_op (tree); extern bool is_lambda_ignored_entity (tree); extern bool lambda_static_thunk_p (tree); +extern bool call_from_lambda_thunk_p (tree); extern tree finish_builtin_launder (location_t, tree, tsubst_flags_t); extern tree cp_build_vec_convert (tree, location_t, tree, diff --git a/gcc/cp/lambda.c b/gcc/cp/lambda.c index 07a5401c97b..1a1647f465e 100644 --- a/gcc/cp/lambda.c +++ b/gcc/cp/lambda.c @@ -1325,6 +1325,13 @@ lambda_static_thunk_p (tree fn) && LAMBDA_TYPE_P (CP_DECL_CONTEXT (fn))); } +bool +call_from_lambda_thunk_p (tree call) +{ + return (CALL_FROM_THUNK_P (call) + && lambda_static_thunk_p (current_function_decl)); +} + /* Returns true iff VAL is a lambda-related declaration which should be ignored by unqualified lookup. */ diff --git a/gcc/cp/pt.c b/gcc/cp/pt.c index 869477f2c2e..f0b6414b0e2 100644 --- a/gcc/cp/pt.c +++ b/gcc/cp/pt.c @@ -19955,7 +19955,7 @@ tsubst_copy_and_build (tree t, /* Stripped-down processing for a call in a thunk. Specifically, in the thunk template for a generic lambda. */ - if (CALL_FROM_THUNK_P (t)) + if (call_from_lambda_thunk_p (t)) { /* Now that we've expanded any packs, the number of call args might be different. */ diff --git a/gcc/cp/typeck.c b/gcc/cp/typeck.c index 9166156a5d5..95b36a92491 100644 --- a/gcc/cp/typeck.c +++ b/gcc/cp/typeck.c @@ -10171,7 +10171,7 @@ check_return_expr (tree retval, bool *no_warning) /* The call in a (lambda) thunk needs no conversions. */ if (TREE_CODE (retval) == CALL_EXPR - && CALL_FROM_THUNK_P (retval)) + && call_from_lambda_thunk_p (retval)) converted = true; /* First convert the value to the function's return type, then diff --git a/gcc/tree-core.h b/gcc/tree-core.h index 0e158784d0e..752bec31c3f 100644 --- a/gcc/tree-core.h +++ b/gcc/tree-core.h @@ -1220,7 +1220,8 @@ struct GTY(()) tree_base { all decls CALL_FROM_THUNK_P and - CALL_ALLOCA_FOR_VAR_P in + CALL_ALLOCA_FOR_VAR_P and + CALL_FROM_NEW_OR_DELETE_P in CALL_EXPR OMP_CLAUSE_LINEAR_VARIABLE_STRIDE in diff --git a/gcc/tree.h b/gcc/tree.h index 5bb6e7bc000..f27a7399a37 100644 --- a/gcc/tree.h +++ b/gcc/tree.h @@ -921,7 +921,8 @@ extern void omp_clause_range_check_failed (const_tree, const char *, int, (TREE_CHECK (NODE, PARM_DECL)->decl_common.decl_nonshareable_flag) /* In a CALL_EXPR, means that the call is the jump from a thunk to the - thunked-to function. */ + thunked-to function. Be careful to avoid using this macro when one of the + next two applies instead. */ #define CALL_FROM_THUNK_P(NODE) (CALL_EXPR_CHECK (NODE)->base.protected_flag) /* In a CALL_EXPR, if the function being called is BUILT_IN_ALLOCA, means that @@ -931,6 +932,12 @@ extern void omp_clause_range_check_failed (const_tree, const char *, int, #define CALL_ALLOCA_FOR_VAR_P(NODE) \ (CALL_EXPR_CHECK (NODE)->base.protected_flag) +/* In a CALL_EXPR, if the function being called is DECL_IS_OPERATOR_NEW_P or + DECL_IS_OPERATOR_DELETE_P, true for allocator calls from C++ new or delete + expressions. */ +#define CALL_FROM_NEW_OR_DELETE_P(NODE) \ + (CALL_EXPR_CHECK (NODE)->base.protected_flag) + /* Used in classes in C++. */ #define TREE_PRIVATE(NODE) ((NODE)->base.private_flag) /* Used in classes in C++. */
On 10/1/20 5:26 AM, Richard Biener wrote: > On Wed, 30 Sep 2020, Jason Merrill wrote: > >> On 9/28/20 3:09 PM, Jason Merrill wrote: >>> On 9/28/20 3:56 AM, Richard Biener wrote: >>>> On Fri, 25 Sep 2020, Jason Merrill wrote: >>>> >>>>> On 9/25/20 2:30 AM, Richard Biener wrote: >>>>>> On Thu, 24 Sep 2020, Jason Merrill wrote: >>>>>> >>>>>>> On 9/24/20 3:43 AM, Richard Biener wrote: >>>>>>>> On Wed, 23 Sep 2020, Jason Merrill wrote: >>>>>>>> >>>>>>>>> On 9/23/20 2:42 PM, Richard Biener wrote: >>>>>>>>>> On September 23, 2020 7:53:18 PM GMT+02:00, Jason Merrill >>>>>>>>>> <jason@redhat.com> >>>>>>>>>> wrote: >>>>>>>>>>> On 9/23/20 4:14 AM, Richard Biener wrote: >>>>>>>>>>>> C++ operator delete, when DECL_IS_REPLACEABLE_OPERATOR_DELETE_P, >>>>>>>>>>>> does not cause the deleted object to be escaped.? It also has no >>>>>>>>>>>> other interesting side-effects for PTA so skip it like we do >>>>>>>>>>>> for BUILT_IN_FREE. >>>>>>>>>>> >>>>>>>>>>> Hmm, this is true of the default implementation, but since the >>>>>>>>>>> function >>>>>>>>>>> >>>>>>>>>>> is replaceable, we don't know what a user definition might do with >>>>>>>>>>> the >>>>>>>>>>> pointer. >>>>>>>>>> >>>>>>>>>> But can the object still be 'used' after delete? Can delete fail / >>>>>>>>>> throw? >>>>>>>>>> >>>>>>>>>> What guarantee does the predicate give us? >>>>>>>>> >>>>>>>>> The deallocation function is called as part of a delete expression in >>>>>>>>> order >>>>>>>>> to >>>>>>>>> release the storage for an object, ending its lifetime (if it was not >>>>>>>>> ended >>>>>>>>> by >>>>>>>>> a destructor), so no, the object can't be used afterward. >>>>>>>> >>>>>>>> OK, but the delete operator can access the object contents if there >>>>>>>> wasn't a destructor ... >>>>>>> >>>>>>>>> A deallocation function that throws has undefined behavior. >>>>>>>> >>>>>>>> OK, so it seems the 'replaceable' operators are the global ones >>>>>>>> (for user-defined/class-specific placement variants I see arbitrary >>>>>>>> extra arguments that we'd possibly need to handle). >>>>>>>> >>>>>>>> I'm happy to revert but I'd like to have a testcase that FAILs >>>>>>>> with the patch ;) >>>>>>>> >>>>>>>> Now, the following aborts: >>>>>>>> >>>>>>>> struct X { >>>>>>>> ???? static struct X saved; >>>>>>>> ???? int *p; >>>>>>>> ???? X() { __builtin_memcpy (this, &saved, sizeof (X)); } >>>>>>>> }; >>>>>>>> void operator delete (void *p) >>>>>>>> { >>>>>>>> ???? __builtin_memcpy (&X::saved, p, sizeof (X)); >>>>>>>> } >>>>>>>> int main() >>>>>>>> { >>>>>>>> ???? int y = 1; >>>>>>>> ???? X *p = new X; >>>>>>>> ???? p->p = &y; >>>>>>>> ???? delete p; >>>>>>>> ???? X *q = new X; >>>>>>>> ???? *(q->p) = 2; >>>>>>>> ???? if (y != 2) >>>>>>>> ?????? __builtin_abort (); >>>>>>>> } >>>>>>>> >>>>>>>> and I could fix this by not making *p but what *p points to escape. >>>>>>>> The testcase is of course maximally awkward, but hey ... ;) >>>>>>>> >>>>>>>> Now this would all be moot if operator delete may not access >>>>>>>> the object (or if the object contents are undefined at that point). >>>>>>>> >>>>>>>> Oh, and the testcase segfaults when compiled with GCC 10 because >>>>>>>> there we elide the new X / delete p pair ... which is invalid then? >>>>>>>> Hmm, we emit >>>>>>>> >>>>>>>> ???? MEM[(struct X *)_8] ={v} {CLOBBER}; >>>>>>>> ???? operator delete (_8, 8); >>>>>>>> >>>>>>>> so the object contents are undefined _before_ calling delete >>>>>>>> even when I do not have a DTOR?? That is, the above, >>>>>>>> w/o -fno-lifetime-dse, makes the PTA patch OK for the testcase. >>>>>>> >>>>>>> Yes, all classes have a destructor, even if it's trivial, so the >>>>>>> object's >>>>>>> lifetime definitely ends before the call to operator delete. This is >>>>>>> less >>>>>>> clear for scalar objects, but treating them similarly would be >>>>>>> consistent >>>>>>> with >>>>>>> other recent changes, so I think it's fine for us to assume that scalar >>>>>>> objects are also invalidated before the call to operator delete. But of >>>>>>> course this doesn't apply to explicit calls to operator delete outside >>>>>>> of a >>>>>>> delete expression. >>>>>> >>>>>> OK, so change the testcase main slightly to >>>>>> >>>>>> int main() >>>>>> { >>>>>> ??? int y = 1; >>>>>> ??? X *p = new X; >>>>>> ??? p->p = &y; >>>>>> ??? ::operator delete(p); >>>>>> ??? X *q = new X; >>>>>> ??? *(q->p) = 2; >>>>>> ??? if (y != 2) >>>>>> ????? __builtin_abort (); >>>>>> } >>>>>> >>>>>> in this case the lifetime of *p does not end before calling >>>>>> ::operator delete() and delete can stash the object contents >>>>>> somewhere before ending its lifetime.? For the very same reason >>>>>> we may not elide a new/delete pair like in >>>>>> >>>>>> int main() >>>>>> { >>>>>> ??? int *p = new int; >>>>>> ??? *p = 1; >>>>>> ??? ::operator delete (p); >>>>>> } >>>>> >>>>> Correct; the permission to elide new/delete pairs are for the expressions, >>>>> not >>>>> the functions. >>>>> >>>>>> which we before the change did not do only because calling >>>>>> operator delete made p escape.? Unfortunately points-to analysis >>>>>> cannot really reconstruct whether delete was called as part of >>>>>> a delete expression or directly (and thus whether object lifetime >>>>>> ended already), neither can DCE.? So I guess we need to mark >>>>>> the operator delete call in some way to make those transforms >>>>>> safe.? At least currently any operator delete call makes the >>>>>> alias guarantee of a operator new call moot by forcing the object >>>>>> to be aliased with all global and escaped memory ... >>>>>> >>>>>> Looks like there are some unallocated flags for CALL_EXPR we could >>>>>> pick but I wonder if we can recycle protected_flag which is >>>>>> >>>>>> ???????? CALL_FROM_THUNK_P and >>>>>> ???????? CALL_ALLOCA_FOR_VAR_P in >>>>>> ???????????? CALL_EXPR >>>>>> >>>>>> for calls to DECL_IS_OPERATOR_{NEW,DELETE}_P, thus whether >>>>>> we have CALL_FROM_THUNK_P for those operators.? Guess picking >>>>>> a new flag is safer. >>>>> >>>>> We won't ever call those operators from a thunk, so it should be OK to >>>>> reuse >>>>> it. >>>>> >>>>>> But, does it seem correct that we need to distinguish >>>>>> delete expressions from plain calls to operator delete? >>>>> >>>>> A reason for that distinction came up in the context of omitting >>>>> new/delete >>>>> pairs: we want to consider the operator first called by the new or delete >>>>> expression, not a call from that first operator to another operator >>>>> new/delete >>>>> and exposed by inlining. >>>>> >>>>> https://gcc.gnu.org/pipermail/gcc-patches/2020-April/543404.html >>>>> >>>>>> In this context I also wonder about non-replaceable operator delete, >>>>>> specifically operator delete in classes - are there any semantic >>>>>> differences between those or why did we choose to only mark >>>>>> the replaceable ones? >>>>> >>>>> The standard says that for omitting a 'new' allocation, the operator new >>>>> has >>>>> to be a replaceable one, but does not say the same about 'delete'; it just >>>>> says that if the allocation was omitted, the delete-expression does not >>>>> call a >>>>> deallocation function.? It may not be necessary to make this distinction >>>>> for >>>>> delete.? And this distinction could be local to the front end. >>>>> >>>>> In the front end, we currently have cxx_replaceable_global_alloc_fn that >>>>> already ignores the replaceability of operator delete.? And we have >>>>> CALL_FROM_NEW_OR_DELETE_P, that would just need to move into the middle >>>>> end. >>>>> And perhaps get renamed to CALL_OMITTABLE_NEW_OR_DELETE_P, and not get set >>>>> for >>>>> calls to non-replaceable operator new. >>>> >>>> CALL_FROM_NEW_OR_DELETE_P indeed looks like the best fit - it's >>>> only evaluated when cxx_replaceable_global_alloc_fn matches in the C++ >>>> FE so could be made to cover only replaceable variants. >>>> >>>> CALL_REPLACEABLE_NEW_OR_DELETE_P () maybe, since we already use >>>> REPLACEABLE for the fndecl flags?? OMITTABLE is too specific >>>> for the PTA case where it really matters whether the object >>>> lifetime ends before the delete call, not whether it can be >>>> omitted (hmm, guess that's not 100% overlap then either...). >>> >>> That seems like good overlap to me, if we agree that object lifetime ends >>> before any delete call from a delete-expression, whether or not the operator >>> delete is replaceable. >>> >>>> Mind doing the C++ side of things recycling protected_flag as suggested? >>> >>> OK. > > Find attached a patch series, your patch plus the GIMPLE side of the fix. > This shows that the C++ FE side is possibly incomplete with for > example g++.dg/pr94314-2.C now FAILing. There we have > > A *a = new A (argc); > delete a; > > being expanded to > > <<cleanup_point <<< Unknown tree: expr_stmt > (void) (a = TARGET_EXPR <D.2357, operator new (1)>;, try > { > A::A ((struct A *) D.2357, argc); > } > catch > { > operator delete (D.2357); > }, (struct A *) D.2357;) >>>>>; > <<cleanup_point <<< Unknown tree: expr_stmt > if (SAVE_EXPR <a> != 0B) > { > try > { > *SAVE_EXPR <a> = {CLOBBER}; > } > finally > { > operator delete ((void *) SAVE_EXPR <a>); > } > } > else > { > <<< Unknown tree: void_cst >>> > } >>>>>; > return <retval> = 0; > > where the operator delete call in the catch {} expression is > not marked as CALL_FROM_NEW_OR_DELETE_P, possibly because this > call is compiler generated. > > So it's probably technically true that CALL_FROM_NEW_OR_DELETE_P is > false but we still expect the same guarantees to hold here, in > particular we expect to elide the new/delete pair? That seems like an oversight in the standard. This first patch sets the flag. The second patch in the file removes consideration of whether an operator delete is replaceable, as I was discussing earlier.
On Thu, 1 Oct 2020, Jason Merrill wrote: > On 10/1/20 5:26 AM, Richard Biener wrote: > > On Wed, 30 Sep 2020, Jason Merrill wrote: > > > >> On 9/28/20 3:09 PM, Jason Merrill wrote: > >>> On 9/28/20 3:56 AM, Richard Biener wrote: > >>>> On Fri, 25 Sep 2020, Jason Merrill wrote: > >>>> > >>>>> On 9/25/20 2:30 AM, Richard Biener wrote: > >>>>>> On Thu, 24 Sep 2020, Jason Merrill wrote: > >>>>>> > >>>>>>> On 9/24/20 3:43 AM, Richard Biener wrote: > >>>>>>>> On Wed, 23 Sep 2020, Jason Merrill wrote: > >>>>>>>> > >>>>>>>>> On 9/23/20 2:42 PM, Richard Biener wrote: > >>>>>>>>>> On September 23, 2020 7:53:18 PM GMT+02:00, Jason Merrill > >>>>>>>>>> <jason@redhat.com> > >>>>>>>>>> wrote: > >>>>>>>>>>> On 9/23/20 4:14 AM, Richard Biener wrote: > >>>>>>>>>>>> C++ operator delete, when DECL_IS_REPLACEABLE_OPERATOR_DELETE_P, > >>>>>>>>>>>> does not cause the deleted object to be escaped.? It also has no > >>>>>>>>>>>> other interesting side-effects for PTA so skip it like we do > >>>>>>>>>>>> for BUILT_IN_FREE. > >>>>>>>>>>> > >>>>>>>>>>> Hmm, this is true of the default implementation, but since the > >>>>>>>>>>> function > >>>>>>>>>>> > >>>>>>>>>>> is replaceable, we don't know what a user definition might do with > >>>>>>>>>>> the > >>>>>>>>>>> pointer. > >>>>>>>>>> > >>>>>>>>>> But can the object still be 'used' after delete? Can delete fail / > >>>>>>>>>> throw? > >>>>>>>>>> > >>>>>>>>>> What guarantee does the predicate give us? > >>>>>>>>> > >>>>>>>>> The deallocation function is called as part of a delete expression > >>>>>>>>> in > >>>>>>>>> order > >>>>>>>>> to > >>>>>>>>> release the storage for an object, ending its lifetime (if it was > >>>>>>>>> not > >>>>>>>>> ended > >>>>>>>>> by > >>>>>>>>> a destructor), so no, the object can't be used afterward. > >>>>>>>> > >>>>>>>> OK, but the delete operator can access the object contents if there > >>>>>>>> wasn't a destructor ... > >>>>>>> > >>>>>>>>> A deallocation function that throws has undefined behavior. > >>>>>>>> > >>>>>>>> OK, so it seems the 'replaceable' operators are the global ones > >>>>>>>> (for user-defined/class-specific placement variants I see arbitrary > >>>>>>>> extra arguments that we'd possibly need to handle). > >>>>>>>> > >>>>>>>> I'm happy to revert but I'd like to have a testcase that FAILs > >>>>>>>> with the patch ;) > >>>>>>>> > >>>>>>>> Now, the following aborts: > >>>>>>>> > >>>>>>>> struct X { > >>>>>>>> ???? static struct X saved; > >>>>>>>> ???? int *p; > >>>>>>>> ???? X() { __builtin_memcpy (this, &saved, sizeof (X)); } > >>>>>>>> }; > >>>>>>>> void operator delete (void *p) > >>>>>>>> { > >>>>>>>> ???? __builtin_memcpy (&X::saved, p, sizeof (X)); > >>>>>>>> } > >>>>>>>> int main() > >>>>>>>> { > >>>>>>>> ???? int y = 1; > >>>>>>>> ???? X *p = new X; > >>>>>>>> ???? p->p = &y; > >>>>>>>> ???? delete p; > >>>>>>>> ???? X *q = new X; > >>>>>>>> ???? *(q->p) = 2; > >>>>>>>> ???? if (y != 2) > >>>>>>>> ?????? __builtin_abort (); > >>>>>>>> } > >>>>>>>> > >>>>>>>> and I could fix this by not making *p but what *p points to escape. > >>>>>>>> The testcase is of course maximally awkward, but hey ... ;) > >>>>>>>> > >>>>>>>> Now this would all be moot if operator delete may not access > >>>>>>>> the object (or if the object contents are undefined at that point). > >>>>>>>> > >>>>>>>> Oh, and the testcase segfaults when compiled with GCC 10 because > >>>>>>>> there we elide the new X / delete p pair ... which is invalid then? > >>>>>>>> Hmm, we emit > >>>>>>>> > >>>>>>>> ???? MEM[(struct X *)_8] ={v} {CLOBBER}; > >>>>>>>> ???? operator delete (_8, 8); > >>>>>>>> > >>>>>>>> so the object contents are undefined _before_ calling delete > >>>>>>>> even when I do not have a DTOR?? That is, the above, > >>>>>>>> w/o -fno-lifetime-dse, makes the PTA patch OK for the testcase. > >>>>>>> > >>>>>>> Yes, all classes have a destructor, even if it's trivial, so the > >>>>>>> object's > >>>>>>> lifetime definitely ends before the call to operator delete. This is > >>>>>>> less > >>>>>>> clear for scalar objects, but treating them similarly would be > >>>>>>> consistent > >>>>>>> with > >>>>>>> other recent changes, so I think it's fine for us to assume that > >>>>>>> scalar > >>>>>>> objects are also invalidated before the call to operator delete. But > >>>>>>> of > >>>>>>> course this doesn't apply to explicit calls to operator delete outside > >>>>>>> of a > >>>>>>> delete expression. > >>>>>> > >>>>>> OK, so change the testcase main slightly to > >>>>>> > >>>>>> int main() > >>>>>> { > >>>>>> ??? int y = 1; > >>>>>> ??? X *p = new X; > >>>>>> ??? p->p = &y; > >>>>>> ??? ::operator delete(p); > >>>>>> ??? X *q = new X; > >>>>>> ??? *(q->p) = 2; > >>>>>> ??? if (y != 2) > >>>>>> ????? __builtin_abort (); > >>>>>> } > >>>>>> > >>>>>> in this case the lifetime of *p does not end before calling > >>>>>> ::operator delete() and delete can stash the object contents > >>>>>> somewhere before ending its lifetime.? For the very same reason > >>>>>> we may not elide a new/delete pair like in > >>>>>> > >>>>>> int main() > >>>>>> { > >>>>>> ??? int *p = new int; > >>>>>> ??? *p = 1; > >>>>>> ??? ::operator delete (p); > >>>>>> } > >>>>> > >>>>> Correct; the permission to elide new/delete pairs are for the > >>>>> expressions, > >>>>> not > >>>>> the functions. > >>>>> > >>>>>> which we before the change did not do only because calling > >>>>>> operator delete made p escape.? Unfortunately points-to analysis > >>>>>> cannot really reconstruct whether delete was called as part of > >>>>>> a delete expression or directly (and thus whether object lifetime > >>>>>> ended already), neither can DCE.? So I guess we need to mark > >>>>>> the operator delete call in some way to make those transforms > >>>>>> safe.? At least currently any operator delete call makes the > >>>>>> alias guarantee of a operator new call moot by forcing the object > >>>>>> to be aliased with all global and escaped memory ... > >>>>>> > >>>>>> Looks like there are some unallocated flags for CALL_EXPR we could > >>>>>> pick but I wonder if we can recycle protected_flag which is > >>>>>> > >>>>>> ???????? CALL_FROM_THUNK_P and > >>>>>> ???????? CALL_ALLOCA_FOR_VAR_P in > >>>>>> ???????????? CALL_EXPR > >>>>>> > >>>>>> for calls to DECL_IS_OPERATOR_{NEW,DELETE}_P, thus whether > >>>>>> we have CALL_FROM_THUNK_P for those operators.? Guess picking > >>>>>> a new flag is safer. > >>>>> > >>>>> We won't ever call those operators from a thunk, so it should be OK to > >>>>> reuse > >>>>> it. > >>>>> > >>>>>> But, does it seem correct that we need to distinguish > >>>>>> delete expressions from plain calls to operator delete? > >>>>> > >>>>> A reason for that distinction came up in the context of omitting > >>>>> new/delete > >>>>> pairs: we want to consider the operator first called by the new or > >>>>> delete > >>>>> expression, not a call from that first operator to another operator > >>>>> new/delete > >>>>> and exposed by inlining. > >>>>> > >>>>> https://gcc.gnu.org/pipermail/gcc-patches/2020-April/543404.html > >>>>> > >>>>>> In this context I also wonder about non-replaceable operator delete, > >>>>>> specifically operator delete in classes - are there any semantic > >>>>>> differences between those or why did we choose to only mark > >>>>>> the replaceable ones? > >>>>> > >>>>> The standard says that for omitting a 'new' allocation, the operator new > >>>>> has > >>>>> to be a replaceable one, but does not say the same about 'delete'; it > >>>>> just > >>>>> says that if the allocation was omitted, the delete-expression does not > >>>>> call a > >>>>> deallocation function.? It may not be necessary to make this distinction > >>>>> for > >>>>> delete.? And this distinction could be local to the front end. > >>>>> > >>>>> In the front end, we currently have cxx_replaceable_global_alloc_fn that > >>>>> already ignores the replaceability of operator delete.? And we have > >>>>> CALL_FROM_NEW_OR_DELETE_P, that would just need to move into the middle > >>>>> end. > >>>>> And perhaps get renamed to CALL_OMITTABLE_NEW_OR_DELETE_P, and not get > >>>>> set > >>>>> for > >>>>> calls to non-replaceable operator new. > >>>> > >>>> CALL_FROM_NEW_OR_DELETE_P indeed looks like the best fit - it's > >>>> only evaluated when cxx_replaceable_global_alloc_fn matches in the C++ > >>>> FE so could be made to cover only replaceable variants. > >>>> > >>>> CALL_REPLACEABLE_NEW_OR_DELETE_P () maybe, since we already use > >>>> REPLACEABLE for the fndecl flags?? OMITTABLE is too specific > >>>> for the PTA case where it really matters whether the object > >>>> lifetime ends before the delete call, not whether it can be > >>>> omitted (hmm, guess that's not 100% overlap then either...). > >>> > >>> That seems like good overlap to me, if we agree that object lifetime ends > >>> before any delete call from a delete-expression, whether or not the > >>> operator > >>> delete is replaceable. > >>> > >>>> Mind doing the C++ side of things recycling protected_flag as suggested? > >>> > >>> OK. > > > > Find attached a patch series, your patch plus the GIMPLE side of the fix. > > This shows that the C++ FE side is possibly incomplete with for > > example g++.dg/pr94314-2.C now FAILing. There we have > > > > A *a = new A (argc); > > delete a; > > > > being expanded to > > > > <<cleanup_point <<< Unknown tree: expr_stmt > > (void) (a = TARGET_EXPR <D.2357, operator new (1)>;, try > > { > > A::A ((struct A *) D.2357, argc); > > } > > catch > > { > > operator delete (D.2357); > > }, (struct A *) D.2357;) >>>>>; > > <<cleanup_point <<< Unknown tree: expr_stmt > > if (SAVE_EXPR <a> != 0B) > > { > > try > > { > > *SAVE_EXPR <a> = {CLOBBER}; > > } > > finally > > { > > operator delete ((void *) SAVE_EXPR <a>); > > } > > } > > else > > { > > <<< Unknown tree: void_cst >>> > > } >>>>>; > > return <retval> = 0; > > > > where the operator delete call in the catch {} expression is > > not marked as CALL_FROM_NEW_OR_DELETE_P, possibly because this > > call is compiler generated. > > > > So it's probably technically true that CALL_FROM_NEW_OR_DELETE_P is > > false but we still expect the same guarantees to hold here, in > > particular we expect to elide the new/delete pair? > > That seems like an oversight in the standard. This first patch sets the flag. > The second patch in the file removes consideration of whether an operator > delete is replaceable, as I was discussing earlier. Thanks. I've re-bootstrapped / tested the series and pushed it to trunk. Richard. From 50125f62cbd726dda1768fc9cda44a34b434b57e Mon Sep 17 00:00:00 2001 From: Jason Merrill <jason@redhat.com> Date: Thu, 1 Oct 2020 10:08:58 +0200 Subject: [PATCH 1/3] c++: Move CALL_FROM_NEW_OR_DELETE_P to tree.h To: gcc-patches@gcc.gnu.org As discussed with richi, we should be able to use TREE_PROTECTED for this flag, since CALL_FROM_THUNK_P will never be set on a call to an operator new or delete. 2020-10-01 Jason Merril <jason@redhat.com> gcc/cp/ChangeLog: * lambda.c (call_from_lambda_thunk_p): New. * cp-gimplify.c (cp_genericize_r): Use it. * pt.c (tsubst_copy_and_build): Use it. * typeck.c (check_return_expr): Use it. * cp-tree.h: Declare it. (CALL_FROM_NEW_OR_DELETE_P): Move to gcc/tree.h. gcc/ChangeLog: * tree.h (CALL_FROM_NEW_OR_DELETE_P): Move from cp-tree.h. --- gcc/cp/cp-gimplify.c | 2 +- gcc/cp/cp-tree.h | 7 +------ gcc/cp/lambda.c | 7 +++++++ gcc/cp/pt.c | 2 +- gcc/cp/typeck.c | 2 +- gcc/tree-core.h | 3 ++- gcc/tree.h | 9 ++++++++- 7 files changed, 21 insertions(+), 11 deletions(-) diff --git a/gcc/cp/cp-gimplify.c b/gcc/cp/cp-gimplify.c index bc8a03c7b41..07549828dc9 100644 --- a/gcc/cp/cp-gimplify.c +++ b/gcc/cp/cp-gimplify.c @@ -962,7 +962,7 @@ cp_genericize_r (tree *stmt_p, int *walk_subtrees, void *data) omp_cxx_notice_variable (wtd->omp_ctx, stmt); /* Don't dereference parms in a thunk, pass the references through. */ - if ((TREE_CODE (stmt) == CALL_EXPR && CALL_FROM_THUNK_P (stmt)) + if ((TREE_CODE (stmt) == CALL_EXPR && call_from_lambda_thunk_p (stmt)) || (TREE_CODE (stmt) == AGGR_INIT_EXPR && AGGR_INIT_FROM_THUNK_P (stmt))) { *walk_subtrees = 0; diff --git a/gcc/cp/cp-tree.h b/gcc/cp/cp-tree.h index 3ccd54ce24b..fda5ffa4036 100644 --- a/gcc/cp/cp-tree.h +++ b/gcc/cp/cp-tree.h @@ -464,7 +464,6 @@ extern GTY(()) tree cp_global_trees[CPTI_MAX]; SWITCH_STMT_NO_BREAK_P (in SWITCH_STMT) LAMBDA_EXPR_CAPTURE_OPTIMIZED (in LAMBDA_EXPR) IMPLICIT_CONV_EXPR_BRACED_INIT (in IMPLICIT_CONV_EXPR) - CALL_FROM_NEW_OR_DELETE_P (in CALL_EXPR) 3: IMPLICIT_RVALUE_P (in NON_LVALUE_EXPR or STATIC_CAST_EXPR) ICS_BAD_FLAG (in _CONV) FN_TRY_BLOCK_P (in TRY_BLOCK) @@ -3839,11 +3838,6 @@ struct GTY(()) lang_decl { should be performed at instantiation time. */ #define KOENIG_LOOKUP_P(NODE) TREE_LANG_FLAG_0 (CALL_EXPR_CHECK (NODE)) -/* In a CALL_EXPR, true for allocator calls from new or delete - expressions. */ -#define CALL_FROM_NEW_OR_DELETE_P(NODE) \ - TREE_LANG_FLAG_2 (CALL_EXPR_CHECK (NODE)) - /* True if the arguments to NODE should be evaluated in left-to-right order regardless of PUSH_ARGS_REVERSED. */ #define CALL_EXPR_ORDERED_ARGS(NODE) \ @@ -7268,6 +7262,7 @@ extern bool lambda_fn_in_template_p (tree); extern void maybe_add_lambda_conv_op (tree); extern bool is_lambda_ignored_entity (tree); extern bool lambda_static_thunk_p (tree); +extern bool call_from_lambda_thunk_p (tree); extern tree finish_builtin_launder (location_t, tree, tsubst_flags_t); extern tree cp_build_vec_convert (tree, location_t, tree, diff --git a/gcc/cp/lambda.c b/gcc/cp/lambda.c index 07a5401c97b..1a1647f465e 100644 --- a/gcc/cp/lambda.c +++ b/gcc/cp/lambda.c @@ -1325,6 +1325,13 @@ lambda_static_thunk_p (tree fn) && LAMBDA_TYPE_P (CP_DECL_CONTEXT (fn))); } +bool +call_from_lambda_thunk_p (tree call) +{ + return (CALL_FROM_THUNK_P (call) + && lambda_static_thunk_p (current_function_decl)); +} + /* Returns true iff VAL is a lambda-related declaration which should be ignored by unqualified lookup. */ diff --git a/gcc/cp/pt.c b/gcc/cp/pt.c index 45b18f6a5ad..72efecff37f 100644 --- a/gcc/cp/pt.c +++ b/gcc/cp/pt.c @@ -19955,7 +19955,7 @@ tsubst_copy_and_build (tree t, /* Stripped-down processing for a call in a thunk. Specifically, in the thunk template for a generic lambda. */ - if (CALL_FROM_THUNK_P (t)) + if (call_from_lambda_thunk_p (t)) { /* Now that we've expanded any packs, the number of call args might be different. */ diff --git a/gcc/cp/typeck.c b/gcc/cp/typeck.c index 9166156a5d5..95b36a92491 100644 --- a/gcc/cp/typeck.c +++ b/gcc/cp/typeck.c @@ -10171,7 +10171,7 @@ check_return_expr (tree retval, bool *no_warning) /* The call in a (lambda) thunk needs no conversions. */ if (TREE_CODE (retval) == CALL_EXPR - && CALL_FROM_THUNK_P (retval)) + && call_from_lambda_thunk_p (retval)) converted = true; /* First convert the value to the function's return type, then diff --git a/gcc/tree-core.h b/gcc/tree-core.h index 0e158784d0e..752bec31c3f 100644 --- a/gcc/tree-core.h +++ b/gcc/tree-core.h @@ -1220,7 +1220,8 @@ struct GTY(()) tree_base { all decls CALL_FROM_THUNK_P and - CALL_ALLOCA_FOR_VAR_P in + CALL_ALLOCA_FOR_VAR_P and + CALL_FROM_NEW_OR_DELETE_P in CALL_EXPR OMP_CLAUSE_LINEAR_VARIABLE_STRIDE in diff --git a/gcc/tree.h b/gcc/tree.h index 5bb6e7bc000..f27a7399a37 100644 --- a/gcc/tree.h +++ b/gcc/tree.h @@ -921,7 +921,8 @@ extern void omp_clause_range_check_failed (const_tree, const char *, int, (TREE_CHECK (NODE, PARM_DECL)->decl_common.decl_nonshareable_flag) /* In a CALL_EXPR, means that the call is the jump from a thunk to the - thunked-to function. */ + thunked-to function. Be careful to avoid using this macro when one of the + next two applies instead. */ #define CALL_FROM_THUNK_P(NODE) (CALL_EXPR_CHECK (NODE)->base.protected_flag) /* In a CALL_EXPR, if the function being called is BUILT_IN_ALLOCA, means that @@ -931,6 +932,12 @@ extern void omp_clause_range_check_failed (const_tree, const char *, int, #define CALL_ALLOCA_FOR_VAR_P(NODE) \ (CALL_EXPR_CHECK (NODE)->base.protected_flag) +/* In a CALL_EXPR, if the function being called is DECL_IS_OPERATOR_NEW_P or + DECL_IS_OPERATOR_DELETE_P, true for allocator calls from C++ new or delete + expressions. */ +#define CALL_FROM_NEW_OR_DELETE_P(NODE) \ + (CALL_EXPR_CHECK (NODE)->base.protected_flag) + /* Used in classes in C++. */ #define TREE_PRIVATE(NODE) ((NODE)->base.private_flag) /* Used in classes in C++. */
diff --git a/gcc/testsuite/g++.dg/cpp1y/new1.C b/gcc/testsuite/g++.dg/cpp1y/new1.C index aa5f647d535..fec0088cb40 100644 --- a/gcc/testsuite/g++.dg/cpp1y/new1.C +++ b/gcc/testsuite/g++.dg/cpp1y/new1.C @@ -69,5 +69,5 @@ test_unused() { delete p; } -/* { dg-final { scan-tree-dump-times "Deleting : operator delete" 5 "cddce1"} } */ -/* { dg-final { scan-tree-dump-times "Deleting : _\\d+ = operator new" 7 "cddce1"} } */ +/* { dg-final { scan-tree-dump-times "Deleting : operator delete" 6 "cddce1"} } */ +/* { dg-final { scan-tree-dump-times "Deleting : _\\d+ = operator new" 8 "cddce1"} } */ diff --git a/gcc/tree-ssa-structalias.c b/gcc/tree-ssa-structalias.c index 44fe52e0f65..f676bf91e95 100644 --- a/gcc/tree-ssa-structalias.c +++ b/gcc/tree-ssa-structalias.c @@ -4857,6 +4857,8 @@ find_func_aliases_for_call (struct function *fn, gcall *t) point for reachable memory of their arguments. */ else if (flags & (ECF_PURE|ECF_LOOPING_CONST_OR_PURE)) handle_pure_call (t, &rhsc); + else if (fndecl && DECL_IS_REPLACEABLE_OPERATOR_DELETE_P (fndecl)) + ; else handle_rhs_call (t, &rhsc); if (gimple_call_lhs (t))