Re: [PATCH v2] libstdc++: Make certain exceptions transaction_safe.

Re: [PATCH v2] libstdc++: Make certain exceptions transaction_safe.

[Dominique d'Humières]

> Addressed these, fixed a problem with using GLIBCXX_WEAK_DEFINITION
> (which is only set on Darwin despite the generic-sounding name -- so
> just use __attribute__((weak)) directly), and also updated
> testsuite_abi.cc so that it knows about CXXABI_1.3.10.
>
> Approved by Jonathan Wakely.  Committed as r232454.
This breaks bootstrap on darwin, see https://gcc.gnu.org/bugzilla/show_bug.cgi?id=69310.

TIA

Dominique

Re: [PATCH v2] libstdc++: Make certain exceptions transaction_safe.

[Torvald Riegel]

On Sat, 2016-01-16 at 10:57 +0100, Dominique d'Humières wrote:
> > Addressed these, fixed a problem with using GLIBCXX_WEAK_DEFINITION
> > (which is only set on Darwin despite the generic-sounding name -- so
> > just use __attribute__((weak)) directly), and also updated
> > testsuite_abi.cc so that it knows about CXXABI_1.3.10.
> >
> > Approved by Jonathan Wakely.  Committed as r232454.
> This breaks bootstrap on darwin, see https://gcc.gnu.org/bugzilla/show_bug.cgi?id=69310.

Can you (or anyone else with access to Darwin) let me know what Darwin
actually supports that would be similar in effect to the alias we use
right now?  I suppose a weak alias would work as well because we define
the C1 constructor anyway?  If so, what's the proper macros / configury
check to use for these?

Is there any documentation for what darwin supports and needs in this
space?

Re: [PATCH v2] libstdc++: Make certain exceptions transaction_safe.

[Torvald Riegel]

[-- Attachment #1: Type: text/plain, Size: 606 bytes --]

On Sat, 2016-01-16 at 10:57 +0100, Dominique d'Humières wrote:
> > Addressed these, fixed a problem with using GLIBCXX_WEAK_DEFINITION
> > (which is only set on Darwin despite the generic-sounding name -- so
> > just use __attribute__((weak)) directly), and also updated
> > testsuite_abi.cc so that it knows about CXXABI_1.3.10.
> >
> > Approved by Jonathan Wakely.  Committed as r232454.
> This breaks bootstrap on darwin, see https://gcc.gnu.org/bugzilla/show_bug.cgi?id=69310.

Please give this patch a try.  I've only tested it on x86_64-linux.
Jon, okay from your side if Darwin testing succeeds?

[-- Attachment #2: libstdc++-darwin-alias.patch --]
[-- Type: text/x-patch, Size: 2021 bytes --]

commit 6987f84f278d2cbf5b828a8c81c1be84b292b1af
Author: Torvald Riegel <triegel@redhat.com>
Date:   Tue Jan 19 18:36:14 2016 +0100

    libstdc: Use weak alias instead of just alias in TM support.
    
    	PR libstdc++/69310
    	* src/c++11/cow-stdexcept.cc: Use weak alias instead of just alias
    	to make Darwin happy.

diff --git a/libstdc++-v3/src/c++11/cow-stdexcept.cc b/libstdc++-v3/src/c++11/cow-stdexcept.cc
index a0f505c..a070747 100644
--- a/libstdc++-v3/src/c++11/cow-stdexcept.cc
+++ b/libstdc++-v3/src/c++11/cow-stdexcept.cc
@@ -364,7 +364,9 @@ _ZGTtNSt##NAME##C1EPKc (CLASS* that, const char* s)			\
      construct the COW string in the latter manually.  Note that the	\
      exception classes will not be declared transaction_safe if the	\
      shared empty _Rep is disabled with --enable-fully-dynamic-string	\
-     (in which case _GLIBCXX_FULLY_DYNAMIC_STRING is nonzero).  */	\
+     (in which case _GLIBCXX_FULLY_DYNAMIC_STRING is nonzero).		\
+     The alias declarations are also declared weak because Darwin	\
+     doesn't support non-weak aliases.  */				\
   CLASS e("");								\
   _ITM_memcpyRnWt(that, &e, sizeof(CLASS));				\
   _txnal_cow_string_C1_for_exceptions(_txnal_##BASE##_get_msg(that),	\
@@ -372,7 +374,7 @@ _ZGTtNSt##NAME##C1EPKc (CLASS* that, const char* s)			\
 }									\
 void									\
 _ZGTtNSt##NAME##C2EPKc (CLASS*, const char*)				\
-  __attribute__((alias ("_ZGTtNSt" #NAME "C1EPKc")));			\
+  __attribute__((weak, alias ("_ZGTtNSt" #NAME "C1EPKc")));		\
 void									\
 _ZGTtNSt##NAME##C1ERKNSt7__cxx1112basic_stringIcSt11char_traitsIcESaIcEEE( \
     CLASS* that, const std::__sso_string& s)				\
@@ -393,7 +395,7 @@ _ZGTtNSt##NAME##D1Ev(CLASS* that)					\
 { _txnal_cow_string_D1(_txnal_##BASE##_get_msg(that)); }		\
 void									\
 _ZGTtNSt##NAME##D2Ev(CLASS*)						\
-__attribute__((alias ("_ZGTtNSt" #NAME "D1Ev")));			\
+__attribute__((weak, alias ("_ZGTtNSt" #NAME "D1Ev")));			\
 void									\
 _ZGTtNSt##NAME##D0Ev(CLASS* that)					\
 {									\

Re: [PATCH v2] libstdc++: Make certain exceptions transaction_safe.

[Jonathan Wakely]

On 19/01/16 20:10 +0100, Torvald Riegel wrote:
>On Sat, 2016-01-16 at 10:57 +0100, Dominique d'Humières wrote:
>> > Addressed these, fixed a problem with using GLIBCXX_WEAK_DEFINITION
>> > (which is only set on Darwin despite the generic-sounding name -- so
>> > just use __attribute__((weak)) directly), and also updated
>> > testsuite_abi.cc so that it knows about CXXABI_1.3.10.
>> >
>> > Approved by Jonathan Wakely.  Committed as r232454.
>> This breaks bootstrap on darwin, see https://gcc.gnu.org/bugzilla/show_bug.cgi?id=69310.
>
>Please give this patch a try.  I've only tested it on x86_64-linux.
>Jon, okay from your side if Darwin testing succeeds?

Yes, OK.

Re: [PATCH v2] libstdc++: Make certain exceptions transaction_safe.

[Dominique d'Humières]

Torvald,

Now that I can bootstrap on darwin, I have found the following failure for libitm.c++/libstdc++-safeexc.C

/opt/gcc/work/libitm/testsuite/libitm.c++/libstdc++-safeexc.C:50:2: error: unsafe function call 'std::underflow_error::underflow_error(const string&)' within atomic transaction
  throw T (what);
  ^~~~~

/opt/gcc/work/libitm/testsuite/libitm.c++/libstdc++-safeexc.C:50:2: error: unsafe function call 'std::overflow_error::overflow_error(const string&)' within atomic transaction
  throw T (what);
  ^~~~~

/opt/gcc/work/libitm/testsuite/libitm.c++/libstdc++-safeexc.C:50:2: error: unsafe function call 'std::range_error::range_error(const string&)' within atomic transaction
  throw T (what);
  ^~~~~

/opt/gcc/work/libitm/testsuite/libitm.c++/libstdc++-safeexc.C:50:2: error: unsafe function call 'std::runtime_error::runtime_error(const string&)' within atomic transaction
  throw T (what);
  ^~~~~

/opt/gcc/work/libitm/testsuite/libitm.c++/libstdc++-safeexc.C:50:2: error: unsafe function call 'std::out_of_range::out_of_range(const string&)' within atomic transaction
  throw T (what);
  ^~~~~

/opt/gcc/work/libitm/testsuite/libitm.c++/libstdc++-safeexc.C:50:2: error: unsafe function call 'std::length_error::length_error(const string&)' within atomic transaction
  throw T (what);
  ^~~~~

/opt/gcc/work/libitm/testsuite/libitm.c++/libstdc++-safeexc.C:50:2: error: unsafe function call 'std::invalid_argument::invalid_argument(const string&)' within atomic transaction
  throw T (what);
  ^~~~~

/opt/gcc/work/libitm/testsuite/libitm.c++/libstdc++-safeexc.C:50:2: error: unsafe function call 'std::domain_error::domain_error(const string&)' within atomic transaction
  throw T (what);
  ^~~~~

/opt/gcc/work/libitm/testsuite/libitm.c++/libstdc++-safeexc.C:50:2: error: unsafe function call 'std::logic_error::logic_error(const string&)' within atomic transaction
  throw T (what);
  ^~~~~

TIA

Dominique

> Le 19 janv. 2016 à 20:20, Jonathan Wakely <jwakely@redhat.com> a écrit :
> 
> On 19/01/16 20:10 +0100, Torvald Riegel wrote:
>> On Sat, 2016-01-16 at 10:57 +0100, Dominique d'Humières wrote:
>>> > Addressed these, fixed a problem with using GLIBCXX_WEAK_DEFINITION
>>> > (which is only set on Darwin despite the generic-sounding name -- so
>>> > just use __attribute__((weak)) directly), and also updated
>>> > testsuite_abi.cc so that it knows about CXXABI_1.3.10.
>>> >
>>> > Approved by Jonathan Wakely.  Committed as r232454.
>>> This breaks bootstrap on darwin, see https://gcc.gnu.org/bugzilla/show_bug.cgi?id=69310.
>> 
>> Please give this patch a try.  I've only tested it on x86_64-linux.
>> Jon, okay from your side if Darwin testing succeeds?
> 
> Yes, OK.

Re: [PATCH v2] libstdc++: Make certain exceptions transaction_safe.

[Torvald Riegel]

On Thu, 2016-01-21 at 11:00 +0100, Dominique d'Humières wrote:
> Torvald,
> 
> Now that I can bootstrap on darwin, I have found the following failure for libitm.c++/libstdc++-safeexc.C
> 
> /opt/gcc/work/libitm/testsuite/libitm.c++/libstdc++-safeexc.C:50:2: error: unsafe function call 'std::underflow_error::underflow_error(const string&)' within atomic transaction
>   throw T (what);
>   ^~~~~

Well, yes, that's my oversight.  The previous fix disabled the support,
so we need to now xfail or disable this test on Darwin.  Same for AIX.
Ignore these failures for now.  I'll work on a fix.

Re: [PATCH v2] libstdc++: Make certain exceptions transaction_safe.

[Torvald Riegel]

[-- Attachment #1: Type: text/plain, Size: 457 bytes --]

On Thu, 2016-01-21 at 11:00 +0100, Dominique d'Humières wrote:
> Torvald,
> 
> Now that I can bootstrap on darwin, I have found the following failure for libitm.c++/libstdc++-safeexc.C
> 
> /opt/gcc/work/libitm/testsuite/libitm.c++/libstdc++-safeexc.C:50:2: error: unsafe function call 'std::underflow_error::underflow_error(const string&)' within atomic transaction
>   throw T (what);
>   ^~~~~

Does this patch fix it (ie, mark the test unsupported)?

[-- Attachment #2: libitm-safeexc-unsupported.patch --]
[-- Type: text/x-patch, Size: 976 bytes --]

commit 259c0cf27d0a88eecc90af1aa500f88f6108cb04
Author: Torvald Riegel <triegel@redhat.com>
Date:   Thu Jan 21 16:21:33 2016 +0100

    libitm: Disable testing transaction-safe exceptions on Darwin and AIX.
    
    	* testsuite/libitm.c++/libstdc++-safeexc.C: Not supported on darwin
    	or AIX.

diff --git a/libitm/testsuite/libitm.c++/libstdc++-safeexc.C b/libitm/testsuite/libitm.c++/libstdc++-safeexc.C
index 3e1655e..55ebd25 100644
--- a/libitm/testsuite/libitm.c++/libstdc++-safeexc.C
+++ b/libitm/testsuite/libitm.c++/libstdc++-safeexc.C
@@ -2,7 +2,10 @@
 // are indeed that.  Thus, this also tests the transactional clones in
 // libstdc++ and libsupc++.
 
-// { dg-do run }
+// Not supported on Darwin nor AIX because those lack the support for
+// weak references to undefined functions that we need in libstdc++ to make
+// exceptions transaction-safe.
+// { dg-do run { target { ! *-*-darwin* powerpc-ibm-aix* } } }
 
 #include <iostream>
 #include <exception>

Re: [PATCH v2] libstdc++: Make certain exceptions transaction_safe.

[Dominique d'Humières]

> Le 21 janv. 2016 à 16:25, Torvald Riegel <triegel@redhat.com> a écrit :
> 
> On Thu, 2016-01-21 at 11:00 +0100, Dominique d'Humières wrote:
>> Torvald,
>> 
>> Now that I can bootstrap on darwin, I have found the following failure for libitm.c++/libstdc++-safeexc.C
>> 
>> /opt/gcc/work/libitm/testsuite/libitm.c++/libstdc++-safeexc.C:50:2: error: unsafe function call 'std::underflow_error::underflow_error(const string&)' within atomic transaction
>>  throw T (what);
>>  ^~~~~
> 
> Does this patch fix it (ie, mark the test unsupported)?

AFAIU the error occurs at compile time and If my memory is correct the patch won’t fix the issue (not tested) and the test should be skipped on *-*-darwin* powerpc-ibm-aix*.

Dominique

Re: [PATCH v2] libstdc++: Make certain exceptions transaction_safe.

[Dominique d'Humières]

> Le 21 janv. 2016 à 18:15, Dominique d'Humières <dominiq@lps.ens.fr> a écrit :
> 
> 
>> Le 21 janv. 2016 à 16:25, Torvald Riegel <triegel@redhat.com> a écrit :
>> 
>> On Thu, 2016-01-21 at 11:00 +0100, Dominique d'Humières wrote:
>>> Torvald,
>>> 
>>> Now that I can bootstrap on darwin, I have found the following failure for libitm.c++/libstdc++-safeexc.C
>>> 
>>> /opt/gcc/work/libitm/testsuite/libitm.c++/libstdc++-safeexc.C:50:2: error: unsafe function call 'std::underflow_error::underflow_error(const string&)' within atomic transaction
>>> throw T (what);
>>> ^~~~~
>> 
>> Does this patch fix it (ie, mark the test unsupported)?
> 
> AFAIU the error occurs at compile time and If my memory is correct the patch won’t fix the issue (not tested) and the test should be skipped on *-*-darwin* powerpc-ibm-aix*.

OK, my memory was not good!-( However the target selection in the patch is wrong, the following one works

// { dg-do run { target { ! { *-*-darwin* powerpc-ibm-aix* } } } }


Dominique

Re: [PATCH v2] libstdc++: Make certain exceptions transaction_safe.

[Mike Stump]

On Jan 21, 2016, at 9:29 AM, Dominique d'Humières <dominiq@lps.ens.fr> wrote:
> // { dg-do run { target { ! { *-*-darwin* powerpc-ibm-aix* } } } }

A comment to hint that this has something to do with weak undefined would be nice.

Re: [PATCH v2] libstdc++: Make certain exceptions transaction_safe.

[Pedro Alves]

On 01/21/2016 06:06 PM, Mike Stump wrote:
> On Jan 21, 2016, at 9:29 AM, Dominique d'Humières <dominiq@lps.ens.fr> wrote:
>> // { dg-do run { target { ! { *-*-darwin* powerpc-ibm-aix* } } } }
> 
> A comment to hint that this has something to do with weak undefined would be nice.
> 

Or come up with some new "dg-require-effective-target weakref-whatnot", making it
self-documenting.

[PATCH] libitm: Disable testing transaction-safe exceptions on Darwin and AIX.

[Torvald Riegel]

[-- Attachment #1: Type: text/plain, Size: 495 bytes --]

On Thu, 2016-01-21 at 10:06 -0800, Mike Stump wrote:
> On Jan 21, 2016, at 9:29 AM, Dominique d'Humières <dominiq@lps.ens.fr> wrote:
> > // { dg-do run { target { ! { *-*-darwin* powerpc-ibm-aix* } } } }
> 
> A comment to hint that this has something to do with weak undefined would be nice.

Here's the patch I prepared (which indeed includes a comment).

OK for trunk?  I'm not quite sure whether this qualifies as a
regression, but having an additional test that now fails is one I guess.

[-- Attachment #2: libitm-safeexc-unsupported.patch --]
[-- Type: text/x-patch, Size: 980 bytes --]

commit 0323fed14832e5744cbc63bcfeeb6728f7f13394
Author: Torvald Riegel <triegel@redhat.com>
Date:   Thu Jan 21 16:21:33 2016 +0100

    libitm: Disable testing transaction-safe exceptions on Darwin and AIX.
    
    	* testsuite/libitm.c++/libstdc++-safeexc.C: Not supported on darwin
    	or AIX.

diff --git a/libitm/testsuite/libitm.c++/libstdc++-safeexc.C b/libitm/testsuite/libitm.c++/libstdc++-safeexc.C
index 3e1655e..20e2e5e 100644
--- a/libitm/testsuite/libitm.c++/libstdc++-safeexc.C
+++ b/libitm/testsuite/libitm.c++/libstdc++-safeexc.C
@@ -2,7 +2,10 @@
 // are indeed that.  Thus, this also tests the transactional clones in
 // libstdc++ and libsupc++.
 
-// { dg-do run }
+// Not supported on Darwin nor AIX because those lack the support for
+// weak references to undefined functions that we need in libstdc++ to make
+// exceptions transaction-safe.
+// { dg-do run { target { ! { *-*-darwin* powerpc-ibm-aix* } } } }
 
 #include <iostream>
 #include <exception>

Re: [PATCH v2] libstdc++: Make certain exceptions transaction_safe.

[Torvald Riegel]

On Thu, 2016-01-21 at 18:12 +0000, Pedro Alves wrote:
> On 01/21/2016 06:06 PM, Mike Stump wrote:
> > On Jan 21, 2016, at 9:29 AM, Dominique d'Humières <dominiq@lps.ens.fr> wrote:
> >> // { dg-do run { target { ! { *-*-darwin* powerpc-ibm-aix* } } } }
> > 
> > A comment to hint that this has something to do with weak undefined would be nice.
> > 
> 
> Or come up with some new "dg-require-effective-target weakref-whatnot", making it
> self-documenting.
> 

It's just one test that needs this right now.  If there should be more
in the future, I agree that a dg-require-... might be nicer.

Re: [PATCH] libitm: Disable testing transaction-safe exceptions on Darwin and AIX.

[Mike Stump]

On Jan 21, 2016, at 10:15 AM, Torvald Riegel <triegel@redhat.com> wrote:
> On Thu, 2016-01-21 at 10:06 -0800, Mike Stump wrote:
>> On Jan 21, 2016, at 9:29 AM, Dominique d'Humières <dominiq@lps.ens.fr> wrote:
>>> // { dg-do run { target { ! { *-*-darwin* powerpc-ibm-aix* } } } }
>> 
>> A comment to hint that this has something to do with weak undefined would be nice.
> 
> Here's the patch I prepared (which indeed includes a comment).
> 
> OK for trunk?  I'm not quite sure whether this qualifies as a
> regression, but having an additional test that now fails is one I guess.
> <libitm-safeexc-unsupported.patch>

A simple testsuite fixup like this is still ok.  From a darwin, AIX perspective it is fine.  If either the transaction or the libstdc++ people like it, I think we’re set.

Re: [PATCH] libitm: Disable testing transaction-safe exceptions on Darwin and AIX.

[Jonathan Wakely]

On 21/01/16 10:19 -0800, Mike Stump wrote:
>On Jan 21, 2016, at 10:15 AM, Torvald Riegel <triegel@redhat.com> wrote:
>> On Thu, 2016-01-21 at 10:06 -0800, Mike Stump wrote:
>>> On Jan 21, 2016, at 9:29 AM, Dominique d'Humières <dominiq@lps.ens.fr> wrote:
>>>> // { dg-do run { target { ! { *-*-darwin* powerpc-ibm-aix* } } } }
>>>
>>> A comment to hint that this has something to do with weak undefined would be nice.
>>
>> Here's the patch I prepared (which indeed includes a comment).
>>
>> OK for trunk?  I'm not quite sure whether this qualifies as a
>> regression, but having an additional test that now fails is one I guess.
>> <libitm-safeexc-unsupported.patch>
>
>A simple testsuite fixup like this is still ok.  From a darwin, AIX perspective it is fine.  If either the transaction or the libstdc++ people like it, I think we’re set.

OK from the libstdc++ side.

Re: [PATCH] libitm: Disable testing transaction-safe exceptions on Darwin and AIX.

[Torvald Riegel]

On Thu, 2016-01-21 at 18:26 +0000, Jonathan Wakely wrote:
> On 21/01/16 10:19 -0800, Mike Stump wrote:
> >On Jan 21, 2016, at 10:15 AM, Torvald Riegel <triegel@redhat.com> wrote:
> >> On Thu, 2016-01-21 at 10:06 -0800, Mike Stump wrote:
> >>> On Jan 21, 2016, at 9:29 AM, Dominique d'Humières <dominiq@lps.ens.fr> wrote:
> >>>> // { dg-do run { target { ! { *-*-darwin* powerpc-ibm-aix* } } } }
> >>>
> >>> A comment to hint that this has something to do with weak undefined would be nice.
> >>
> >> Here's the patch I prepared (which indeed includes a comment).
> >>
> >> OK for trunk?  I'm not quite sure whether this qualifies as a
> >> regression, but having an additional test that now fails is one I guess.
> >> <libitm-safeexc-unsupported.patch>
> >
> >A simple testsuite fixup like this is still ok.  From a darwin, AIX perspective it is fine.  If either the transaction or the libstdc++ people like it, I think we’re set.
> 
> OK from the libstdc++ side.
> 

Committed.