[Bug c/31128] New: ICE with __builtin_stack_restore

[Bug c/31128] New: ICE with __builtin_stack_restore

[nicholas at mxc dot ca]

[-- Warning: decoded text below may be mangled, UTF-8 assumed --]
[-- Attachment #1: Type: text/plain, Size: 2286 bytes --]

In a delta reduction from an ICE compiling code from LLVM's C backend, I get
the following one-line testcase:

  void f() { __builtin_stack_restore(f); }

$ gcc-4.1 b2.c -c
b2.c: In function ‘foo’:
b2.c:2: internal compiler error: in lhd_set_decl_assembler_name, at
langhooks.c:165
Please submit a full bug report,
with preprocessed source if appropriate.
See <URL:http://gcc.gnu.org/bugs.html> for instructions.
For Debian GNU/Linux specific bug reporting instructions,
see <URL:file:///usr/share/doc/gcc-4.1/README.Bugs>.
Preprocessed source stored into /tmp/cc6cllgG.out file, please attach this to
your bugreport.

This fails with 4.1.2 and 4.0.4:

$ gcc-4.1 -v
Using built-in specs.
Target: i486-linux-gnu
Configured with: ../src/configure -v
--enable-languages=c,c++,fortran,objc,obj-c++,treelang --prefix=/usr
--enable-shared --with-system-zlib --libexecdir=/usr/lib
--without-included-gettext --enable-threads=posix --enable-nls
--program-suffix=-4.1 --enable-__cxa_atexit --enable-clocale=gnu
--enable-libstdcxx-debug --enable-mpfr --with-tune=i686
--enable-checking=release i486-linux-gnu
Thread model: posix
gcc version 4.1.2 20061115 (prerelease) (Debian 4.1.1-21)
$ gcc-4.0 -v
Using built-in specs.
Target: i486-linux-gnu
Configured with: ../src/configure -v --enable-languages=c,c++,java
--prefix=/usr --enable-shared --with-system-zlib --libexecdir=/usr/lib
--without-included-gettext --enable-threads=posix --enable-nls
--program-suffix=-4.0 --enable-__cxa_atexit --enable-clocale=gnu
--enable-libstdcxx-debug --enable-java-awt=gtk-default --enable-gtk-cairo
--with-java-home=/usr/lib/jvm/java-1.4.2-gcj-4.0-1.4.2.0/jre --with-tune=i686
--enable-checking=release i486-linux-gnu
Thread model: posix
gcc version 4.0.4 20060904 (prerelease) (Debian 4.0.3-7)

but passes with gcc 3.4 and 3.3.

This bug may be related to bug 30095, except that this bug uses C not fortran,
and this bug occurs on x86 not ppc.


-- 
           Summary: ICE with __builtin_stack_restore
           Product: gcc
           Version: 4.1.2
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: c
        AssignedTo: unassigned at gcc dot gnu dot org
        ReportedBy: nicholas at mxc dot ca


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=31128

[Bug c/31128] ICE with __builtin_stack_restore

[pinskia at gcc dot gnu dot org]

------- Comment #1 from pinskia at gcc dot gnu dot org  2007-03-11 01:51 -------
I get a different ICE on the mainline with checking enabled:
t.c:1: internal compiler error: tree check: expected tree that contains 'decl
with RTL' structure, have 'addr_expr' in expand_stack_restore, at stmt.c:2012

Really this builtin should not be exposed to the user.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=31128

[Bug c/31128] __builtin_stack_restore/__builtin_stack_save should not be exposed to the user

[pinskia at gcc dot gnu dot org]

------- Comment #2 from pinskia at gcc dot gnu dot org  2007-03-11 01:58 -------
I have a patch to force this not to be exposed to the user, these builtin was
never documented so they should not be exposed and they are never used outside
of the internals of the compiler either.


-- 

pinskia at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
         AssignedTo|unassigned at gcc dot gnu   |pinskia at gcc dot gnu dot
                   |dot org                     |org
             Status|UNCONFIRMED                 |ASSIGNED
     Ever Confirmed|0                           |1
   Last reconfirmed|0000-00-00 00:00:00         |2007-03-11 01:58:22
               date|                            |


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=31128

[Bug c/31128] __builtin_stack_restore/__builtin_stack_save should not be exposed to the user

[nicholas at mxc dot ca]

------- Comment #3 from nicholas at mxc dot ca  2007-03-11 02:16 -------
That's fair; we were relying on an undocumented builtin (see llvm.org/PR1028).
If you remove it, can you provide a substitute?


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=31128

[Bug c/31128] __builtin_stack_restore/__builtin_stack_save should not be exposed to the user

[pinskia at gcc dot gnu dot org]

------- Comment #4 from pinskia at gcc dot gnu dot org  2007-03-11 02:20 -------
The solution I have for you is to use VLAs correctly and {}, basically going
lowered VLAs of __builtin_stack_restore/__builtin_stack_save/__builtin_alloca
to back VLAs.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=31128

[Bug c/31128] __builtin_stack_restore/__builtin_stack_save should not be exposed to the user

[pinskia at gcc dot gnu dot org]

------- Comment #5 from pinskia at gcc dot gnu dot org  2007-03-11 02:57 -------
Created an attachment (id=13187)
 --> (http://gcc.gnu.org/bugzilla/attachment.cgi?id=13187&action=view)
patch which hides the stack functions


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=31128

[Bug c/31128] __builtin_stack_restore/__builtin_stack_save should not be exposed to the user

[sabre at nondot dot org]

------- Comment #6 from sabre at nondot dot org  2007-05-28 17:44 -------
This is very useful for compilers generating C code (e.g. LLVM, and various
other source -> C compilers).  Why remove it?  These compilers are generating
partially structured code, that don't have syntactic blocks required for VLAs.


-- 

sabre at nondot dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |sabre at nondot dot org


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=31128

[Bug c/31128] __builtin_stack_restore/__builtin_stack_save should not be exposed to the user

[pinskia at gcc dot gnu dot org]

------- Comment #7 from pinskia at gcc dot gnu dot org  2007-05-28 19:08 -------
(In reply to comment #6)
> This is very useful for compilers generating C code (e.g. LLVM, and various
> other source -> C compilers).  Why remove it?  These compilers are generating
> partially structured code, that don't have syntactic blocks required for VLAs.

Because there are better ways of implementing that in the generated C code. 
You can create a fake VLA (that is of size 0).
Like:
int f(int t1)
{
  {
     int t = 0;
     int a[t];
     int *b = __builtin_alloca(t1);
  }
}


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=31128

[Bug c/31128] __builtin_stack_restore/__builtin_stack_save should not be exposed to the user

[sabre at nondot dot org]

------- Comment #8 from sabre at nondot dot org  2007-05-29 15:14 -------
Right, you could do that, but it is a) not guaranteed to work going forward,
and b) expects properly structured (e.g. nested) control flow.

If I had b, I could just make a vla :)

-Chris


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=31128

[Bug c/31128] __builtin_stack_restore/__builtin_stack_save should not be exposed to the user

[pinskia at gcc dot gnu dot org]

------- Comment #9 from pinskia at gcc dot gnu dot org  2007-11-17 18:21 -------
*** Bug 34134 has been marked as a duplicate of this bug. ***


-- 

pinskia at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |edwintorok at gmail dot com


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=31128

[Bug c/31128] __builtin_stack_restore/__builtin_stack_save should not be exposed to the user

[jakub at gcc dot gnu dot org]

------- Comment #10 from jakub at gcc dot gnu dot org  2007-11-17 19:25 -------
Created an attachment (id=14573)
 --> (http://gcc.gnu.org/bugzilla/attachment.cgi?id=14573&action=view)
gcc43-pr34134.patch

Patch to avoid the ICE.

I don't mind the patch which disables these builtins either, but
1) it might be better to add the space at the end of builtin's name rather than
   in the middle
2) have you posted it to gcc-patches?


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=31128

[Bug c/31128] __builtin_stack_restore/__builtin_stack_save should not be exposed to the user

[sabre at nondot dot org]

------- Comment #11 from sabre at nondot dot org  2007-11-28 20:57 -------
Please don't disable these.  There are a variety of compilers that optionally
or only generate C code, particularly for the functional languages.  This is
useful functionality for these compilers.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=31128

[Bug c/31128] __builtin_stack_restore/__builtin_stack_save should not be exposed to the user

[jakub at gcc dot gnu dot org]

------- Comment #12 from jakub at gcc dot gnu dot org  2007-12-04 22:01 -------
Fixed on the trunk.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=31128

[Bug c/31128] __builtin_stack_restore/__builtin_stack_save should not be exposed to the user

[pinskia at gcc dot gnu dot org]

------- Comment #13 from pinskia at gcc dot gnu dot org  2008-01-03 17:48 -------
I am not going to work on this anymore.


-- 

pinskia at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
         AssignedTo|pinskia at gcc dot gnu dot  |unassigned at gcc dot gnu
                   |org                         |dot org
             Status|ASSIGNED                    |NEW


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=31128

[Bug c/31128] __builtin_stack_restore/__builtin_stack_save should not be exposed to the user

[edmar at freescale dot com]

------- Comment #14 from edmar at freescale dot com  2008-01-24 18:57 -------
(In reply to comment #10)
> Created an attachment (id=14573)
 --> (http://gcc.gnu.org/bugzilla/attachment.cgi?id=14573&action=view) [edit]
> gcc43-pr34134.patch
> 
> Patch to avoid the ICE.
> 
> I don't mind the patch which disables these builtins either, but
> 1) it might be better to add the space at the end of builtin's name rather than
>    in the middle
> 2) have you posted it to gcc-patches?
> 

I still see a problem on PowerPC. From dejagnu logfile:

Executing on host: /tmp/edmar/build_cde/job_28/obj_gcc/gcc/xgcc
-B/tmp/edmar/build_cde/job_28/obj_gcc/gcc/   -O0  -w -fno-show-column -c
-o 20071117-1.o
/tmp/edmar/build_cde/job_28/gcc/gcc/testsuite/gcc.c-torture/compile/20071117-1.c
   (timeout = 300)
/tmp/edmar/build_cde/job_28/gcc/gcc/testsuite/gcc.c-torture/compile/20071117-1.c:
In function 'foo':
/tmp/edmar/build_cde/job_28/gcc/gcc/testsuite/gcc.c-torture/compile/20071117-1.c:13:
error: unrecognizable insn:
(insn 29 28 30 3
/tmp/edmar/build_cde/job_28/gcc/gcc/testsuite/gcc.c-torture/compile/20071117-1.c:10
(set (mem/c:SI (mem/f/c/i:SI (plus:SI (reg/f:SI 115 virtual-stack-vars)
                    (const_int 4 [0x4])) [0 p+0 S4 A32]) [0 S4 A8])
        (reg:SI 134)) -1 (nil))
/tmp/edmar/build_cde/job_28/gcc/gcc/testsuite/gcc.c-torture/compile/20071117-1.c:13:
internal compiler error: in extract_insn, at recog.c:1990


-- 

edmar at freescale dot com changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |edmar at freescale dot com


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=31128

[Bug c/31128] __builtin_stack_restore/__builtin_stack_save should not be exposed to the user

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=31128

Andrew Pinski <pinskia at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           See Also|                            |https://gcc.gnu.org/bugzill
                   |                            |a/show_bug.cgi?id=34393

--- Comment #15 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
(In reply to Edmar Wienskoski from comment #14)
> (In reply to comment #10)
> > Created an attachment (id=14573) [edit]
> > gcc43-pr34134.patch
> > 
> > Patch to avoid the ICE.
> > 
> > I don't mind the patch which disables these builtins either, but
> > 1) it might be better to add the space at the end of builtin's name rather than
> >    in the middle
> > 2) have you posted it to gcc-patches?
> > 
> 
> I still see a problem on PowerPC. From dejagnu logfile:

That was PR 34393 .