* [Bug libstdc++/105417] [11/12 Regression] powerpc64le-linux abilist changes based on --with-long-double-format=
2022-04-28 8:27 [Bug libstdc++/105417] New: [11/12 Regression] powerpc64le-linux abilist changes based on --with-long-double-format= jakub at gcc dot gnu.org
@ 2022-04-28 8:27 ` jakub at gcc dot gnu.org
2022-04-28 8:29 ` redi at gcc dot gnu.org
` (10 subsequent siblings)
11 siblings, 0 replies; 13+ messages in thread
From: jakub at gcc dot gnu.org @ 2022-04-28 8:27 UTC (permalink / raw)
To: gcc-bugs
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105417
Jakub Jelinek <jakub at gcc dot gnu.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
Last reconfirmed| |2022-04-28
Status|UNCONFIRMED |NEW
Ever confirmed|0 |1
Target Milestone|--- |11.4
^ permalink raw reply [flat|nested] 13+ messages in thread* [Bug libstdc++/105417] [11/12 Regression] powerpc64le-linux abilist changes based on --with-long-double-format=
2022-04-28 8:27 [Bug libstdc++/105417] New: [11/12 Regression] powerpc64le-linux abilist changes based on --with-long-double-format= jakub at gcc dot gnu.org
2022-04-28 8:27 ` [Bug libstdc++/105417] " jakub at gcc dot gnu.org
@ 2022-04-28 8:29 ` redi at gcc dot gnu.org
2022-04-28 8:33 ` jakub at gcc dot gnu.org
` (9 subsequent siblings)
11 siblings, 0 replies; 13+ messages in thread
From: redi at gcc dot gnu.org @ 2022-04-28 8:29 UTC (permalink / raw)
To: gcc-bugs
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105417
--- Comment #1 from Jonathan Wakely <redi at gcc dot gnu.org> ---
(In reply to Jakub Jelinek from comment #0)
> The removed symbols for the ieee configuration in 11.3 got fixed on the
> trunk,
That was r12-7225-g813415650235b8
^ permalink raw reply [flat|nested] 13+ messages in thread* [Bug libstdc++/105417] [11/12 Regression] powerpc64le-linux abilist changes based on --with-long-double-format=
2022-04-28 8:27 [Bug libstdc++/105417] New: [11/12 Regression] powerpc64le-linux abilist changes based on --with-long-double-format= jakub at gcc dot gnu.org
2022-04-28 8:27 ` [Bug libstdc++/105417] " jakub at gcc dot gnu.org
2022-04-28 8:29 ` redi at gcc dot gnu.org
@ 2022-04-28 8:33 ` jakub at gcc dot gnu.org
2022-04-28 13:21 ` jakub at gcc dot gnu.org
` (8 subsequent siblings)
11 siblings, 0 replies; 13+ messages in thread
From: jakub at gcc dot gnu.org @ 2022-04-28 8:33 UTC (permalink / raw)
To: gcc-bugs
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105417
--- Comment #2 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
So I think we want that backported to 11.4. And I think we just should export
the other symbols with the 3.4.29 version, adding 3.4.30 aliases for those
wouldn't improve things.
^ permalink raw reply [flat|nested] 13+ messages in thread* [Bug libstdc++/105417] [11/12 Regression] powerpc64le-linux abilist changes based on --with-long-double-format=
2022-04-28 8:27 [Bug libstdc++/105417] New: [11/12 Regression] powerpc64le-linux abilist changes based on --with-long-double-format= jakub at gcc dot gnu.org
` (2 preceding siblings ...)
2022-04-28 8:33 ` jakub at gcc dot gnu.org
@ 2022-04-28 13:21 ` jakub at gcc dot gnu.org
2022-04-29 10:46 ` [Bug libstdc++/105417] [11/12/13 " redi at gcc dot gnu.org
` (7 subsequent siblings)
11 siblings, 0 replies; 13+ messages in thread
From: jakub at gcc dot gnu.org @ 2022-04-28 13:21 UTC (permalink / raw)
To: gcc-bugs
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105417
--- Comment #3 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
>From what I can see, the --with-long-double-format=ibm libstdc++.a has
following _M_extract_int symbols (for simplicity only listing Ij symbols,
I[lmtxy] are present/absent next to it:
compatibility-ldbl-alt128.o:
0000000000000000 W
_ZNKSt17__gnu_cxx_ieee1287num_getIcSt19istreambuf_iteratorIcSt11char_traitsIcEEE14_M_extract_intIjEES4_S4_S4_RSt8ios_baseRSt12_Ios_IostateRT_
0000000000000000 W
_ZNKSt17__gnu_cxx_ieee1287num_getIwSt19istreambuf_iteratorIwSt11char_traitsIwEEE14_M_extract_intIjEES4_S4_S4_RSt8ios_baseRSt12_Ios_IostateRT_
cxx11-locale-inst.o:
0000000000000000 W
_ZNKSt17__gnu_cxx_ldbl1287num_getIcSt19istreambuf_iteratorIcSt11char_traitsIcEEE14_M_extract_intB5cxx11IjEES4_S4_S4_RSt8ios_baseRSt12_Ios_IostateRT_
cxx11-wlocale-inst.o:
0000000000000000 W
_ZNKSt17__gnu_cxx_ldbl1287num_getIwSt19istreambuf_iteratorIwSt11char_traitsIwEEE14_M_extract_intB5cxx11IjEES4_S4_S4_RSt8ios_baseRSt12_Ios_IostateRT_
locale-inst.o:
0000000000000000 W
_ZNKSt17__gnu_cxx_ldbl1287num_getIcSt19istreambuf_iteratorIcSt11char_traitsIcEEE14_M_extract_intIjEES4_S4_S4_RSt8ios_baseRSt12_Ios_IostateRT_
0000000000000000 W
_ZNKSt7num_getIcSt19istreambuf_iteratorIcSt11char_traitsIcEEE14_M_extract_intIjEES3_S3_S3_RSt8ios_baseRSt12_Ios_IostateRT_
wlocale-inst.o:
0000000000000000 W
_ZNKSt17__gnu_cxx_ldbl1287num_getIwSt19istreambuf_iteratorIwSt11char_traitsIwEEE14_M_extract_intIjEES4_S4_S4_RSt8ios_baseRSt12_Ios_IostateRT_
0000000000000000 W
_ZNKSt7num_getIwSt19istreambuf_iteratorIwSt11char_traitsIwEEE14_M_extract_intIjEES3_S3_S3_RSt8ios_baseRSt12_Ios_IostateRT_
while --with-long-double-format=ieee libstdc++.a has:
compatibility-ldbl-alt128.o:
0000000000000000 W
_ZNKSt17__gnu_cxx_ldbl1287num_getIcSt19istreambuf_iteratorIcSt11char_traitsIcEEE14_M_extract_intIjEES4_S4_S4_RSt8ios_baseRSt12_Ios_IostateRT_
0000000000000000 W
_ZNKSt17__gnu_cxx_ldbl1287num_getIwSt19istreambuf_iteratorIwSt11char_traitsIwEEE14_M_extract_intIjEES4_S4_S4_RSt8ios_baseRSt12_Ios_IostateRT_
0000000000000000 W
_ZNKSt7num_getIcSt19istreambuf_iteratorIcSt11char_traitsIcEEE14_M_extract_intIjEES3_S3_S3_RSt8ios_baseRSt12_Ios_IostateRT_
0000000000000000 W
_ZNKSt7num_getIwSt19istreambuf_iteratorIwSt11char_traitsIwEEE14_M_extract_intIjEES3_S3_S3_RSt8ios_baseRSt12_Ios_IostateRT_
cxx11-locale-inst.o:
0000000000000000 W
_ZNKSt17__gnu_cxx_ieee1287num_getIcSt19istreambuf_iteratorIcSt11char_traitsIcEEE14_M_extract_intB5cxx11IjEES4_S4_S4_RSt8ios_baseRSt12_Ios_IostateRT_
cxx11-wlocale-inst.o:
0000000000000000 W
_ZNKSt17__gnu_cxx_ieee1287num_getIwSt19istreambuf_iteratorIwSt11char_traitsIwEEE14_M_extract_intB5cxx11IjEES4_S4_S4_RSt8ios_baseRSt12_Ios_IostateRT_
locale-inst.o:
0000000000000000 W
_ZNKSt17__gnu_cxx_ieee1287num_getIcSt19istreambuf_iteratorIcSt11char_traitsIcEEE14_M_extract_intIjEES4_S4_S4_RSt8ios_baseRSt12_Ios_IostateRT_
wlocale-inst.o:
0000000000000000 W
_ZNKSt17__gnu_cxx_ieee1287num_getIwSt19istreambuf_iteratorIwSt11char_traitsIwEEE14_M_extract_intIjEES4_S4_S4_RSt8ios_baseRSt12_Ios_IostateRT_
Now, the symbols in compatibility-ldbl-alt128.o and {,w}locale-inst.o match
between the 2 builds, basically for each narrow and wide we export all of
std::num_get, std::__gnu_cxx_ldbl128::num_get and
std::__gnu_cxx_ieee128::num_get symbols.
But the cxx11 ABI symbols are not.
On x86_64, I see we have
_ZNKSt7num_getIcSt19istreambuf_iteratorIcSt11char_traitsIcEEE14_M_extract_intB5cxx11IjEES3_S3_S3_RSt8ios_baseRSt12_Ios_IostateRT_
in cxx11-locale-inst.o and
_ZNKSt7num_getIwSt19istreambuf_iteratorIwSt11char_traitsIwEEE14_M_extract_intB5cxx11IjEES3_S3_S3_RSt8ios_baseRSt12_Ios_IostateRT_
in cxx11-wlocale-inst.o, but we don't export those, similarly we don't export
those
_ZNKSt17__gnu_cxx_ldbl1287num_getIcSt19istreambuf_iteratorIcSt11char_traitsIcEEE14_M_extract_intB5cxx11IjEES4_S4_S4_RSt8ios_baseRSt12_Ios_IostateRT_
and
_ZNKSt17__gnu_cxx_ldbl1287num_getIwSt19istreambuf_iteratorIwSt11char_traitsIwEEE14_M_extract_intB5cxx11IjEES4_S4_S4_RSt8ios_baseRSt12_Ios_IostateRT_
on targets which switched long double from double to something wider in the
past (including powerpc64*), but we do export the __gnu_cxx_ieee128 ones, which
is
why we get those extra symbols on --with-long-double-format=ieee builds.
Are those cxx11 symbols never needed and so it was a mistake that they've been
exported?
Or is it an omission that they are not exported?
^ permalink raw reply [flat|nested] 13+ messages in thread* [Bug libstdc++/105417] [11/12/13 Regression] powerpc64le-linux abilist changes based on --with-long-double-format=
2022-04-28 8:27 [Bug libstdc++/105417] New: [11/12 Regression] powerpc64le-linux abilist changes based on --with-long-double-format= jakub at gcc dot gnu.org
` (3 preceding siblings ...)
2022-04-28 13:21 ` jakub at gcc dot gnu.org
@ 2022-04-29 10:46 ` redi at gcc dot gnu.org
2022-04-29 10:49 ` redi at gcc dot gnu.org
` (6 subsequent siblings)
11 siblings, 0 replies; 13+ messages in thread
From: redi at gcc dot gnu.org @ 2022-04-29 10:46 UTC (permalink / raw)
To: gcc-bugs
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105417
--- Comment #4 from Jonathan Wakely <redi at gcc dot gnu.org> ---
(In reply to Jakub Jelinek from comment #3)
> Are those cxx11 symbols never needed and so it was a mistake that they've
> been exported?
I thought they were needed, but I've now convinced myself they aren't. The
public headers do not declare explicit instantiations for those symbols, so
user code doesn't know that they're (sometimes) defined in libstdc++.so and so
user code will contain implicit instantiations of them. So the definitions
exported from the library are never needed outside the library itself.
^ permalink raw reply [flat|nested] 13+ messages in thread* [Bug libstdc++/105417] [11/12/13 Regression] powerpc64le-linux abilist changes based on --with-long-double-format=
2022-04-28 8:27 [Bug libstdc++/105417] New: [11/12 Regression] powerpc64le-linux abilist changes based on --with-long-double-format= jakub at gcc dot gnu.org
` (4 preceding siblings ...)
2022-04-29 10:46 ` [Bug libstdc++/105417] [11/12/13 " redi at gcc dot gnu.org
@ 2022-04-29 10:49 ` redi at gcc dot gnu.org
2022-04-29 10:56 ` jakub at gcc dot gnu.org
` (5 subsequent siblings)
11 siblings, 0 replies; 13+ messages in thread
From: redi at gcc dot gnu.org @ 2022-04-29 10:49 UTC (permalink / raw)
To: gcc-bugs
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105417
--- Comment #5 from Jonathan Wakely <redi at gcc dot gnu.org> ---
Created attachment 52905
--> https://gcc.gnu.org/bugzilla/attachment.cgi?id=52905&action=edit
Add alias symbols for the missing exports
^ permalink raw reply [flat|nested] 13+ messages in thread* [Bug libstdc++/105417] [11/12/13 Regression] powerpc64le-linux abilist changes based on --with-long-double-format=
2022-04-28 8:27 [Bug libstdc++/105417] New: [11/12 Regression] powerpc64le-linux abilist changes based on --with-long-double-format= jakub at gcc dot gnu.org
` (5 preceding siblings ...)
2022-04-29 10:49 ` redi at gcc dot gnu.org
@ 2022-04-29 10:56 ` jakub at gcc dot gnu.org
2022-04-29 13:00 ` redi at gcc dot gnu.org
` (4 subsequent siblings)
11 siblings, 0 replies; 13+ messages in thread
From: jakub at gcc dot gnu.org @ 2022-04-29 10:56 UTC (permalink / raw)
To: gcc-bugs
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105417
--- Comment #6 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
Comment on attachment 52905
--> https://gcc.gnu.org/bugzilla/attachment.cgi?id=52905
Add alias symbols for the missing exports
Shouldn't the #if 1 be wropped?
On the other side, perhaps wrap the wchar_t stuff in
#ifdef _GLIBCXX_USE_WCHAR_T
just for consistency?
^ permalink raw reply [flat|nested] 13+ messages in thread* [Bug libstdc++/105417] [11/12/13 Regression] powerpc64le-linux abilist changes based on --with-long-double-format=
2022-04-28 8:27 [Bug libstdc++/105417] New: [11/12 Regression] powerpc64le-linux abilist changes based on --with-long-double-format= jakub at gcc dot gnu.org
` (6 preceding siblings ...)
2022-04-29 10:56 ` jakub at gcc dot gnu.org
@ 2022-04-29 13:00 ` redi at gcc dot gnu.org
2022-04-29 13:41 ` cvs-commit at gcc dot gnu.org
` (3 subsequent siblings)
11 siblings, 0 replies; 13+ messages in thread
From: redi at gcc dot gnu.org @ 2022-04-29 13:00 UTC (permalink / raw)
To: gcc-bugs
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105417
--- Comment #7 from Jonathan Wakely <redi at gcc dot gnu.org> ---
(In reply to Jakub Jelinek from comment #6)
> Shouldn't the #if 1 be wropped?
Yeah, that's gone in the real patch I'm testing.
> On the other side, perhaps wrap the wchar_t stuff in
> #ifdef _GLIBCXX_USE_WCHAR_T
> just for consistency?
Yes, I suppose it should. That macro should always be defined for
powerpc64le-*-linux-gnu because we know we have a Glibc that isn't from the
last millennium, but it's possible somebody wants to use --disable-wchar_t for
some reason.
^ permalink raw reply [flat|nested] 13+ messages in thread* [Bug libstdc++/105417] [11/12/13 Regression] powerpc64le-linux abilist changes based on --with-long-double-format=
2022-04-28 8:27 [Bug libstdc++/105417] New: [11/12 Regression] powerpc64le-linux abilist changes based on --with-long-double-format= jakub at gcc dot gnu.org
` (7 preceding siblings ...)
2022-04-29 13:00 ` redi at gcc dot gnu.org
@ 2022-04-29 13:41 ` cvs-commit at gcc dot gnu.org
2022-04-29 13:44 ` cvs-commit at gcc dot gnu.org
` (2 subsequent siblings)
11 siblings, 0 replies; 13+ messages in thread
From: cvs-commit at gcc dot gnu.org @ 2022-04-29 13:41 UTC (permalink / raw)
To: gcc-bugs
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105417
--- Comment #8 from CVS Commits <cvs-commit at gcc dot gnu.org> ---
The master branch has been updated by Jonathan Wakely <redi@gcc.gnu.org>:
https://gcc.gnu.org/g:bb7cf39b05a216431a431499d0c36a6034f6acc4
commit r13-45-gbb7cf39b05a216431a431499d0c36a6034f6acc4
Author: Jonathan Wakely <jwakely@redhat.com>
Date: Fri Apr 29 12:17:13 2022 +0100
libstdc++: Add missing exports for ppc64le --with-long-double-format=ibm
[PR105417]
The --with-long-double-abi=ibm build is missing some exports that are
present in the --with-long-double-abi=ieee build. Those symbols never
should have been exported at all, but now that they have been, they
should be exported consistently by both ibm and ieee.
This simply defines them as aliases for equivalent symbols that are
already present. The abi-tag on num_get::_M_extract_int isn't really
needed, because it only uses a std::string as a local variable, not in
the return type or function parameters, so it's safe to define the
_M_extract_int[abi:cxx11] symbols as aliases for the corresponding
function without the abi-tag.
This causes some new symbols to be added to the GLIBCXX_3.4.29 version
for the ibm long double build mode, but there is no advantage to adding
them to 3.4.30 for that build. That would just create more
inconsistencies.
libstdc++-v3/ChangeLog:
PR libstdc++/105417
* config/abi/post/powerpc64-linux-gnu/baseline_symbols.txt:
Regenerate.
* src/c++11/compatibility-ldbl-alt128.cc [_GLIBCXX_USE_DUAL_ABI]:
Define __gnu_ieee128::num_get<C>::_M_extract_int[abi:cxx11]<I>
symbols as aliases for corresponding symbols without abi-tag.
^ permalink raw reply [flat|nested] 13+ messages in thread* [Bug libstdc++/105417] [11/12/13 Regression] powerpc64le-linux abilist changes based on --with-long-double-format=
2022-04-28 8:27 [Bug libstdc++/105417] New: [11/12 Regression] powerpc64le-linux abilist changes based on --with-long-double-format= jakub at gcc dot gnu.org
` (8 preceding siblings ...)
2022-04-29 13:41 ` cvs-commit at gcc dot gnu.org
@ 2022-04-29 13:44 ` cvs-commit at gcc dot gnu.org
2022-04-29 14:18 ` [Bug libstdc++/105417] [11 " redi at gcc dot gnu.org
2023-05-29 10:06 ` jakub at gcc dot gnu.org
11 siblings, 0 replies; 13+ messages in thread
From: cvs-commit at gcc dot gnu.org @ 2022-04-29 13:44 UTC (permalink / raw)
To: gcc-bugs
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105417
--- Comment #9 from CVS Commits <cvs-commit at gcc dot gnu.org> ---
The releases/gcc-12 branch has been updated by Jonathan Wakely
<redi@gcc.gnu.org>:
https://gcc.gnu.org/g:621650f64fb6679c457c33abf27c925f28bddc62
commit r12-8321-g621650f64fb6679c457c33abf27c925f28bddc62
Author: Jonathan Wakely <jwakely@redhat.com>
Date: Fri Apr 29 12:17:13 2022 +0100
libstdc++: Add missing exports for ppc64le --with-long-double-format=ibm
[PR105417]
The --with-long-double-abi=ibm build is missing some exports that are
present in the --with-long-double-abi=ieee build. Those symbols never
should have been exported at all, but now that they have been, they
should be exported consistently by both ibm and ieee.
This simply defines them as aliases for equivalent symbols that are
already present. The abi-tag on num_get::_M_extract_int isn't really
needed, because it only uses a std::string as a local variable, not in
the return type or function parameters, so it's safe to define the
_M_extract_int[abi:cxx11] symbols as aliases for the corresponding
function without the abi-tag.
This causes some new symbols to be added to the GLIBCXX_3.4.29 version
for the ibm long double build mode, but there is no advantage to adding
them to 3.4.30 for that build. That would just create more
inconsistencies.
libstdc++-v3/ChangeLog:
PR libstdc++/105417
* config/abi/post/powerpc64-linux-gnu/baseline_symbols.txt:
Regenerate.
* src/c++11/compatibility-ldbl-alt128.cc [_GLIBCXX_USE_DUAL_ABI]:
Define __gnu_ieee128::num_get<C>::_M_extract_int[abi:cxx11]<I>
symbols as aliases for corresponding symbols without abi-tag.
(cherry picked from commit bb7cf39b05a216431a431499d0c36a6034f6acc4)
^ permalink raw reply [flat|nested] 13+ messages in thread* [Bug libstdc++/105417] [11 Regression] powerpc64le-linux abilist changes based on --with-long-double-format=
2022-04-28 8:27 [Bug libstdc++/105417] New: [11/12 Regression] powerpc64le-linux abilist changes based on --with-long-double-format= jakub at gcc dot gnu.org
` (9 preceding siblings ...)
2022-04-29 13:44 ` cvs-commit at gcc dot gnu.org
@ 2022-04-29 14:18 ` redi at gcc dot gnu.org
2023-05-29 10:06 ` jakub at gcc dot gnu.org
11 siblings, 0 replies; 13+ messages in thread
From: redi at gcc dot gnu.org @ 2022-04-29 14:18 UTC (permalink / raw)
To: gcc-bugs
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105417
Jonathan Wakely <redi at gcc dot gnu.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
Summary|[11/12/13 Regression] |[11 Regression]
|powerpc64le-linux abilist |powerpc64le-linux abilist
|changes based on |changes based on
|--with-long-double-format= |--with-long-double-format=
--- Comment #10 from Jonathan Wakely <redi at gcc dot gnu.org> ---
Fixed for trunk and 12, backport to 11 still needed.
^ permalink raw reply [flat|nested] 13+ messages in thread* [Bug libstdc++/105417] [11 Regression] powerpc64le-linux abilist changes based on --with-long-double-format=
2022-04-28 8:27 [Bug libstdc++/105417] New: [11/12 Regression] powerpc64le-linux abilist changes based on --with-long-double-format= jakub at gcc dot gnu.org
` (10 preceding siblings ...)
2022-04-29 14:18 ` [Bug libstdc++/105417] [11 " redi at gcc dot gnu.org
@ 2023-05-29 10:06 ` jakub at gcc dot gnu.org
11 siblings, 0 replies; 13+ messages in thread
From: jakub at gcc dot gnu.org @ 2023-05-29 10:06 UTC (permalink / raw)
To: gcc-bugs
https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105417
Jakub Jelinek <jakub at gcc dot gnu.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
Target Milestone|11.4 |11.5
--- Comment #11 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
GCC 11.4 is being released, retargeting bugs to GCC 11.5.
^ permalink raw reply [flat|nested] 13+ messages in thread