[Bug other/37280] New: weak symbol regression breaks linux kernel

[Bug other/37280] New: weak symbol regression breaks linux kernel

[ak at muc dot de]

The linux kernel doesn't build anymore with gcc version 4.4.0 20080826

The problem seems to be that one extern variable (kallsyms_token_index)
declared with  __attribute__((weak)) does not get a ".weak" in the assembler
output
anymore (compared the gcc 4.3)

I attached a preprocessed test case.

Build options:
-Wall -Wundef -Wstrict-prototypes -Wno-trigraphs -fno-strict-aliasing
-fno-common -Werror-implicit-function-declaration -Os -m64 -mtune=generic
-mno-red-zone -mcmodel=kernel -funit-at-a-time -maccumulate-outgoing-args -pipe
-Wno-sign-compare -fno-asynchronous-unwind-tables -mno-sse -mno-mmx -mno-sse2
-mno-3dnow -Iinclude/asm-x86/mach-default -Wframe-larger-than=2048
-fno-stack-protector -fomit-frame-pointer --param large-stack-frame=100 --param
large-stack-frame-growth=100 -Wdeclaration-after-statement -Wno-pointer-sign
-fverbose-asm

Built with gcc43:

% grep kallsyms_token_index kernel/kallsyms.s
 movzwl  kallsyms_token_index(%rax,%rax), %eax   # kallsyms_token_index, tmp118
        movzwl  kallsyms_token_index(%rax,%rax), %eax   # kallsyms_token_index,
tmp124
        movzwl  kallsyms_token_index(%rax,%rax), %eax   # kallsyms_token_index,
tmp105
        movzwl  kallsyms_token_index(%rax,%rax), %eax   # kallsyms_token_index,
tmp102
        movzwl  kallsyms_token_index(%rax,%rax), %eax   # kallsyms_token_index,
tmp106
        movzwl  kallsyms_token_index(%rax,%rax), %eax   # kallsyms_token_index,
tmp93
        .weak   kallsyms_token_index


With 4.4:
  movzwl  kallsyms_token_index(%rdx,%rdx), %edx   # kallsyms_token_index,
tmp116
        movzwl  kallsyms_token_index(%rdi,%rdi), %edi   # kallsyms_token_index,
tmp122
        movzwl  kallsyms_token_index(%rdx,%rdx), %edx   # kallsyms_token_index,
tmp103
        movzwl  kallsyms_token_index(%rdx,%rdx), %edx   # kallsyms_token_index,
tmp100
        movzwl  kallsyms_token_index(%rcx,%rcx), %ecx   # kallsyms_token_index,
tmp103
        movzwl  kallsyms_token_index(%rcx,%rcx), %ecx   # kallsyms_token_index,
tmp91


Note no ".weak" in the second case. This breaks the build because it relies
on this symbol being weak on the first link pass.


-- 
           Summary: weak symbol regression breaks linux kernel
           Product: gcc
           Version: 4.4.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: other
        AssignedTo: unassigned at gcc dot gnu dot org
        ReportedBy: ak at muc dot de
  GCC host triplet: x86_64-linux
GCC target triplet: x86_64-linux


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=37280

[Bug other/37280] weak symbol regression breaks linux kernel

[hp at gcc dot gnu dot org]

------- Comment #1 from hp at gcc dot gnu dot org  2008-08-29 16:02 -------
(In reply to comment #0)
> I attached a preprocessed test case.

Where?


-- 

hp at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |hp at gcc dot gnu dot org


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=37280

[Bug other/37280] weak symbol regression breaks linux kernel

[andi-gcc at firstfloor dot org]

------- Comment #1 from hp at gcc dot gnu dot org  2008-08-29 16:02 -------
(In reply to comment #0)
> I attached a preprocessed test case.

Where?


------- Comment #2 from andi-gcc at firstfloor dot org  2008-08-29 16:02 -------
Created an attachment (id=16164)
 --> (http://gcc.gnu.org/bugzilla/attachment.cgi?id=16164&action=view)
preprocessed test case


-- 

hp at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |hp at gcc dot gnu dot org


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=37280

[Bug other/37280] weak symbol regression breaks linux kernel

[hp at gcc dot gnu dot org]

------- Comment #3 from hp at gcc dot gnu dot org  2008-08-29 16:30 -------
This is fixed by the patch in PR37170.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=37280

[Bug other/37280] weak symbol regression breaks linux kernel

[hp at gcc dot gnu dot org]

------- Comment #4 from hp at gcc dot gnu dot org  2008-08-29 17:03 -------
A wee bit shorter:
extern int kallsyms_token_index[] __attribute__((weak));
extern int kallsyms_token_table[] __attribute__((weak));
void kallsyms_expand_symbol(int *result)
{
  int len = *result;
  int *tptr;
  while(len) {
    tptr = &kallsyms_token_table[ kallsyms_token_index[*result] ];
    len--;
    while (*tptr) tptr++;
    *tptr = 1;
  }
 *result = 0;
}


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=37280

[Bug middle-end/37280] [4.4 Regression] weak symbol regression breaks linux kernel

[rguenth at gcc dot gnu dot org]

-- 

rguenth at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
  BugsThisDependsOn|                            |37170
          Component|other                       |middle-end
           Keywords|                            |wrong-code
            Summary|weak symbol regression      |[4.4 Regression] weak symbol
                   |breaks linux kernel         |regression breaks linux
                   |                            |kernel
   Target Milestone|---                         |4.4.0


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=37280

[Bug middle-end/37280] [4.4 Regression] weak symbol regression breaks linux kernel

[happyarch at gmail dot com]

------- Comment #5 from happyarch at gmail dot com  2008-09-03 22:58 -------
Hi,

i tried your patch, but it doesn't work for my lfs64(linux from scratch
pure-64bit)
http://gcc.gnu.org/ml/gcc-patches/2008-08/msg01407.html

...

  CC      init/version.o
  LD      init/built-in.o
  LD      .tmp_vmlinux1
kernel/built-in.o: In function `update_iter':
kallsyms.c:(.text+0x2fb86): undefined reference to `kallsyms_token_index'
kallsyms.c:(.text+0x2fbb5): undefined reference to `kallsyms_token_index'
kernel/built-in.o: In function `lookup_symbol_attrs':
(.text+0x2fdcc): undefined reference to `kallsyms_token_index'
kernel/built-in.o: In function `lookup_symbol_name':
(.text+0x2fea1): undefined reference to `kallsyms_token_index'
kernel/built-in.o: In function `kallsyms_lookup':
(.text+0x2ff92): undefined reference to `kallsyms_token_index'
kernel/built-in.o:(.text+0x30171): more undefined references to
`kallsyms_token_index' follow
make: *** [.tmp_vmlinux1] Error 1
bash-3.2$ cc -v
Using built-in specs.
Target: x86_64-unknown-linux-gnu
Configured with: ../gcc/configure --prefix=/usr --libexecdir=/usr/lib
--enable-shared --enable-threads=posix --enable-__cxa_atexit
--enable-clocale=gnu --enable-languages=c,c++ --enable-static
--disable-multilib
Thread model: posix
gcc version 4.4.0 20080903 (experimental) (GCC) 
bash-3.2$ pwd
/usr/src/linux-2.6.27-rc5


TIA


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=37280

[Bug middle-end/37280] [4.4 Regression] weak symbol regression breaks linux kernel

[hp at gcc dot gnu dot org]

------- Comment #6 from hp at gcc dot gnu dot org  2008-09-03 23:24 -------
(In reply to comment #5)
> Hi,
> 
> i tried your patch, but it doesn't work for my lfs64(linux from scratch
> pure-64bit)
> http://gcc.gnu.org/ml/gcc-patches/2008-08/msg01407.html

You're a few versions behind.  This is the current one.  See also PR in the
"depended on" box, PR37170:
<http://gcc.gnu.org/ml/gcc-patches/2008-09/msg00195.html>


-- 

hp at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
         AssignedTo|unassigned at gcc dot gnu   |hp at gcc dot gnu dot org
                   |dot org                     |
             Status|UNCONFIRMED                 |ASSIGNED
     Ever Confirmed|0                           |1
           Keywords|                            |patch
   Last reconfirmed|0000-00-00 00:00:00         |2008-09-03 23:24:28
               date|                            |


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=37280

[Bug middle-end/37280] [4.4 Regression] weak symbol regression breaks linux kernel

[happyarch at gmail dot com]

------- Comment #7 from happyarch at gmail dot com  2008-09-04 23:07 -------
(In reply to comment #6)
> (In reply to comment #5)
> > Hi,
> > 
> > i tried your patch, but it doesn't work for my lfs64(linux from scratch
> > pure-64bit)
> > http://gcc.gnu.org/ml/gcc-patches/2008-08/msg01407.html
> 
> You're a few versions behind.  This is the current one.  See also PR in the
> "depended on" box, PR37170:
> <http://gcc.gnu.org/ml/gcc-patches/2008-09/msg00195.html>
> 

Thanks it works like charm!


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=37280

[Bug middle-end/37280] [4.4 Regression] weak symbol regression breaks linux kernel

[jsm28 at gcc dot gnu dot org]

-- 

jsm28 at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Priority|P3                          |P1


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=37280

[Bug middle-end/37280] [4.4 Regression] weak symbol regression breaks linux kernel

[chaoyingfu at gcc dot gnu dot org]

------- Comment #8 from chaoyingfu at gcc dot gnu dot org  2008-09-16 01:11 -------
Hi,

  For MIPS targets (mipsisa32r2-sde-elf), ".weak" is still not emitted.
  Thanks!

Ex:
# cat 152.c
extern int arr[] __attribute__((weak));

int test(int i)
{
  return arr[i];
}

# ~/dev/binutils3/build-all2/gcc/cc1 -quiet 152.c -o 152.s -O2
# cat 152.s
        .file   1 "152.c"
        .section .mdebug.abi32
        .previous
        .gnu_attribute 4, 1
        .text
        .align  2
        .globl  test
        .set    nomips16
        .ent    test
        .type   test, @function
test:
        .frame  $sp,0,$31               # vars= 0, regs= 0/0, args= 0, gp= 0
        .mask   0x00000000,0
        .fmask  0x00000000,0
        .set    noreorder
        .set    nomacro

        lui     $2,%hi(arr)
        addiu   $2,$2,%lo(arr)
        sll     $4,$4,2
        addu    $4,$4,$2
        j       $31
        lw      $2,0($4)

        .set    macro
        .set    reorder
        .end    test
        .size   test, .-test
        .ident  "GCC: (GNU) 4.4.0 20080915 (experimental) [trunk revision
131984]"


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=37280

[Bug middle-end/37280] [4.4 Regression] weak symbol regression breaks linux kernel

[chaoyingfu at gcc dot gnu dot org]

------- Comment #9 from chaoyingfu at gcc dot gnu dot org  2008-09-16 18:54 -------
Hi,

  I tried the following patch.
http://gcc.gnu.org/ml/gcc-patches/2008-09/msg00195.html

  ".weak symbol" can be emitted, but with two same lines.  Thanks!

Ex:
        .file   1 "152.c"
        .section .mdebug.abi32
        .previous
        .gnu_attribute 4, 1
        .text
        .align  2
        .globl  test
        .set    nomips16
        .ent    test
        .type   test, @function
test:
        .frame  $sp,0,$31               # vars= 0, regs= 0/0, args= 0, gp= 0
        .mask   0x00000000,0
        .fmask  0x00000000,0
        .set    noreorder
        .set    nomacro

        lui     $2,%hi(arr)
        addiu   $2,$2,%lo(arr)
        sll     $4,$4,2
        addu    $4,$4,$2
        j       $31
        lw      $2,0($4)

        .set    macro
        .set    reorder
        .end    test
        .size   test, .-test
        .weak   arr
        .weak   arr
        .ident  "GCC: (GNU) 4.4.0 20080915 (experimental) [trunk revision
131984
]"


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=37280

[Bug middle-end/37280] [4.4 Regression] weak symbol regression breaks linux kernel

[aldot at gcc dot gnu dot org]

------- Comment #10 from aldot at gcc dot gnu dot org  2008-09-16 19:15 -------
(In reply to comment #9)
> Hi,
> 
>   I tried the following patch.
> http://gcc.gnu.org/ml/gcc-patches/2008-09/msg00195.html
> 
>   ".weak symbol" can be emitted, but with two same lines.  Thanks!

So that would bring you into the same situation as PR31537 is in (FE emits
duplicate asm directives, quick workaround in the BE for weakrefs is attached
to PR31537).


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=37280

[Bug middle-end/37280] [4.4 Regression] weak symbol regression breaks linux kernel

[hp at gcc dot gnu dot org]

------- Comment #11 from hp at gcc dot gnu dot org  2008-09-22 01:55 -------
Subject: Bug 37280

Author: hp
Date: Mon Sep 22 01:54:03 2008
New Revision: 140539

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=140539
Log:
        PR middle-end/37170
        PR middle-end/37280
        * final.c (mark_symbol_ref_as_used): New helper function.
        (output_operand): Instead of just looking inside MEMs for
        SYMBOL_REFs, use new helper function and for_each_rtx.
        * varasm.c (assemble_external): Move #ifndef ASM_OUTPUT_EXTERNAL
        to after weak-handling.  Don't mark decls with TREE_STATIC as weak.
        Make head comment more general.
        * config/darwin.c (machopic_output_indirection): Handle weak
        references here, like in assemble_external.

Modified:
    trunk/gcc/ChangeLog
    trunk/gcc/config/darwin.c
    trunk/gcc/final.c
    trunk/gcc/varasm.c


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=37280

[Bug middle-end/37280] [4.4 Regression] weak symbol regression breaks linux kernel

[hp at gcc dot gnu dot org]

------- Comment #12 from hp at gcc dot gnu dot org  2008-09-22 01:56 -------
Subject: Bug 37280

Author: hp
Date: Mon Sep 22 01:54:41 2008
New Revision: 140540

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=140540
Log:
        PR middle-end/37170
        PR middle-end/37280
        * gcc.dg/weak/weak-15.c, gcc.dg/weak/weak-16.c,
        g++.dg/ext/inline1.C: New tests.

Added:
    trunk/gcc/testsuite/g++.dg/ext/inline1.C
    trunk/gcc/testsuite/gcc.dg/weak/weak-15.c
    trunk/gcc/testsuite/gcc.dg/weak/weak-16.c
Modified:
    trunk/gcc/testsuite/ChangeLog


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=37280

[Bug middle-end/37280] [4.4 Regression] weak symbol regression breaks linux kernel

[hp at gcc dot gnu dot org]

------- Comment #13 from hp at gcc dot gnu dot org  2008-09-22 02:12 -------
(In reply to comment #10)
> (In reply to comment #9)
> >   ".weak symbol" can be emitted, but with two same lines.  Thanks!
> So that would bring you into the same situation as PR31537 is in (FE emits
> duplicate asm directives, quick workaround in the BE for weakrefs is attached
> to PR31537).

Pre-existing wart.
Better use a hashtable than a linked list (i.e. tree) for weak_decls.
I suggest opening a separate PR for that; this one's now FIXED.


-- 

hp at gcc dot gnu dot org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|ASSIGNED                    |RESOLVED
         Resolution|                            |FIXED


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=37280