Re: [jira] Created: (STDCXX-1022) [MSVC x86 / optimized] ICE in std::__make_heap()
Farid Zaripov-2 wrote:
>
>> Did you happen to reduce this to a small test case and send
>> it to Microsoft?
>
> Not yet. It's difficult to create the small test.
>
>> Do you see the ICE in a specific test (or example), maybe only for a
>> specific instantiation of the template, or does it happen regardless
>> of the actual types of the template arguments?
>
> Today I've made some investigations:
>
> 1. the 25.heap, 25.partial.sort and 25.sort tests from trunk and 4.3.x
> branch
> are compiled withoud ICE;
>
> 2. the only difference between preprocessed 25.heap tests from 4.2.x
> branch and trunk is
>
> -
> Comparing files 42x.i and trunk.i
> * 42x.i
> void __declspec (noreturn)
> __rw_assert_fail (const char*, const char*, int, const char*)
> * trunk.i
> void
> __rw_assert_fail (const char*, const char*, int, const char*)
> *
> -
>
> 3. removing _RWSTD_NORETURN from declaration of the __rw_assert_fail()
> eliminates
> the ICE in tests from 4.2.x branch;
>
>
Thanks for looking into it! That's too bad about noreturn. It can help
the compiler generate better code so it would be a shame to have to
remove it. A test case might help us come up with a workaround and get
Microsoft to fix their compiler :)
>
> 4. The __rw_assert_fail() is used in RandomAccessIter methods thus
> RW_ASSERT macro;
>
> So I guess, that MSVC issues ICE when processing "infinite" for-loop
> (loop without condition)
> with __declspec (noreturn) function inside the loop.
>
>
>> Btw., the reason I ask is so we can come up with a better/cleaner
>> workaround, hopefully one that will be as efficient as the original
>> code and won't need an #ifdef.
>
> I've just added condition in for-loop, to make the loop finite.
> We can rewrite the loop the same way, as it's implemented in Dinkumware
> STL
> (it's also fixes the ICE):
>
>
...
>
> _EXPORT
> template
> void __make_heap (_RandomAccessIter __first, _RandomAccessIter __last,
> _Compare __comp, _Dist*)
> {
> _RWSTD_ASSERT_RANGE (__first, __last);
>
> const _Dist __dist = __last - __first;
>
> for (_Dist __parent = __dist / 2; 0 < __parent; ) {
> --__parent;
> __adjust_heap (__first, __parent, __dist, *(__first + __parent),
>__comp);
> }
> }
>
>
This is close to what I had in mind when I saw the #ifdef but is it
the same thing? IIUC, in the original version, the loop will execute
at least once. In this version, it may not execute at all (i.e., if
__parent is zero). Or does it not matter? If not, this would seem
like an improvement over the original code irrespective of the ICE,
which is exactly what I was hoping for! :)
Martin
--
View this message in context:
http://www.nabble.com/Re%3A--jira--Created%3A-%28STDCXX-1022%29--MSVC-x86---optimized--ICE-in-std%3A%3A__make_heap%28%29-tp20532273p20553378.html
Sent from the stdcxx-dev mailing list archive at Nabble.com.
Re: svn commit: r713762 - /stdcxx/branches/4.2.x/include/rw/_config.h
Farid Zaripov-2 wrote: > >> I wonder if it would be worthwhile to give users the ability to >> decide whether to enable TLS in case they don't need LoadLibrary() >> or not. What do you think? > > Hmm. I only can say, that nor MSVC run-time, nor STLport nor boost > libraries are not > using the implicit TLS. > > Currently the TLS variables are used in 3 places only: > - exception's what-buffer; > - table for random number generator; > - buffer for __rw_tmpbuf(). > > Right. But there might be other opportunities for TLS (e.g., in locale or maybe in some of the C++ 0x facilities?) > > We need to check what would be if an exception object is created in one > thread (i.e. thus "throw new exception();") > and after catch() the pointer passed to another thread and there deleted? > > Ouch! Tricky! I hadn't thought of this when I implemented it. I think we either need to get this case to work or disable TLS for exceptions, but we can't have it crash (which, I assume, is what happens in this case?) > > The same issue with __rw_tmpbuf(). What would be if we getting the > temporary buffer in one thread using > get_temporary_buffer(), and releasing it, using return_temporary_buffer(), > in another thread? > > That would also be a problem. In this case, though, I think it would be sufficient to document it as a restriction of the API. IMO, getting this to work would be more trouble than it's worth. Martin -- View this message in context: http://www.nabble.com/Re%3A-svn-commit%3A-r713762stdcxx-branches-4.2.x-include-rw-_config.h-tp20518249p20553291.html Sent from the stdcxx-dev mailing list archive at Nabble.com.
Re: svn commit: r713762 - /stdcxx/branches/4.2.x/include/rw/_config.h
> I wonder if it would be worthwhile to give users the ability to > decide whether to enable TLS in case they don't need LoadLibrary() > or not. What do you think? Hmm. I only can say, that nor MSVC run-time, nor STLport nor boost libraries are not using the implicit TLS. Currently the TLS variables are used in 3 places only: - exception's what-buffer; - table for random number generator; - buffer for __rw_tmpbuf(). We need to check what would be if an exception object is created in one thread (i.e. thus "throw new exception();") and after catch() the pointer passed to another thread and there deleted? The same issue with __rw_tmpbuf(). What would be if we getting the temporary buffer in one thread using get_temporary_buffer(), and releasing it, using return_temporary_buffer(), in another thread? Farid.
Re: [jira] Created: (STDCXX-1022) [MSVC x86 / optimized] ICE in std::__make_heap()
> Did you happen to reduce this to a small test case and send
> it to Microsoft?
Not yet. It's difficult to create the small test.
> Do you see the ICE in a specific test (or example), maybe only for a
> specific instantiation of the template, or does it happen regardless
> of the actual types of the template arguments?
Today I've made some investigations:
1. the 25.heap, 25.partial.sort and 25.sort tests from trunk and 4.3.x branch
are compiled withoud ICE;
2. the only difference between preprocessed 25.heap tests from 4.2.x branch and
trunk is
-
Comparing files 42x.i and trunk.i
* 42x.i
void __declspec (noreturn)
__rw_assert_fail (const char*, const char*, int, const char*)
* trunk.i
void
__rw_assert_fail (const char*, const char*, int, const char*)
*
-
3. removing _RWSTD_NORETURN from declaration of the __rw_assert_fail()
eliminates
the ICE in tests from 4.2.x branch;
4. The __rw_assert_fail() is used in RandomAccessIter methods thus RW_ASSERT
macro;
So I guess, that MSVC issues ICE when processing "infinite" for-loop (loop
without condition)
with __declspec (noreturn) function inside the loop.
> Btw., the reason I ask is so we can come up with a better/cleaner
> workaround, hopefully one that will be as efficient as the original
> code and won't need an #ifdef.
I've just added condition in for-loop, to make the loop finite.
We can rewrite the loop the same way, as it's implemented in Dinkumware STL
(it's also fixes the ICE):
---
template inline
void _Make_heap(_RanIt _First, _RanIt _Last, _Pr _Pred, _Diff *, _Ty *)
{ // make nontrivial [_First, _Last) into a heap, using _Pred
_Diff _Bottom = _Last - _First;
for (_Diff _Hole = _Bottom / 2; 0 < _Hole; )
{ // reheap top half, bottom to top
--_Hole;
std::_Adjust_heap(_First, _Hole, _Bottom,
_Ty(*(_First + _Hole)), _Pred);
}
}
---
-
_EXPORT
template
void __make_heap (_RandomAccessIter __first, _RandomAccessIter __last,
_Compare __comp, _Dist*)
{
_RWSTD_ASSERT_RANGE (__first, __last);
const _Dist __dist = __last - __first;
for (_Dist __parent = __dist / 2; 0 < __parent; ) {
--__parent;
__adjust_heap (__first, __parent, __dist, *(__first + __parent),
__comp);
}
}
-
Farid.
