[Bug c/34484] New: pulls in allegedly unneeded floatingpoint exception access funcs

[Bug c/34484] New: pulls in allegedly unneeded floatingpoint exception access funcs

[aldot at gcc dot gnu dot org]

My libc is configured to omit any FP support (UCLIBC_HAS_FLOATS is not set)
but the recent libbid updates seems to unconditionally pull in floatingpoint
accessor functions thus breaking bootstrap -- since feclearexcept et al are
thus unavailable. My notes on this read:

--------8<--------
Follows:
Precedes:

    do not pull in allegedly unneeded floatingpoint exception access funcs

      HJL's recent update of libbid would pull in Floating-point exception
      handling, although __GCC_FLOAT_NOT_NEEDED is defined.               

      Prevent pulling in feclearexcept, feraiseexcept et al for now.
      FIXME: revisit
--------8<--------
A proposed patch is in this thread:
http://gcc.gnu.org/ml/gcc/2007-12/msg00149.html


-- 
           Summary: pulls in allegedly unneeded floatingpoint exception
                    access funcs
           Product: gcc
           Version: 4.3.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: c
        AssignedTo: hjl at lucon dot org
        ReportedBy: aldot at gcc dot gnu dot org


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=34484

[Bug c/34484] [4.3 Regression] pulls in allegedly unneeded floatingpoint exception access funcs

[aldot at gcc dot gnu dot org]

-- 

aldot at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
      Known to work|                            |4.2.2
            Summary|pulls in allegedly unneeded |[4.3 Regression] pulls in
                   |floatingpoint exception     |allegedly unneeded
                   |access funcs                |floatingpoint exception
                   |                            |access funcs
   Target Milestone|---                         |4.3.0


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=34484

[Bug target/34484] pulls in allegedly unneeded floatingpoint exception access funcs

[pinskia at gcc dot gnu dot org]

-- 

pinskia at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |pinskia at gcc dot gnu dot
                   |                            |org
          Component|c                           |target
 GCC target triplet|                            |i386-pc-linux-gnu
            Summary|[4.3 Regression] pulls in   |pulls in allegedly unneeded
                   |allegedly unneeded          |floatingpoint exception
                   |floatingpoint exception     |access funcs
                   |access funcs                |
   Target Milestone|4.3.0                       |---


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=34484

[Bug target/34484] pulls in allegedly unneeded floatingpoint exception access funcs

[pinskia at gcc dot gnu dot org]

------- Comment #1 from pinskia at gcc dot gnu dot org  2007-12-15 20:07 -------
Note it is unusual to disable fp support for x86.  So really I think it is a
bug that you can disable it.  Actually I think we should declare this as
invalid.


-- 

pinskia at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
 GCC target triplet|i386-pc-linux-gnu           |i386-pc-linux-uclibc


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=34484

[Bug target/34484] pulls in allegedly unneeded floatingpoint exception access funcs

[aldot at gcc dot gnu dot org]

------- Comment #2 from aldot at gcc dot gnu dot org  2007-12-15 20:15 -------
IIRC i initially experienced this not on i386 but with mips or arm.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=34484

[Bug target/34484] pulls in allegedly unneeded floatingpoint exception access funcs

[pinskia at gcc dot gnu dot org]

------- Comment #3 from pinskia at gcc dot gnu dot org  2007-12-15 20:22 -------
Even then you have to have some kind of fp support for GCC to work correctly. 
Even if it is soft fp.


-- 

pinskia at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
 GCC target triplet|i386-pc-linux-uclibc        |*-pc-linux-uclibc


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=34484

[Bug target/34484] pulls in allegedly unneeded floatingpoint exception access funcs

[hjl at lucon dot org]

------- Comment #4 from hjl at lucon dot org  2007-12-15 22:05 -------
Where is __GCC_FLOAT_NOT_NEEDED documented? Is this a supported gcc macro?


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=34484

[Bug target/34484] pulls in allegedly unneeded floatingpoint exception access funcs

[hjl at lucon dot org]

------- Comment #5 from hjl at lucon dot org  2007-12-27 16:56 -------
Any update on this?


-- 

hjl at lucon dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |WAITING


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=34484

[Bug target/34484] pulls in allegedly unneeded floatingpoint exception access funcs

[uros at gcc dot gnu dot org]

------- Comment #6 from uros at gcc dot gnu dot org  2008-01-18 09:56 -------
Subject: Bug 34484

Author: uros
Date: Fri Jan 18 09:55:15 2008
New Revision: 131626

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=131626
Log:
        PR debug/34484
        * dwarf2out.c (dwarf2out_switch_text_section): Do not guard with
        DWARF2_DEBUGGING_INFO.
        (dwarf2out_note_section_used): Ditto.  Add prototype.
        (have_multiple_function_sections, text_section_used,
        cold_text_section_used, *cold_text_sections): Move declarations
        before their uses.


Modified:
    trunk/gcc/ChangeLog
    trunk/gcc/dwarf2out.c


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=34484

[Bug target/34484] pulls in allegedly unneeded floatingpoint exception access funcs

[ubizjak at gmail dot com]

------- Comment #7 from ubizjak at gmail dot com  2008-01-18 11:54 -------
Sorry, wrong PR number in the ChangeLog.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=34484

[Bug target/34484] pulls in allegedly unneeded floatingpoint exception access funcs

[aldot at gcc dot gnu dot org]

------- Comment #8 from aldot at gcc dot gnu dot org  2008-01-19 21:38 -------
>From FSFChangelog.10:
Mon Feb 12 20:42:11 1996  Randy Smith  <randys@camaro.osf.org>

        * i386/x-osfrose (XCFLAGS{,_NODEBUG}): Remove $(SHLIB).
        (XCFLAGS): New variable.
        (libdir, mandir, bindir): Delete.
        * i386/t-osf: New file.
        * i860/paragon.h (STARTFILE_SPEC): Make gcc find crt0.o, not loader.
        (LIB_SPEC): Remove /usr/lib.
        * Makefile.in (TCFLAGS): New variable.
        (GCC_CFLAGS): Add $(TCFLAGS).
        (LIBGCC2_CFLAGS): Add -D for __GCC_FLOAT_NOT_NEEDED.
        (libgcc1-test): Remove -nostdlib.
        (float.h-cross): Don't give error #ifdef __GCC_FLOAT_NOT_NEEDED.
        * enquire.c: Define __GCC_FLOAT_NOT_NEEEDED.
        * configure (i[3456]86-*-osfrose): Add t-osf as tmake_file.


Nowadays we compile libgcc with it (for the very same reasons, i suspect):
[from gcc/Makefile.in]
# Options to use when compiling libgcc2.a.
#
LIBGCC2_DEBUG_CFLAGS = -g
LIBGCC2_CFLAGS = -O2 $(LIBGCC2_INCLUDES) $(GCC_CFLAGS) $(TARGET_LIBGCC2_CFLAGS) 
\
                 $(LIBGCC2_DEBUG_CFLAGS) $(GTHREAD_FLAGS) \
                 -DIN_LIBGCC2 -D__GCC_FLOAT_NOT_NEEDED \
                 $(INHIBIT_LIBC_CFLAGS)

Current trunk still fails of you don't have a fenv.h or none of the respective
functions provided in fenv.h.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=34484

[Bug target/34484] pulls in allegedly unneeded floatingpoint exception access funcs

[hjl dot tools at gmail dot com]

------- Comment #9 from hjl dot tools at gmail dot com  2008-03-14 01:17 -------
DFP needs floating point exception support. If your C library doesn't
support it, you can disable DFP.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=34484

[Bug target/34484] pulls in allegedly unneeded floatingpoint exception access funcs

[aldot at gcc dot gnu dot org]

------- Comment #10 from aldot at gcc dot gnu dot org  2008-06-24 11:19 -------
(In reply to comment #9)
> DFP needs floating point exception support. If your C library doesn't
> support it, you can disable DFP.

libdecnumber's configury should thus check whether fenv.h is available or not
and default to no if no fenv.h was found.

Looks like the enable_decimal_float handling is duplicated both in
gcc/configure.ac as well as libdecnumber/configure.ac. Is there a reason for
this duplication or is it pure oversight?


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=34484

[Bug target/34484] pulls in allegedly unneeded floatingpoint exception access funcs

[aldot at gcc dot gnu dot org]

------- Comment #11 from aldot at gcc dot gnu dot org  2008-06-24 11:21 -------
libdecnumber hunk could look like this:

Index: libdecnumber/configure.ac
===================================================================
--- libdecnumber/configure.ac   (revision 137059)
+++ libdecnumber/configure.ac   (working copy)
@@ -49,6 +49,7 @@ ACX_PROG_CC_WARNINGS_ARE_ERRORS([manual]

 # Checks for header files.
 AC_CHECK_HEADERS(ctype.h stddef.h string.h stdio.h)
+AC_CHECK_HEADER([fenv.h],[],[libdecnumber_cv_fenv=no])
 GCC_HEADER_STDINT(gstdint.h)

 # Checks for typedefs, structures, and compiler characteristics.
@@ -99,10 +100,13 @@ Valid choices are 'yes', 'bid', 'dpd', a
   esac
 ])

+if test x$libdecnumber_cv_fenv = xno; then
+  enable_decimal_float=no
+elif test x$enable_decimal_float = xyes -o x$enable_decimal_float = xno; then
 # x86's use BID format instead of DPD
 # In theory --enable-decimal-float=no should not compile anything
 # For the sake of simplicity, just use the default format in this directory
-if test x$enable_decimal_float = xyes -o x$enable_decimal_float = xno; then
+
   case $target in
     i?86*-*-linux* | x86_64*-*-linux*)
       enable_decimal_float=bid


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=34484

[Bug target/34484] pulls in allegedly unneeded floatingpoint exception access funcs

[joseph at codesourcery dot com]

------- Comment #12 from joseph at codesourcery dot com  2008-06-24 12:36 -------
Subject: Re:  pulls in allegedly unneeded floatingpoint
 exception access funcs

On Tue, 24 Jun 2008, aldot at gcc dot gnu dot org wrote:

> libdecnumber hunk could look like this:

libdecnumber's configure.ac is for the host, not the target.  As such, it 
needs to remain in sync with gcc's.  I'm pretty sure it's deliberate that 
it still needs to select a format even with --disable-decimal-float.

The logic selecting for the target is in libgcc/configure.ac.  That's the 
only place disabling without <fenv.h> might make sense (and if an explicit 
--enable-decimal-float={yes,bid,dpd} is given without <fenv.h>, give an 
error).  The only reason this really makes sense is that you can 
legitimately have both glibc and uClibc multilibs, some with <fenv.h> and 
some without, and libgcc's configure is run once for each multilib.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=34484