incubator-stdcxx-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Martin Sebor (JIRA)" <>
Subject [jira] Commented: (STDCXX-509) RW libstd is sensitive to link order
Date Thu, 09 Aug 2007 01:00:59 GMT


Martin Sebor commented on STDCXX-509:

See the thread in the archive for reference:

Below are the relevant symbols in 4.1.3 (on Linux x86_64), with 'B' denoting BSS data (uninitialized
data section).

$ nm -C ../lib/ | egrep "__rw_(flt|dbl|ldbl)_"
0000000000062214 t global constructors keyed to _ZN4__rw17__rw_dbl_infinityE
0000000000213ea0 B __rw::__rw_dbl_qNaN
0000000000213ec0 B __rw::__rw_dbl_sNaN
0000000000213ea8 B __rw::__rw_flt_qNaN
0000000000213ec8 B __rw::__rw_flt_sNaN
0000000000213eb0 B __rw::__rw_ldbl_qNaN
0000000000213ed0 B __rw::__rw_ldbl_sNaN
00000000000bae90 r __rw::__rw_dbl_inf_bits
0000000000213e80 B __rw::__rw_dbl_infinity
00000000000bae98 r __rw::__rw_flt_inf_bits
0000000000213e88 B __rw::__rw_flt_infinity
00000000000baeb0 r __rw::__rw_dbl_qNaN_bits
00000000000baed0 r __rw::__rw_dbl_sNaN_bits
00000000000baeb8 r __rw::__rw_flt_qNaN_bits
00000000000baed8 r __rw::__rw_flt_sNaN_bits
00000000000baea0 r __rw::__rw_ldbl_inf_bits
0000000000213e90 B __rw::__rw_ldbl_infinity
0000000000213ee0 B __rw::__rw_dbl_denorm_min
0000000000213ee8 B __rw::__rw_flt_denorm_min
00000000000baec0 r __rw::__rw_ldbl_qNaN_bits
00000000000baee0 r __rw::__rw_ldbl_sNaN_bits
0000000000213ef0 B __rw::__rw_ldbl_denorm_min
00000000000baef0 r __rw::__rw_dbl_denorm_min_bits
00000000000baef8 r __rw::__rw_flt_denorm_min_bits
00000000000baf00 r __rw::__rw_ldbl_denorm_min_bits

Here are the symbols on trunk after the committed change, with 'R' denoting statically initialized
symbols in the read-only section:

$ nm -C ../lib/ | egrep "__rw_(flt|dbl|ldbl)_"
00000000000c2510 R __rw_dbl_denorm_min
00000000000c24b0 R __rw_dbl_infinity
00000000000c24d0 R __rw_dbl_qNaN
00000000000c24f0 R __rw_dbl_sNaN
00000000000c2518 R __rw_flt_denorm_min
00000000000c24b8 R __rw_flt_infinity
00000000000c24d8 R __rw_flt_qNaN
00000000000c24f8 R __rw_flt_sNaN
00000000000c2520 R __rw_ldbl_denorm_min
00000000000c24c0 R __rw_ldbl_infinity
00000000000c24e0 R __rw_ldbl_qNaN
00000000000c2500 R __rw_ldbl_sNaN

In addition to making the symbols statically initialized, the change also reduced the size
of the limits_bits.o object file in which they are defined from 12K to 5K:

$ ls -l ../lib/limits_bits.o ../lib/
-rwxr-xr-x  1 sebor devel 4870602 Aug  7 22:58 ../lib/
-rw-r--r--  1 sebor devel   12288 Aug  7 22:58 ../lib/limits_bits.o

$ ls -l ../lib/limits_bits.o ../lib/
-rwxr-xr-x  1 sebor devel 5007014 Aug  9 00:49 ../lib/
-rw-r--r--  1 sebor devel    5224 Aug  9 00:49 ../lib/limits_bits.o

> RW libstd is sensitive to link order
> ------------------------------------
>                 Key: STDCXX-509
>                 URL:
>             Project: C++ Standard Library
>          Issue Type: Improvement
>          Components: Build
>    Affects Versions: 4.1.2
>         Environment: gcc3.2, linux AS3.0
>            Reporter: Ravi K Inampudi
>            Assignee: Martin Sebor
>             Fix For: 4.1.2
>         Attachments: limits_bits.cpp.patch
> A customer discovered a problem with std::numeric_limits<double> in RW libstd.
Placing RW libstd on linkline *before* libFoo results in the program printing "0" instead
of "inf". The problem doesn't happen with native gcc STL.
> We reproduced the problem with the latest version of standard library downloaded from
Apache. From our understanding, when creating the and using the native standard
library is creating an implicit dependency on the libsupc++ (compiler dependent). This implicit
dependency is initializing the static variable when is loaded. If we use RW libstd
instread of libstdc++(native gcc stl), the program will print "0" instead of "inf".
>  # make shared lib
> gcc -I/amd/homes/dean/temp/stdlib-cxx/stdcxx-4.1.3/include/ansi -pthread -D_RWSTD_USE_CONFIG
-I/nfs/homes/dean/temp/stdlib-cxx/12d/include -I/amd/homes/dean/temp/stdlib-cxx/stdcxx-4.1.3/include
-I/amd/homes/dean/temp/stdlib-cxx/stdcxx-4.1.3/examples/include -pedantic -nostdinc++ -O2
-Wall -W -Wcast-qual -Winline -Wshadow -Wwrite-strings -Wno-long-long -Wcast-align -fPIC -shared
-o foo.C 
>  # make binary.
> gcc -pthread -o mainRW main.C -L/nfs/homes/dean/temp/stdlib-cxx/12d/lib -lstd12d -lsupc++
-lm -lFooRW 
> However, using the RW standard library, if such dependency is created explicitly then
it prints "inf". For example, the following set of commands fixes the problem:
>  # make shared lib
> gcc -I/amd/homes/dean/temp/stdlib-cxx/stdcxx-4.1.3/include/ansi -pthread -D_RWSTD_USE_CONFIG
-I/nfs/homes/dean/temp/stdlib-cxx/12d/include -I/amd/homes/dean/temp/stdlib-cxx/stdcxx-4.1.3/include
-I/amd/homes/dean/temp/stdlib-cxx/stdcxx-4.1.3/examples/include -pedantic -nostdinc++ -O2
-Wall -W -Wcast-qual -Winline -Wshadow -Wwrite-strings -Wno-long-long -Wcast-align -fPIC -shared
-o foo.C -lstd12d
>  # make binary.
> gcc -pthread -o mainRW main.C -L/nfs/homes/dean/temp/stdlib-cxx/12d/lib -lstd12d -lsupc++
-lm -lFooRW 
> But the customer never links their shared libs(i.e libFoo in this example) with RW libstd.
 They only link binaries with RW libstd! But it makes RW libstd sensitive to link order. And
they think the problem is in limits.cpp file:
> > <snip>
> > static union {
> > char _C_bits [sizeof (double)];
> > double _C_inf;
> > } __rw_dbl_inf_bits = { _RWSTD_DBL_INF_BITS };
> >
> > _RWSTD_EXPORT extern const double __rw_dbl_infinity =
> > __rw_dbl_inf_bits._C_inf; </snip>
> >
> > __rw_dbl_infinity ends up in the uninitialized data section of
> >
> >
> > nm -C| grep __rw_dbl_infinity
> > 000937f8 B __rw::__rw_dbl_infinity
> >
> If the symbol was initialized in data section, the link order wouldn't matter.
> Environment: RWSP6, gcc3.2, linux AS3.0
> Martin Sebor's Comments:
> Because the initializer of __rw_dbl_infinity is not a constant expresssion [expr.const]
the symbol is initialized dynamically (at runtime) rather than statically (i.e., at load time).
It seems that the compiler should be able to initialize it statically anyway, even if it's
not required to. Regardless, we should avoid making the assumption that it will and change
the initialization of the extern constants to use constant expressions instead.

This message is automatically generated by JIRA.
You can reply to this email to add a comment to the issue online.

View raw message