Skip to contentSkip to conversation

Add support for compiling with -ffunction-sections -fdata-sections

PATCH Superseded Tim Abbott 17 participants 99 messages Apr 5, 2009 – Apr 30, 2009

Subsystems (23)
lore
Share & tools
lore

Tim Abbott

author
#0 lore raw

v2 cover Add support for compiling with -ffunction-sections -fdata-sections

Hi Linus,

Here is a new version of the -ffunction-sections -fdata-sections patch series based on your current master. Please apply.

When I sent a version of this patch series at the start of the merge window, Stephen Rothwell pointed out that the first patch conflicted with several changes in linux-next, and it should be resent near the end of the merge window once those changes were merged. There are now no conflicting changes staged in linux-next[0], so now is probably the optimal time to merge this patch series.

[0] There is one trivial merge conflict with next-20090403 because e5083a63b6a8546c5fe1e571fe529e3939787ec2 was merged into master since then, but this will presumably disappear when next catches up tomorrow.

-Tim Abbott

--

The purpose of this patch series is to make it possible to build the kernel with "gcc -ffunction-sections -fdata-sections". There are two major applications for this functionality: decreasing vmlinux image size with --gc-sections, and Ksplice.

The original motivation for this functionality was to allow using the linker's unused section garbage collection support (ld --gc-sections) in order to get a smaller vmlinux image for embedded systems. People have been developing patches for supporting building the kernel with -ffunction-sections -fdata-sections for this purpose for a few years now (e.g. [1]). The most recent previous set of patches for --gc-sections was by Denys Vlasenko, and saved 10% on vmlinux size with CONFIG_MODULES=n and 1% with CONFIG_MODULES=y [2,3,4]. The primary source of complexity in the various patch series for doing section garbage collection has been in the patches adding support for compiling the kernel with -ffunction-sections -fdata-sections, so merging this patch series should be a big step towards those significant savings in kernel image size.

[1] <http://lkml.org/lkml/2006/6/4/169>
[2] <http://lkml.org/lkml/2007/9/5/90>
[3] <http://lkml.org/lkml/2007/9/7/110>
[4] <http://lkml.org/lkml/2008/7/1/499>

Support for building the kernel with -ffunction-sections -fdata-sections is the only significant prerequisite change currently required by Ksplice. Ksplice is a system for updating the Linux kernel without rebooting [5].

[5] <http://lkml.org/lkml/2009/2/24/362> is the latest version of the Ksplice patch series

Patches 1/4, 2/4, and 3/4 are independent of each other, but all three are prerequisites for patch 4/4.

Anders Kaseorg (1):
modpost: Check the section flags, not name, to catch missing

    "ax"/"aw"

Denys Vlasenko (1):
modpost: Support objects with more than 64k sections

Tim Abbott (1):
Make section names compatible with -ffunction-sections
-fdata-sections

Waseem Daher (1):
x86: Add an option to compile with -ffunction-sections
-fdata-sections

 Documentation/mutex-design.txt              |    4 +-
 Makefile                                    |    4 +
 arch/alpha/kernel/head.S                    |    2 +-
 arch/alpha/kernel/init_task.c               |    2 +-
 arch/alpha/kernel/vmlinux.lds.S             |   14 ++--
 arch/arm/kernel/head-nommu.S                |    2 +-
 arch/arm/kernel/head.S                      |    2 +-
 arch/arm/kernel/init_task.c                 |    2 +-
 arch/arm/kernel/vmlinux.lds.S               |   16 ++--
 arch/arm/mm/proc-v6.S                       |    2 +-
 arch/arm/mm/proc-v7.S                       |    2 +-
 arch/arm/mm/tlb-v6.S                        |    2 +-
 arch/arm/mm/tlb-v7.S                        |    2 +-
 arch/avr32/kernel/init_task.c               |    2 +-
 arch/avr32/kernel/vmlinux.lds.S             |    6 +-
 arch/avr32/mm/init.c                        |    2 +-
 arch/blackfin/kernel/vmlinux.lds.S          |    2 +-
 arch/cris/kernel/process.c                  |    2 +-
 arch/cris/kernel/vmlinux.lds.S              |    2 +-
 arch/frv/kernel/break.S                     |    4 +-
 arch/frv/kernel/entry.S                     |    2 +-
 arch/frv/kernel/head-mmu-fr451.S            |    2 +-
 arch/frv/kernel/head-uc-fr401.S             |    2 +-
 arch/frv/kernel/head-uc-fr451.S             |    2 +-
 arch/frv/kernel/head-uc-fr555.S             |    2 +-
 arch/frv/kernel/head.S                      |    4 +-
 arch/frv/kernel/init_task.c                 |    2 +-
 arch/frv/kernel/vmlinux.lds.S               |   18 ++--
 arch/frv/mm/tlb-miss.S                      |    2 +-
 arch/h8300/boot/compressed/head.S           |    2 +-
 arch/h8300/boot/compressed/vmlinux.lds      |    2 +-
 arch/h8300/kernel/init_task.c               |    2 +-
 arch/h8300/kernel/vmlinux.lds.S             |    2 +-
 arch/ia64/include/asm/asmmacro.h            |   12 +-
 arch/ia64/include/asm/cache.h               |    2 +-
 arch/ia64/include/asm/percpu.h              |    2 +-
 arch/ia64/kernel/Makefile.gate              |    2 +-
 arch/ia64/kernel/gate-data.S                |    2 +-
 arch/ia64/kernel/gate.S                     |    8 +-
 arch/ia64/kernel/gate.lds.S                 |   10 +-
 arch/ia64/kernel/head.S                     |    2 +-
 arch/ia64/kernel/init_task.c                |    4 +-
 arch/ia64/kernel/ivt.S                      |    2 +-
 arch/ia64/kernel/minstate.h                 |    4 +-
 arch/ia64/kernel/paravirtentry.S            |    2 +-
 arch/ia64/kernel/vmlinux.lds.S              |   44 ++++----
 arch/ia64/kvm/vmm_ivt.S                     |    2 +-
 arch/ia64/xen/gate-data.S                   |    2 +-
 arch/ia64/xen/xensetup.S                    |    2 +-
 arch/m32r/kernel/head.S                     |    2 +-
 arch/m32r/kernel/init_task.c                |    2 +-
 arch/m32r/kernel/vmlinux.lds.S              |    8 +-
 arch/m68k/kernel/head.S                     |    2 +-
 arch/m68k/kernel/process.c                  |    2 +-
 arch/m68k/kernel/sun3-head.S                |    2 +-
 arch/m68k/kernel/vmlinux-std.lds            |    6 +-
 arch/m68k/kernel/vmlinux-sun3.lds           |    4 +-
 arch/m68knommu/kernel/init_task.c           |    2 +-
 arch/m68knommu/kernel/vmlinux.lds.S         |    6 +-
 arch/m68knommu/platform/68360/head-ram.S    |    2 +-
 arch/m68knommu/platform/68360/head-rom.S    |    2 +-
 arch/mips/kernel/init_task.c                |    2 +-
 arch/mips/kernel/vmlinux.lds.S              |    8 +-
 arch/mips/lasat/image/head.S                |    2 +-
 arch/mips/lasat/image/romscript.normal      |    2 +-
 arch/mn10300/kernel/head.S                  |    2 +-
 arch/mn10300/kernel/init_task.c             |    2 +-
 arch/mn10300/kernel/vmlinux.lds.S           |   16 ++--
 arch/parisc/include/asm/cache.h             |    2 +-
 arch/parisc/include/asm/system.h            |    2 +-
 arch/parisc/kernel/head.S                   |    2 +-
 arch/parisc/kernel/init_task.c              |    8 +-
 arch/parisc/kernel/vmlinux.lds.S            |   26 +++---
 arch/powerpc/include/asm/cache.h            |    2 +-
 arch/powerpc/include/asm/page_64.h          |    2 +-
 arch/powerpc/include/asm/ppc_asm.h          |    4 +-
 arch/powerpc/kernel/head_32.S               |    2 +-
 arch/powerpc/kernel/head_40x.S              |    2 +-
 arch/powerpc/kernel/head_44x.S              |    2 +-
 arch/powerpc/kernel/head_8xx.S              |    2 +-
 arch/powerpc/kernel/head_fsl_booke.S        |    2 +-
 arch/powerpc/kernel/init_task.c             |    2 +-
 arch/powerpc/kernel/machine_kexec_64.c      |    2 +-
 arch/powerpc/kernel/vdso.c                  |    2 +-
 arch/powerpc/kernel/vdso32/vdso32_wrapper.S |    2 +-
 arch/powerpc/kernel/vdso64/vdso64_wrapper.S |    2 +-
 arch/powerpc/kernel/vmlinux.lds.S           |   22 ++--
 arch/s390/include/asm/cache.h               |    2 +-
 arch/s390/kernel/head.S                     |    2 +-
 arch/s390/kernel/init_task.c                |    2 +-
 arch/s390/kernel/vdso.c                     |    2 +-
 arch/s390/kernel/vdso32/vdso32_wrapper.S    |    2 +-
 arch/s390/kernel/vdso64/vdso64_wrapper.S    |    2 +-
 arch/s390/kernel/vmlinux.lds.S              |   20 ++--
 arch/sh/include/asm/cache.h                 |    2 +-
 arch/sh/kernel/cpu/sh5/entry.S              |    4 +-
 arch/sh/kernel/head_32.S                    |    2 +-
 arch/sh/kernel/head_64.S                    |    2 +-
 arch/sh/kernel/init_task.c                  |    2 +-
 arch/sh/kernel/irq.c                        |    4 +-
 arch/sh/kernel/vmlinux_32.lds.S             |   14 ++--
 arch/sh/kernel/vmlinux_64.lds.S             |   14 ++--
 arch/sparc/boot/btfixupprep.c               |    4 +-
 arch/sparc/include/asm/cache.h              |    2 +-
 arch/sparc/kernel/head_32.S                 |    4 +-
 arch/sparc/kernel/head_64.S                 |    2 +-
 arch/sparc/kernel/init_task.c               |    2 +-
 arch/sparc/kernel/vmlinux.lds.S             |   14 ++--
 arch/um/include/asm/common.lds.S            |    4 +-
 arch/um/kernel/dyn.lds.S                    |    4 +-
 arch/um/kernel/init_task.c                  |    4 +-
 arch/um/kernel/uml.lds.S                    |    4 +-
 arch/x86/Kconfig                            |    1 +
 arch/x86/boot/compressed/head_32.S          |    2 +-
 arch/x86/boot/compressed/head_64.S          |    2 +-
 arch/x86/boot/compressed/relocs.c           |    2 +-
 arch/x86/boot/compressed/vmlinux.scr        |    2 +-
 arch/x86/boot/compressed/vmlinux_32.lds     |   14 ++-
 arch/x86/boot/compressed/vmlinux_64.lds     |   10 +-
 arch/x86/include/asm/cache.h                |    4 +-
 arch/x86/kernel/acpi/wakeup_32.S            |    2 +-
 arch/x86/kernel/head_32.S                   |    6 +-
 arch/x86/kernel/head_64.S                   |    4 +-
 arch/x86/kernel/init_task.c                 |    4 +-
 arch/x86/kernel/setup_percpu.c              |    2 +-
 arch/x86/kernel/traps.c                     |    2 +-
 arch/x86/kernel/vmlinux_32.lds.S            |   29 +++---
 arch/x86/kernel/vmlinux_64.lds.S            |   27 +++---
 arch/xtensa/kernel/head.S                   |    2 +-
 arch/xtensa/kernel/init_task.c              |    2 +-
 arch/xtensa/kernel/vmlinux.lds.S            |    6 +-
 include/asm-frv/init.h                      |    8 +-
 include/asm-generic/vmlinux.lds.h           |   35 ++++---
 include/linux/cache.h                       |    2 +-
 include/linux/init.h                        |    8 +-
 include/linux/linkage.h                     |    4 +-
 include/linux/percpu.h                      |    2 +-
 include/linux/spinlock.h                    |    2 +-
 kernel/module.c                             |    2 +-
 lib/Kconfig.debug                           |   18 +++
 scripts/mod/file2alias.c                    |    6 +-
 scripts/mod/modpost.c                       |  155 ++++++++++++++++-----------
 scripts/mod/modpost.h                       |   43 ++++++++
 scripts/recordmcount.pl                     |    6 +-
 144 files changed, 513 insertions(+), 407 deletions(-)
author
#1 lore raw

PATCH v2 1/4 Make section names compatible with -ffunction-sections -fdata-sections

The purpose of this patch is to make it possible to build the kernel with "gcc -ffunction-sections -fdata-sections". This is a key step towards being able to compile the kernel using the --gc-sections linker option, which can be used to decrease the vmlinux size by garbage collecting unused functions. Also, Ksplice's 'run-pre matching' process is much simpler if the original kernel was compiled with -ffunction-sections and -fdata-sections.

Currently, the kernel uses a number of "magic" section names such as ".data.nosave" and ".text.head".

The problem is that with -ffunction-sections -fdata-sections, gcc places code like

static void head(...) {...}

in the .text.head section, and code like

static int nosave = 1;

in the .data.nosave section, causing code to be inappropriately placed in the "magic" sections. This patch renames all "magic" section names used by the kernel to use ".." rather than "." as the delimiter between the section prefix (e.g. ".text") and suffix (e.g. "head"), so that ".data.nosave" becomes ".data..nosave", etc. The key property of these names is that there are no collisions between the kernel's "magic" sections and the sections generated by gcc's -ffunction-sections and -fdata-sections options. One can then reference the sections generated by -ffunction-sections using a linker script pattern such as *(.text.[A-Za-z$_]*)

This patch is based on an some earlier patches by Denys Vlasenko. Those earlier patches used section names like ".kernel.text.head". Because these names did not begin with e.g. ".text", the assembler did not automatically use the correct section flags for them, and so one needed to explicitly declare the sections flags when declaring a section in assembly files using code such as:

.section .kernel.text.head, "ax", @progbits

These explicit section flags in assembly files turned out to be a common source of confusion, and so I rewrote the patch with this new section naming scheme suggested by Anders Kaseorg.

---
 Documentation/mutex-design.txt              |    4 +-
 arch/alpha/kernel/head.S                    |    2 +-
 arch/alpha/kernel/init_task.c               |    2 +-
 arch/alpha/kernel/vmlinux.lds.S             |   14 ++++----
 arch/arm/kernel/head-nommu.S                |    2 +-
 arch/arm/kernel/head.S                      |    2 +-
 arch/arm/kernel/init_task.c                 |    2 +-
 arch/arm/kernel/vmlinux.lds.S               |   16 +++++-----
 arch/arm/mm/proc-v6.S                       |    2 +-
 arch/arm/mm/proc-v7.S                       |    2 +-
 arch/arm/mm/tlb-v6.S                        |    2 +-
 arch/arm/mm/tlb-v7.S                        |    2 +-
 arch/avr32/kernel/init_task.c               |    2 +-
 arch/avr32/kernel/vmlinux.lds.S             |    6 ++--
 arch/avr32/mm/init.c                        |    2 +-
 arch/blackfin/kernel/vmlinux.lds.S          |    2 +-
 arch/cris/kernel/process.c                  |    2 +-
 arch/cris/kernel/vmlinux.lds.S              |    2 +-
 arch/frv/kernel/break.S                     |    4 +-
 arch/frv/kernel/entry.S                     |    2 +-
 arch/frv/kernel/head-mmu-fr451.S            |    2 +-
 arch/frv/kernel/head-uc-fr401.S             |    2 +-
 arch/frv/kernel/head-uc-fr451.S             |    2 +-
 arch/frv/kernel/head-uc-fr555.S             |    2 +-
 arch/frv/kernel/head.S                      |    4 +-
 arch/frv/kernel/init_task.c                 |    2 +-
 arch/frv/kernel/vmlinux.lds.S               |   18 +++++-----
 arch/frv/mm/tlb-miss.S                      |    2 +-
 arch/h8300/boot/compressed/head.S           |    2 +-
 arch/h8300/boot/compressed/vmlinux.lds      |    2 +-
 arch/h8300/kernel/init_task.c               |    2 +-
 arch/h8300/kernel/vmlinux.lds.S             |    2 +-
 arch/ia64/include/asm/asmmacro.h            |   12 ++++----
 arch/ia64/include/asm/cache.h               |    2 +-
 arch/ia64/include/asm/percpu.h              |    2 +-
 arch/ia64/kernel/Makefile.gate              |    2 +-
 arch/ia64/kernel/gate-data.S                |    2 +-
 arch/ia64/kernel/gate.S                     |    8 ++--
 arch/ia64/kernel/gate.lds.S                 |   10 +++---
 arch/ia64/kernel/head.S                     |    2 +-
 arch/ia64/kernel/init_task.c                |    4 +-
 arch/ia64/kernel/ivt.S                      |    2 +-
 arch/ia64/kernel/minstate.h                 |    4 +-
 arch/ia64/kernel/paravirtentry.S            |    2 +-
 arch/ia64/kernel/vmlinux.lds.S              |   44 +++++++++++++-------------
 arch/ia64/kvm/vmm_ivt.S                     |    2 +-
 arch/ia64/xen/gate-data.S                   |    2 +-
 arch/ia64/xen/xensetup.S                    |    2 +-
 arch/m32r/kernel/head.S                     |    2 +-
 arch/m32r/kernel/init_task.c                |    2 +-
 arch/m32r/kernel/vmlinux.lds.S              |    8 ++--
 arch/m68k/kernel/head.S                     |    2 +-
 arch/m68k/kernel/process.c                  |    2 +-
 arch/m68k/kernel/sun3-head.S                |    2 +-
 arch/m68k/kernel/vmlinux-std.lds            |    6 ++--
 arch/m68k/kernel/vmlinux-sun3.lds           |    4 +-
 arch/m68knommu/kernel/init_task.c           |    2 +-
 arch/m68knommu/kernel/vmlinux.lds.S         |    6 ++--
 arch/m68knommu/platform/68360/head-ram.S    |    2 +-
 arch/m68knommu/platform/68360/head-rom.S    |    2 +-
 arch/mips/kernel/init_task.c                |    2 +-
 arch/mips/kernel/vmlinux.lds.S              |    8 ++--
 arch/mips/lasat/image/head.S                |    2 +-
 arch/mips/lasat/image/romscript.normal      |    2 +-
 arch/mn10300/kernel/head.S                  |    2 +-
 arch/mn10300/kernel/init_task.c             |    2 +-
 arch/mn10300/kernel/vmlinux.lds.S           |   16 +++++-----
 arch/parisc/include/asm/cache.h             |    2 +-
 arch/parisc/include/asm/system.h            |    2 +-
 arch/parisc/kernel/head.S                   |    2 +-
 arch/parisc/kernel/init_task.c              |    8 ++--
 arch/parisc/kernel/vmlinux.lds.S            |   26 ++++++++--------
 arch/powerpc/include/asm/cache.h            |    2 +-
 arch/powerpc/include/asm/page_64.h          |    2 +-
 arch/powerpc/include/asm/ppc_asm.h          |    4 +-
 arch/powerpc/kernel/head_32.S               |    2 +-
 arch/powerpc/kernel/head_40x.S              |    2 +-
 arch/powerpc/kernel/head_44x.S              |    2 +-
 arch/powerpc/kernel/head_8xx.S              |    2 +-
 arch/powerpc/kernel/head_fsl_booke.S        |    2 +-
 arch/powerpc/kernel/init_task.c             |    2 +-
 arch/powerpc/kernel/machine_kexec_64.c      |    2 +-
 arch/powerpc/kernel/vdso.c                  |    2 +-
 arch/powerpc/kernel/vdso32/vdso32_wrapper.S |    2 +-
 arch/powerpc/kernel/vdso64/vdso64_wrapper.S |    2 +-
 arch/powerpc/kernel/vmlinux.lds.S           |   22 +++++++-------
 arch/s390/include/asm/cache.h               |    2 +-
 arch/s390/kernel/head.S                     |    2 +-
 arch/s390/kernel/init_task.c                |    2 +-
 arch/s390/kernel/vdso.c                     |    2 +-
 arch/s390/kernel/vdso32/vdso32_wrapper.S    |    2 +-
 arch/s390/kernel/vdso64/vdso64_wrapper.S    |    2 +-
 arch/s390/kernel/vmlinux.lds.S              |   20 ++++++------
 arch/sh/include/asm/cache.h                 |    2 +-
 arch/sh/kernel/cpu/sh5/entry.S              |    4 +-
 arch/sh/kernel/head_32.S                    |    2 +-
 arch/sh/kernel/head_64.S                    |    2 +-
 arch/sh/kernel/init_task.c                  |    2 +-
 arch/sh/kernel/irq.c                        |    4 +-
 arch/sh/kernel/vmlinux_32.lds.S             |   14 ++++----
 arch/sh/kernel/vmlinux_64.lds.S             |   14 ++++----
 arch/sparc/boot/btfixupprep.c               |    4 +-
 arch/sparc/include/asm/cache.h              |    2 +-
 arch/sparc/kernel/head_32.S                 |    4 +-
 arch/sparc/kernel/head_64.S                 |    2 +-
 arch/sparc/kernel/init_task.c               |    2 +-
 arch/sparc/kernel/vmlinux.lds.S             |   14 ++++----
 arch/um/include/asm/common.lds.S            |    4 +-
 arch/um/kernel/dyn.lds.S                    |    4 +-
 arch/um/kernel/init_task.c                  |    4 +-
 arch/um/kernel/uml.lds.S                    |    4 +-
 arch/x86/boot/compressed/head_32.S          |    2 +-
 arch/x86/boot/compressed/head_64.S          |    2 +-
 arch/x86/boot/compressed/relocs.c           |    2 +-
 arch/x86/boot/compressed/vmlinux.scr        |    2 +-
 arch/x86/boot/compressed/vmlinux_32.lds     |   14 ++++++--
 arch/x86/boot/compressed/vmlinux_64.lds     |   10 ++++--
 arch/x86/include/asm/cache.h                |    4 +-
 arch/x86/kernel/acpi/wakeup_32.S            |    2 +-
 arch/x86/kernel/head_32.S                   |    6 ++--
 arch/x86/kernel/head_64.S                   |    4 +-
 arch/x86/kernel/init_task.c                 |    4 +-
 arch/x86/kernel/setup_percpu.c              |    2 +-
 arch/x86/kernel/traps.c                     |    2 +-
 arch/x86/kernel/vmlinux_32.lds.S            |   28 ++++++++--------
 arch/x86/kernel/vmlinux_64.lds.S            |   26 ++++++++--------
 arch/xtensa/kernel/head.S                   |    2 +-
 arch/xtensa/kernel/init_task.c              |    2 +-
 arch/xtensa/kernel/vmlinux.lds.S            |    6 ++--
 include/asm-frv/init.h                      |    8 ++--
 include/asm-generic/vmlinux.lds.h           |   30 +++++++++---------
 include/linux/cache.h                       |    2 +-
 include/linux/init.h                        |    8 ++--
 include/linux/linkage.h                     |    4 +-
 include/linux/percpu.h                      |    2 +-
 include/linux/spinlock.h                    |    2 +-
 kernel/module.c                             |    2 +-
 scripts/mod/modpost.c                       |   12 ++++----
 scripts/recordmcount.pl                     |    6 ++--
 139 files changed, 353 insertions(+), 345 deletions(-)
Documentation/mutex-design.txt +2 -2
index aa60d1f..c91ccc0 100644
@@ -66,14 +66,14 @@ of advantages of mutexes:
 
     c0377ccb <mutex_lock>:
     c0377ccb:       f0 ff 08                lock decl (%eax)
-    c0377cce:       78 0e                   js     c0377cde <.text.lock.mutex>
+    c0377cce:       78 0e                   js     c0377cde <.text..lock.mutex>
     c0377cd0:       c3                      ret
 
    the unlocking fastpath is equally tight:
 
     c0377cd1 <mutex_unlock>:
     c0377cd1:       f0 ff 00                lock incl (%eax)
-    c0377cd4:       7e 0f                   jle    c0377ce5 <.text.lock.mutex+0x7>
+    c0377cd4:       7e 0f                   jle    c0377ce5 <.text..lock.mutex+0x7>
     c0377cd6:       c3                      ret
 
  - 'struct mutex' semantics are well-defined and are enforced if
arch/alpha/kernel/head.S +1 -1
index 7ac1f13..16293d4 100644
@@ -10,7 +10,7 @@
 #include <asm/system.h>
 #include <asm/asm-offsets.h>
 
-.section .text.head, "ax"
+.section .text..head, "ax"
 .globl swapper_pg_dir
 .globl _stext
 swapper_pg_dir=SWAPPER_PGD
arch/alpha/kernel/init_task.c +1 -1
index c2938e5..7929755 100644
@@ -17,5 +17,5 @@ EXPORT_SYMBOL(init_mm);
 EXPORT_SYMBOL(init_task);
 
 union thread_union init_thread_union
-	__attribute__((section(".data.init_thread")))
+	__attribute__((section(".data..init_thread")))
 	= { INIT_THREAD_INFO(init_task) };
arch/alpha/kernel/vmlinux.lds.S +7 -7
index ef37fc1..511f8ca 100644
@@ -16,7 +16,7 @@ SECTIONS
 
 	_text = .;	/* Text and read-only data */
 	.text : {
-	*(.text.head)
+	*(.text..head)
 		TEXT_TEXT
 		SCHED_TEXT
 		LOCK_TEXT
@@ -93,18 +93,18 @@ SECTIONS
 	/* Freed after init ends here */
 
 	/* Note 2 page alignment above.  */
-	.data.init_thread : {
-		*(.data.init_thread)
+	.data..init_thread : {
+		*(.data..init_thread)
 	}
 
 	. = ALIGN(PAGE_SIZE);
-	.data.page_aligned : {
-		*(.data.page_aligned)
+	.data..page_aligned : {
+		*(.data..page_aligned)
 	}
 
 	. = ALIGN(64);
-	.data.cacheline_aligned : {
-		*(.data.cacheline_aligned)
+	.data..cacheline_aligned : {
+		*(.data..cacheline_aligned)
 	}
 
 	_data = .;
arch/arm/kernel/head-nommu.S +1 -1
index cc87e17..fcd93f5 100644
@@ -32,7 +32,7 @@
  * numbers for r1.
  *
  */
-	.section ".text.head", "ax"
+	.section ".text..head", "ax"
 ENTRY(stext)
 	msr	cpsr_c, #PSR_F_BIT | PSR_I_BIT | SVC_MODE @ ensure svc mode
 						@ and irqs disabled
arch/arm/kernel/head.S +1 -1
index 21e17dc..705e759 100644
@@ -74,7 +74,7 @@
  * crap here - that's what the boot loader (or in extreme, well justified
  * circumstances, zImage) is for.
  */
-	.section ".text.head", "ax"
+	.section ".text..head", "ax"
 ENTRY(stext)
 	msr	cpsr_c, #PSR_F_BIT | PSR_I_BIT | SVC_MODE @ ensure svc mode
 						@ and irqs disabled
arch/arm/kernel/init_task.c +1 -1
index e859af3..0b81b8f 100644
@@ -29,7 +29,7 @@ EXPORT_SYMBOL(init_mm);
  * The things we do for performance..
  */
 union thread_union init_thread_union
-	__attribute__((__section__(".data.init_task"))) =
+	__attribute__((__section__(".data..init_task"))) =
 		{ INIT_THREAD_INFO(init_task) };
 
 /*
arch/arm/kernel/vmlinux.lds.S +8 -8
index c90f272..89c97c3 100644
@@ -23,10 +23,10 @@ SECTIONS
 #else
 	. = PAGE_OFFSET + TEXT_OFFSET;
 #endif
-	.text.head : {
+	.text..head : {
 		_stext = .;
 		_sinittext = .;
-		*(.text.head)
+		*(.text..head)
 	}
 
 	.init : {			/* Init code and data		*/
@@ -66,9 +66,9 @@ SECTIONS
 		. = ALIGN(4096);
 		__per_cpu_load = .;
 		__per_cpu_start = .;
-			*(.data.percpu.page_aligned)
-			*(.data.percpu)
-			*(.data.percpu.shared_aligned)
+			*(.data..percpu.page_aligned)
+			*(.data..percpu)
+			*(.data..percpu.shared_aligned)
 		__per_cpu_end = .;
 #ifndef CONFIG_XIP_KERNEL
 		__init_begin = _stext;
@@ -146,7 +146,7 @@ SECTIONS
 		 * first, the init task union, aligned
 		 * to an 8192 byte boundary.
 		 */
-		*(.data.init_task)
+		*(.data..init_task)
 
 #ifdef CONFIG_XIP_KERNEL
 		. = ALIGN(4096);
@@ -158,7 +158,7 @@ SECTIONS
 
 		. = ALIGN(4096);
 		__nosave_begin = .;
-		*(.data.nosave)
+		*(.data..nosave)
 		. = ALIGN(4096);
 		__nosave_end = .;
 
@@ -166,7 +166,7 @@ SECTIONS
 		 * then the cacheline aligned data
 		 */
 		. = ALIGN(32);
-		*(.data.cacheline_aligned)
+		*(.data..cacheline_aligned)
 
 		/*
 		 * The exception fixup table (might need resorting at runtime)
arch/arm/mm/proc-v6.S +1 -1
index f0cc599..39d8dd1 100644
@@ -132,7 +132,7 @@ cpu_v6_name:
 	.asciz	"ARMv6-compatible processor"
 	.align
 
-	.section ".text.init", #alloc, #execinstr
+	.section ".text..init", #alloc, #execinstr
 
 /*
  *	__v6_setup
arch/arm/mm/proc-v7.S +1 -1
index d1ebec4..f03c14c 100644
@@ -153,7 +153,7 @@ cpu_v7_name:
 	.ascii	"ARMv7 Processor"
 	.align
 
-	.section ".text.init", #alloc, #execinstr
+	.section ".text..init", #alloc, #execinstr
 
 /*
  *	__v7_setup
arch/arm/mm/tlb-v6.S +1 -1
index 20f84bb..7e9a62a 100644
@@ -87,7 +87,7 @@ ENTRY(v6wbi_flush_kern_tlb_range)
 	mcr	p15, 0, r2, c7, c5, 4		@ prefetch flush
 	mov	pc, lr
 
-	.section ".text.init", #alloc, #execinstr
+	.section ".text..init", #alloc, #execinstr
 
 	.type	v6wbi_tlb_fns, #object
 ENTRY(v6wbi_tlb_fns)
arch/arm/mm/tlb-v7.S +1 -1
index 24ba510..875a0bd 100644
@@ -80,7 +80,7 @@ ENTRY(v7wbi_flush_kern_tlb_range)
 	mov	pc, lr
 ENDPROC(v7wbi_flush_kern_tlb_range)
 
-	.section ".text.init", #alloc, #execinstr
+	.section ".text..init", #alloc, #execinstr
 
 	.type	v7wbi_tlb_fns, #object
 ENTRY(v7wbi_tlb_fns)
arch/avr32/kernel/init_task.c +1 -1
index 993d56e..4678bc6 100644
@@ -23,7 +23,7 @@ EXPORT_SYMBOL(init_mm);
  * Initial thread structure. Must be aligned on an 8192-byte boundary.
  */
 union thread_union init_thread_union
-	__attribute__((__section__(".data.init_task"))) =
+	__attribute__((__section__(".data..init_task"))) =
 		{ INIT_THREAD_INFO(init_task) };
 
 /*
arch/avr32/kernel/vmlinux.lds.S +3 -3
index 7910d41..5e73a02 100644
@@ -95,15 +95,15 @@ SECTIONS
 		/*
 		 * First, the init task union, aligned to an 8K boundary.
 		 */
-		*(.data.init_task)
+		*(.data..init_task)
 
 		/* Then, the page-aligned data */
 		. = ALIGN(PAGE_SIZE);
-		*(.data.page_aligned)
+		*(.data..page_aligned)
 
 		/* Then, the cacheline aligned data */
 		. = ALIGN(L1_CACHE_BYTES);
-		*(.data.cacheline_aligned)
+		*(.data..cacheline_aligned)
 
 		/* And the rest... */
 		*(.data.rel*)
arch/avr32/mm/init.c +1 -1
index e819fa6..533a011 100644
@@ -24,7 +24,7 @@
 #include <asm/setup.h>
 #include <asm/sections.h>
 
-#define __page_aligned	__attribute__((section(".data.page_aligned")))
+#define __page_aligned	__attribute__((section(".data..page_aligned")))
 
 DEFINE_PER_CPU(struct mmu_gather, mmu_gathers);
 
arch/blackfin/kernel/vmlinux.lds.S +1 -1
index 27952ae..f3ef8f1 100644
@@ -94,7 +94,7 @@ SECTIONS
 		__sdata = .;
 		/* This gets done first, so the glob doesn't suck it in */
 		. = ALIGN(32);
-		*(.data.cacheline_aligned)
+		*(.data..cacheline_aligned)
 
 #if !L1_DATA_A_LENGTH
 		. = ALIGN(32);
arch/cris/kernel/process.c +1 -1
index 4df0b32..0f0b4bd 100644
@@ -50,7 +50,7 @@ EXPORT_SYMBOL(init_mm);
  * "init_task" linker map entry..
  */
 union thread_union init_thread_union 
-	__attribute__((__section__(".data.init_task"))) =
+	__attribute__((__section__(".data..init_task"))) =
 		{ INIT_THREAD_INFO(init_task) };
 
 /*
arch/cris/kernel/vmlinux.lds.S +1 -1
index 0d2adfc..f603bc8 100644
@@ -68,7 +68,7 @@ SECTIONS
 	_edata = . ;
 
 	. = ALIGN(PAGE_SIZE);	/* init_task and stack, must be aligned. */
-	.data.init_task : { *(.data.init_task) }
+	.data..init_task : { *(.data..init_task) }
 
 	. = ALIGN(PAGE_SIZE);		/* Init code and data. */
 	__init_begin = .;
arch/frv/kernel/break.S +2 -2
index bd0bdf9..cbb6958 100644
@@ -21,7 +21,7 @@
 #
 # the break handler has its own stack
 #
-	.section	.bss.stack
+	.section	.bss..stack
 	.globl		__break_user_context
 	.balign		THREAD_SIZE
 __break_stack:
@@ -63,7 +63,7 @@ __break_trace_through_exceptions:
 # entry point for Break Exceptions/Interrupts
 #
 ###############################################################################
-	.section	.text.break
+	.section	.text..break
 	.balign		4
 	.globl		__entry_break
 __entry_break:
arch/frv/kernel/entry.S +1 -1
index 99060ab..fe5479b 100644
@@ -38,7 +38,7 @@
 
 #define nr_syscalls ((syscall_table_size)/4)
 
-	.section	.text.entry
+	.section	.text..entry
 	.balign		4
 
 .macro LEDS val
arch/frv/kernel/head-mmu-fr451.S +1 -1
index c8f210d..5900967 100644
@@ -31,7 +31,7 @@
 #define __400_LCR	0xfe000100
 #define __400_LSBR	0xfe000c00
 
-	.section	.text.init,"ax"
+	.section	.text..init,"ax"
 	.balign		4
 
 ###############################################################################
arch/frv/kernel/head-uc-fr401.S +1 -1
index ee282be..ecba176 100644
@@ -30,7 +30,7 @@
 #define __400_LCR	0xfe000100
 #define __400_LSBR	0xfe000c00
 
-	.section	.text.init,"ax"
+	.section	.text..init,"ax"
 	.balign		4
 
 ###############################################################################
arch/frv/kernel/head-uc-fr451.S +1 -1
index b10d9c8..ed9dd17 100644
@@ -30,7 +30,7 @@
 #define __400_LCR	0xfe000100
 #define __400_LSBR	0xfe000c00
 
-	.section	.text.init,"ax"
+	.section	.text..init,"ax"
 	.balign		4
 
 ###############################################################################
arch/frv/kernel/head-uc-fr555.S +1 -1
index 39937c1..ee46b0b 100644
@@ -29,7 +29,7 @@
 #define __551_LCR	0xfeff1100
 #define __551_LSBR	0xfeff1c00
 
-	.section	.text.init,"ax"
+	.section	.text..init,"ax"
 	.balign		4
 
 ###############################################################################
arch/frv/kernel/head.S +2 -2
index fecf751..35e6391 100644
@@ -27,7 +27,7 @@
 #   command line string
 #
 ###############################################################################
-	.section	.text.head,"ax"
+	.section	.text..head,"ax"
 	.balign		4
 
 	.globl		_boot, __head_reference
@@ -541,7 +541,7 @@ __head_end:
 	.size		_boot, .-_boot
 
 	# provide a point for GDB to place a break
-	.section	.text.start,"ax"
+	.section	.text..start,"ax"
 	.globl		_start
 	.balign		4
 _start:
arch/frv/kernel/init_task.c +1 -1
index 29429a8..f57ec19 100644
@@ -24,7 +24,7 @@ EXPORT_SYMBOL(init_mm);
  * "init_task" linker map entry..
  */
 union thread_union init_thread_union
-	__attribute__((__section__(".data.init_task"))) =
+	__attribute__((__section__(".data..init_task"))) =
 		{ INIT_THREAD_INFO(init_task) };
 
 /*
arch/frv/kernel/vmlinux.lds.S +9 -9
index b95c4ea..a7bcf9b 100644
@@ -26,7 +26,7 @@ SECTIONS
 
   _sinittext = .;
   .init.text : {
-	*(.text.head)
+	*(.text..head)
 #ifndef CONFIG_DEBUG_INFO
 	INIT_TEXT
 	EXIT_TEXT
@@ -71,13 +71,13 @@ SECTIONS
 
   /* put sections together that have massive alignment issues */
   . = ALIGN(THREAD_SIZE);
-  .data.init_task : {
+  .data..init_task : {
 	  /* init task record & stack */
-	  *(.data.init_task)
+	  *(.data..init_task)
   }
 
   . = ALIGN(L1_CACHE_BYTES);
-  .data.cacheline_aligned : { *(.data.cacheline_aligned) }
+  .data..cacheline_aligned : { *(.data..cacheline_aligned) }
 
   .trap : {
 	/* trap table management - read entry-table.S before modifying */
@@ -94,10 +94,10 @@ SECTIONS
   _text = .;
   _stext = .;
   .text : {
-	*(.text.start)
-	*(.text.entry)
-	*(.text.break)
-	*(.text.tlbmiss)
+	*(.text..start)
+	*(.text..entry)
+	*(.text..break)
+	*(.text..tlbmiss)
 	TEXT_TEXT
 	SCHED_TEXT
 	LOCK_TEXT
@@ -152,7 +152,7 @@ SECTIONS
 
   .sbss		: { *(.sbss .sbss.*) }
   .bss		: { *(.bss .bss.*) }
-  .bss.stack	: { *(.bss) }
+  .bss..stack	: { *(.bss) }
 
   __bss_stop = .;
   _end = . ;
arch/frv/mm/tlb-miss.S +1 -1
index 0764348..0f41b41 100644
@@ -16,7 +16,7 @@
 #include <asm/highmem.h>
 #include <asm/spr-regs.h>
 
-	.section	.text.tlbmiss
+	.section	.text..tlbmiss
 	.balign		4
 
 	.globl		__entry_insn_mmu_miss
arch/h8300/boot/compressed/head.S +1 -1
index 985a81a..10e9a2d 100644
@@ -9,7 +9,7 @@
 
 #define SRAM_START 0xff4000
 
-	.section	.text.startup
+	.section	.text..startup
 	.global	startup
 startup:
 	mov.l	#SRAM_START+0x8000, sp
arch/h8300/boot/compressed/vmlinux.lds +1 -1
index 65e2a0d..a0a3a0e 100644
@@ -4,7 +4,7 @@ SECTIONS
         {
         __stext = . ;
 	__text = .;
-	       *(.text.startup)
+	       *(.text..startup)
 	       *(.text)
         __etext = . ;
         }
arch/h8300/kernel/init_task.c +1 -1
index cb5dc55..fb473b1 100644
@@ -36,6 +36,6 @@ EXPORT_SYMBOL(init_task);
  * "init_task" linker map entry..
  */
 union thread_union init_thread_union
-	__attribute__((__section__(".data.init_task"))) =
+	__attribute__((__section__(".data..init_task"))) =
 		{ INIT_THREAD_INFO(init_task) };
 
arch/h8300/kernel/vmlinux.lds.S +1 -1
index 43a87b9..c0e3635 100644
@@ -101,7 +101,7 @@ SECTIONS
 	___data_start = . ;
 
 	. = ALIGN(0x2000) ;
-		*(.data.init_task)
+		*(.data..init_task)
 	. = ALIGN(0x4) ;
 		DATA_DATA
 	. = ALIGN(0x4) ;
arch/ia64/include/asm/asmmacro.h +6 -6
index c1642fd..3ab6d75 100644
@@ -70,12 +70,12 @@ name:
  * path (ivt.S - TLB miss processing) or in places where it might not be
  * safe to use a "tpa" instruction (mca_asm.S - error recovery).
  */
-	.section ".data.patch.vtop", "a"	// declare section & section attributes
+	.section ".data..patch.vtop", "a"	// declare section & section attributes
 	.previous
 
 #define	LOAD_PHYSICAL(pr, reg, obj)		\
 [1:](pr)movl reg = obj;				\
-	.xdata4 ".data.patch.vtop", 1b-.
+	.xdata4 ".data..patch.vtop", 1b-.
 
 /*
  * For now, we always put in the McKinley E9 workaround.  On CPUs that don't need it,
@@ -84,11 +84,11 @@ name:
 #define DO_MCKINLEY_E9_WORKAROUND
 
 #ifdef DO_MCKINLEY_E9_WORKAROUND
-	.section ".data.patch.mckinley_e9", "a"
+	.section ".data..patch.mckinley_e9", "a"
 	.previous
 /* workaround for Itanium 2 Errata 9: */
 # define FSYS_RETURN					\
-	.xdata4 ".data.patch.mckinley_e9", 1f-.;	\
+	.xdata4 ".data..patch.mckinley_e9", 1f-.;	\
 1:{ .mib;						\
 	nop.m 0;					\
 	mov r16=ar.pfs;					\
@@ -107,11 +107,11 @@ name:
  * If physical stack register size is different from DEF_NUM_STACK_REG,
  * dynamically patch the kernel for correct size.
  */
-	.section ".data.patch.phys_stack_reg", "a"
+	.section ".data..patch.phys_stack_reg", "a"
 	.previous
 #define LOAD_PHYS_STACK_REG_SIZE(reg)			\
 [1:]	adds reg=IA64_NUM_PHYS_STACK_REG*8+8,r0;	\
-	.xdata4 ".data.patch.phys_stack_reg", 1b-.
+	.xdata4 ".data..patch.phys_stack_reg", 1b-.
 
 /*
  * Up until early 2004, use of .align within a function caused bad unwind info.
arch/ia64/include/asm/cache.h +1 -1
index e7482bd..988254a 100644
@@ -24,6 +24,6 @@
 # define SMP_CACHE_BYTES	(1 << 3)
 #endif
 
-#define __read_mostly __attribute__((__section__(".data.read_mostly")))
+#define __read_mostly __attribute__((__section__(".data..read_mostly")))
 
 #endif /* _ASM_IA64_CACHE_H */
arch/ia64/include/asm/percpu.h +1 -1
index 30cf465..35d9aeb 100644
@@ -31,7 +31,7 @@ extern void *per_cpu_init(void);
 
 #endif	/* SMP */
 
-#define PER_CPU_BASE_SECTION ".data.percpu"
+#define PER_CPU_BASE_SECTION ".data..percpu"
 
 /*
  * Be extremely careful when taking the address of this variable!  Due to virtual
arch/ia64/kernel/Makefile.gate +1 -1
index 1d87f84..af90125 100644
@@ -21,7 +21,7 @@ GATECFLAGS_gate-syms.o = -r
 $(obj)/gate-syms.o: $(obj)/gate.lds $(obj)/gate.o FORCE
 	$(call if_changed,gate)
 
-# gate-data.o contains the gate DSO image as data in section .data.gate.
+# gate-data.o contains the gate DSO image as data in section .data..gate.
 # We must build gate.so before we can assemble it.
 # Note: kbuild does not track this dependency due to usage of .incbin
 $(obj)/gate-data.o: $(obj)/gate.so
arch/ia64/kernel/gate-data.S +1 -1
index 258c0a3..b3ef1c7 100644
@@ -1,3 +1,3 @@
-	.section .data.gate, "aw"
+	.section .data..gate, "aw"
 
 	.incbin "arch/ia64/kernel/gate.so"
arch/ia64/kernel/gate.S +4 -4
index cf5e0a1..245d3e1 100644
@@ -21,18 +21,18 @@
  * to targets outside the shared object) and to avoid multi-phase kernel builds, we
  * simply create minimalistic "patch lists" in special ELF sections.
  */
-	.section ".data.patch.fsyscall_table", "a"
+	.section ".data..patch.fsyscall_table", "a"
 	.previous
 #define LOAD_FSYSCALL_TABLE(reg)			\
 [1:]	movl reg=0;					\
-	.xdata4 ".data.patch.fsyscall_table", 1b-.
+	.xdata4 ".data..patch.fsyscall_table", 1b-.
 
-	.section ".data.patch.brl_fsys_bubble_down", "a"
+	.section ".data..patch.brl_fsys_bubble_down", "a"
 	.previous
 #define BRL_COND_FSYS_BUBBLE_DOWN(pr)			\
 [1:](pr)brl.cond.sptk 0;				\
 	;;						\
-	.xdata4 ".data.patch.brl_fsys_bubble_down", 1b-.
+	.xdata4 ".data..patch.brl_fsys_bubble_down", 1b-.
 
 GLOBAL_ENTRY(__kernel_syscall_via_break)
 	.prologue
arch/ia64/kernel/gate.lds.S +5 -5
index 88c64ed..d32b085 100644
@@ -33,21 +33,21 @@ SECTIONS
 	 */
 	. = GATE_ADDR + 0x600;
 
-	.data.patch		: {
+	.data..patch		: {
 		__paravirt_start_gate_mckinley_e9_patchlist = .;
-		*(.data.patch.mckinley_e9)
+		*(.data..patch.mckinley_e9)
 		__paravirt_end_gate_mckinley_e9_patchlist = .;
 
 		__paravirt_start_gate_vtop_patchlist = .;
-		*(.data.patch.vtop)
+		*(.data..patch.vtop)
 		__paravirt_end_gate_vtop_patchlist = .;
 
 		__paravirt_start_gate_fsyscall_patchlist = .;
-		*(.data.patch.fsyscall_table)
+		*(.data..patch.fsyscall_table)
 		__paravirt_end_gate_fsyscall_patchlist = .;
 
 		__paravirt_start_gate_brl_fsys_bubble_down_patchlist = .;
-		*(.data.patch.brl_fsys_bubble_down)
+		*(.data..patch.brl_fsys_bubble_down)
 		__paravirt_end_gate_brl_fsys_bubble_down_patchlist = .;
 	}						:readable
 
arch/ia64/kernel/head.S +1 -1
index 23f846d..c0739fc 100644
@@ -181,7 +181,7 @@ swapper_pg_dir:
 halt_msg:
 	stringz "Halting kernel\n"
 
-	.section .text.head,"ax"
+	.section .text..head,"ax"
 
 	.global start_ap
 
arch/ia64/kernel/init_task.c +2 -2
index 5b0e830..8a5028c 100644
@@ -27,7 +27,7 @@ EXPORT_SYMBOL(init_mm);
  * Initial task structure.
  *
  * We need to make sure that this is properly aligned due to the way process stacks are
- * handled. This is done by having a special ".data.init_task" section...
+ * handled. This is done by having a special ".data..init_task" section...
  */
 #define init_thread_info	init_task_mem.s.thread_info
 
@@ -37,7 +37,7 @@ union {
 		struct thread_info thread_info;
 	} s;
 	unsigned long stack[KERNEL_STACK_SIZE/sizeof (unsigned long)];
-} init_task_mem asm ("init_task") __attribute__((section(".data.init_task"))) = {{
+} init_task_mem asm ("init_task") __attribute__((section(".data..init_task"))) = {{
 	.task =		INIT_TASK(init_task_mem.s.task),
 	.thread_info =	INIT_THREAD_INFO(init_task_mem.s.task)
 }};
arch/ia64/kernel/ivt.S +1 -1
index ec9a5fd..0c14512 100644
@@ -83,7 +83,7 @@
 	mov r19=n;;			/* prepare to save predicates */		\
 	br.sptk.many dispatch_to_fault_handler
 
-	.section .text.ivt,"ax"
+	.section .text..ivt,"ax"
 
 	.align 32768	// align on 32KB boundary
 	.global ia64_ivt
arch/ia64/kernel/minstate.h +2 -2
index 292e214..d56753a 100644
@@ -16,7 +16,7 @@
 #define ACCOUNT_SYS_ENTER
 #endif
 
-.section ".data.patch.rse", "a"
+.section ".data..patch.rse", "a"
 .previous
 
 /*
@@ -215,7 +215,7 @@
 (pUStk) extr.u r17=r18,3,6;			\
 (pUStk)	sub r16=r18,r22;			\
 [1:](pKStk)	br.cond.sptk.many 1f;		\
-	.xdata4 ".data.patch.rse",1b-.		\
+	.xdata4 ".data..patch.rse",1b-.		\
 	;;					\
 	cmp.ge p6,p7 = 33,r17;			\
 	;;					\
arch/ia64/kernel/paravirtentry.S +1 -1
index 6158560..92d880c 100644
@@ -28,7 +28,7 @@
 #include "entry.h"
 
 #define DATA8(sym, init_value)			\
-	.pushsection .data.read_mostly ;	\
+	.pushsection .data..read_mostly ;	\
 	.align 8 ;				\
 	.global sym ;				\
 	sym: ;					\
arch/ia64/kernel/vmlinux.lds.S +22 -22

105 lines not on this page · raw message

arch/ia64/kvm/vmm_ivt.S +1 -1
index 3ef1a01..c5ab23b 100644
@@ -104,7 +104,7 @@ GLOBAL_ENTRY(kvm_vmm_panic)
 	br.call.sptk.many b6=vmm_panic_handler;
 END(kvm_vmm_panic)
 
-    .section .text.ivt,"ax"
+    .section .text..ivt,"ax"
 
     .align 32768    // align on 32KB boundary
     .global kvm_ia64_ivt
arch/ia64/xen/gate-data.S +1 -1
index 7d4830a..6f95b6b 100644
@@ -1,3 +1,3 @@
-	.section .data.gate.xen, "aw"
+	.section .data..gate.xen, "aw"
 
 	.incbin "arch/ia64/xen/gate.so"
arch/ia64/xen/xensetup.S +1 -1
index 28fed1f..28def53 100644
@@ -14,7 +14,7 @@
 #include <linux/init.h>
 #include <xen/interface/elfnote.h>
 
-	.section .data.read_mostly
+	.section .data..read_mostly
 	.align 8
 	.global xen_domain_type
 xen_domain_type:
arch/m32r/kernel/head.S +1 -1
index 9091606..2bac669 100644
@@ -23,7 +23,7 @@ __INITDATA
 /*
  * References to members of the boot_cpu_data structure.
  */
-.section .text.head, "ax"
+.section .text..head, "ax"
 	.global	start_kernel
 	.global __bss_start
 	.global _end
arch/m32r/kernel/init_task.c +1 -1
index 016885c..7c24cec 100644
@@ -25,7 +25,7 @@ EXPORT_SYMBOL(init_mm);
  * "init_task" linker map entry..
  */
 union thread_union init_thread_union
-	__attribute__((__section__(".data.init_task"))) =
+	__attribute__((__section__(".data..init_task"))) =
 		{ INIT_THREAD_INFO(init_task) };
 
 /*
arch/m32r/kernel/vmlinux.lds.S +4 -4
index 9db05df..2bcc37e 100644
@@ -27,7 +27,7 @@ SECTIONS
   _text = .;			/* Text and read-only data */
   .boot : { *(.boot) } = 0
   .text : {
-	*(.text.head)
+	*(.text..head)
 	TEXT_TEXT
 	SCHED_TEXT
 	LOCK_TEXT
@@ -57,17 +57,17 @@ SECTIONS
 
   . = ALIGN(4096);
   __nosave_begin = .;
-  .data_nosave : { *(.data.nosave) }
+  .data_nosave : { *(.data..nosave) }
   . = ALIGN(4096);
   __nosave_end = .;
 
   . = ALIGN(32);
-  .data.cacheline_aligned : { *(.data.cacheline_aligned) }
+  .data..cacheline_aligned : { *(.data..cacheline_aligned) }
 
   _edata = .;			/* End of data section */
 
   . = ALIGN(8192);		/* init_task */
-  .data.init_task : { *(.data.init_task) }
+  .data..init_task : { *(.data..init_task) }
 
   /* will be freed after init */
   . = ALIGN(4096);		/* Init code and data */
arch/m68k/kernel/head.S +1 -1
index f513f53..bdb247e 100644
@@ -577,7 +577,7 @@ func_define	putn,1
 #endif
 .endm
 
-.section ".text.head","ax"
+.section ".text..head","ax"
 ENTRY(_stext)
 /*
  * Version numbers of the bootinfo interface
arch/m68k/kernel/process.c +1 -1
index ec37fb5..4b83a2e 100644
@@ -47,7 +47,7 @@ struct mm_struct init_mm = INIT_MM(init_mm);
 EXPORT_SYMBOL(init_mm);
 
 union thread_union init_thread_union
-__attribute__((section(".data.init_task"), aligned(THREAD_SIZE)))
+__attribute__((section(".data..init_task"), aligned(THREAD_SIZE)))
        = { INIT_THREAD_INFO(init_task) };
 
 /* initial task structure */
arch/m68k/kernel/sun3-head.S +1 -1
index aad0159..5a6714b 100644
@@ -29,7 +29,7 @@ kernel_pmd_table:              .skip 0x2000
 .globl kernel_pg_dir
 .equ    kernel_pg_dir,kernel_pmd_table
 
-	.section .text.head
+	.section .text..head
 ENTRY(_stext)
 ENTRY(_start)
 
arch/m68k/kernel/vmlinux-std.lds +3 -3
index f846d4e..35b6da8 100644
@@ -12,7 +12,7 @@ SECTIONS
   . = 0x1000;
   _text = .;			/* Text and read-only data */
   .text : {
-	*(.text.head)
+	*(.text..head)
 	TEXT_TEXT
 	SCHED_TEXT
 	LOCK_TEXT
@@ -35,7 +35,7 @@ SECTIONS
 	}
 
   . = ALIGN(16);
-  .data.cacheline_aligned : { *(.data.cacheline_aligned) }
+  .data..cacheline_aligned : { *(.data..cacheline_aligned) }
 
   .bss : { *(.bss) }		/* BSS */
 
@@ -78,7 +78,7 @@ SECTIONS
   . = ALIGN(8192);
   __init_end = .;
 
-  .data.init_task : { *(.data.init_task) }	/* The initial task and kernel stack */
+  .data..init_task : { *(.data..init_task) }	/* The initial task and kernel stack */
 
   _end = . ;
 
arch/m68k/kernel/vmlinux-sun3.lds +2 -2
index d9368c0..e6ce561 100644
@@ -12,7 +12,7 @@ SECTIONS
   . = 0xE002000;
   _text = .;			/* Text and read-only data */
   .text : {
-	*(.text.head)
+	*(.text..head)
 	TEXT_TEXT
 	SCHED_TEXT
 	LOCK_TEXT
@@ -70,7 +70,7 @@ __init_begin = .;
 #endif
 	. = ALIGN(PAGE_SIZE);
 	__init_end = .;
-	.data.init.task : { *(.data.init_task) }
+	.data..init.task : { *(.data..init_task) }
 
 
   .bss : { *(.bss) }		/* BSS */
arch/m68knommu/kernel/init_task.c +1 -1
index fe282de..1ea36d6 100644
@@ -36,6 +36,6 @@ EXPORT_SYMBOL(init_task);
  * "init_task" linker map entry..
  */
 union thread_union init_thread_union
-	__attribute__((__section__(".data.init_task"))) =
+	__attribute__((__section__(".data..init_task"))) =
 		{ INIT_THREAD_INFO(init_task) };
 
arch/m68knommu/kernel/vmlinux.lds.S +3 -3
index 69ba9b1..bdf9b1c 100644
@@ -55,7 +55,7 @@ SECTIONS {
 	.romvec : {
 		__rom_start = . ;
 		_romvec = .;
-		*(.data.initvect)
+		*(.data..initvect)
 	} > romvec
 #endif
 
@@ -66,7 +66,7 @@ SECTIONS {
 		TEXT_TEXT
 		SCHED_TEXT
 		LOCK_TEXT
-        	*(.text.lock)
+		*(.text..lock)
 
 		. = ALIGN(16);          /* Exception table              */
 		__start___ex_table = .;
@@ -148,7 +148,7 @@ SECTIONS {
 		_sdata = . ;
 		DATA_DATA
 		. = ALIGN(8192) ;
-		*(.data.init_task)
+		*(.data..init_task)
 		_edata = . ;
 	} > DATA
 
arch/m68knommu/platform/68360/head-ram.S +1 -1
index 2ef0624..8eb94fb 100644
@@ -280,7 +280,7 @@ _dprbase:
      * and then overwritten as needed.
      */
  
-.section ".data.initvect","awx"
+.section ".data..initvect","awx"
     .long   RAMEND	/* Reset: Initial Stack Pointer                 - 0.  */
     .long   _start      /* Reset: Initial Program Counter               - 1.  */
     .long   buserr      /* Bus Error                                    - 2.  */
arch/m68knommu/platform/68360/head-rom.S +1 -1
index 62ecf41..97510e5 100644
@@ -291,7 +291,7 @@ _dprbase:
      * and then overwritten as needed.
      */
  
-.section ".data.initvect","awx"
+.section ".data..initvect","awx"
     .long   RAMEND	/* Reset: Initial Stack Pointer                 - 0.  */
     .long   _start      /* Reset: Initial Program Counter               - 1.  */
     .long   buserr      /* Bus Error                                    - 2.  */
arch/mips/kernel/init_task.c +1 -1
index 149cd91..acb71af 100644
@@ -26,7 +26,7 @@ EXPORT_SYMBOL(init_mm);
  * The things we do for performance..
  */
 union thread_union init_thread_union
-	__attribute__((__section__(".data.init_task"),
+	__attribute__((__section__(".data..init_task"),
 	               __aligned__(THREAD_SIZE))) =
 		{ INIT_THREAD_INFO(init_task) };
 
arch/mips/kernel/vmlinux.lds.S +4 -4
index 58738c8..6b302c7 100644
@@ -77,7 +77,7 @@ SECTIONS
 		 * object file alignment.  Using 32768
 		 */
 		. = ALIGN(_PAGE_SIZE);
-		*(.data.init_task)
+		*(.data..init_task)
 
 		DATA_DATA
 		CONSTRUCTORS
@@ -99,14 +99,14 @@ SECTIONS
 	. = ALIGN(_PAGE_SIZE);
 	.data_nosave : {
 		__nosave_begin = .;
-		*(.data.nosave)
+		*(.data..nosave)
 	}
 	. = ALIGN(_PAGE_SIZE);
 	__nosave_end = .;
 
 	. = ALIGN(1 << CONFIG_MIPS_L1_CACHE_SHIFT);
-	.data.cacheline_aligned : {
-		*(.data.cacheline_aligned)
+	.data..cacheline_aligned : {
+		*(.data..cacheline_aligned)
 	}
 	_edata =  .;			/* End of data section */
 
arch/mips/lasat/image/head.S +1 -1
index efb95f2..e0ecda9 100644
@@ -1,7 +1,7 @@
 #include <asm/lasat/head.h>
 
 	.text
-	.section .text.start, "ax"
+	.section .text..start, "ax"
 	.set noreorder
 	.set mips3
 
arch/mips/lasat/image/romscript.normal +1 -1
index 988f8ad..0864c96 100644
@@ -4,7 +4,7 @@ SECTIONS
 {
   .text :
   {
-    *(.text.start)
+    *(.text..start)
   }
 
   /* Data in ROM */
arch/mn10300/kernel/head.S +1 -1
index 606bd8c..dd0db5f 100644
@@ -19,7 +19,7 @@
 #include <asm/param.h>
 #include <asm/unit/serial.h>
 
-	.section .text.head,"ax"
+	.section .text..head,"ax"
 
 ###############################################################################
 #
arch/mn10300/kernel/init_task.c +1 -1
index 5ac3566..599840f 100644
@@ -31,7 +31,7 @@ EXPORT_SYMBOL(init_mm);
  * "init_task" linker map entry..
  */
 union thread_union init_thread_union
-	__attribute__((__section__(".data.init_task"))) =
+	__attribute__((__section__(".data..init_task"))) =
 		{ INIT_THREAD_INFO(init_task) };
 
 /*
arch/mn10300/kernel/vmlinux.lds.S +8 -8
index b825966..938792a 100644
@@ -28,7 +28,7 @@ SECTIONS
   _text = .;			/* Text and read-only data */
   .text : {
 	*(
-	.text.head
+	.text..head
 	.text
 	)
 	TEXT_TEXT
@@ -58,25 +58,25 @@ SECTIONS
 
   . = ALIGN(PAGE_SIZE);
   __nosave_begin = .;
-  .data_nosave : { *(.data.nosave) }
+  .data_nosave : { *(.data..nosave) }
   . = ALIGN(PAGE_SIZE);
   __nosave_end = .;
 
   . = ALIGN(PAGE_SIZE);
-  .data.page_aligned : { *(.data.idt) }
+  .data..page_aligned : { *(.data..idt) }
 
   . = ALIGN(32);
-  .data.cacheline_aligned : { *(.data.cacheline_aligned) }
+  .data..cacheline_aligned : { *(.data..cacheline_aligned) }
 
   /* rarely changed data like cpu maps */
   . = ALIGN(32);
-  .data.read_mostly : AT(ADDR(.data.read_mostly)) {
-	*(.data.read_mostly)
+  .data..read_mostly : AT(ADDR(.data..read_mostly)) {
+	*(.data..read_mostly)
 	_edata = .;		/* End of data section */
   }
 
   . = ALIGN(THREAD_SIZE);	/* init_task */
-  .data.init_task : { *(.data.init_task) }
+  .data..init_task : { *(.data..init_task) }
 
   /* might get freed after init */
   . = ALIGN(PAGE_SIZE);
@@ -134,7 +134,7 @@ SECTIONS
 
   __bss_start = .;		/* BSS */
   .bss : {
-	*(.bss.page_aligned)
+	*(.bss..page_aligned)
 	*(.bss)
   }
   . = ALIGN(4);
arch/parisc/include/asm/cache.h +1 -1
index 32c2cca..45effe6 100644
@@ -28,7 +28,7 @@
 
 #define SMP_CACHE_BYTES L1_CACHE_BYTES
 
-#define __read_mostly __attribute__((__section__(".data.read_mostly")))
+#define __read_mostly __attribute__((__section__(".data..read_mostly")))
 
 void parisc_cache_init(void);	/* initializes cache-flushing */
 void disable_sr_hashing_asm(int); /* low level support for above */
arch/parisc/include/asm/system.h +1 -1
index ee80c92..7ccaf29 100644
@@ -174,7 +174,7 @@ static inline void set_eiem(unsigned long val)
 })
 
 #ifdef CONFIG_SMP
-# define __lock_aligned __attribute__((__section__(".data.lock_aligned")))
+# define __lock_aligned __attribute__((__section__(".data..lock_aligned")))
 #endif
 
 #define arch_align_stack(x) (x)
arch/parisc/kernel/head.S +1 -1
index 0e3d9f9..4dbdf0e 100644
@@ -345,7 +345,7 @@ smp_slave_stext:
 ENDPROC(stext)
 
 #ifndef CONFIG_64BIT
-	.section .data.read_mostly
+	.section .data..read_mostly
 
 	.align	4
 	.export	$global$,data
arch/parisc/kernel/init_task.c +4 -4
index 1e25a45..14394c0 100644
@@ -48,7 +48,7 @@ EXPORT_SYMBOL(init_mm);
  * "init_task" linker map entry..
  */
 union thread_union init_thread_union
-	__attribute__((aligned(128))) __attribute__((__section__(".data.init_task"))) =
+	__attribute__((aligned(128))) __attribute__((__section__(".data..init_task"))) =
 		{ INIT_THREAD_INFO(init_task) };
 
 #if PT_NLEVELS == 3
@@ -57,11 +57,11 @@ union thread_union init_thread_union
  * guarantee that global objects will be laid out in memory in the same order 
  * as the order of declaration, so put these in different sections and use
  * the linker script to order them. */
-pmd_t pmd0[PTRS_PER_PMD] __attribute__ ((__section__ (".data.vm0.pmd"), aligned(PAGE_SIZE)));
+pmd_t pmd0[PTRS_PER_PMD] __attribute__ ((__section__ (".data..vm0.pmd"), aligned(PAGE_SIZE)));
 #endif
 
-pgd_t swapper_pg_dir[PTRS_PER_PGD] __attribute__ ((__section__ (".data.vm0.pgd"), aligned(PAGE_SIZE)));
-pte_t pg0[PT_INITIAL * PTRS_PER_PTE] __attribute__ ((__section__ (".data.vm0.pte"), aligned(PAGE_SIZE)));
+pgd_t swapper_pg_dir[PTRS_PER_PGD] __attribute__ ((__section__ (".data..vm0.pgd"), aligned(PAGE_SIZE)));
+pte_t pg0[PT_INITIAL * PTRS_PER_PTE] __attribute__ ((__section__ (".data..vm0.pte"), aligned(PAGE_SIZE)));
 
 /*
  * Initial task structure.
arch/parisc/kernel/vmlinux.lds.S +13 -13
index fd2cc4f..9f5184f 100644
@@ -96,8 +96,8 @@ SECTIONS
 
 	/* rarely changed data like cpu maps */
 	. = ALIGN(16);
-	.data.read_mostly : {
-		*(.data.read_mostly)
+	.data..read_mostly : {
+		*(.data..read_mostly)
 	}
 
 	. = ALIGN(L1_CACHE_BYTES);
@@ -108,14 +108,14 @@ SECTIONS
 	}
 
 	. = ALIGN(L1_CACHE_BYTES);
-	.data.cacheline_aligned : {
-		*(.data.cacheline_aligned)
+	.data..cacheline_aligned : {
+		*(.data..cacheline_aligned)
 	}
 
 	/* PA-RISC locks requires 16-byte alignment */
 	. = ALIGN(16);
-	.data.lock_aligned : {
-		*(.data.lock_aligned)
+	.data..lock_aligned : {
+		*(.data..lock_aligned)
 	}
 
 	/* nosave data is really only used for software suspend...it's here
@@ -124,7 +124,7 @@ SECTIONS
 	. = ALIGN(PAGE_SIZE);
 	__nosave_begin = .;
 	.data_nosave : {
-		*(.data.nosave)
+		*(.data..nosave)
 	}
 	. = ALIGN(PAGE_SIZE);
 	__nosave_end = .;
@@ -136,10 +136,10 @@ SECTIONS
 	__bss_start = .;
 	/* page table entries need to be PAGE_SIZE aligned */
 	. = ALIGN(PAGE_SIZE);
-	.data.vmpages : {
-		*(.data.vm0.pmd)
-		*(.data.vm0.pgd)
-		*(.data.vm0.pte)
+	.data..vmpages : {
+		*(.data..vm0.pmd)
+		*(.data..vm0.pgd)
+		*(.data..vm0.pte)
 	}
 	.bss : {
 		*(.bss)
@@ -151,8 +151,8 @@ SECTIONS
 	/* assembler code expects init_task to be 16k aligned */
 	. = ALIGN(16384);
 	/* init_task */
-	.data.init_task : {
-		*(.data.init_task)
+	.data..init_task : {
+		*(.data..init_task)
 	}
 
 #ifdef CONFIG_64BIT
arch/powerpc/include/asm/cache.h +1 -1
index 81de6eb..3f41ab9 100644
@@ -38,7 +38,7 @@ extern struct ppc64_caches ppc64_caches;
 #endif /* __powerpc64__ && ! __ASSEMBLY__ */
 
 #if !defined(__ASSEMBLY__)
-#define __read_mostly __attribute__((__section__(".data.read_mostly")))
+#define __read_mostly __attribute__((__section__(".data..read_mostly")))
 #endif
 
 #endif /* __KERNEL__ */
arch/powerpc/include/asm/page_64.h +1 -1
index 043bfdf..2307910 100644
@@ -157,7 +157,7 @@ do {						\
 #else
 #define __page_aligned \
 	__attribute__((__aligned__(PAGE_SIZE), \
-		__section__(".data.page_aligned")))
+		__section__(".data..page_aligned")))
 #endif
 
 #define VM_DATA_DEFAULT_FLAGS \
arch/powerpc/include/asm/ppc_asm.h +2 -2
index f59a666..676ed60 100644
@@ -189,7 +189,7 @@ name: \
 GLUE(.,name):
 
 #define _INIT_GLOBAL(name) \
-	.section ".text.init.refok"; \
+	.section ".text..init.refok"; \
 	.align 2 ; \
 	.globl name; \
 	.globl GLUE(.,name); \
@@ -229,7 +229,7 @@ name: \
 GLUE(.,name):
 
 #define _INIT_STATIC(name) \
-	.section ".text.init.refok"; \
+	.section ".text..init.refok"; \
 	.align 2 ; \
 	.section ".opd","aw"; \
 name: \
arch/powerpc/kernel/head_32.S +1 -1
index 54e68c1..5ce5dfa 100644
@@ -50,7 +50,7 @@
 	mtspr	SPRN_DBAT##n##L,RB;	\
 1:
 
-	.section	.text.head, "ax"
+	.section	.text..head, "ax"
 	.stabs	"arch/powerpc/kernel/",N_SO,0,0,0f
 	.stabs	"head_32.S",N_SO,0,0,0f
 0:
arch/powerpc/kernel/head_40x.S +1 -1
index 56d8e5d..de69ecd 100644
@@ -52,7 +52,7 @@
  *
  * This is all going to change RSN when we add bi_recs.......  -- Dan
  */
-	.section	.text.head, "ax"
+	.section	.text..head, "ax"
 _ENTRY(_stext);
 _ENTRY(_start);
 
arch/powerpc/kernel/head_44x.S +1 -1
index b56fecc..ab0b339 100644
@@ -50,7 +50,7 @@
  *   r7 - End of kernel command line string
  *
  */
-	.section	.text.head, "ax"
+	.section	.text..head, "ax"
 _ENTRY(_stext);
 _ENTRY(_start);
 	/*
arch/powerpc/kernel/head_8xx.S +1 -1
index 3c9452d..f6d1e67 100644
@@ -38,7 +38,7 @@
 #else
 #define DO_8xx_CPU6(val, reg)
 #endif
-	.section	.text.head, "ax"
+	.section	.text..head, "ax"
 _ENTRY(_stext);
 _ENTRY(_start);
 
arch/powerpc/kernel/head_fsl_booke.S +1 -1
index 4c22620..123bfec 100644
@@ -53,7 +53,7 @@
  *   r7 - End of kernel command line string
  *
  */
-	.section	.text.head, "ax"
+	.section	.text..head, "ax"
 _ENTRY(_stext);
 _ENTRY(_start);
 	/*
arch/powerpc/kernel/init_task.c +1 -1
index 688b329..1162c3c 100644
@@ -21,7 +21,7 @@ EXPORT_SYMBOL(init_mm);
  * "init_task" linker map entry..
  */
 union thread_union init_thread_union 
-	__attribute__((__section__(".data.init_task"))) =
+	__attribute__((__section__(".data..init_task"))) =
 		{ INIT_THREAD_INFO(init_task) };
 
 /*
arch/powerpc/kernel/machine_kexec_64.c +1 -1
index 49e705f..815a13c 100644
@@ -250,7 +250,7 @@ static void kexec_prepare_cpus(void)
  * current, but that audit has not been performed.
  */
 static union thread_union kexec_stack
-	__attribute__((__section__(".data.init_task"))) = { };
+	__attribute__((__section__(".data..init_task"))) = { };
 
 /* Our assembly helper, in kexec_stub.S */
 extern NORET_TYPE void kexec_sequence(void *newstack, unsigned long start,
arch/powerpc/kernel/vdso.c +1 -1
index ad06d5c..6685af8 100644
@@ -74,7 +74,7 @@ static int vdso_ready;
 static union {
 	struct vdso_data	data;
 	u8			page[PAGE_SIZE];
-} vdso_data_store __attribute__((__section__(".data.page_aligned")));
+} vdso_data_store __attribute__((__section__(".data..page_aligned")));
 struct vdso_data *vdso_data = &vdso_data_store.data;
 
 /* Format of the patch table */
arch/powerpc/kernel/vdso32/vdso32_wrapper.S +1 -1
index 556f0ca..10f61ac 100644
@@ -1,7 +1,7 @@
 #include <linux/init.h>
 #include <asm/page.h>
 
-	.section ".data.page_aligned"
+	.section ".data..page_aligned"
 
 	.globl vdso32_start, vdso32_end
 	.balign PAGE_SIZE
arch/powerpc/kernel/vdso64/vdso64_wrapper.S +1 -1
index 0529cb9..3c1cc59 100644
@@ -1,7 +1,7 @@
 #include <linux/init.h>
 #include <asm/page.h>
 
-	.section ".data.page_aligned"
+	.section ".data..page_aligned"
 
 	.globl vdso64_start, vdso64_end
 	.balign PAGE_SIZE
arch/powerpc/kernel/vmlinux.lds.S +11 -11
index b9ef164..cedaad0 100644
@@ -52,9 +52,9 @@ SECTIONS
 	/* Text and gots */
 	.text : AT(ADDR(.text) - LOAD_OFFSET) {
 		ALIGN_FUNCTION();
-		*(.text.head)
+		*(.text..head)
 		_text = .;
-		*(.text .fixup .text.init.refok .exit.text.refok __ftr_alt_*)
+		*(.text .fixup .text..init.refok .text..exit.refok __ftr_alt_*)
 		SCHED_TEXT
 		LOCK_TEXT
 		KPROBES_TEXT
@@ -254,28 +254,28 @@ SECTIONS
 #else
 	. = ALIGN(16384);
 #endif
-	.data.init_task : AT(ADDR(.data.init_task) - LOAD_OFFSET) {
-		*(.data.init_task)
+	.data..init_task : AT(ADDR(.data..init_task) - LOAD_OFFSET) {
+		*(.data..init_task)
 	}
 
 	. = ALIGN(PAGE_SIZE);
-	.data.page_aligned : AT(ADDR(.data.page_aligned) - LOAD_OFFSET) {
-		*(.data.page_aligned)
+	.data..page_aligned : AT(ADDR(.data..page_aligned) - LOAD_OFFSET) {
+		*(.data..page_aligned)
 	}
 
-	.data.cacheline_aligned : AT(ADDR(.data.cacheline_aligned) - LOAD_OFFSET) {
-		*(.data.cacheline_aligned)
+	.data..cacheline_aligned : AT(ADDR(.data..cacheline_aligned) - LOAD_OFFSET) {
+		*(.data..cacheline_aligned)
 	}
 
 	. = ALIGN(L1_CACHE_BYTES);
-	.data.read_mostly : AT(ADDR(.data.read_mostly) - LOAD_OFFSET) {
-		*(.data.read_mostly)
+	.data..read_mostly : AT(ADDR(.data..read_mostly) - LOAD_OFFSET) {
+		*(.data..read_mostly)
 	}
 
 	. = ALIGN(PAGE_SIZE);
 	.data_nosave : AT(ADDR(.data_nosave) - LOAD_OFFSET) {
 		__nosave_begin = .;
-		*(.data.nosave)
+		*(.data..nosave)
 		. = ALIGN(PAGE_SIZE);
 		__nosave_end = .;
 	}
arch/s390/include/asm/cache.h +1 -1
index 9b86681..24aafa6 100644
@@ -14,6 +14,6 @@
 #define L1_CACHE_BYTES     256
 #define L1_CACHE_SHIFT     8
 
-#define __read_mostly __attribute__((__section__(".data.read_mostly")))
+#define __read_mostly __attribute__((__section__(".data..read_mostly")))
 
 #endif
arch/s390/kernel/head.S +1 -1
index 1046c2c..d0cc437 100644
@@ -35,7 +35,7 @@
 #define ARCH_OFFSET	0
 #endif
 
-.section ".text.head","ax"
+.section ".text..head","ax"
 #ifndef CONFIG_IPL
 	.org   0
 	.long  0x00080000,0x80000000+startup	# Just a restart PSW
arch/s390/kernel/init_task.c +1 -1
index 7db95c0..7cfd9af 100644
@@ -30,7 +30,7 @@ EXPORT_SYMBOL(init_mm);
  * "init_task" linker map entry..
  */
 union thread_union init_thread_union 
-	__attribute__((__section__(".data.init_task"))) =
+	__attribute__((__section__(".data..init_task"))) =
 		{ INIT_THREAD_INFO(init_task) };
 
 /*
arch/s390/kernel/vdso.c +1 -1
index 89b2e7f..e4c0609 100644
@@ -64,7 +64,7 @@ __setup("vdso=", vdso_setup);
 static union {
 	struct vdso_data	data;
 	u8			page[PAGE_SIZE];
-} vdso_data_store __attribute__((__section__(".data.page_aligned")));
+} vdso_data_store __attribute__((__section__(".data..page_aligned")));
 struct vdso_data *vdso_data = &vdso_data_store.data;
 
 /*
arch/s390/kernel/vdso32/vdso32_wrapper.S +1 -1
index 61639a8..025cb77 100644
@@ -1,7 +1,7 @@
 #include <linux/init.h>
 #include <asm/page.h>
 
-	.section ".data.page_aligned"
+	.section ".data..page_aligned"
 
 	.globl vdso32_start, vdso32_end
 	.balign PAGE_SIZE
arch/s390/kernel/vdso64/vdso64_wrapper.S +1 -1
index d8e2ac1..d26126d 100644
@@ -1,7 +1,7 @@
 #include <linux/init.h>
 #include <asm/page.h>
 
-	.section ".data.page_aligned"
+	.section ".data..page_aligned"
 
 	.globl vdso64_start, vdso64_end
 	.balign PAGE_SIZE
arch/s390/kernel/vmlinux.lds.S +10 -10
index 7a2063e..c1f97b1 100644
@@ -29,7 +29,7 @@ SECTIONS
 	. = 0x00000000;
 	.text : {
 	_text = .;		/* Text and read-only data */
-		*(.text.head)
+		*(.text..head)
 	TEXT_TEXT
 		SCHED_TEXT
 		LOCK_TEXT
@@ -66,30 +66,30 @@ SECTIONS
 	. = ALIGN(PAGE_SIZE);
 	.data_nosave : {
 	__nosave_begin = .;
-		*(.data.nosave)
+		*(.data..nosave)
 	}
 	. = ALIGN(PAGE_SIZE);
 	__nosave_end = .;
 
 	. = ALIGN(PAGE_SIZE);
-	.data.page_aligned : {
-		*(.data.idt)
+	.data..page_aligned : {
+		*(.data..idt)
 	}
 
 	. = ALIGN(0x100);
-	.data.cacheline_aligned : {
-		*(.data.cacheline_aligned)
+	.data..cacheline_aligned : {
+		*(.data..cacheline_aligned)
 	}
 
 	. = ALIGN(0x100);
-	.data.read_mostly : {
-		*(.data.read_mostly)
+	.data..read_mostly : {
+		*(.data..read_mostly)
 	}
 	_edata = .;		/* End of data section */
 
 	. = ALIGN(THREAD_SIZE);	/* init_task */
-	.data.init_task : {
-		*(.data.init_task)
+	.data..init_task : {
+		*(.data..init_task)
 	}
 
 	/* will be freed after init */
arch/sh/include/asm/cache.h +1 -1
index 02df18e..455a9a9 100644
@@ -14,7 +14,7 @@
 
 #define L1_CACHE_BYTES		(1 << L1_CACHE_SHIFT)
 
-#define __read_mostly __attribute__((__section__(".data.read_mostly")))
+#define __read_mostly __attribute__((__section__(".data..read_mostly")))
 
 #ifndef __ASSEMBLY__
 struct cache_info {
arch/sh/kernel/cpu/sh5/entry.S +2 -2
index e640c63..433de32 100644
@@ -2058,10 +2058,10 @@ asm_uaccess_end:
 
 
 /*
- * --- .text.init Section
+ * --- .text..init Section
  */
 
-	.section	.text.init, "ax"
+	.section	.text..init, "ax"
 
 /*
  * void trap_init (void)
arch/sh/kernel/head_32.S +1 -1
index 788605f..4065946 100644
@@ -40,7 +40,7 @@ ENTRY(empty_zero_page)
 1:
 	.skip	PAGE_SIZE - empty_zero_page - 1b
 
-	.section	.text.head, "ax"
+	.section	.text..head, "ax"
 
 /*
  * Condition at the entry of _stext:
arch/sh/kernel/head_64.S +1 -1
index 7ccfb99..c3543d1 100644
@@ -110,7 +110,7 @@ empty_bad_pte_table:
 fpu_in_use:	.quad	0
 
 
-	.section	.text.head, "ax"
+	.section	.text..head, "ax"
 	.balign L1_CACHE_BYTES
 /*
  * Condition at the entry of __stext:
arch/sh/kernel/init_task.c +1 -1
index 80c35ff..af29479 100644
@@ -21,7 +21,7 @@ EXPORT_SYMBOL(init_mm);
  * "init_task" linker map entry..
  */
 union thread_union init_thread_union
-	__attribute__((__section__(".data.init_task"))) =
+	__attribute__((__section__(".data..init_task"))) =
 		{ INIT_THREAD_INFO(init_task) };
 
 /*
arch/sh/kernel/irq.c +2 -2
index 3f1372e..63cd9a0 100644
@@ -158,10 +158,10 @@ asmlinkage int do_IRQ(unsigned int irq, struct pt_regs *regs)
 
 #ifdef CONFIG_IRQSTACKS
 static char softirq_stack[NR_CPUS * THREAD_SIZE]
-		__attribute__((__section__(".bss.page_aligned")));
+		__attribute__((__section__(".bss..page_aligned")));
 
 static char hardirq_stack[NR_CPUS * THREAD_SIZE]
-		__attribute__((__section__(".bss.page_aligned")));
+		__attribute__((__section__(".bss..page_aligned")));
 
 /*
  * allocate per-cpu stacks for hardirq and for softirq processing
arch/sh/kernel/vmlinux_32.lds.S +7 -7
index d0b2a71..a7b1be0 100644
@@ -31,7 +31,7 @@ SECTIONS
 	} = 0
 
 	.text : {
-		*(.text.head)
+		*(.text..head)
 		TEXT_TEXT
 		SCHED_TEXT
 		LOCK_TEXT
@@ -61,19 +61,19 @@ SECTIONS
 
 	. = ALIGN(THREAD_SIZE);
 	.data : {			/* Data */
-		*(.data.init_task)
+		*(.data..init_task)
 
 		. = ALIGN(L1_CACHE_BYTES);
-		*(.data.cacheline_aligned)
+		*(.data..cacheline_aligned)
 
 		. = ALIGN(L1_CACHE_BYTES);
-		*(.data.read_mostly)
+		*(.data..read_mostly)
 
 		. = ALIGN(PAGE_SIZE);
-		*(.data.page_aligned)
+		*(.data..page_aligned)
 
 		__nosave_begin = .;
-		*(.data.nosave)
+		*(.data..nosave)
 		. = ALIGN(PAGE_SIZE);
 		__nosave_end = .;
 
@@ -131,7 +131,7 @@ SECTIONS
 	.bss : {
 		__init_end = .;
 		__bss_start = .;		/* BSS */
-		*(.bss.page_aligned)
+		*(.bss..page_aligned)
 		*(.bss)
 		*(COMMON)
 		. = ALIGN(4);
arch/sh/kernel/vmlinux_64.lds.S +7 -7
index 33fa464..c7d5f5b 100644
@@ -42,7 +42,7 @@ SECTIONS
 	} = 0
 
 	.text : C_PHYS(.text) {
-		*(.text.head)
+		*(.text..head)
 		TEXT_TEXT
 		*(.text64)
 		*(.text..SHmedia32)
@@ -70,19 +70,19 @@ SECTIONS
 
 	. = ALIGN(THREAD_SIZE);
 	.data : C_PHYS(.data) {			/* Data */
-		*(.data.init_task)
+		*(.data..init_task)
 
 		. = ALIGN(L1_CACHE_BYTES);
-		*(.data.cacheline_aligned)
+		*(.data..cacheline_aligned)
 
 		. = ALIGN(L1_CACHE_BYTES);
-		*(.data.read_mostly)
+		*(.data..read_mostly)
 
 		. = ALIGN(PAGE_SIZE);
-		*(.data.page_aligned)
+		*(.data..page_aligned)
 
 		__nosave_begin = .;
-		*(.data.nosave)
+		*(.data..nosave)
 		. = ALIGN(PAGE_SIZE);
 		__nosave_end = .;
 
@@ -140,7 +140,7 @@ SECTIONS
 	.bss : C_PHYS(.bss) {
 		__init_end = .;
 		__bss_start = .;		/* BSS */
-		*(.bss.page_aligned)
+		*(.bss..page_aligned)
 		*(.bss)
 		*(COMMON)
 		. = ALIGN(4);
arch/sparc/boot/btfixupprep.c +2 -2
index 52a4208..e899f2a 100644
@@ -171,7 +171,7 @@ main1:
 			}
 		} else if (buffer[nbase+4] != '_')
 			continue;
-		if (!strcmp (sect, ".text.exit"))
+		if (!strcmp (sect, ".text..exit"))
 			continue;
 		if (strcmp (sect, ".text") &&
 		    strcmp (sect, ".init.text") &&
@@ -325,7 +325,7 @@ main1:
 		(*rr)->next = NULL;
 	}
 	printf("! Generated by btfixupprep. Do not edit.\n\n");
-	printf("\t.section\t\".data.init\",#alloc,#write\n\t.align\t4\n\n");
+	printf("\t.section\t\".data..init\",#alloc,#write\n\t.align\t4\n\n");
 	printf("\t.global\t___btfixup_start\n___btfixup_start:\n\n");
 	for (i = 0; i < last; i++) {
 		f = array + i;
arch/sparc/include/asm/cache.h +1 -1
index 41f85ae..2909f0a 100644
@@ -19,7 +19,7 @@
 
 #define SMP_CACHE_BYTES (1 << SMP_CACHE_BYTES_SHIFT)
 
-#define __read_mostly __attribute__((__section__(".data.read_mostly")))
+#define __read_mostly __attribute__((__section__(".data..read_mostly")))
 
 #ifdef CONFIG_SPARC32
 #include <asm/asi.h>
arch/sparc/kernel/head_32.S +2 -2
index f0b4b51..0ca3dc5 100644
@@ -72,7 +72,7 @@ sun4e_notsup:
 	.align 4
 
 	/* The Sparc trap table, bootloader gives us control at _start. */
-	.section .text.head,"ax"
+	.section .text..head,"ax"
 	.globl	start, _stext, _start, __stext
 	.globl  trapbase
 _start:   /* danger danger */
@@ -735,7 +735,7 @@ go_to_highmem:
 		 nop
 
 /* The code above should be at beginning and we have to take care about
- * short jumps, as branching to .text.init section from .text is usually
+ * short jumps, as branching to .text..init section from .text is usually
  * impossible */
 		__INIT
 /* Acquire boot time privileged register values, this will help debugging.
arch/sparc/kernel/head_64.S +1 -1
index 3a1b7bf..709353f 100644
@@ -467,7 +467,7 @@ jump_to_sun4u_init:
 	jmpl    %g2 + %g0, %g0
 	 nop
 
-	.section	.text.init.refok
+	.section	.text..init.refok
 sun4u_init:
 	BRANCH_IF_SUN4V(g1, sun4v_init)
 
arch/sparc/kernel/init_task.c +1 -1
index f28cb82..7f7a468 100644
@@ -22,5 +22,5 @@ EXPORT_SYMBOL(init_task);
  * in etrap.S which assumes it.
  */
 union thread_union init_thread_union
-	__attribute__((section (".data.init_task")))
+	__attribute__((section (".data..init_task")))
 	= { INIT_THREAD_INFO(init_task) };
arch/sparc/kernel/vmlinux.lds.S +7 -7
index 7626708..66eddbe 100644
@@ -41,7 +41,7 @@ SECTIONS
 	.text TEXTSTART :
 	{
 		_text = .;
-		*(.text.head)
+		*(.text..head)
 		TEXT_TEXT
 		SCHED_TEXT
 		LOCK_TEXT
@@ -59,20 +59,20 @@ SECTIONS
 		*(.data1)
 	}
 	. = ALIGN(SMP_CACHE_BYTES);
-	.data.cacheline_aligned : {
-		*(.data.cacheline_aligned)
+	.data..cacheline_aligned : {
+		*(.data..cacheline_aligned)
 	}
 	. = ALIGN(SMP_CACHE_BYTES);
-	.data.read_mostly : {
-		*(.data.read_mostly)
+	.data..read_mostly : {
+		*(.data..read_mostly)
 	}
 	/* End of data section */
 	_edata = .;
 
 	/* init_task */
 	. = ALIGN(THREAD_SIZE);
-	.data.init_task : {
-		*(.data.init_task)
+	.data..init_task : {
+		*(.data..init_task)
 	}
 	.fixup : {
 		__start___fixup = .;
arch/um/include/asm/common.lds.S +2 -2
index cb02486..b623606 100644
@@ -49,9 +49,9 @@
   }
 
   . = ALIGN(32);
-  .data.percpu : {
+  .data..percpu : {
 	__per_cpu_start = . ;
-	*(.data.percpu)
+	*(.data..percpu)
 	__per_cpu_end = . ;
   }
 	
arch/um/kernel/dyn.lds.S +2 -2
index 9975e1a..a415658 100644
@@ -97,9 +97,9 @@ SECTIONS
   .fini_array     : { *(.fini_array) }
   .data           : {
     . = ALIGN(KERNEL_STACK_SIZE);		/* init_task */
-    *(.data.init_task)
+    *(.data..init_task)
     . = ALIGN(KERNEL_STACK_SIZE);
-    *(.data.init_irqstack)
+    *(.data..init_irqstack)
     DATA_DATA
     *(.data.* .gnu.linkonce.d.*)
     SORT(CONSTRUCTORS)
arch/um/kernel/init_task.c +2 -2
index 806d381..eed073e 100644
@@ -34,9 +34,9 @@ EXPORT_SYMBOL(init_task);
  */
 
 union thread_union init_thread_union
-	__attribute__((__section__(".data.init_task"))) =
+	__attribute__((__section__(".data..init_task"))) =
 		{ INIT_THREAD_INFO(init_task) };
 
 union thread_union cpu0_irqstack
-	__attribute__((__section__(".data.init_irqstack"))) =
+	__attribute__((__section__(".data..init_irqstack"))) =
 		{ INIT_THREAD_INFO(init_task) };
arch/um/kernel/uml.lds.S +2 -2
index 11b8352..7b03348 100644
@@ -53,9 +53,9 @@ SECTIONS
   .data    :
   {
     . = ALIGN(KERNEL_STACK_SIZE);		/* init_task */
-    *(.data.init_task)
+    *(.data..init_task)
     . = ALIGN(KERNEL_STACK_SIZE);
-    *(.data.init_irqstack)
+    *(.data..init_irqstack)
     DATA_DATA
     *(.gnu.linkonce.d*)
     CONSTRUCTORS
arch/x86/boot/compressed/head_32.S +1 -1
index 3a8a866..b755fb8 100644
@@ -29,7 +29,7 @@
 #include <asm/boot.h>
 #include <asm/asm-offsets.h>
 
-.section ".text.head","ax",@progbits
+.section ".text..head","ax",@progbits
 ENTRY(startup_32)
 	cld
 	/* test KEEP_SEGMENTS flag to see if the bootloader is asking
arch/x86/boot/compressed/head_64.S +1 -1
index ed4a829..779e06d 100644
@@ -33,7 +33,7 @@
 #include <asm/processor-flags.h>
 #include <asm/asm-offsets.h>
 
-.section ".text.head"
+.section ".text..head"
 	.code32
 ENTRY(startup_32)
 	cld
arch/x86/boot/compressed/relocs.c +1 -1
index 857e492..af8b9c5 100644
@@ -559,7 +559,7 @@ static void emit_relocs(int as_text)
 		/* Print the relocations in a form suitable that
 		 * gas will like.
 		 */
-		printf(".section \".data.reloc\",\"a\"\n");
+		printf(".section \".data..reloc\",\"a\"\n");
 		printf(".balign 4\n");
 		for (i = 0; i < reloc_count; i++) {
 			printf("\t .long 0x%08lx\n", relocs[i]);
arch/x86/boot/compressed/vmlinux.scr +1 -1
index f02382a..862d748 100644
@@ -1,6 +1,6 @@
 SECTIONS
 {
-  .rodata.compressed : {
+  .rodata..compressed : {
 	input_len = .;
 	LONG(input_data_end - input_data) input_data = .;
 	*(.data)
arch/x86/boot/compressed/vmlinux_32.lds +10 -4
index bb3c483..d70318a 100644
@@ -7,13 +7,13 @@ SECTIONS
 	 * address 0.
 	 */
 	. = 0;
-	.text.head : {
+	.text..head : {
 		_head = . ;
-		*(.text.head)
+		*(.text..head)
 		_ehead = . ;
 	}
-	.rodata.compressed : {
-		*(.rodata.compressed)
+	.rodata..compressed : {
+		*(.rodata..compressed)
 	}
 	.text :	{
 		_text = .; 	/* Text */
@@ -21,6 +21,10 @@ SECTIONS
 		*(.text.*)
 		_etext = . ;
 	}
+	.got : {
+		*(.got)
+		*(.got.*)
+	}
 	.rodata : {
 		_rodata = . ;
 		*(.rodata)	 /* read-only data */
@@ -40,4 +44,6 @@ SECTIONS
 		*(COMMON)
 		_end = . ;
 	}
+	/* Be bold, and discard everything not explicitly mentioned */
+	/DISCARD/ : { *(*) }
 }
arch/x86/boot/compressed/vmlinux_64.lds +6 -4
index bef1ac8..d3d1468 100644
@@ -7,13 +7,13 @@ SECTIONS
 	 * address 0.
 	 */
 	. = 0;
-	.text.head : {
+	.text..head : {
 		_head = . ;
-		*(.text.head)
+		*(.text..head)
 		_ehead = . ;
 	}
-	.rodata.compressed : {
-		*(.rodata.compressed)
+	.rodata..compressed : {
+		*(.rodata..compressed)
 	}
 	.text :	{
 		_text = .; 	/* Text */
@@ -45,4 +45,6 @@ SECTIONS
 		. = . + 4096 * 6;
 		_ebss = .;
 	}
+	/* Be bold, and discard everything not explicitly mentioned */
+	/DISCARD/ : { *(*) }
 }
arch/x86/include/asm/cache.h +2 -2
index 5d367ca..6b94d49 100644
@@ -5,7 +5,7 @@
 #define L1_CACHE_SHIFT	(CONFIG_X86_L1_CACHE_SHIFT)
 #define L1_CACHE_BYTES	(1 << L1_CACHE_SHIFT)
 
-#define __read_mostly __attribute__((__section__(".data.read_mostly")))
+#define __read_mostly __attribute__((__section__(".data..read_mostly")))
 
 #ifdef CONFIG_X86_VSMP
 /* vSMP Internode cacheline shift */
@@ -13,7 +13,7 @@
 #ifdef CONFIG_SMP
 #define __cacheline_aligned_in_smp					\
 	__attribute__((__aligned__(1 << (INTERNODE_CACHE_SHIFT))))	\
-	__attribute__((__section__(".data.page_aligned")))
+	__attribute__((__section__(".data..page_aligned")))
 #endif
 #endif
 
arch/x86/kernel/acpi/wakeup_32.S +1 -1
index 8ded418..13ab720 100644
@@ -1,4 +1,4 @@
-	.section .text.page_aligned
+	.section .text..page_aligned
 #include <linux/linkage.h>
 #include <asm/segment.h>
 #include <asm/page_types.h>
arch/x86/kernel/head_32.S +3 -3
index 3068388..4a4bfef 100644
@@ -80,7 +80,7 @@ RESERVE_BRK(pagetables, INIT_MAP_SIZE)
  * any particular GDT layout, because we load our own as soon as we
  * can.
  */
-.section .text.head,"ax",@progbits
+.section .text..head,"ax",@progbits
 ENTRY(startup_32)
 	/* test KEEP_SEGMENTS flag to see if the bootloader is asking
 		us to not reload segments */
@@ -618,7 +618,7 @@ ENTRY(_stext)
 /*
  * BSS section
  */
-.section ".bss.page_aligned","wa"
+.section ".bss..page_aligned","wa"
 	.align PAGE_SIZE_asm
 #ifdef CONFIG_X86_PAE
 swapper_pg_pmd:
@@ -636,7 +636,7 @@ ENTRY(empty_zero_page)
  * This starts the data section.
  */
 #ifdef CONFIG_X86_PAE
-.section ".data.page_aligned","wa"
+.section ".data..page_aligned","wa"
 	/* Page-aligned for the benefit of paravirt? */
 	.align PAGE_SIZE_asm
 ENTRY(swapper_pg_dir)
arch/x86/kernel/head_64.S +2 -2
index 54b29bb..05b8013 100644
@@ -41,7 +41,7 @@ L4_START_KERNEL = pgd_index(__START_KERNEL_map)
 L3_START_KERNEL = pud_index(__START_KERNEL_map)
 
 	.text
-	.section .text.head
+	.section .text..head
 	.code64
 	.globl startup_64
 startup_64:
@@ -419,7 +419,7 @@ ENTRY(phys_base)
 ENTRY(idt_table)
 	.skip IDT_ENTRIES * 16
 
-	.section .bss.page_aligned, "aw", @nobits
+	.section .bss..page_aligned, "aw", @nobits
 	.align PAGE_SIZE
 ENTRY(empty_zero_page)
 	.skip PAGE_SIZE
arch/x86/kernel/init_task.c +2 -2
index df3bf26..103b1c5 100644
@@ -22,7 +22,7 @@ struct mm_struct init_mm = INIT_MM(init_mm);
  * "init_task" linker map entry..
  */
 union thread_union init_thread_union
-	__attribute__((__section__(".data.init_task"))) =
+	__attribute__((__section__(".data..init_task"))) =
 		{ INIT_THREAD_INFO(init_task) };
 
 /*
@@ -36,7 +36,7 @@ EXPORT_SYMBOL(init_task);
 /*
  * per-CPU TSS segments. Threads are completely 'soft' on Linux,
  * no more per-task TSS's. The TSS size is kept cacheline-aligned
- * so they are allowed to end up in the .data.cacheline_aligned
+ * so they are allowed to end up in the .data..cacheline_aligned
  * section. Since TSS's are completely CPU-local, we want them
  * on exact cacheline boundaries, to eliminate cacheline ping-pong.
  */
arch/x86/kernel/setup_percpu.c +1 -1
index 3a97a4c..9548f29 100644
@@ -407,7 +407,7 @@ void __init setup_per_cpu_areas(void)
 #endif
 #endif
 		/*
-		 * Up to this point, the boot CPU has been using .data.init
+		 * Up to this point, the boot CPU has been using .data..init
 		 * area.  Reload any changed state for the boot CPU.
 		 */
 		if (cpu == boot_cpu_id)
arch/x86/kernel/traps.c +1 -1
index a1d2883..2a88259 100644
@@ -77,7 +77,7 @@ char ignore_fpu_irq;
  * for this.
  */
 gate_desc idt_table[256]
-	__attribute__((__section__(".data.idt"))) = { { { { 0, 0 } } }, };
+	__attribute__((__section__(".data..idt"))) = { { { { 0, 0 } } }, };
 #endif
 
 DECLARE_BITMAP(used_vectors, NR_VECTORS);
arch/x86/kernel/vmlinux_32.lds.S +14 -14
index 62ad500..5342ea8 100644
@@ -31,15 +31,15 @@ SECTIONS
   . = LOAD_OFFSET + LOAD_PHYSICAL_ADDR;
   phys_startup_32 = startup_32 - LOAD_OFFSET;
 
-  .text.head : AT(ADDR(.text.head) - LOAD_OFFSET) {
+  .text..head : AT(ADDR(.text..head) - LOAD_OFFSET) {
   	_text = .;			/* Text and read-only data */
-	*(.text.head)
+	*(.text..head)
   } :text = 0x9090
 
   /* read-only */
   .text : AT(ADDR(.text) - LOAD_OFFSET) {
 	. = ALIGN(PAGE_SIZE); /* not really needed, already page aligned */
-	*(.text.page_aligned)
+	*(.text..page_aligned)
 	TEXT_TEXT
 	SCHED_TEXT
 	LOCK_TEXT
@@ -71,32 +71,32 @@ SECTIONS
   . = ALIGN(PAGE_SIZE);
   .data_nosave : AT(ADDR(.data_nosave) - LOAD_OFFSET) {
   	__nosave_begin = .;
-	*(.data.nosave)
+	*(.data..nosave)
   	. = ALIGN(PAGE_SIZE);
   	__nosave_end = .;
   }
 
   . = ALIGN(PAGE_SIZE);
-  .data.page_aligned : AT(ADDR(.data.page_aligned) - LOAD_OFFSET) {
-	*(.data.page_aligned)
-	*(.data.idt)
+  .data..page_aligned : AT(ADDR(.data..page_aligned) - LOAD_OFFSET) {
+	*(.data..page_aligned)
+	*(.data..idt)
   }
 
   . = ALIGN(32);
-  .data.cacheline_aligned : AT(ADDR(.data.cacheline_aligned) - LOAD_OFFSET) {
-	*(.data.cacheline_aligned)
+  .data..cacheline_aligned : AT(ADDR(.data..cacheline_aligned) - LOAD_OFFSET) {
+	*(.data..cacheline_aligned)
   }
 
   /* rarely changed data like cpu maps */
   . = ALIGN(32);
-  .data.read_mostly : AT(ADDR(.data.read_mostly) - LOAD_OFFSET) {
-	*(.data.read_mostly)
+  .data..read_mostly : AT(ADDR(.data..read_mostly) - LOAD_OFFSET) {
+	*(.data..read_mostly)
 	_edata = .;		/* End of data section */
   }
 
   . = ALIGN(THREAD_SIZE);	/* init_task */
-  .data.init_task : AT(ADDR(.data.init_task) - LOAD_OFFSET) {
-	*(.data.init_task)
+  .data..init_task : AT(ADDR(.data..init_task) - LOAD_OFFSET) {
+	*(.data..init_task)
   }
 
   /* might get freed after init */
@@ -185,7 +185,7 @@ SECTIONS
   .bss : AT(ADDR(.bss) - LOAD_OFFSET) {
 	__init_end = .;
 	__bss_start = .;		/* BSS */
-	*(.bss.page_aligned)
+	*(.bss..page_aligned)
 	*(.bss)
 	. = ALIGN(4);
 	__bss_stop = .;
arch/x86/kernel/vmlinux_64.lds.S +13 -13
index c874250..98016a2 100644
@@ -32,7 +32,7 @@ SECTIONS
   .text :  AT(ADDR(.text) - LOAD_OFFSET) {
 	_text = .;			/* Text and read-only data */
 	/* First the code that has to be first for bootstrapping */
-	*(.text.head)
+	*(.text..head)
 	_stext = .;
 	/* Then the rest */
 	TEXT_TEXT
@@ -65,19 +65,19 @@ SECTIONS
 	} :data
 
 
-  .data.cacheline_aligned : AT(ADDR(.data.cacheline_aligned) - LOAD_OFFSET) {
+  .data..cacheline_aligned : AT(ADDR(.data..cacheline_aligned) - LOAD_OFFSET) {
 	. = ALIGN(PAGE_SIZE);
 	. = ALIGN(CONFIG_X86_L1_CACHE_BYTES);
-	*(.data.cacheline_aligned)
+	*(.data..cacheline_aligned)
   }
   . = ALIGN(CONFIG_X86_INTERNODE_CACHE_BYTES);
-  .data.read_mostly : AT(ADDR(.data.read_mostly) - LOAD_OFFSET) {
-  	*(.data.read_mostly)
+  .data..read_mostly : AT(ADDR(.data..read_mostly) - LOAD_OFFSET) {
+	*(.data..read_mostly)
   }
 
 #define VSYSCALL_ADDR (-10*1024*1024)
-#define VSYSCALL_PHYS_ADDR ((LOADADDR(.data.read_mostly) + SIZEOF(.data.read_mostly) + 4095) & ~(4095))
-#define VSYSCALL_VIRT_ADDR ((ADDR(.data.read_mostly) + SIZEOF(.data.read_mostly) + 4095) & ~(4095))
+#define VSYSCALL_PHYS_ADDR ((LOADADDR(.data..read_mostly) + SIZEOF(.data..read_mostly) + 4095) & ~(4095))
+#define VSYSCALL_VIRT_ADDR ((ADDR(.data..read_mostly) + SIZEOF(.data..read_mostly) + 4095) & ~(4095))
 
 #define VLOAD_OFFSET (VSYSCALL_ADDR - VSYSCALL_PHYS_ADDR)
 #define VLOAD(x) (ADDR(x) - VLOAD_OFFSET)
@@ -125,14 +125,14 @@ SECTIONS
 #undef VVIRT_OFFSET
 #undef VVIRT
 
-  .data.init_task : AT(ADDR(.data.init_task) - LOAD_OFFSET) {
+  .data..init_task : AT(ADDR(.data..init_task) - LOAD_OFFSET) {
 	. = ALIGN(THREAD_SIZE);	/* init_task */
-	*(.data.init_task)
+	*(.data..init_task)
   }:data.init
 
-  .data.page_aligned : AT(ADDR(.data.page_aligned) - LOAD_OFFSET) {
+  .data..page_aligned : AT(ADDR(.data..page_aligned) - LOAD_OFFSET) {
 	. = ALIGN(PAGE_SIZE);
-	*(.data.page_aligned)
+	*(.data..page_aligned)
   }
 
   .smp_locks : AT(ADDR(.smp_locks) - LOAD_OFFSET) {
@@ -236,7 +236,7 @@ SECTIONS
   .data_nosave : AT(ADDR(.data_nosave) - LOAD_OFFSET) {
 	. = ALIGN(PAGE_SIZE);
 	__nosave_begin = .;
-	*(.data.nosave)
+	*(.data..nosave)
 	. = ALIGN(PAGE_SIZE);
 	__nosave_end = .;
   } :data.init2 /* use another section data.init2, see PERCPU_VADDR() above */
@@ -244,7 +244,7 @@ SECTIONS
   .bss : AT(ADDR(.bss) - LOAD_OFFSET) {
 	. = ALIGN(PAGE_SIZE);
 	__bss_start = .;		/* BSS */
-	*(.bss.page_aligned)
+	*(.bss..page_aligned)
 	*(.bss)
 	__bss_stop = .;
   }
arch/xtensa/kernel/head.S +1 -1
index 0817f9d..5e093b2 100644
@@ -234,7 +234,7 @@ should_never_return:
  * BSS section
  */
 	
-.section ".bss.page_aligned", "w"
+.section ".bss..page_aligned", "w"
 #ifdef CONFIG_MMU
 ENTRY(swapper_pg_dir)
 	.fill	PAGE_SIZE, 1, 0
arch/xtensa/kernel/init_task.c +1 -1
index e07f5c9..d07db9b 100644
@@ -28,7 +28,7 @@ struct mm_struct init_mm = INIT_MM(init_mm);
 EXPORT_SYMBOL(init_mm);
 
 union thread_union init_thread_union
-	__attribute__((__section__(".data.init_task"))) =
+	__attribute__((__section__(".data..init_task"))) =
 { INIT_THREAD_INFO(init_task) };
 
 struct task_struct init_task = INIT_TASK(init_task);
arch/xtensa/kernel/vmlinux.lds.S +3 -3
index c1be9a4..4c1cfc9 100644
@@ -124,14 +124,14 @@ SECTIONS
     DATA_DATA
     CONSTRUCTORS
     . = ALIGN(XCHAL_ICACHE_LINESIZE);
-    *(.data.cacheline_aligned)
+    *(.data..cacheline_aligned)
   }
 
   _edata = .;
 
   /* The initial task */
   . = ALIGN(8192);
-  .data.init_task : { *(.data.init_task) }
+  .data..init_task : { *(.data..init_task) }
 
   /* Initialization code and data: */
 
@@ -262,7 +262,7 @@ SECTIONS
 
   /* BSS section */
   _bss_start = .;
-  .bss : { *(.bss.page_aligned) *(.bss) }
+  .bss : { *(.bss..page_aligned) *(.bss) }
   _bss_end = .;
 
   _end = .;
include/asm-frv/init.h +4 -4
index 8b15838..4d21473 100644
@@ -1,12 +1,12 @@
 #ifndef _ASM_INIT_H
 #define _ASM_INIT_H
 
-#define __init __attribute__ ((__section__ (".text.init")))
-#define __initdata __attribute__ ((__section__ (".data.init")))
+#define __init __attribute__ ((__section__ (".text..init")))
+#define __initdata __attribute__ ((__section__ (".data..init")))
 /* For assembly routines */
-#define __INIT		.section	".text.init",#alloc,#execinstr
+#define __INIT		.section	".text..init",#alloc,#execinstr
 #define __FINIT		.previous
-#define __INITDATA	.section	".data.init",#alloc,#write
+#define __INITDATA	.section	".data..init",#alloc,#write
 
 #endif
 
include/asm-generic/vmlinux.lds.h +15 -15
index 7fa660f..37feab0 100644
@@ -88,7 +88,7 @@
 /* .data section */
 #define DATA_DATA							\
 	*(.data)							\
-	*(.data.init.refok)						\
+	*(.data..init.refok)						\
 	*(.ref.data)							\
 	DEV_KEEP(init.data)						\
 	DEV_KEEP(exit.data)						\
@@ -287,8 +287,8 @@
 		*(.text.hot)						\
 		*(.text)						\
 		*(.ref.text)						\
-		*(.text.init.refok)					\
-		*(.exit.text.refok)					\
+		*(.text..init.refok)					\
+		*(.text..exit.refok)					\
 	DEV_KEEP(init.text)						\
 	DEV_KEEP(exit.text)						\
 	CPU_KEEP(init.text)						\
@@ -475,16 +475,16 @@
  */
 #define PERCPU_VADDR(vaddr, phdr)					\
 	VMLINUX_SYMBOL(__per_cpu_load) = .;				\
-	.data.percpu vaddr : AT(VMLINUX_SYMBOL(__per_cpu_load)		\
+	.data..percpu vaddr : AT(VMLINUX_SYMBOL(__per_cpu_load)		\
 				- LOAD_OFFSET) {			\
 		VMLINUX_SYMBOL(__per_cpu_start) = .;			\
-		*(.data.percpu.first)					\
-		*(.data.percpu.page_aligned)				\
-		*(.data.percpu)						\
-		*(.data.percpu.shared_aligned)				\
+		*(.data..percpu.first)					\
+		*(.data..percpu.page_aligned)				\
+		*(.data..percpu)						\
+		*(.data..percpu.shared_aligned)				\
 		VMLINUX_SYMBOL(__per_cpu_end) = .;			\
 	} phdr								\
-	. = VMLINUX_SYMBOL(__per_cpu_load) + SIZEOF(.data.percpu);
+	. = VMLINUX_SYMBOL(__per_cpu_load) + SIZEOF(.data..percpu);
 
 /**
  * PERCPU - define output section for percpu area, simple version
@@ -496,17 +496,17 @@
  *
  * This macro is equivalent to ALIGN(align); PERCPU_VADDR( , ) except
  * that __per_cpu_load is defined as a relative symbol against
- * .data.percpu which is required for relocatable x86_32
+ * .data..percpu which is required for relocatable x86_32
  * configuration.
  */
 #define PERCPU(align)							\
 	. = ALIGN(align);						\
-	.data.percpu	: AT(ADDR(.data.percpu) - LOAD_OFFSET) {	\
+	.data..percpu	: AT(ADDR(.data..percpu) - LOAD_OFFSET) {	\
 		VMLINUX_SYMBOL(__per_cpu_load) = .;			\
 		VMLINUX_SYMBOL(__per_cpu_start) = .;			\
-		*(.data.percpu.first)					\
-		*(.data.percpu.page_aligned)				\
-		*(.data.percpu)						\
-		*(.data.percpu.shared_aligned)				\
+		*(.data..percpu.first)					\
+		*(.data..percpu.page_aligned)				\
+		*(.data..percpu)						\
+		*(.data..percpu.shared_aligned)				\
 		VMLINUX_SYMBOL(__per_cpu_end) = .;			\
 	}
include/linux/cache.h +1 -1
index 97e2488..4c57065 100644
@@ -31,7 +31,7 @@
 #ifndef __cacheline_aligned
 #define __cacheline_aligned					\
   __attribute__((__aligned__(SMP_CACHE_BYTES),			\
-		 __section__(".data.cacheline_aligned")))
+		 __section__(".data..cacheline_aligned")))
 #endif /* __cacheline_aligned */
 
 #ifndef __cacheline_aligned_in_smp
include/linux/init.h +4 -4
index 68cb026..f87423d 100644
@@ -62,9 +62,9 @@
 
 /* backward compatibility note
  *  A few places hardcode the old section names:
- *  .text.init.refok
- *  .data.init.refok
- *  .exit.text.refok
+ *  .text..init.refok
+ *  .data..init.refok
+ *  .text..exit.refok
  *  They should be converted to use the defines from this file
  */
 
@@ -301,7 +301,7 @@ void __init parse_early_param(void);
 #endif
 
 /* Data marked not to be saved by software suspend */
-#define __nosavedata __section(.data.nosave)
+#define __nosavedata __section(.data..nosave)
 
 /* This means "can be init if no module support, otherwise module load
    may call it." */
include/linux/linkage.h +2 -2
index fee9e59..bddaee6 100644
@@ -18,8 +18,8 @@
 # define asmregparm
 #endif
 
-#define __page_aligned_data	__section(.data.page_aligned) __aligned(PAGE_SIZE)
-#define __page_aligned_bss	__section(.bss.page_aligned) __aligned(PAGE_SIZE)
+#define __page_aligned_data	__section(.data..page_aligned) __aligned(PAGE_SIZE)
+#define __page_aligned_bss	__section(.bss..page_aligned) __aligned(PAGE_SIZE)
 
 /*
  * This is used by architectures to keep arguments on the stack
include/linux/percpu.h +1 -1
index ee5615d..d208c5b 100644
@@ -11,7 +11,7 @@
 
 #ifndef PER_CPU_BASE_SECTION
 #ifdef CONFIG_SMP
-#define PER_CPU_BASE_SECTION ".data.percpu"
+#define PER_CPU_BASE_SECTION ".data..percpu"
 #else
 #define PER_CPU_BASE_SECTION ".data"
 #endif
include/linux/spinlock.h +1 -1
index 252b245..1a67c69 100644
@@ -60,7 +60,7 @@
 /*
  * Must define these before including other files, inline functions need them
  */
-#define LOCK_SECTION_NAME ".text.lock."KBUILD_BASENAME
+#define LOCK_SECTION_NAME ".text..lock."KBUILD_BASENAME
 
 #define LOCK_SECTION_START(extra)               \
         ".subsection 1\n\t"                     \
kernel/module.c +1 -1
index c268a77..9c0404d 100644
@@ -525,7 +525,7 @@ static unsigned int find_pcpusec(Elf_Ehdr *hdr,
 				 Elf_Shdr *sechdrs,
 				 const char *secstrings)
 {
-	return find_sec(hdr, sechdrs, secstrings, ".data.percpu");
+	return find_sec(hdr, sechdrs, secstrings, ".data..percpu");
 }
 
 static void percpu_modcopy(void *pcpudest, const void *from, unsigned long size)
scripts/mod/modpost.c +6 -6
index 8cc7061..4a3c375 100644
@@ -797,9 +797,9 @@ static const char *data_sections[] = { DATA_SECTIONS, NULL };
 /* sections that may refer to an init/exit section with no warning */
 static const char *initref_sections[] =
 {
-	".text.init.refok*",
-	".exit.text.refok*",
-	".data.init.refok*",
+	".text..init.refok*",
+	".text..exit.refok*",
+	".data..init.refok*",
 	NULL
 };
 
@@ -918,7 +918,7 @@ static int section_mismatch(const char *fromsec, const char *tosec)
  * Pattern 0:
  *   Do not warn if funtion/data are marked with __init_refok/__initdata_refok.
  *   The pattern is identified by:
- *   fromsec = .text.init.refok* | .data.init.refok*
+ *   fromsec = .text..init.refok* | .data..init.refok*
  *
  * Pattern 1:
  *   If a module parameter is declared __initdata and permissions=0
@@ -942,8 +942,8 @@ static int section_mismatch(const char *fromsec, const char *tosec)
  *           *probe_one, *_console, *_timer
  *
  * Pattern 3:
- *   Whitelist all refereces from .text.head to .init.data
- *   Whitelist all refereces from .text.head to .init.text
+ *   Whitelist all references from .text..head to .init.data
+ *   Whitelist all references from .text..head to .init.text
  *
  * Pattern 4:
  *   Some symbols belong to init section but still it is ok to reference
scripts/recordmcount.pl +3 -3
index 409596e..a638a1c 100755
@@ -26,7 +26,7 @@
 # which will also be the location of that section after final link.
 # e.g.
 #
-#  .section ".text.sched"
+#  .section ".text..sched"
 #  .globl my_func
 #  my_func:
 #        [...]
@@ -39,7 +39,7 @@
 #        [...]
 #
 # Both relocation offsets for the mcounts in the above example will be
-# offset from .text.sched. If we make another file called tmp.s with:
+# offset from .text..sched. If we make another file called tmp.s with:
 #
 #  .section __mcount_loc
 #  .quad  my_func + 0x5
@@ -51,7 +51,7 @@
 # But this gets hard if my_func is not globl (a static function).
 # In such a case we have:
 #
-#  .section ".text.sched"
+#  .section ".text..sched"
 #  my_func:
 #        [...]
 #        call mcount  (offset: 0x5)
Signature · 1 line
1.6.2.1
author
#2 lore raw

PATCH v2 2/4 modpost: Check the section flags, not name, to catch missing "ax"/"aw"

From: Anders Kaseorg <[email protected]>

When you put
.section ".foo"
in an assembly file instead of
.section "foo", "ax"
, one of the possible symptoms is that modpost will see an ld-generated section name ".foo.1" in section_rel() or section_rela(). But this heuristic has two problems: it will miss a bad section that has no relocations, and it will incorrectly flag many gcc-generated sections as bad when compiling with -ffunction-sections -fdata-sections.

So instead of checking whether the section name matches a particular pattern, we directly check for a missing SHF_ALLOC in the section flags.

---
 scripts/mod/modpost.c |   49 ++++++++++++++++++-------------------------------
 1 files changed, 18 insertions(+), 31 deletions(-)
scripts/mod/modpost.c +19 -32
index 4a3c375..41c2229 100644
@@ -716,41 +716,27 @@ int match(const char *sym, const char * const pat[])
 
 /* sections that we do not want to do full section mismatch check on */
 static const char *section_white_list[] =
-	{ ".debug*", ".stab*", ".note*", ".got*", ".toc*", NULL };
+	{ ".comment", ".debug*", ".stab*", ".note*", ".got*", ".toc*", NULL };
 
 /*
- * Is this section one we do not want to check?
- * This is often debug sections.
- * If we are going to check this section then
- * test if section name ends with a dot and a number.
- * This is used to find sections where the linker have
- * appended a dot-number to make the name unique.
+ * This is used to find sections missing the SHF_ALLOC flag.
  * The cause of this is often a section specified in assembler
- * without "ax" / "aw" and the same section used in .c
- * code where gcc add these.
+ * without "ax" / "aw".
  */
-static int check_section(const char *modname, const char *sec)
-{
-	const char *e = sec + strlen(sec) - 1;
-	if (match(sec, section_white_list))
-		return 1;
-
-	if (*e && isdigit(*e)) {
-		/* consume all digits */
-		while (*e && e != sec && isdigit(*e))
-			e--;
-		if (*e == '.' && !strstr(sec, ".linkonce")) {
-			warn("%s (%s): unexpected section name.\n"
-			     "The (.[number]+) following section name are "
-			     "ld generated and not expected.\n"
-			     "Did you forget to use \"ax\"/\"aw\" "
-			     "in a .S file?\n"
-			     "Note that for example <linux/init.h> contains\n"
-			     "section definitions for use in .S files.\n\n",
-			     modname, sec);
-		}
+static void check_section(const char *modname, struct elf_info *elf,
+                          Elf_Shdr *sechdr)
+{
+	const char *sec = sech_name(elf, sechdr);
+
+	if (sechdr->sh_type == SHT_PROGBITS &&
+	    !(sechdr->sh_flags & SHF_ALLOC) &&
+	    !match(sec, section_white_list)) {
+		warn("%s (%s): unexpected non-allocatable section.\n"
+		     "Did you forget to use \"ax\"/\"aw\" in a .S file?\n"
+		     "Note that for example <linux/init.h> contains\n"
+		     "section definitions for use in .S files.\n\n",
+		     modname, sec);
 	}
-	return 0;
 }
 
 
@@ -1377,7 +1363,7 @@ static void section_rela(const char *modname, struct elf_info *elf,
 	fromsec = sech_name(elf, sechdr);
 	fromsec += strlen(".rela");
 	/* if from section (name) is know good then skip it */
-	if (check_section(modname, fromsec))
+	if (match(fromsec, section_white_list))
 		return;
 
 	for (rela = start; rela < stop; rela++) {
@@ -1421,7 +1407,7 @@ static void section_rel(const char *modname, struct elf_info *elf,
 	fromsec = sech_name(elf, sechdr);
 	fromsec += strlen(".rel");
 	/* if from section (name) is know good then skip it */
-	if (check_section(modname, fromsec))
+	if (match(fromsec, section_white_list))
 		return;
 
 	for (rel = start; rel < stop; rel++) {
@@ -1484,6 +1470,7 @@ static void check_sec_ref(struct module *mod, const char *modname,
 
 	/* Walk through all sections */
 	for (i = 0; i < elf->hdr->e_shnum; i++) {
+		check_section(modname, elf, &elf->sechdrs[i]);
 		/* We want to process only relocation sections and not .init */
 		if (sechdrs[i].sh_type == SHT_RELA)
 			section_rela(modname, elf, &elf->sechdrs[i]);
Signature · 1 line
1.6.2.1
author
#3 lore raw

PATCH v2 3/4 modpost: Support objects with more than 64k sections

From: Denys Vlasenko <[email protected]>

This patch makes modpost able to process object files with more than 64k sections. Needed for huge kernel builds (allyesconfig, for example) with -ffunction-sections.

---
 scripts/mod/file2alias.c |    6 +-
 scripts/mod/modpost.c    |   94 ++++++++++++++++++++++++++++++++-------------
 scripts/mod/modpost.h    |   43 +++++++++++++++++++++
 3 files changed, 113 insertions(+), 30 deletions(-)
scripts/mod/file2alias.c +3 -3
index a334428..49c669e 100644
@@ -761,16 +761,16 @@ void handle_moddevtable(struct module *mod, struct elf_info *info,
 	char *zeros = NULL;
 
 	/* We're looking for a section relative symbol */
-	if (!sym->st_shndx || sym->st_shndx >= info->hdr->e_shnum)
+	if (!sym->st_shndx || get_secindex(info, sym) >= info->num_sections)
 		return;
 
 	/* Handle all-NULL symbols allocated into .bss */
-	if (info->sechdrs[sym->st_shndx].sh_type & SHT_NOBITS) {
+	if (info->sechdrs[get_secindex(info, sym)].sh_type & SHT_NOBITS) {
 		zeros = calloc(1, sym->st_size);
 		symval = zeros;
 	} else {
 		symval = (void *)info->hdr
-			+ info->sechdrs[sym->st_shndx].sh_offset
+			+ info->sechdrs[get_secindex(info, sym)].sh_offset
 			+ sym->st_value;
 	}
 
scripts/mod/modpost.c +67 -27

211 lines not on this page · raw message

scripts/mod/modpost.h +43 -0

51 lines not on this page · raw message

Signature · 1 line
1.6.2.1
author
#4 lore raw

PATCH v2 4/4 x86: Add an option to compile with -ffunction-sections -fdata-sections

From: Waseem Daher <[email protected]>

This patch makes it possible to link and boot an x86 kernel with -ffunction-sections and -fdata-sections enabled.

Modpost currently warns whenever it sees a section with a name matching [.][0-9]+$ because they are often caused by section flag mismatch errors. When compiling with -ffunction-sections -fdata-sections, gcc places various classes of local symbols in sections with names such as .rodata.__func__.12345, causing these warnings to be printed spuriously. The simplest fix is to disable the warning when CONFIG_FUNCTION_DATA_SECTIONS is enabled.

[[email protected]: modpost support]
[[email protected]: depend on x86, update CONFIG_FUNCTION_TRACER conflict]
---
 Makefile                          |    4 ++++
 arch/x86/Kconfig                  |    1 +
 arch/x86/kernel/vmlinux_32.lds.S  |    1 +
 arch/x86/kernel/vmlinux_64.lds.S  |    1 +
 include/asm-generic/vmlinux.lds.h |    5 ++++-
 lib/Kconfig.debug                 |   18 ++++++++++++++++++
 6 files changed, 29 insertions(+), 1 deletions(-)
Makefile +4 -0
index c6307b6..939104e 100644
@@ -552,6 +552,10 @@ ifdef CONFIG_FUNCTION_TRACER
 KBUILD_CFLAGS	+= -pg
 endif
 
+ifdef CONFIG_FUNCTION_DATA_SECTIONS
+KBUILD_CFLAGS	+= -ffunction-sections -fdata-sections
+endif
+
 # We trigger additional mismatches with less inlining
 ifdef CONFIG_DEBUG_SECTION_MISMATCH
 KBUILD_CFLAGS += $(call cc-option, -fno-inline-functions-called-once)
arch/x86/Kconfig +1 -0
index 5b2196a..032d75d 100644
@@ -28,6 +28,7 @@ config X86
 	select HAVE_KPROBES
 	select ARCH_WANT_OPTIONAL_GPIOLIB
 	select ARCH_WANT_FRAME_POINTERS
+	select HAVE_FUNCTION_DATA_SECTIONS
 	select HAVE_KRETPROBES
 	select HAVE_FTRACE_MCOUNT_RECORD
 	select HAVE_DYNAMIC_FTRACE
arch/x86/kernel/vmlinux_32.lds.S +1 -0
index 5342ea8..21c764e 100644
@@ -187,6 +187,7 @@ SECTIONS
 	__bss_start = .;		/* BSS */
 	*(.bss..page_aligned)
 	*(.bss)
+	*(.bss.[A-Za-z$_]*)		/* handle -fdata-sections */
 	. = ALIGN(4);
 	__bss_stop = .;
   }
arch/x86/kernel/vmlinux_64.lds.S +1 -0
index 98016a2..e7dac7d 100644
@@ -246,6 +246,7 @@ SECTIONS
 	__bss_start = .;		/* BSS */
 	*(.bss..page_aligned)
 	*(.bss)
+	*(.bss.[A-Za-z$_]*)	/* handle -fdata-sections */
 	__bss_stop = .;
   }
 
include/asm-generic/vmlinux.lds.h +4 -1
index 37feab0..70a2a67 100644
@@ -88,6 +88,7 @@
 /* .data section */
 #define DATA_DATA							\
 	*(.data)							\
+	*(.data.[A-Za-z$_]*)	/* handle -fdata-sections */		\
 	*(.data..init.refok)						\
 	*(.ref.data)							\
 	DEV_KEEP(init.data)						\
@@ -119,7 +120,8 @@
 	. = ALIGN((align));						\
 	.rodata           : AT(ADDR(.rodata) - LOAD_OFFSET) {		\
 		VMLINUX_SYMBOL(__start_rodata) = .;			\
-		*(.rodata) *(.rodata.*)					\
+		*(.rodata)						\
+		*(.rodata.[A-Za-z$_]*)	/* handle -fdata-sections */	\
 		*(__vermagic)		/* Kernel version magic */	\
 		*(__markers_strings)	/* Markers: strings */		\
 		*(__tracepoints_strings)/* Tracepoints: strings */	\
@@ -286,6 +288,7 @@
 		ALIGN_FUNCTION();					\
 		*(.text.hot)						\
 		*(.text)						\
+		*(.text.[A-Za-z$_]*)	/* handle -ffunction-sections */\
 		*(.ref.text)						\
 		*(.text..init.refok)					\
 		*(.text..exit.refok)					\
lib/Kconfig.debug +18 -0
index 9638d99..45fe0c2 100644
@@ -591,6 +591,24 @@ config FRAME_POINTER
 	  larger and slower, but it gives very useful debugging information
 	  in case of kernel bugs. (precise oopses/stacktraces/warnings)
 
+config HAVE_FUNCTION_DATA_SECTIONS
+	bool
+
+config FUNCTION_DATA_SECTIONS
+	bool "Compile with -ffunction-sections -fdata-sections"
+	depends on HAVE_FUNCTION_DATA_SECTIONS
+	depends on !FUNCTION_TRACER
+	help
+	  If you say Y here the compiler will give each function
+	  and data structure its own ELF section.
+
+	  This option conflicts with CONFIG_FUNCTION_TRACER, which
+	  enables profiling code generation, because current GCC does
+	  not support compiling with -ffunction-sections -pg (see
+	  <http://gcc.gnu.org/ml/gcc-help/2008-11/msg00128.html>).
+
+	  If unsure, say N.
+
 config BOOT_PRINTK_DELAY
 	bool "Delay each boot printk message by N milliseconds"
 	depends on DEBUG_KERNEL && PRINTK && GENERIC_CALIBRATE_DELAY
Signature · 1 line
1.6.2.1
author
#5 lore raw
In reply to Tim Abbott

Linus,

I see that this -ffunction-sections -fdata-sections compatability patch series didn't get merged for 2.6.30-rc1. Is there a problem with the patches that needs to be addressed? Or are you planning to merge it sometime later in this release cycle (Stephen Rothwell mentioned just after -rc1 as a good time for these patches)?

It still applies to both linux-next and 2.6.30-rc1.

Thanks,
-Tim Abbott

#6 lore raw
Quoted 4 lines · On Wednesday 08 April 2009 10:04:04 Tim Abbott wrote:

Linus,

I see that this -ffunction-sections -fdata-sections compatability patch series didn't get merged for 2.6.30-rc1.

Yep, it'd be good if Linus took this. Pending scattered changes are painful for everyone.

Thanks,
Rusty.

author
#7 lore raw
In reply to Tim Abbott

Is there a problem that is preventing this -ffunction-sections patch series from being merged?

Support for building the kernel with -ffunction-sections is significant progress for decreasing the kernel size on embedded systems using section garbage collection. Support for -ffunction-sections is also the primary change to the kernel needed as a prerequisite for the Ksplice rebootless update system.

It still applies to 2.6.30-rc2 and linux-next.

Thanks,
-Tim Abbott

#8 lore raw
Quoted 3 lines · Tim Abbott

Is there a problem that is preventing this -ffunction-sections patch series from being merged?

Quite frankly, every time I see something like this, I shudder.

Why?

The bugs in the toolchains tend to be infinite and _really_ annoying. We'll end up with everything from "gcc-xyz cannot handle it" to "binutils-abc which was pre-released by RH/Ubuntu/SUSE will SIGSEGV".

And those are the _good_ cases. The bad cases are just silently miscompiled/linked stuff under certain circumstances.

In other words, I'm not going to merge it without a _lot_ of people pushing me on it with nice numbers etc. Because I absolutely hate the tools issues that I'm convinced will happen the moment I merge it.

Linus

author
#9 lore raw
Quoted 2 lines · Linus Torvalds

Because I absolutely hate the tools issues that I'm convinced will happen the moment I merge it.

I assume you're only worried about toolchain problems for people who are actually using the -ffunction-sections option. Would it help if the -ffunction-sections compilation option were marked as experimental until proven otherwise?

If you're not willing to merge even an experimental option for -ffunction-sections, would you at least be willing to merge the first three patches in the patch series? Compiling with -ffunction-sections would not be supported by the mainline kernel, so any toolchain issues with it would not be your problem. But any vendor that wants to take advantage of -ffunction-sections would still be able to use it without having to maintain 300 lines of scattered changes to the kernel.

-Tim Abbott

#10 lore raw
Quoted 5 lines · Tim Abbott

I assume you're only worried about toolchain problems for people who are actually using the -ffunction-sections option. Would it help if the -ffunction-sections compilation option were marked as experimental until proven otherwise?

The thing is, people will enable them, and then maybe the compiler _appears_ to work, and things don't boot, and people spend tons of time chasing down somethign that just turns out to be a tools issue and not a kernel issue at all. And nobody happens to realize that what's up is that the person who reported the regression had enabled an experimental feature.

Quoted 7 lines

If you're not willing to merge even an experimental option for -ffunction-sections, would you at least be willing to merge the first three patches in the patch series? Compiling with -ffunction-sections would not be supported by the mainline kernel, so any toolchain issues with it would not be your problem. But any vendor that wants to take advantage of -ffunction-sections would still be able to use it without having to maintain 300 lines of scattered changes to the kernel.

Are there any advantages outside of the size things?

Do we end up packing data better?

I'd like to have some more champions of this code, in other words.

I'd be ok with merging it, but I haven't really gotten a strong feeling that anybody is going to enable it or use it.

Linus

#11 lore raw
Quoted 10 lines · Linus Torvalds

Are there any advantages outside of the size things?

Do we end up packing data better?

I'd like to have some more champions of this code, in other words.

I'd be ok with merging it, but I haven't really gotten a strong feeling that anybody is going to enable it or use it.

It may be a long haul to get people to actually use it, but there is a fairly major advantage near the end, and that is that we can structure the source code in a way that makes sense for reading it, rather than what will end up using what.

Assuming it does work correctly.

-hpa

Signature · 2 lines
H. Peter Anvin, Intel Open Source Technology Center
I work for Intel.  I don't speak on their behalf.
#12 lore raw
Quoted 12 lines · H. Peter Anvin
Quoted 10 lines · Linus Torvalds

Are there any advantages outside of the size things?

Do we end up packing data better?

I'd like to have some more champions of this code, in other words.

I'd be ok with merging it, but I haven't really gotten a strong feeling that anybody is going to enable it or use it.

Quoted 6 lines

It may be a long haul to get people to actually use it, but there is a fairly major advantage near the end, and that is that we can structure the source code in a way that makes sense for reading it, rather than what will end up using what.

Assuming it does work correctly.

It's not like it's known to be broken. I use it all the time for the stuff I build.
--
vda

#13 lore raw
In reply to Linus Torvalds
Quoted 16 lines · On Mon, 20 Apr 2009 17:35:32 -0700 (PDT) Linus Torvalds…
Quoted 5 lines · Tim Abbott

I assume you're only worried about toolchain problems for people who are actually using the -ffunction-sections option. Would it help if the -ffunction-sections compilation option were marked as experimental until proven otherwise?

The thing is, people will enable them, and then maybe the compiler _appears_ to work, and things don't boot, and people spend tons of time chasing down somethign that just turns out to be a tools issue and not a kernel issue at all. And nobody happens to realize that what's up is that the person who reported the regression had enabled an experimental feature.

Quoted 10 lines
Quoted 8 lines

If you're not willing to merge even an experimental option for -ffunction-sections, would you at least be willing to merge the first three patches in the patch series? Compiling with -ffunction-sections would not be supported by the mainline kernel, so any toolchain issues with it would not be your problem. But any vendor that wants to take advantage of -ffunction-sections would still be able to use it without having to maintain 300 lines of scattered changes to the kernel.

Are there any advantages outside of the size things?

we can probably drop a ton of ifdefs...

an alternative approach would be to have a special build target that uses this, but makes a 2nd vmlinux, and then runs a diff on the symbols against the regular one... and finds which functions got dropped. Some patient soul can then go about and stick in ifdefs and stuff...

(or delete the totally not-used-by-anyone functions entirely)

Signature · 3 lines
Arjan van de Ven 	Intel Open Source Technology Centre
For development, discussion and tips for power savings, 
visit http://www.lesswatts.org
#14 lore raw
Quoted 4 lines · Arjan van de Ven

an alternative approach would be to have a special build target that uses this, but makes a 2nd vmlinux, and then runs a diff on the symbols against the regular one... and finds which functions got dropped. Some patient soul can then go about and stick in ifdefs and stuff...

We wouldn’t even need a 2nd vmlinux for that. Once we can compile with -ffunction-sections -fdata-sections, we can just ask the linker to print the list of unused sections that it removed (--gc-sections --print-gc-sections).

Anders

#15 lore raw
In reply to Arjan van de Ven
Quoted 1 line

we can probably drop a ton of ifdefs...

That would require always forcing it to on.

Quoted 4 lines

an alternative approach would be to have a special build target that uses this, but makes a 2nd vmlinux, and then runs a diff on the symbols against the regular one... and finds which functions got dropped. Some patient soul can then go about and stick in ifdefs and stuff...

I bet that would result in horrible looking code. If the toolchain based approach can be made to work (and from the reports of people who tried it it sounds like it can) that would be vastly preferable.

-Andi

Signature · 1 line
[email protected] -- Speaking for myself only.
author
#16 lore raw
In reply to Linus Torvalds
Quoted 1 line · Linus Torvalds

Are there any advantages outside of the size things?

The main reason I'm interested in these patches is that they are needed by the Ksplice rebootless update system.

The Ksplice developers have been working to make Ksplice mergeable into mainline for about a year now. I think it would be best for everyone if Ksplice were merged before it gets shipped as a feature of a distribution or as a widely deployed standalone product.

Ksplice, Inc., the company behind Ksplice, has been doing deployments of a commercial rebootless update service based on Ksplice for several months now. Once we make a variant of this service freely available for a community Linux distribution later this year, we anticipate that many deployed Linux systems will be running Ksplice. We'd like to be good community members and merge Ksplice before that happens.

We broke off this -ffunction-sections patch series from the main Ksplice patch series because it has significant independent interest and thus could be merged first.

We are not asking you to make a decision at this time on whether or not Ksplice should be merged. However, this -ffunction-sections patch series is essentially the entire portion of the Ksplice patch series that touches the rest of the kernel, so merging this patch series will make it much easier to forward port Ksplice to new kernels or to merge Ksplice in the future.

-Tim Abbott

#17 lore raw
In reply to Linus Torvalds
Quoted 1 line

Are there any advantages outside of the size things?

Yes ksplice needs it too.

Quoted 1 line

Do we end up packing data better?

There used to be tools to do that (use profile feedback to pack code), but I'm not sure they would help the kernel too much because they were mostly aimed at demand paging. But such tools definitely would need this infrastructure. Iirc there are some architectures othat benefit from hold/cold partioning during execution.

Quoted 1 line

I'd like to have some more champions of this code, in other words.

I definitely like the size angle -- letting the linker eliminate code without ifdef jungles certainly sounds attractive to me.

Also if we find broken toolchains I suppose it wouldn't be too difficult to black list them.

-Andi

Signature · 1 line
[email protected] -- Speaking for myself only.
author
#18 lore raw

v3 cover Add support for compiling with -ffunction-sections -fdata-sections

In reply to Linus Torvalds
Quoted 2 lines · Linus Torvalds

I'd be ok with merging it, but I haven't really gotten a strong feeling that anybody is going to enable it or use it.

Linus,

Would you be willing to merge the first 3 patches in this patch series at this point? I'm resending these patches to resolve some conflicts that was recently introduced.

To summarize the discussion, there are three uses for this code that have significant interest.

One is decreasing the kernel size. Denys Vlasenko has numbers that using the linker's --gc-sections option (for which these patches are a prerequisite) saves 10% of the kernel size when CONFIG_MODULES=n, and 1% when CONFIG_MODULES=y. Paul Mundt thinks many embedded users would use this if it were available.

A related use is cleaning up a lot of ifdefs that exist only to decrease the kernel size. Several developers commented favorably on the direction of using section garbage collection for dead code removal.

Finally, Ksplice depends on the option added by this patch series.

The first 3 patches in this patch series fix problems associated with using -ffunction-sections. As there are some architectures currently using -ffunction-sections, these should ideally be merged now, especially the first patch which involves scattered changes.

-Tim Abbott

author
#19 lore raw

PATCH v3 1/3 Make section names compatible with -ffunction-sections -fdata-sections

The purpose of this patch is to make it possible to build the kernel with "gcc -ffunction-sections -fdata-sections". This is a key step towards being able to compile the kernel using the --gc-sections linker option, which can be used to decrease the vmlinux size by garbage collecting unused functions. Also, Ksplice's 'run-pre matching' process is much simpler if the original kernel was compiled with -ffunction-sections and -fdata-sections.

Currently, the kernel uses a number of "magic" section names such as ".data.nosave" and ".text.head".

The problem is that with -ffunction-sections -fdata-sections, gcc places code like

static void head(...) {...}

in the .text.head section, and code like

static int nosave = 1;

in the .data.nosave section, causing code to be inappropriately placed in the "magic" sections. This patch renames all "magic" section names used by the kernel to use ".." rather than "." as the delimiter between the section prefix (e.g. ".text") and suffix (e.g. "head"), so that ".data.nosave" becomes ".data..nosave", etc. The key property of these names is that there are no collisions between the kernel's "magic" sections and the sections generated by gcc's -ffunction-sections and -fdata-sections options. One can then reference the sections generated by -ffunction-sections using a linker script pattern such as *(.text.[A-Za-z$_]*)

This patch is based on an some earlier patches by Denys Vlasenko. Those earlier patches used section names like ".kernel.text.head". Because these names did not begin with e.g. ".text", the assembler did not automatically use the correct section flags for them, and so one needed to explicitly declare the sections flags when declaring a section in assembly files using code such as:

.section .kernel.text.head, "ax", @progbits

These explicit section flags in assembly files turned out to be a common source of confusion, and so I rewrote the patch with this new section naming scheme suggested by Anders Kaseorg.

---
 Documentation/mutex-design.txt              |    4 +-
 arch/alpha/kernel/head.S                    |    2 +-
 arch/alpha/kernel/init_task.c               |    2 +-
 arch/alpha/kernel/vmlinux.lds.S             |   14 ++++----
 arch/arm/kernel/head-nommu.S                |    2 +-
 arch/arm/kernel/head.S                      |    2 +-
 arch/arm/kernel/init_task.c                 |    2 +-
 arch/arm/kernel/vmlinux.lds.S               |   16 +++++-----
 arch/arm/mm/proc-v6.S                       |    2 +-
 arch/arm/mm/proc-v7.S                       |    2 +-
 arch/arm/mm/tlb-v6.S                        |    2 +-
 arch/arm/mm/tlb-v7.S                        |    2 +-
 arch/avr32/kernel/init_task.c               |    2 +-
 arch/avr32/kernel/vmlinux.lds.S             |    6 ++--
 arch/avr32/mm/init.c                        |    2 +-
 arch/blackfin/kernel/vmlinux.lds.S          |    2 +-
 arch/cris/kernel/process.c                  |    2 +-
 arch/cris/kernel/vmlinux.lds.S              |    2 +-
 arch/frv/include/asm/init.h                 |    8 ++--
 arch/frv/kernel/break.S                     |    4 +-
 arch/frv/kernel/entry.S                     |    2 +-
 arch/frv/kernel/head-mmu-fr451.S            |    2 +-
 arch/frv/kernel/head-uc-fr401.S             |    2 +-
 arch/frv/kernel/head-uc-fr451.S             |    2 +-
 arch/frv/kernel/head-uc-fr555.S             |    2 +-
 arch/frv/kernel/head.S                      |    4 +-
 arch/frv/kernel/init_task.c                 |    2 +-
 arch/frv/kernel/vmlinux.lds.S               |   18 +++++-----
 arch/frv/mm/tlb-miss.S                      |    2 +-
 arch/h8300/boot/compressed/head.S           |    2 +-
 arch/h8300/boot/compressed/vmlinux.lds      |    2 +-
 arch/h8300/kernel/init_task.c               |    2 +-
 arch/h8300/kernel/vmlinux.lds.S             |    2 +-
 arch/ia64/include/asm/asmmacro.h            |   12 ++++----
 arch/ia64/include/asm/cache.h               |    2 +-
 arch/ia64/include/asm/percpu.h              |    2 +-
 arch/ia64/kernel/Makefile.gate              |    2 +-
 arch/ia64/kernel/gate-data.S                |    2 +-
 arch/ia64/kernel/gate.S                     |    8 ++--
 arch/ia64/kernel/gate.lds.S                 |   10 +++---
 arch/ia64/kernel/head.S                     |    2 +-
 arch/ia64/kernel/init_task.c                |    4 +-
 arch/ia64/kernel/ivt.S                      |    2 +-
 arch/ia64/kernel/minstate.h                 |    4 +-
 arch/ia64/kernel/paravirtentry.S            |    2 +-
 arch/ia64/kernel/vmlinux.lds.S              |   44 +++++++++++++-------------
 arch/ia64/kvm/vmm_ivt.S                     |    2 +-
 arch/ia64/xen/gate-data.S                   |    2 +-
 arch/ia64/xen/xensetup.S                    |    2 +-
 arch/m32r/kernel/head.S                     |    2 +-
 arch/m32r/kernel/init_task.c                |    2 +-
 arch/m32r/kernel/vmlinux.lds.S              |    8 ++--
 arch/m68k/kernel/head.S                     |    2 +-
 arch/m68k/kernel/process.c                  |    2 +-
 arch/m68k/kernel/sun3-head.S                |    2 +-
 arch/m68k/kernel/vmlinux-std.lds            |    6 ++--
 arch/m68k/kernel/vmlinux-sun3.lds           |    4 +-
 arch/m68knommu/kernel/init_task.c           |    2 +-
 arch/m68knommu/kernel/vmlinux.lds.S         |    6 ++--
 arch/m68knommu/platform/68360/head-ram.S    |    2 +-
 arch/m68knommu/platform/68360/head-rom.S    |    2 +-
 arch/microblaze/kernel/init_task.c          |    2 +-
 arch/microblaze/kernel/vmlinux.lds.S        |    4 +-
 arch/mips/kernel/init_task.c                |    2 +-
 arch/mips/kernel/vmlinux.lds.S              |    8 ++--
 arch/mips/lasat/image/head.S                |    2 +-
 arch/mips/lasat/image/romscript.normal      |    2 +-
 arch/mn10300/kernel/head.S                  |    2 +-
 arch/mn10300/kernel/init_task.c             |    2 +-
 arch/mn10300/kernel/vmlinux.lds.S           |   16 +++++-----
 arch/parisc/include/asm/cache.h             |    2 +-
 arch/parisc/include/asm/system.h            |    2 +-
 arch/parisc/kernel/head.S                   |    2 +-
 arch/parisc/kernel/init_task.c              |    8 ++--
 arch/parisc/kernel/vmlinux.lds.S            |   26 ++++++++--------
 arch/powerpc/include/asm/cache.h            |    2 +-
 arch/powerpc/include/asm/page_64.h          |    2 +-
 arch/powerpc/include/asm/ppc_asm.h          |    4 +-
 arch/powerpc/kernel/head_32.S               |    2 +-
 arch/powerpc/kernel/head_40x.S              |    2 +-
 arch/powerpc/kernel/head_44x.S              |    2 +-
 arch/powerpc/kernel/head_8xx.S              |    2 +-
 arch/powerpc/kernel/head_fsl_booke.S        |    2 +-
 arch/powerpc/kernel/init_task.c             |    2 +-
 arch/powerpc/kernel/machine_kexec_64.c      |    2 +-
 arch/powerpc/kernel/vdso.c                  |    2 +-
 arch/powerpc/kernel/vdso32/vdso32_wrapper.S |    2 +-
 arch/powerpc/kernel/vdso64/vdso64_wrapper.S |    2 +-
 arch/powerpc/kernel/vmlinux.lds.S           |   22 +++++++-------
 arch/s390/include/asm/cache.h               |    2 +-
 arch/s390/kernel/head.S                     |    2 +-
 arch/s390/kernel/init_task.c                |    2 +-
 arch/s390/kernel/vdso.c                     |    2 +-
 arch/s390/kernel/vdso32/vdso32_wrapper.S    |    2 +-
 arch/s390/kernel/vdso64/vdso64_wrapper.S    |    2 +-
 arch/s390/kernel/vmlinux.lds.S              |   20 ++++++------
 arch/sh/include/asm/cache.h                 |    2 +-
 arch/sh/kernel/cpu/sh5/entry.S              |    4 +-
 arch/sh/kernel/head_32.S                    |    2 +-
 arch/sh/kernel/head_64.S                    |    2 +-
 arch/sh/kernel/init_task.c                  |    2 +-
 arch/sh/kernel/irq.c                        |    4 +-
 arch/sh/kernel/vmlinux_32.lds.S             |   14 ++++----
 arch/sh/kernel/vmlinux_64.lds.S             |   14 ++++----
 arch/sparc/boot/btfixupprep.c               |    4 +-
 arch/sparc/include/asm/cache.h              |    2 +-
 arch/sparc/kernel/head_32.S                 |    4 +-
 arch/sparc/kernel/head_64.S                 |    2 +-
 arch/sparc/kernel/init_task.c               |    2 +-
 arch/sparc/kernel/vmlinux.lds.S             |   14 ++++----
 arch/um/include/asm/common.lds.S            |    4 +-
 arch/um/kernel/dyn.lds.S                    |    4 +-
 arch/um/kernel/init_task.c                  |    4 +-
 arch/um/kernel/uml.lds.S                    |    4 +-
 arch/x86/boot/compressed/head_32.S          |    2 +-
 arch/x86/boot/compressed/head_64.S          |    2 +-
 arch/x86/boot/compressed/relocs.c           |    2 +-
 arch/x86/boot/compressed/vmlinux.scr        |    2 +-
 arch/x86/boot/compressed/vmlinux_32.lds     |   14 ++++++--
 arch/x86/boot/compressed/vmlinux_64.lds     |   10 ++++--
 arch/x86/include/asm/cache.h                |    4 +-
 arch/x86/kernel/acpi/wakeup_32.S            |    2 +-
 arch/x86/kernel/head_32.S                   |    6 ++--
 arch/x86/kernel/head_64.S                   |    4 +-
 arch/x86/kernel/init_task.c                 |    4 +-
 arch/x86/kernel/setup_percpu.c              |    2 +-
 arch/x86/kernel/traps.c                     |    2 +-
 arch/x86/kernel/vmlinux_32.lds.S            |   28 ++++++++--------
 arch/x86/kernel/vmlinux_64.lds.S            |   26 ++++++++--------
 arch/xtensa/kernel/head.S                   |    2 +-
 arch/xtensa/kernel/init_task.c              |    2 +-
 arch/xtensa/kernel/vmlinux.lds.S            |    6 ++--
 include/asm-generic/percpu.h                |    2 +-
 include/asm-generic/vmlinux.lds.h           |   30 +++++++++---------
 include/linux/cache.h                       |    2 +-
 include/linux/init.h                        |    8 ++--
 include/linux/linkage.h                     |    4 +-
 include/linux/spinlock.h                    |    2 +-
 kernel/module.c                             |    2 +-
 scripts/mod/modpost.c                       |   12 ++++----
 scripts/recordmcount.pl                     |    6 ++--
 141 files changed, 356 insertions(+), 348 deletions(-)
Documentation/mutex-design.txt +2 -2
index aa60d1f..c91ccc0 100644
@@ -66,14 +66,14 @@ of advantages of mutexes:
 
     c0377ccb <mutex_lock>:
     c0377ccb:       f0 ff 08                lock decl (%eax)
-    c0377cce:       78 0e                   js     c0377cde <.text.lock.mutex>
+    c0377cce:       78 0e                   js     c0377cde <.text..lock.mutex>
     c0377cd0:       c3                      ret
 
    the unlocking fastpath is equally tight:
 
     c0377cd1 <mutex_unlock>:
     c0377cd1:       f0 ff 00                lock incl (%eax)
-    c0377cd4:       7e 0f                   jle    c0377ce5 <.text.lock.mutex+0x7>
+    c0377cd4:       7e 0f                   jle    c0377ce5 <.text..lock.mutex+0x7>
     c0377cd6:       c3                      ret
 
  - 'struct mutex' semantics are well-defined and are enforced if
arch/alpha/kernel/head.S +1 -1
index 7ac1f13..16293d4 100644
@@ -10,7 +10,7 @@
 #include <asm/system.h>
 #include <asm/asm-offsets.h>
 
-.section .text.head, "ax"
+.section .text..head, "ax"
 .globl swapper_pg_dir
 .globl _stext
 swapper_pg_dir=SWAPPER_PGD
arch/alpha/kernel/init_task.c +1 -1
index c2938e5..7929755 100644
@@ -17,5 +17,5 @@ EXPORT_SYMBOL(init_mm);
 EXPORT_SYMBOL(init_task);
 
 union thread_union init_thread_union
-	__attribute__((section(".data.init_thread")))
+	__attribute__((section(".data..init_thread")))
 	= { INIT_THREAD_INFO(init_task) };
arch/alpha/kernel/vmlinux.lds.S +7 -7
index ef37fc1..511f8ca 100644
@@ -16,7 +16,7 @@ SECTIONS
 
 	_text = .;	/* Text and read-only data */
 	.text : {
-	*(.text.head)
+	*(.text..head)
 		TEXT_TEXT
 		SCHED_TEXT
 		LOCK_TEXT
@@ -93,18 +93,18 @@ SECTIONS
 	/* Freed after init ends here */
 
 	/* Note 2 page alignment above.  */
-	.data.init_thread : {
-		*(.data.init_thread)
+	.data..init_thread : {
+		*(.data..init_thread)
 	}
 
 	. = ALIGN(PAGE_SIZE);
-	.data.page_aligned : {
-		*(.data.page_aligned)
+	.data..page_aligned : {
+		*(.data..page_aligned)
 	}
 
 	. = ALIGN(64);
-	.data.cacheline_aligned : {
-		*(.data.cacheline_aligned)
+	.data..cacheline_aligned : {
+		*(.data..cacheline_aligned)
 	}
 
 	_data = .;
arch/arm/kernel/head-nommu.S +1 -1
index cc87e17..fcd93f5 100644
@@ -32,7 +32,7 @@
  * numbers for r1.
  *
  */
-	.section ".text.head", "ax"
+	.section ".text..head", "ax"
 ENTRY(stext)
 	msr	cpsr_c, #PSR_F_BIT | PSR_I_BIT | SVC_MODE @ ensure svc mode
 						@ and irqs disabled
arch/arm/kernel/head.S +1 -1
index 21e17dc..705e759 100644
@@ -74,7 +74,7 @@
  * crap here - that's what the boot loader (or in extreme, well justified
  * circumstances, zImage) is for.
  */
-	.section ".text.head", "ax"
+	.section ".text..head", "ax"
 ENTRY(stext)
 	msr	cpsr_c, #PSR_F_BIT | PSR_I_BIT | SVC_MODE @ ensure svc mode
 						@ and irqs disabled
arch/arm/kernel/init_task.c +1 -1
index e859af3..0b81b8f 100644
@@ -29,7 +29,7 @@ EXPORT_SYMBOL(init_mm);
  * The things we do for performance..
  */
 union thread_union init_thread_union
-	__attribute__((__section__(".data.init_task"))) =
+	__attribute__((__section__(".data..init_task"))) =
 		{ INIT_THREAD_INFO(init_task) };
 
 /*
arch/arm/kernel/vmlinux.lds.S +8 -8
index c90f272..89c97c3 100644
@@ -23,10 +23,10 @@ SECTIONS
 #else
 	. = PAGE_OFFSET + TEXT_OFFSET;
 #endif
-	.text.head : {
+	.text..head : {
 		_stext = .;
 		_sinittext = .;
-		*(.text.head)
+		*(.text..head)
 	}
 
 	.init : {			/* Init code and data		*/
@@ -66,9 +66,9 @@ SECTIONS
 		. = ALIGN(4096);
 		__per_cpu_load = .;
 		__per_cpu_start = .;
-			*(.data.percpu.page_aligned)
-			*(.data.percpu)
-			*(.data.percpu.shared_aligned)
+			*(.data..percpu.page_aligned)
+			*(.data..percpu)
+			*(.data..percpu.shared_aligned)
 		__per_cpu_end = .;
 #ifndef CONFIG_XIP_KERNEL
 		__init_begin = _stext;
@@ -146,7 +146,7 @@ SECTIONS
 		 * first, the init task union, aligned
 		 * to an 8192 byte boundary.
 		 */
-		*(.data.init_task)
+		*(.data..init_task)
 
 #ifdef CONFIG_XIP_KERNEL
 		. = ALIGN(4096);
@@ -158,7 +158,7 @@ SECTIONS
 
 		. = ALIGN(4096);
 		__nosave_begin = .;
-		*(.data.nosave)
+		*(.data..nosave)
 		. = ALIGN(4096);
 		__nosave_end = .;
 
@@ -166,7 +166,7 @@ SECTIONS
 		 * then the cacheline aligned data
 		 */
 		. = ALIGN(32);
-		*(.data.cacheline_aligned)
+		*(.data..cacheline_aligned)
 
 		/*
 		 * The exception fixup table (might need resorting at runtime)
arch/arm/mm/proc-v6.S +1 -1
index f0cc599..39d8dd1 100644
@@ -132,7 +132,7 @@ cpu_v6_name:
 	.asciz	"ARMv6-compatible processor"
 	.align
 
-	.section ".text.init", #alloc, #execinstr
+	.section ".text..init", #alloc, #execinstr
 
 /*
  *	__v6_setup
arch/arm/mm/proc-v7.S +1 -1
index d1ebec4..f03c14c 100644
@@ -153,7 +153,7 @@ cpu_v7_name:
 	.ascii	"ARMv7 Processor"
 	.align
 
-	.section ".text.init", #alloc, #execinstr
+	.section ".text..init", #alloc, #execinstr
 
 /*
  *	__v7_setup
arch/arm/mm/tlb-v6.S +1 -1
index 20f84bb..7e9a62a 100644
@@ -87,7 +87,7 @@ ENTRY(v6wbi_flush_kern_tlb_range)
 	mcr	p15, 0, r2, c7, c5, 4		@ prefetch flush
 	mov	pc, lr
 
-	.section ".text.init", #alloc, #execinstr
+	.section ".text..init", #alloc, #execinstr
 
 	.type	v6wbi_tlb_fns, #object
 ENTRY(v6wbi_tlb_fns)
arch/arm/mm/tlb-v7.S +1 -1
index 24ba510..875a0bd 100644
@@ -80,7 +80,7 @@ ENTRY(v7wbi_flush_kern_tlb_range)
 	mov	pc, lr
 ENDPROC(v7wbi_flush_kern_tlb_range)
 
-	.section ".text.init", #alloc, #execinstr
+	.section ".text..init", #alloc, #execinstr
 
 	.type	v7wbi_tlb_fns, #object
 ENTRY(v7wbi_tlb_fns)
arch/avr32/kernel/init_task.c +1 -1
index 993d56e..4678bc6 100644
@@ -23,7 +23,7 @@ EXPORT_SYMBOL(init_mm);
  * Initial thread structure. Must be aligned on an 8192-byte boundary.
  */
 union thread_union init_thread_union
-	__attribute__((__section__(".data.init_task"))) =
+	__attribute__((__section__(".data..init_task"))) =
 		{ INIT_THREAD_INFO(init_task) };
 
 /*
arch/avr32/kernel/vmlinux.lds.S +3 -3
index 7910d41..5e73a02 100644
@@ -95,15 +95,15 @@ SECTIONS
 		/*
 		 * First, the init task union, aligned to an 8K boundary.
 		 */
-		*(.data.init_task)
+		*(.data..init_task)
 
 		/* Then, the page-aligned data */
 		. = ALIGN(PAGE_SIZE);
-		*(.data.page_aligned)
+		*(.data..page_aligned)
 
 		/* Then, the cacheline aligned data */
 		. = ALIGN(L1_CACHE_BYTES);
-		*(.data.cacheline_aligned)
+		*(.data..cacheline_aligned)
 
 		/* And the rest... */
 		*(.data.rel*)
arch/avr32/mm/init.c +1 -1
index e819fa6..533a011 100644
@@ -24,7 +24,7 @@
 #include <asm/setup.h>
 #include <asm/sections.h>
 
-#define __page_aligned	__attribute__((section(".data.page_aligned")))
+#define __page_aligned	__attribute__((section(".data..page_aligned")))
 
 DEFINE_PER_CPU(struct mmu_gather, mmu_gathers);
 
arch/blackfin/kernel/vmlinux.lds.S +1 -1
index 27952ae..f3ef8f1 100644
@@ -94,7 +94,7 @@ SECTIONS
 		__sdata = .;
 		/* This gets done first, so the glob doesn't suck it in */
 		. = ALIGN(32);
-		*(.data.cacheline_aligned)
+		*(.data..cacheline_aligned)
 
 #if !L1_DATA_A_LENGTH
 		. = ALIGN(32);
arch/cris/kernel/process.c +1 -1
index 4df0b32..0f0b4bd 100644
@@ -50,7 +50,7 @@ EXPORT_SYMBOL(init_mm);
  * "init_task" linker map entry..
  */
 union thread_union init_thread_union 
-	__attribute__((__section__(".data.init_task"))) =
+	__attribute__((__section__(".data..init_task"))) =
 		{ INIT_THREAD_INFO(init_task) };
 
 /*
arch/cris/kernel/vmlinux.lds.S +1 -1
index 0d2adfc..f603bc8 100644
@@ -68,7 +68,7 @@ SECTIONS
 	_edata = . ;
 
 	. = ALIGN(PAGE_SIZE);	/* init_task and stack, must be aligned. */
-	.data.init_task : { *(.data.init_task) }
+	.data..init_task : { *(.data..init_task) }
 
 	. = ALIGN(PAGE_SIZE);		/* Init code and data. */
 	__init_begin = .;
arch/frv/include/asm/init.h +4 -4
index 8b15838..4d21473 100644
@@ -1,12 +1,12 @@
 #ifndef _ASM_INIT_H
 #define _ASM_INIT_H
 
-#define __init __attribute__ ((__section__ (".text.init")))
-#define __initdata __attribute__ ((__section__ (".data.init")))
+#define __init __attribute__ ((__section__ (".text..init")))
+#define __initdata __attribute__ ((__section__ (".data..init")))
 /* For assembly routines */
-#define __INIT		.section	".text.init",#alloc,#execinstr
+#define __INIT		.section	".text..init",#alloc,#execinstr
 #define __FINIT		.previous
-#define __INITDATA	.section	".data.init",#alloc,#write
+#define __INITDATA	.section	".data..init",#alloc,#write
 
 #endif
 
arch/frv/kernel/break.S +2 -2
index bd0bdf9..cbb6958 100644
@@ -21,7 +21,7 @@
 #
 # the break handler has its own stack
 #
-	.section	.bss.stack
+	.section	.bss..stack
 	.globl		__break_user_context
 	.balign		THREAD_SIZE
 __break_stack:
@@ -63,7 +63,7 @@ __break_trace_through_exceptions:
 # entry point for Break Exceptions/Interrupts
 #
 ###############################################################################
-	.section	.text.break
+	.section	.text..break
 	.balign		4
 	.globl		__entry_break
 __entry_break:
arch/frv/kernel/entry.S +1 -1
index 99060ab..fe5479b 100644
@@ -38,7 +38,7 @@
 
 #define nr_syscalls ((syscall_table_size)/4)
 
-	.section	.text.entry
+	.section	.text..entry
 	.balign		4
 
 .macro LEDS val
arch/frv/kernel/head-mmu-fr451.S +1 -1
index c8f210d..5900967 100644
@@ -31,7 +31,7 @@
 #define __400_LCR	0xfe000100
 #define __400_LSBR	0xfe000c00
 
-	.section	.text.init,"ax"
+	.section	.text..init,"ax"
 	.balign		4
 
 ###############################################################################
arch/frv/kernel/head-uc-fr401.S +1 -1
index ee282be..ecba176 100644
@@ -30,7 +30,7 @@
 #define __400_LCR	0xfe000100
 #define __400_LSBR	0xfe000c00
 
-	.section	.text.init,"ax"
+	.section	.text..init,"ax"
 	.balign		4
 
 ###############################################################################
arch/frv/kernel/head-uc-fr451.S +1 -1
index b10d9c8..ed9dd17 100644
@@ -30,7 +30,7 @@
 #define __400_LCR	0xfe000100
 #define __400_LSBR	0xfe000c00
 
-	.section	.text.init,"ax"
+	.section	.text..init,"ax"
 	.balign		4
 
 ###############################################################################
arch/frv/kernel/head-uc-fr555.S +1 -1
index 39937c1..ee46b0b 100644
@@ -29,7 +29,7 @@
 #define __551_LCR	0xfeff1100
 #define __551_LSBR	0xfeff1c00
 
-	.section	.text.init,"ax"
+	.section	.text..init,"ax"
 	.balign		4
 
 ###############################################################################
arch/frv/kernel/head.S +2 -2
index fecf751..35e6391 100644
@@ -27,7 +27,7 @@
 #   command line string
 #
 ###############################################################################
-	.section	.text.head,"ax"
+	.section	.text..head,"ax"
 	.balign		4
 
 	.globl		_boot, __head_reference
@@ -541,7 +541,7 @@ __head_end:
 	.size		_boot, .-_boot
 
 	# provide a point for GDB to place a break
-	.section	.text.start,"ax"
+	.section	.text..start,"ax"
 	.globl		_start
 	.balign		4
 _start:
arch/frv/kernel/init_task.c +1 -1
index 29429a8..f57ec19 100644
@@ -24,7 +24,7 @@ EXPORT_SYMBOL(init_mm);
  * "init_task" linker map entry..
  */
 union thread_union init_thread_union
-	__attribute__((__section__(".data.init_task"))) =
+	__attribute__((__section__(".data..init_task"))) =
 		{ INIT_THREAD_INFO(init_task) };
 
 /*
arch/frv/kernel/vmlinux.lds.S +9 -9
index b95c4ea..a7bcf9b 100644
@@ -26,7 +26,7 @@ SECTIONS
 
   _sinittext = .;
   .init.text : {
-	*(.text.head)
+	*(.text..head)
 #ifndef CONFIG_DEBUG_INFO
 	INIT_TEXT
 	EXIT_TEXT
@@ -71,13 +71,13 @@ SECTIONS
 
   /* put sections together that have massive alignment issues */
   . = ALIGN(THREAD_SIZE);
-  .data.init_task : {
+  .data..init_task : {
 	  /* init task record & stack */
-	  *(.data.init_task)
+	  *(.data..init_task)
   }
 
   . = ALIGN(L1_CACHE_BYTES);
-  .data.cacheline_aligned : { *(.data.cacheline_aligned) }
+  .data..cacheline_aligned : { *(.data..cacheline_aligned) }
 
   .trap : {
 	/* trap table management - read entry-table.S before modifying */
@@ -94,10 +94,10 @@ SECTIONS
   _text = .;
   _stext = .;
   .text : {
-	*(.text.start)
-	*(.text.entry)
-	*(.text.break)
-	*(.text.tlbmiss)
+	*(.text..start)
+	*(.text..entry)
+	*(.text..break)
+	*(.text..tlbmiss)
 	TEXT_TEXT
 	SCHED_TEXT
 	LOCK_TEXT
@@ -152,7 +152,7 @@ SECTIONS
 
   .sbss		: { *(.sbss .sbss.*) }
   .bss		: { *(.bss .bss.*) }
-  .bss.stack	: { *(.bss) }
+  .bss..stack	: { *(.bss) }
 
   __bss_stop = .;
   _end = . ;
arch/frv/mm/tlb-miss.S +1 -1
index 7f392bc..f3ac019 100644
@@ -15,7 +15,7 @@
 #include <asm/pgtable.h>
 #include <asm/spr-regs.h>
 
-	.section	.text.tlbmiss
+	.section	.text..tlbmiss
 	.balign		4
 
 	.globl		__entry_insn_mmu_miss
arch/h8300/boot/compressed/head.S +1 -1
index 985a81a..10e9a2d 100644
@@ -9,7 +9,7 @@
 
 #define SRAM_START 0xff4000
 
-	.section	.text.startup
+	.section	.text..startup
 	.global	startup
 startup:
 	mov.l	#SRAM_START+0x8000, sp
arch/h8300/boot/compressed/vmlinux.lds +1 -1
index 65e2a0d..a0a3a0e 100644
@@ -4,7 +4,7 @@ SECTIONS
         {
         __stext = . ;
 	__text = .;
-	       *(.text.startup)
+	       *(.text..startup)
 	       *(.text)
         __etext = . ;
         }
arch/h8300/kernel/init_task.c +1 -1
index cb5dc55..fb473b1 100644
@@ -36,6 +36,6 @@ EXPORT_SYMBOL(init_task);
  * "init_task" linker map entry..
  */
 union thread_union init_thread_union
-	__attribute__((__section__(".data.init_task"))) =
+	__attribute__((__section__(".data..init_task"))) =
 		{ INIT_THREAD_INFO(init_task) };
 
arch/h8300/kernel/vmlinux.lds.S +1 -1
index 43a87b9..c0e3635 100644
@@ -101,7 +101,7 @@ SECTIONS
 	___data_start = . ;
 
 	. = ALIGN(0x2000) ;
-		*(.data.init_task)
+		*(.data..init_task)
 	. = ALIGN(0x4) ;
 		DATA_DATA
 	. = ALIGN(0x4) ;
arch/ia64/include/asm/asmmacro.h +6 -6
index c1642fd..3ab6d75 100644
@@ -70,12 +70,12 @@ name:
  * path (ivt.S - TLB miss processing) or in places where it might not be
  * safe to use a "tpa" instruction (mca_asm.S - error recovery).
  */
-	.section ".data.patch.vtop", "a"	// declare section & section attributes
+	.section ".data..patch.vtop", "a"	// declare section & section attributes
 	.previous
 
 #define	LOAD_PHYSICAL(pr, reg, obj)		\
 [1:](pr)movl reg = obj;				\
-	.xdata4 ".data.patch.vtop", 1b-.
+	.xdata4 ".data..patch.vtop", 1b-.
 
 /*
  * For now, we always put in the McKinley E9 workaround.  On CPUs that don't need it,
@@ -84,11 +84,11 @@ name:
 #define DO_MCKINLEY_E9_WORKAROUND
 
 #ifdef DO_MCKINLEY_E9_WORKAROUND
-	.section ".data.patch.mckinley_e9", "a"
+	.section ".data..patch.mckinley_e9", "a"
 	.previous
 /* workaround for Itanium 2 Errata 9: */
 # define FSYS_RETURN					\
-	.xdata4 ".data.patch.mckinley_e9", 1f-.;	\
+	.xdata4 ".data..patch.mckinley_e9", 1f-.;	\
 1:{ .mib;						\
 	nop.m 0;					\
 	mov r16=ar.pfs;					\
@@ -107,11 +107,11 @@ name:
  * If physical stack register size is different from DEF_NUM_STACK_REG,
  * dynamically patch the kernel for correct size.
  */
-	.section ".data.patch.phys_stack_reg", "a"
+	.section ".data..patch.phys_stack_reg", "a"
 	.previous
 #define LOAD_PHYS_STACK_REG_SIZE(reg)			\
 [1:]	adds reg=IA64_NUM_PHYS_STACK_REG*8+8,r0;	\
-	.xdata4 ".data.patch.phys_stack_reg", 1b-.
+	.xdata4 ".data..patch.phys_stack_reg", 1b-.
 
 /*
  * Up until early 2004, use of .align within a function caused bad unwind info.
arch/ia64/include/asm/cache.h +1 -1
index e7482bd..988254a 100644
@@ -24,6 +24,6 @@
 # define SMP_CACHE_BYTES	(1 << 3)
 #endif
 
-#define __read_mostly __attribute__((__section__(".data.read_mostly")))
+#define __read_mostly __attribute__((__section__(".data..read_mostly")))
 
 #endif /* _ASM_IA64_CACHE_H */
arch/ia64/include/asm/percpu.h +1 -1
index 30cf465..35d9aeb 100644
@@ -31,7 +31,7 @@ extern void *per_cpu_init(void);
 
 #endif	/* SMP */
 
-#define PER_CPU_BASE_SECTION ".data.percpu"
+#define PER_CPU_BASE_SECTION ".data..percpu"
 
 /*
  * Be extremely careful when taking the address of this variable!  Due to virtual
arch/ia64/kernel/Makefile.gate +1 -1
index 1d87f84..af90125 100644
@@ -21,7 +21,7 @@ GATECFLAGS_gate-syms.o = -r
 $(obj)/gate-syms.o: $(obj)/gate.lds $(obj)/gate.o FORCE
 	$(call if_changed,gate)
 
-# gate-data.o contains the gate DSO image as data in section .data.gate.
+# gate-data.o contains the gate DSO image as data in section .data..gate.
 # We must build gate.so before we can assemble it.
 # Note: kbuild does not track this dependency due to usage of .incbin
 $(obj)/gate-data.o: $(obj)/gate.so
arch/ia64/kernel/gate-data.S +1 -1
index 258c0a3..b3ef1c7 100644
@@ -1,3 +1,3 @@
-	.section .data.gate, "aw"
+	.section .data..gate, "aw"
 
 	.incbin "arch/ia64/kernel/gate.so"
arch/ia64/kernel/gate.S +4 -4
index cf5e0a1..245d3e1 100644
@@ -21,18 +21,18 @@
  * to targets outside the shared object) and to avoid multi-phase kernel builds, we
  * simply create minimalistic "patch lists" in special ELF sections.
  */
-	.section ".data.patch.fsyscall_table", "a"
+	.section ".data..patch.fsyscall_table", "a"
 	.previous
 #define LOAD_FSYSCALL_TABLE(reg)			\
 [1:]	movl reg=0;					\
-	.xdata4 ".data.patch.fsyscall_table", 1b-.
+	.xdata4 ".data..patch.fsyscall_table", 1b-.
 
-	.section ".data.patch.brl_fsys_bubble_down", "a"
+	.section ".data..patch.brl_fsys_bubble_down", "a"
 	.previous
 #define BRL_COND_FSYS_BUBBLE_DOWN(pr)			\
 [1:](pr)brl.cond.sptk 0;				\
 	;;						\
-	.xdata4 ".data.patch.brl_fsys_bubble_down", 1b-.
+	.xdata4 ".data..patch.brl_fsys_bubble_down", 1b-.
 
 GLOBAL_ENTRY(__kernel_syscall_via_break)
 	.prologue
arch/ia64/kernel/gate.lds.S +5 -5
index 88c64ed..d32b085 100644
@@ -33,21 +33,21 @@ SECTIONS
 	 */
 	. = GATE_ADDR + 0x600;
 
-	.data.patch		: {
+	.data..patch		: {
 		__paravirt_start_gate_mckinley_e9_patchlist = .;
-		*(.data.patch.mckinley_e9)
+		*(.data..patch.mckinley_e9)
 		__paravirt_end_gate_mckinley_e9_patchlist = .;
 
 		__paravirt_start_gate_vtop_patchlist = .;
-		*(.data.patch.vtop)
+		*(.data..patch.vtop)
 		__paravirt_end_gate_vtop_patchlist = .;
 
 		__paravirt_start_gate_fsyscall_patchlist = .;
-		*(.data.patch.fsyscall_table)
+		*(.data..patch.fsyscall_table)
 		__paravirt_end_gate_fsyscall_patchlist = .;
 
 		__paravirt_start_gate_brl_fsys_bubble_down_patchlist = .;
-		*(.data.patch.brl_fsys_bubble_down)
+		*(.data..patch.brl_fsys_bubble_down)
 		__paravirt_end_gate_brl_fsys_bubble_down_patchlist = .;
 	}						:readable
 
arch/ia64/kernel/head.S +1 -1
index 23f846d..c0739fc 100644
@@ -181,7 +181,7 @@ swapper_pg_dir:
 halt_msg:
 	stringz "Halting kernel\n"
 
-	.section .text.head,"ax"
+	.section .text..head,"ax"
 
 	.global start_ap
 
arch/ia64/kernel/init_task.c +2 -2
index 5b0e830..8a5028c 100644
@@ -27,7 +27,7 @@ EXPORT_SYMBOL(init_mm);
  * Initial task structure.
  *
  * We need to make sure that this is properly aligned due to the way process stacks are
- * handled. This is done by having a special ".data.init_task" section...
+ * handled. This is done by having a special ".data..init_task" section...
  */
 #define init_thread_info	init_task_mem.s.thread_info
 
@@ -37,7 +37,7 @@ union {
 		struct thread_info thread_info;
 	} s;
 	unsigned long stack[KERNEL_STACK_SIZE/sizeof (unsigned long)];
-} init_task_mem asm ("init_task") __attribute__((section(".data.init_task"))) = {{
+} init_task_mem asm ("init_task") __attribute__((section(".data..init_task"))) = {{
 	.task =		INIT_TASK(init_task_mem.s.task),
 	.thread_info =	INIT_THREAD_INFO(init_task_mem.s.task)
 }};
arch/ia64/kernel/ivt.S +1 -1
index ec9a5fd..0c14512 100644
@@ -83,7 +83,7 @@
 	mov r19=n;;			/* prepare to save predicates */		\
 	br.sptk.many dispatch_to_fault_handler
 
-	.section .text.ivt,"ax"
+	.section .text..ivt,"ax"
 
 	.align 32768	// align on 32KB boundary
 	.global ia64_ivt
arch/ia64/kernel/minstate.h +2 -2
index 292e214..d56753a 100644
@@ -16,7 +16,7 @@
 #define ACCOUNT_SYS_ENTER
 #endif
 
-.section ".data.patch.rse", "a"
+.section ".data..patch.rse", "a"
 .previous
 
 /*
@@ -215,7 +215,7 @@
 (pUStk) extr.u r17=r18,3,6;			\
 (pUStk)	sub r16=r18,r22;			\
 [1:](pKStk)	br.cond.sptk.many 1f;		\
-	.xdata4 ".data.patch.rse",1b-.		\
+	.xdata4 ".data..patch.rse",1b-.		\
 	;;					\
 	cmp.ge p6,p7 = 33,r17;			\
 	;;					\
arch/ia64/kernel/paravirtentry.S +1 -1
index 6158560..92d880c 100644
@@ -28,7 +28,7 @@
 #include "entry.h"
 
 #define DATA8(sym, init_value)			\
-	.pushsection .data.read_mostly ;	\
+	.pushsection .data..read_mostly ;	\
 	.align 8 ;				\
 	.global sym ;				\
 	sym: ;					\
arch/ia64/kernel/vmlinux.lds.S +22 -22

105 lines not on this page · raw message

arch/ia64/kvm/vmm_ivt.S +1 -1
index 3ef1a01..c5ab23b 100644
@@ -104,7 +104,7 @@ GLOBAL_ENTRY(kvm_vmm_panic)
 	br.call.sptk.many b6=vmm_panic_handler;
 END(kvm_vmm_panic)
 
-    .section .text.ivt,"ax"
+    .section .text..ivt,"ax"
 
     .align 32768    // align on 32KB boundary
     .global kvm_ia64_ivt
arch/ia64/xen/gate-data.S +1 -1
index 7d4830a..6f95b6b 100644
@@ -1,3 +1,3 @@
-	.section .data.gate.xen, "aw"
+	.section .data..gate.xen, "aw"
 
 	.incbin "arch/ia64/xen/gate.so"
arch/ia64/xen/xensetup.S +1 -1
index 28fed1f..28def53 100644
@@ -14,7 +14,7 @@
 #include <linux/init.h>
 #include <xen/interface/elfnote.h>
 
-	.section .data.read_mostly
+	.section .data..read_mostly
 	.align 8
 	.global xen_domain_type
 xen_domain_type:
arch/m32r/kernel/head.S +1 -1
index 9091606..2bac669 100644
@@ -23,7 +23,7 @@ __INITDATA
 /*
  * References to members of the boot_cpu_data structure.
  */
-.section .text.head, "ax"
+.section .text..head, "ax"
 	.global	start_kernel
 	.global __bss_start
 	.global _end
arch/m32r/kernel/init_task.c +1 -1
index 016885c..7c24cec 100644
@@ -25,7 +25,7 @@ EXPORT_SYMBOL(init_mm);
  * "init_task" linker map entry..
  */
 union thread_union init_thread_union
-	__attribute__((__section__(".data.init_task"))) =
+	__attribute__((__section__(".data..init_task"))) =
 		{ INIT_THREAD_INFO(init_task) };
 
 /*
arch/m32r/kernel/vmlinux.lds.S +4 -4
index 9db05df..2bcc37e 100644
@@ -27,7 +27,7 @@ SECTIONS
   _text = .;			/* Text and read-only data */
   .boot : { *(.boot) } = 0
   .text : {
-	*(.text.head)
+	*(.text..head)
 	TEXT_TEXT
 	SCHED_TEXT
 	LOCK_TEXT
@@ -57,17 +57,17 @@ SECTIONS
 
   . = ALIGN(4096);
   __nosave_begin = .;
-  .data_nosave : { *(.data.nosave) }
+  .data_nosave : { *(.data..nosave) }
   . = ALIGN(4096);
   __nosave_end = .;
 
   . = ALIGN(32);
-  .data.cacheline_aligned : { *(.data.cacheline_aligned) }
+  .data..cacheline_aligned : { *(.data..cacheline_aligned) }
 
   _edata = .;			/* End of data section */
 
   . = ALIGN(8192);		/* init_task */
-  .data.init_task : { *(.data.init_task) }
+  .data..init_task : { *(.data..init_task) }
 
   /* will be freed after init */
   . = ALIGN(4096);		/* Init code and data */
arch/m68k/kernel/head.S +1 -1
index f513f53..bdb247e 100644
@@ -577,7 +577,7 @@ func_define	putn,1
 #endif
 .endm
 
-.section ".text.head","ax"
+.section ".text..head","ax"
 ENTRY(_stext)
 /*
  * Version numbers of the bootinfo interface
arch/m68k/kernel/process.c +1 -1
index ec37fb5..4b83a2e 100644
@@ -47,7 +47,7 @@ struct mm_struct init_mm = INIT_MM(init_mm);
 EXPORT_SYMBOL(init_mm);
 
 union thread_union init_thread_union
-__attribute__((section(".data.init_task"), aligned(THREAD_SIZE)))
+__attribute__((section(".data..init_task"), aligned(THREAD_SIZE)))
        = { INIT_THREAD_INFO(init_task) };
 
 /* initial task structure */
arch/m68k/kernel/sun3-head.S +1 -1
index aad0159..5a6714b 100644
@@ -29,7 +29,7 @@ kernel_pmd_table:              .skip 0x2000
 .globl kernel_pg_dir
 .equ    kernel_pg_dir,kernel_pmd_table
 
-	.section .text.head
+	.section .text..head
 ENTRY(_stext)
 ENTRY(_start)
 
arch/m68k/kernel/vmlinux-std.lds +3 -3
index f846d4e..35b6da8 100644
@@ -12,7 +12,7 @@ SECTIONS
   . = 0x1000;
   _text = .;			/* Text and read-only data */
   .text : {
-	*(.text.head)
+	*(.text..head)
 	TEXT_TEXT
 	SCHED_TEXT
 	LOCK_TEXT
@@ -35,7 +35,7 @@ SECTIONS
 	}
 
   . = ALIGN(16);
-  .data.cacheline_aligned : { *(.data.cacheline_aligned) }
+  .data..cacheline_aligned : { *(.data..cacheline_aligned) }
 
   .bss : { *(.bss) }		/* BSS */
 
@@ -78,7 +78,7 @@ SECTIONS
   . = ALIGN(8192);
   __init_end = .;
 
-  .data.init_task : { *(.data.init_task) }	/* The initial task and kernel stack */
+  .data..init_task : { *(.data..init_task) }	/* The initial task and kernel stack */
 
   _end = . ;
 
arch/m68k/kernel/vmlinux-sun3.lds +2 -2
index d9368c0..e6ce561 100644
@@ -12,7 +12,7 @@ SECTIONS
   . = 0xE002000;
   _text = .;			/* Text and read-only data */
   .text : {
-	*(.text.head)
+	*(.text..head)
 	TEXT_TEXT
 	SCHED_TEXT
 	LOCK_TEXT
@@ -70,7 +70,7 @@ __init_begin = .;
 #endif
 	. = ALIGN(PAGE_SIZE);
 	__init_end = .;
-	.data.init.task : { *(.data.init_task) }
+	.data..init.task : { *(.data..init_task) }
 
 
   .bss : { *(.bss) }		/* BSS */
arch/m68knommu/kernel/init_task.c +1 -1
index fe282de..1ea36d6 100644
@@ -36,6 +36,6 @@ EXPORT_SYMBOL(init_task);
  * "init_task" linker map entry..
  */
 union thread_union init_thread_union
-	__attribute__((__section__(".data.init_task"))) =
+	__attribute__((__section__(".data..init_task"))) =
 		{ INIT_THREAD_INFO(init_task) };
 
arch/m68knommu/kernel/vmlinux.lds.S +3 -3
index 69ba9b1..bdf9b1c 100644
@@ -55,7 +55,7 @@ SECTIONS {
 	.romvec : {
 		__rom_start = . ;
 		_romvec = .;
-		*(.data.initvect)
+		*(.data..initvect)
 	} > romvec
 #endif
 
@@ -66,7 +66,7 @@ SECTIONS {
 		TEXT_TEXT
 		SCHED_TEXT
 		LOCK_TEXT
-        	*(.text.lock)
+		*(.text..lock)
 
 		. = ALIGN(16);          /* Exception table              */
 		__start___ex_table = .;
@@ -148,7 +148,7 @@ SECTIONS {
 		_sdata = . ;
 		DATA_DATA
 		. = ALIGN(8192) ;
-		*(.data.init_task)
+		*(.data..init_task)
 		_edata = . ;
 	} > DATA
 
arch/m68knommu/platform/68360/head-ram.S +1 -1
index 2ef0624..8eb94fb 100644
@@ -280,7 +280,7 @@ _dprbase:
      * and then overwritten as needed.
      */
  
-.section ".data.initvect","awx"
+.section ".data..initvect","awx"
     .long   RAMEND	/* Reset: Initial Stack Pointer                 - 0.  */
     .long   _start      /* Reset: Initial Program Counter               - 1.  */
     .long   buserr      /* Bus Error                                    - 2.  */
arch/m68knommu/platform/68360/head-rom.S +1 -1
index 62ecf41..97510e5 100644
@@ -291,7 +291,7 @@ _dprbase:
      * and then overwritten as needed.
      */
  
-.section ".data.initvect","awx"
+.section ".data..initvect","awx"
     .long   RAMEND	/* Reset: Initial Stack Pointer                 - 0.  */
     .long   _start      /* Reset: Initial Program Counter               - 1.  */
     .long   buserr      /* Bus Error                                    - 2.  */
arch/microblaze/kernel/init_task.c +1 -1
index 48eb9fb..70a23dc 100644
@@ -22,7 +22,7 @@ struct mm_struct init_mm = INIT_MM(init_mm);
 EXPORT_SYMBOL(init_mm);
 
 union thread_union init_thread_union
-	__attribute__((__section__(".data.init_task"))) =
+	__attribute__((__section__(".data..init_task"))) =
 { INIT_THREAD_INFO(init_task) };
 
 struct task_struct init_task = INIT_TASK(init_task);
arch/microblaze/kernel/vmlinux.lds.S +2 -2
index 840385e..e332495 100644
@@ -66,7 +66,7 @@ SECTIONS {
 		*(.data)
 	}
 	. = ALIGN(32);
-	.data.cacheline_aligned : { *(.data.cacheline_aligned) }
+	.data..cacheline_aligned : { *(.data..cacheline_aligned) }
 	_edata = . ;
 
 	/* Reserve some low RAM for r0 based memory references */
@@ -76,7 +76,7 @@ SECTIONS {
 
 	/* The initial task */
 	. = ALIGN(8192);
-	.data.init_task : { *(.data.init_task) }
+	.data..init_task : { *(.data..init_task) }
 
 	/* Under the microblaze ABI, .sdata and .sbss must be contiguous */
 	. = ALIGN(8);
arch/mips/kernel/init_task.c +1 -1
index 149cd91..acb71af 100644
@@ -26,7 +26,7 @@ EXPORT_SYMBOL(init_mm);
  * The things we do for performance..
  */
 union thread_union init_thread_union
-	__attribute__((__section__(".data.init_task"),
+	__attribute__((__section__(".data..init_task"),
 	               __aligned__(THREAD_SIZE))) =
 		{ INIT_THREAD_INFO(init_task) };
 
arch/mips/kernel/vmlinux.lds.S +4 -4
index 58738c8..6b302c7 100644
@@ -77,7 +77,7 @@ SECTIONS
 		 * object file alignment.  Using 32768
 		 */
 		. = ALIGN(_PAGE_SIZE);
-		*(.data.init_task)
+		*(.data..init_task)
 
 		DATA_DATA
 		CONSTRUCTORS
@@ -99,14 +99,14 @@ SECTIONS
 	. = ALIGN(_PAGE_SIZE);
 	.data_nosave : {
 		__nosave_begin = .;
-		*(.data.nosave)
+		*(.data..nosave)
 	}
 	. = ALIGN(_PAGE_SIZE);
 	__nosave_end = .;
 
 	. = ALIGN(1 << CONFIG_MIPS_L1_CACHE_SHIFT);
-	.data.cacheline_aligned : {
-		*(.data.cacheline_aligned)
+	.data..cacheline_aligned : {
+		*(.data..cacheline_aligned)
 	}
 	_edata =  .;			/* End of data section */
 
arch/mips/lasat/image/head.S +1 -1
index efb95f2..e0ecda9 100644
@@ -1,7 +1,7 @@
 #include <asm/lasat/head.h>
 
 	.text
-	.section .text.start, "ax"
+	.section .text..start, "ax"
 	.set noreorder
 	.set mips3
 
arch/mips/lasat/image/romscript.normal +1 -1
index 988f8ad..0864c96 100644
@@ -4,7 +4,7 @@ SECTIONS
 {
   .text :
   {
-    *(.text.start)
+    *(.text..start)
   }
 
   /* Data in ROM */
arch/mn10300/kernel/head.S +1 -1
index 8a8309f..3330c43 100644
@@ -19,7 +19,7 @@
 #include <asm/param.h>
 #include <unit/serial.h>
 
-	.section .text.head,"ax"
+	.section .text..head,"ax"
 
 ###############################################################################
 #
arch/mn10300/kernel/init_task.c +1 -1
index 5ac3566..599840f 100644
@@ -31,7 +31,7 @@ EXPORT_SYMBOL(init_mm);
  * "init_task" linker map entry..
  */
 union thread_union init_thread_union
-	__attribute__((__section__(".data.init_task"))) =
+	__attribute__((__section__(".data..init_task"))) =
 		{ INIT_THREAD_INFO(init_task) };
 
 /*
arch/mn10300/kernel/vmlinux.lds.S +8 -8
index b825966..938792a 100644
@@ -28,7 +28,7 @@ SECTIONS
   _text = .;			/* Text and read-only data */
   .text : {
 	*(
-	.text.head
+	.text..head
 	.text
 	)
 	TEXT_TEXT
@@ -58,25 +58,25 @@ SECTIONS
 
   . = ALIGN(PAGE_SIZE);
   __nosave_begin = .;
-  .data_nosave : { *(.data.nosave) }
+  .data_nosave : { *(.data..nosave) }
   . = ALIGN(PAGE_SIZE);
   __nosave_end = .;
 
   . = ALIGN(PAGE_SIZE);
-  .data.page_aligned : { *(.data.idt) }
+  .data..page_aligned : { *(.data..idt) }
 
   . = ALIGN(32);
-  .data.cacheline_aligned : { *(.data.cacheline_aligned) }
+  .data..cacheline_aligned : { *(.data..cacheline_aligned) }
 
   /* rarely changed data like cpu maps */
   . = ALIGN(32);
-  .data.read_mostly : AT(ADDR(.data.read_mostly)) {
-	*(.data.read_mostly)
+  .data..read_mostly : AT(ADDR(.data..read_mostly)) {
+	*(.data..read_mostly)
 	_edata = .;		/* End of data section */
   }
 
   . = ALIGN(THREAD_SIZE);	/* init_task */
-  .data.init_task : { *(.data.init_task) }
+  .data..init_task : { *(.data..init_task) }
 
   /* might get freed after init */
   . = ALIGN(PAGE_SIZE);
@@ -134,7 +134,7 @@ SECTIONS
 
   __bss_start = .;		/* BSS */
   .bss : {
-	*(.bss.page_aligned)
+	*(.bss..page_aligned)
 	*(.bss)
   }
   . = ALIGN(4);
arch/parisc/include/asm/cache.h +1 -1
index 32c2cca..45effe6 100644
@@ -28,7 +28,7 @@
 
 #define SMP_CACHE_BYTES L1_CACHE_BYTES
 
-#define __read_mostly __attribute__((__section__(".data.read_mostly")))
+#define __read_mostly __attribute__((__section__(".data..read_mostly")))
 
 void parisc_cache_init(void);	/* initializes cache-flushing */
 void disable_sr_hashing_asm(int); /* low level support for above */
arch/parisc/include/asm/system.h +1 -1
index ee80c92..7ccaf29 100644
@@ -174,7 +174,7 @@ static inline void set_eiem(unsigned long val)
 })
 
 #ifdef CONFIG_SMP
-# define __lock_aligned __attribute__((__section__(".data.lock_aligned")))
+# define __lock_aligned __attribute__((__section__(".data..lock_aligned")))
 #endif
 
 #define arch_align_stack(x) (x)
arch/parisc/kernel/head.S +1 -1
index 0e3d9f9..4dbdf0e 100644
@@ -345,7 +345,7 @@ smp_slave_stext:
 ENDPROC(stext)
 
 #ifndef CONFIG_64BIT
-	.section .data.read_mostly
+	.section .data..read_mostly
 
 	.align	4
 	.export	$global$,data
arch/parisc/kernel/init_task.c +4 -4
index 1e25a45..14394c0 100644
@@ -48,7 +48,7 @@ EXPORT_SYMBOL(init_mm);
  * "init_task" linker map entry..
  */
 union thread_union init_thread_union
-	__attribute__((aligned(128))) __attribute__((__section__(".data.init_task"))) =
+	__attribute__((aligned(128))) __attribute__((__section__(".data..init_task"))) =
 		{ INIT_THREAD_INFO(init_task) };
 
 #if PT_NLEVELS == 3
@@ -57,11 +57,11 @@ union thread_union init_thread_union
  * guarantee that global objects will be laid out in memory in the same order 
  * as the order of declaration, so put these in different sections and use
  * the linker script to order them. */
-pmd_t pmd0[PTRS_PER_PMD] __attribute__ ((__section__ (".data.vm0.pmd"), aligned(PAGE_SIZE)));
+pmd_t pmd0[PTRS_PER_PMD] __attribute__ ((__section__ (".data..vm0.pmd"), aligned(PAGE_SIZE)));
 #endif
 
-pgd_t swapper_pg_dir[PTRS_PER_PGD] __attribute__ ((__section__ (".data.vm0.pgd"), aligned(PAGE_SIZE)));
-pte_t pg0[PT_INITIAL * PTRS_PER_PTE] __attribute__ ((__section__ (".data.vm0.pte"), aligned(PAGE_SIZE)));
+pgd_t swapper_pg_dir[PTRS_PER_PGD] __attribute__ ((__section__ (".data..vm0.pgd"), aligned(PAGE_SIZE)));
+pte_t pg0[PT_INITIAL * PTRS_PER_PTE] __attribute__ ((__section__ (".data..vm0.pte"), aligned(PAGE_SIZE)));
 
 /*
  * Initial task structure.
arch/parisc/kernel/vmlinux.lds.S +13 -13
index fd2cc4f..9f5184f 100644
@@ -96,8 +96,8 @@ SECTIONS
 
 	/* rarely changed data like cpu maps */
 	. = ALIGN(16);
-	.data.read_mostly : {
-		*(.data.read_mostly)
+	.data..read_mostly : {
+		*(.data..read_mostly)
 	}
 
 	. = ALIGN(L1_CACHE_BYTES);
@@ -108,14 +108,14 @@ SECTIONS
 	}
 
 	. = ALIGN(L1_CACHE_BYTES);
-	.data.cacheline_aligned : {
-		*(.data.cacheline_aligned)
+	.data..cacheline_aligned : {
+		*(.data..cacheline_aligned)
 	}
 
 	/* PA-RISC locks requires 16-byte alignment */
 	. = ALIGN(16);
-	.data.lock_aligned : {
-		*(.data.lock_aligned)
+	.data..lock_aligned : {
+		*(.data..lock_aligned)
 	}
 
 	/* nosave data is really only used for software suspend...it's here
@@ -124,7 +124,7 @@ SECTIONS
 	. = ALIGN(PAGE_SIZE);
 	__nosave_begin = .;
 	.data_nosave : {
-		*(.data.nosave)
+		*(.data..nosave)
 	}
 	. = ALIGN(PAGE_SIZE);
 	__nosave_end = .;
@@ -136,10 +136,10 @@ SECTIONS
 	__bss_start = .;
 	/* page table entries need to be PAGE_SIZE aligned */
 	. = ALIGN(PAGE_SIZE);
-	.data.vmpages : {
-		*(.data.vm0.pmd)
-		*(.data.vm0.pgd)
-		*(.data.vm0.pte)
+	.data..vmpages : {
+		*(.data..vm0.pmd)
+		*(.data..vm0.pgd)
+		*(.data..vm0.pte)
 	}
 	.bss : {
 		*(.bss)
@@ -151,8 +151,8 @@ SECTIONS
 	/* assembler code expects init_task to be 16k aligned */
 	. = ALIGN(16384);
 	/* init_task */
-	.data.init_task : {
-		*(.data.init_task)
+	.data..init_task : {
+		*(.data..init_task)
 	}
 
 #ifdef CONFIG_64BIT
arch/powerpc/include/asm/cache.h +1 -1
index 81de6eb..3f41ab9 100644
@@ -38,7 +38,7 @@ extern struct ppc64_caches ppc64_caches;
 #endif /* __powerpc64__ && ! __ASSEMBLY__ */
 
 #if !defined(__ASSEMBLY__)
-#define __read_mostly __attribute__((__section__(".data.read_mostly")))
+#define __read_mostly __attribute__((__section__(".data..read_mostly")))
 #endif
 
 #endif /* __KERNEL__ */
arch/powerpc/include/asm/page_64.h +1 -1
index 043bfdf..2307910 100644
@@ -157,7 +157,7 @@ do {						\
 #else
 #define __page_aligned \
 	__attribute__((__aligned__(PAGE_SIZE), \
-		__section__(".data.page_aligned")))
+		__section__(".data..page_aligned")))
 #endif
 
 #define VM_DATA_DEFAULT_FLAGS \
arch/powerpc/include/asm/ppc_asm.h +2 -2
index f59a666..676ed60 100644
@@ -189,7 +189,7 @@ name: \
 GLUE(.,name):
 
 #define _INIT_GLOBAL(name) \
-	.section ".text.init.refok"; \
+	.section ".text..init.refok"; \
 	.align 2 ; \
 	.globl name; \
 	.globl GLUE(.,name); \
@@ -229,7 +229,7 @@ name: \
 GLUE(.,name):
 
 #define _INIT_STATIC(name) \
-	.section ".text.init.refok"; \
+	.section ".text..init.refok"; \
 	.align 2 ; \
 	.section ".opd","aw"; \
 name: \
arch/powerpc/kernel/head_32.S +1 -1
index 54e68c1..5ce5dfa 100644
@@ -50,7 +50,7 @@
 	mtspr	SPRN_DBAT##n##L,RB;	\
 1:
 
-	.section	.text.head, "ax"
+	.section	.text..head, "ax"
 	.stabs	"arch/powerpc/kernel/",N_SO,0,0,0f
 	.stabs	"head_32.S",N_SO,0,0,0f
 0:
arch/powerpc/kernel/head_40x.S +1 -1
index 56d8e5d..de69ecd 100644
@@ -52,7 +52,7 @@
  *
  * This is all going to change RSN when we add bi_recs.......  -- Dan
  */
-	.section	.text.head, "ax"
+	.section	.text..head, "ax"
 _ENTRY(_stext);
 _ENTRY(_start);
 
arch/powerpc/kernel/head_44x.S +1 -1
index b56fecc..ab0b339 100644
@@ -50,7 +50,7 @@
  *   r7 - End of kernel command line string
  *
  */
-	.section	.text.head, "ax"
+	.section	.text..head, "ax"
 _ENTRY(_stext);
 _ENTRY(_start);
 	/*
arch/powerpc/kernel/head_8xx.S +1 -1
index 3c9452d..f6d1e67 100644
@@ -38,7 +38,7 @@
 #else
 #define DO_8xx_CPU6(val, reg)
 #endif
-	.section	.text.head, "ax"
+	.section	.text..head, "ax"
 _ENTRY(_stext);
 _ENTRY(_start);
 
arch/powerpc/kernel/head_fsl_booke.S +1 -1
index 4c22620..123bfec 100644
@@ -53,7 +53,7 @@
  *   r7 - End of kernel command line string
  *
  */
-	.section	.text.head, "ax"
+	.section	.text..head, "ax"
 _ENTRY(_stext);
 _ENTRY(_start);
 	/*
arch/powerpc/kernel/init_task.c +1 -1
index 688b329..1162c3c 100644
@@ -21,7 +21,7 @@ EXPORT_SYMBOL(init_mm);
  * "init_task" linker map entry..
  */
 union thread_union init_thread_union 
-	__attribute__((__section__(".data.init_task"))) =
+	__attribute__((__section__(".data..init_task"))) =
 		{ INIT_THREAD_INFO(init_task) };
 
 /*
arch/powerpc/kernel/machine_kexec_64.c +1 -1
index 49e705f..815a13c 100644
@@ -250,7 +250,7 @@ static void kexec_prepare_cpus(void)
  * current, but that audit has not been performed.
  */
 static union thread_union kexec_stack
-	__attribute__((__section__(".data.init_task"))) = { };
+	__attribute__((__section__(".data..init_task"))) = { };
 
 /* Our assembly helper, in kexec_stub.S */
 extern NORET_TYPE void kexec_sequence(void *newstack, unsigned long start,
arch/powerpc/kernel/vdso.c +1 -1
index ad06d5c..6685af8 100644
@@ -74,7 +74,7 @@ static int vdso_ready;
 static union {
 	struct vdso_data	data;
 	u8			page[PAGE_SIZE];
-} vdso_data_store __attribute__((__section__(".data.page_aligned")));
+} vdso_data_store __attribute__((__section__(".data..page_aligned")));
 struct vdso_data *vdso_data = &vdso_data_store.data;
 
 /* Format of the patch table */
arch/powerpc/kernel/vdso32/vdso32_wrapper.S +1 -1
index 556f0ca..10f61ac 100644
@@ -1,7 +1,7 @@
 #include <linux/init.h>
 #include <asm/page.h>
 
-	.section ".data.page_aligned"
+	.section ".data..page_aligned"
 
 	.globl vdso32_start, vdso32_end
 	.balign PAGE_SIZE
arch/powerpc/kernel/vdso64/vdso64_wrapper.S +1 -1
index 0529cb9..3c1cc59 100644
@@ -1,7 +1,7 @@
 #include <linux/init.h>
 #include <asm/page.h>
 
-	.section ".data.page_aligned"
+	.section ".data..page_aligned"
 
 	.globl vdso64_start, vdso64_end
 	.balign PAGE_SIZE
arch/powerpc/kernel/vmlinux.lds.S +11 -11
index b9ef164..cedaad0 100644
@@ -52,9 +52,9 @@ SECTIONS
 	/* Text and gots */
 	.text : AT(ADDR(.text) - LOAD_OFFSET) {
 		ALIGN_FUNCTION();
-		*(.text.head)
+		*(.text..head)
 		_text = .;
-		*(.text .fixup .text.init.refok .exit.text.refok __ftr_alt_*)
+		*(.text .fixup .text..init.refok .text..exit.refok __ftr_alt_*)
 		SCHED_TEXT
 		LOCK_TEXT
 		KPROBES_TEXT
@@ -254,28 +254,28 @@ SECTIONS
 #else
 	. = ALIGN(16384);
 #endif
-	.data.init_task : AT(ADDR(.data.init_task) - LOAD_OFFSET) {
-		*(.data.init_task)
+	.data..init_task : AT(ADDR(.data..init_task) - LOAD_OFFSET) {
+		*(.data..init_task)
 	}
 
 	. = ALIGN(PAGE_SIZE);
-	.data.page_aligned : AT(ADDR(.data.page_aligned) - LOAD_OFFSET) {
-		*(.data.page_aligned)
+	.data..page_aligned : AT(ADDR(.data..page_aligned) - LOAD_OFFSET) {
+		*(.data..page_aligned)
 	}
 
-	.data.cacheline_aligned : AT(ADDR(.data.cacheline_aligned) - LOAD_OFFSET) {
-		*(.data.cacheline_aligned)
+	.data..cacheline_aligned : AT(ADDR(.data..cacheline_aligned) - LOAD_OFFSET) {
+		*(.data..cacheline_aligned)
 	}
 
 	. = ALIGN(L1_CACHE_BYTES);
-	.data.read_mostly : AT(ADDR(.data.read_mostly) - LOAD_OFFSET) {
-		*(.data.read_mostly)
+	.data..read_mostly : AT(ADDR(.data..read_mostly) - LOAD_OFFSET) {
+		*(.data..read_mostly)
 	}
 
 	. = ALIGN(PAGE_SIZE);
 	.data_nosave : AT(ADDR(.data_nosave) - LOAD_OFFSET) {
 		__nosave_begin = .;
-		*(.data.nosave)
+		*(.data..nosave)
 		. = ALIGN(PAGE_SIZE);
 		__nosave_end = .;
 	}
arch/s390/include/asm/cache.h +1 -1
index 9b86681..24aafa6 100644
@@ -14,6 +14,6 @@
 #define L1_CACHE_BYTES     256
 #define L1_CACHE_SHIFT     8
 
-#define __read_mostly __attribute__((__section__(".data.read_mostly")))
+#define __read_mostly __attribute__((__section__(".data..read_mostly")))
 
 #endif
arch/s390/kernel/head.S +1 -1
index bba1449..e899eda 100644
@@ -35,7 +35,7 @@
 #define ARCH_OFFSET	0
 #endif
 
-.section ".text.head","ax"
+.section ".text..head","ax"
 #ifndef CONFIG_IPL
 	.org   0
 	.long  0x00080000,0x80000000+startup	# Just a restart PSW
arch/s390/kernel/init_task.c +1 -1
index 7db95c0..7cfd9af 100644
@@ -30,7 +30,7 @@ EXPORT_SYMBOL(init_mm);
  * "init_task" linker map entry..
  */
 union thread_union init_thread_union 
-	__attribute__((__section__(".data.init_task"))) =
+	__attribute__((__section__(".data..init_task"))) =
 		{ INIT_THREAD_INFO(init_task) };
 
 /*
arch/s390/kernel/vdso.c +1 -1
index 89b2e7f..e4c0609 100644
@@ -64,7 +64,7 @@ __setup("vdso=", vdso_setup);
 static union {
 	struct vdso_data	data;
 	u8			page[PAGE_SIZE];
-} vdso_data_store __attribute__((__section__(".data.page_aligned")));
+} vdso_data_store __attribute__((__section__(".data..page_aligned")));
 struct vdso_data *vdso_data = &vdso_data_store.data;
 
 /*
arch/s390/kernel/vdso32/vdso32_wrapper.S +1 -1
index 61639a8..025cb77 100644
@@ -1,7 +1,7 @@
 #include <linux/init.h>
 #include <asm/page.h>
 
-	.section ".data.page_aligned"
+	.section ".data..page_aligned"
 
 	.globl vdso32_start, vdso32_end
 	.balign PAGE_SIZE
arch/s390/kernel/vdso64/vdso64_wrapper.S +1 -1
index d8e2ac1..d26126d 100644
@@ -1,7 +1,7 @@
 #include <linux/init.h>
 #include <asm/page.h>
 
-	.section ".data.page_aligned"
+	.section ".data..page_aligned"
 
 	.globl vdso64_start, vdso64_end
 	.balign PAGE_SIZE
arch/s390/kernel/vmlinux.lds.S +10 -10
index 7a2063e..c1f97b1 100644
@@ -29,7 +29,7 @@ SECTIONS
 	. = 0x00000000;
 	.text : {
 	_text = .;		/* Text and read-only data */
-		*(.text.head)
+		*(.text..head)
 	TEXT_TEXT
 		SCHED_TEXT
 		LOCK_TEXT
@@ -66,30 +66,30 @@ SECTIONS
 	. = ALIGN(PAGE_SIZE);
 	.data_nosave : {
 	__nosave_begin = .;
-		*(.data.nosave)
+		*(.data..nosave)
 	}
 	. = ALIGN(PAGE_SIZE);
 	__nosave_end = .;
 
 	. = ALIGN(PAGE_SIZE);
-	.data.page_aligned : {
-		*(.data.idt)
+	.data..page_aligned : {
+		*(.data..idt)
 	}
 
 	. = ALIGN(0x100);
-	.data.cacheline_aligned : {
-		*(.data.cacheline_aligned)
+	.data..cacheline_aligned : {
+		*(.data..cacheline_aligned)
 	}
 
 	. = ALIGN(0x100);
-	.data.read_mostly : {
-		*(.data.read_mostly)
+	.data..read_mostly : {
+		*(.data..read_mostly)
 	}
 	_edata = .;		/* End of data section */
 
 	. = ALIGN(THREAD_SIZE);	/* init_task */
-	.data.init_task : {
-		*(.data.init_task)
+	.data..init_task : {
+		*(.data..init_task)
 	}
 
 	/* will be freed after init */
arch/sh/include/asm/cache.h +1 -1
index 02df18e..455a9a9 100644
@@ -14,7 +14,7 @@
 
 #define L1_CACHE_BYTES		(1 << L1_CACHE_SHIFT)
 
-#define __read_mostly __attribute__((__section__(".data.read_mostly")))
+#define __read_mostly __attribute__((__section__(".data..read_mostly")))
 
 #ifndef __ASSEMBLY__
 struct cache_info {
arch/sh/kernel/cpu/sh5/entry.S +2 -2
index e640c63..433de32 100644
@@ -2058,10 +2058,10 @@ asm_uaccess_end:
 
 
 /*
- * --- .text.init Section
+ * --- .text..init Section
  */
 
-	.section	.text.init, "ax"
+	.section	.text..init, "ax"
 
 /*
  * void trap_init (void)
arch/sh/kernel/head_32.S +1 -1
index 788605f..4065946 100644
@@ -40,7 +40,7 @@ ENTRY(empty_zero_page)
 1:
 	.skip	PAGE_SIZE - empty_zero_page - 1b
 
-	.section	.text.head, "ax"
+	.section	.text..head, "ax"
 
 /*
  * Condition at the entry of _stext:
arch/sh/kernel/head_64.S +1 -1
index 7ccfb99..c3543d1 100644
@@ -110,7 +110,7 @@ empty_bad_pte_table:
 fpu_in_use:	.quad	0
 
 
-	.section	.text.head, "ax"
+	.section	.text..head, "ax"
 	.balign L1_CACHE_BYTES
 /*
  * Condition at the entry of __stext:
arch/sh/kernel/init_task.c +1 -1
index 80c35ff..af29479 100644
@@ -21,7 +21,7 @@ EXPORT_SYMBOL(init_mm);
  * "init_task" linker map entry..
  */
 union thread_union init_thread_union
-	__attribute__((__section__(".data.init_task"))) =
+	__attribute__((__section__(".data..init_task"))) =
 		{ INIT_THREAD_INFO(init_task) };
 
 /*
arch/sh/kernel/irq.c +2 -2
index 3f1372e..63cd9a0 100644
@@ -158,10 +158,10 @@ asmlinkage int do_IRQ(unsigned int irq, struct pt_regs *regs)
 
 #ifdef CONFIG_IRQSTACKS
 static char softirq_stack[NR_CPUS * THREAD_SIZE]
-		__attribute__((__section__(".bss.page_aligned")));
+		__attribute__((__section__(".bss..page_aligned")));
 
 static char hardirq_stack[NR_CPUS * THREAD_SIZE]
-		__attribute__((__section__(".bss.page_aligned")));
+		__attribute__((__section__(".bss..page_aligned")));
 
 /*
  * allocate per-cpu stacks for hardirq and for softirq processing
arch/sh/kernel/vmlinux_32.lds.S +7 -7
index d0b2a71..a7b1be0 100644
@@ -31,7 +31,7 @@ SECTIONS
 	} = 0
 
 	.text : {
-		*(.text.head)
+		*(.text..head)
 		TEXT_TEXT
 		SCHED_TEXT
 		LOCK_TEXT
@@ -61,19 +61,19 @@ SECTIONS
 
 	. = ALIGN(THREAD_SIZE);
 	.data : {			/* Data */
-		*(.data.init_task)
+		*(.data..init_task)
 
 		. = ALIGN(L1_CACHE_BYTES);
-		*(.data.cacheline_aligned)
+		*(.data..cacheline_aligned)
 
 		. = ALIGN(L1_CACHE_BYTES);
-		*(.data.read_mostly)
+		*(.data..read_mostly)
 
 		. = ALIGN(PAGE_SIZE);
-		*(.data.page_aligned)
+		*(.data..page_aligned)
 
 		__nosave_begin = .;
-		*(.data.nosave)
+		*(.data..nosave)
 		. = ALIGN(PAGE_SIZE);
 		__nosave_end = .;
 
@@ -131,7 +131,7 @@ SECTIONS
 	.bss : {
 		__init_end = .;
 		__bss_start = .;		/* BSS */
-		*(.bss.page_aligned)
+		*(.bss..page_aligned)
 		*(.bss)
 		*(COMMON)
 		. = ALIGN(4);
arch/sh/kernel/vmlinux_64.lds.S +7 -7
index 33fa464..c7d5f5b 100644
@@ -42,7 +42,7 @@ SECTIONS
 	} = 0
 
 	.text : C_PHYS(.text) {
-		*(.text.head)
+		*(.text..head)
 		TEXT_TEXT
 		*(.text64)
 		*(.text..SHmedia32)
@@ -70,19 +70,19 @@ SECTIONS
 
 	. = ALIGN(THREAD_SIZE);
 	.data : C_PHYS(.data) {			/* Data */
-		*(.data.init_task)
+		*(.data..init_task)
 
 		. = ALIGN(L1_CACHE_BYTES);
-		*(.data.cacheline_aligned)
+		*(.data..cacheline_aligned)
 
 		. = ALIGN(L1_CACHE_BYTES);
-		*(.data.read_mostly)
+		*(.data..read_mostly)
 
 		. = ALIGN(PAGE_SIZE);
-		*(.data.page_aligned)
+		*(.data..page_aligned)
 
 		__nosave_begin = .;
-		*(.data.nosave)
+		*(.data..nosave)
 		. = ALIGN(PAGE_SIZE);
 		__nosave_end = .;
 
@@ -140,7 +140,7 @@ SECTIONS
 	.bss : C_PHYS(.bss) {
 		__init_end = .;
 		__bss_start = .;		/* BSS */
-		*(.bss.page_aligned)
+		*(.bss..page_aligned)
 		*(.bss)
 		*(COMMON)
 		. = ALIGN(4);
arch/sparc/boot/btfixupprep.c +2 -2
index 52a4208..e899f2a 100644
@@ -171,7 +171,7 @@ main1:
 			}
 		} else if (buffer[nbase+4] != '_')
 			continue;
-		if (!strcmp (sect, ".text.exit"))
+		if (!strcmp (sect, ".text..exit"))
 			continue;
 		if (strcmp (sect, ".text") &&
 		    strcmp (sect, ".init.text") &&
@@ -325,7 +325,7 @@ main1:
 		(*rr)->next = NULL;
 	}
 	printf("! Generated by btfixupprep. Do not edit.\n\n");
-	printf("\t.section\t\".data.init\",#alloc,#write\n\t.align\t4\n\n");
+	printf("\t.section\t\".data..init\",#alloc,#write\n\t.align\t4\n\n");
 	printf("\t.global\t___btfixup_start\n___btfixup_start:\n\n");
 	for (i = 0; i < last; i++) {
 		f = array + i;
arch/sparc/include/asm/cache.h +1 -1
index 41f85ae..2909f0a 100644
@@ -19,7 +19,7 @@
 
 #define SMP_CACHE_BYTES (1 << SMP_CACHE_BYTES_SHIFT)
 
-#define __read_mostly __attribute__((__section__(".data.read_mostly")))
+#define __read_mostly __attribute__((__section__(".data..read_mostly")))
 
 #ifdef CONFIG_SPARC32
 #include <asm/asi.h>
arch/sparc/kernel/head_32.S +2 -2
index f0b4b51..0ca3dc5 100644
@@ -72,7 +72,7 @@ sun4e_notsup:
 	.align 4
 
 	/* The Sparc trap table, bootloader gives us control at _start. */
-	.section .text.head,"ax"
+	.section .text..head,"ax"
 	.globl	start, _stext, _start, __stext
 	.globl  trapbase
 _start:   /* danger danger */
@@ -735,7 +735,7 @@ go_to_highmem:
 		 nop
 
 /* The code above should be at beginning and we have to take care about
- * short jumps, as branching to .text.init section from .text is usually
+ * short jumps, as branching to .text..init section from .text is usually
  * impossible */
 		__INIT
 /* Acquire boot time privileged register values, this will help debugging.
arch/sparc/kernel/head_64.S +1 -1
index 3a1b7bf..709353f 100644
@@ -467,7 +467,7 @@ jump_to_sun4u_init:
 	jmpl    %g2 + %g0, %g0
 	 nop
 
-	.section	.text.init.refok
+	.section	.text..init.refok
 sun4u_init:
 	BRANCH_IF_SUN4V(g1, sun4v_init)
 
arch/sparc/kernel/init_task.c +1 -1
index f28cb82..7f7a468 100644
@@ -22,5 +22,5 @@ EXPORT_SYMBOL(init_task);
  * in etrap.S which assumes it.
  */
 union thread_union init_thread_union
-	__attribute__((section (".data.init_task")))
+	__attribute__((section (".data..init_task")))
 	= { INIT_THREAD_INFO(init_task) };
arch/sparc/kernel/vmlinux.lds.S +7 -7
index 7626708..66eddbe 100644
@@ -41,7 +41,7 @@ SECTIONS
 	.text TEXTSTART :
 	{
 		_text = .;
-		*(.text.head)
+		*(.text..head)
 		TEXT_TEXT
 		SCHED_TEXT
 		LOCK_TEXT
@@ -59,20 +59,20 @@ SECTIONS
 		*(.data1)
 	}
 	. = ALIGN(SMP_CACHE_BYTES);
-	.data.cacheline_aligned : {
-		*(.data.cacheline_aligned)
+	.data..cacheline_aligned : {
+		*(.data..cacheline_aligned)
 	}
 	. = ALIGN(SMP_CACHE_BYTES);
-	.data.read_mostly : {
-		*(.data.read_mostly)
+	.data..read_mostly : {
+		*(.data..read_mostly)
 	}
 	/* End of data section */
 	_edata = .;
 
 	/* init_task */
 	. = ALIGN(THREAD_SIZE);
-	.data.init_task : {
-		*(.data.init_task)
+	.data..init_task : {
+		*(.data..init_task)
 	}
 	.fixup : {
 		__start___fixup = .;
arch/um/include/asm/common.lds.S +2 -2
index cb02486..b623606 100644
@@ -49,9 +49,9 @@
   }
 
   . = ALIGN(32);
-  .data.percpu : {
+  .data..percpu : {
 	__per_cpu_start = . ;
-	*(.data.percpu)
+	*(.data..percpu)
 	__per_cpu_end = . ;
   }
 	
arch/um/kernel/dyn.lds.S +2 -2
index 9975e1a..a415658 100644
@@ -97,9 +97,9 @@ SECTIONS
   .fini_array     : { *(.fini_array) }
   .data           : {
     . = ALIGN(KERNEL_STACK_SIZE);		/* init_task */
-    *(.data.init_task)
+    *(.data..init_task)
     . = ALIGN(KERNEL_STACK_SIZE);
-    *(.data.init_irqstack)
+    *(.data..init_irqstack)
     DATA_DATA
     *(.data.* .gnu.linkonce.d.*)
     SORT(CONSTRUCTORS)
arch/um/kernel/init_task.c +2 -2
index 806d381..eed073e 100644
@@ -34,9 +34,9 @@ EXPORT_SYMBOL(init_task);
  */
 
 union thread_union init_thread_union
-	__attribute__((__section__(".data.init_task"))) =
+	__attribute__((__section__(".data..init_task"))) =
 		{ INIT_THREAD_INFO(init_task) };
 
 union thread_union cpu0_irqstack
-	__attribute__((__section__(".data.init_irqstack"))) =
+	__attribute__((__section__(".data..init_irqstack"))) =
 		{ INIT_THREAD_INFO(init_task) };
arch/um/kernel/uml.lds.S +2 -2
index 11b8352..7b03348 100644
@@ -53,9 +53,9 @@ SECTIONS
   .data    :
   {
     . = ALIGN(KERNEL_STACK_SIZE);		/* init_task */
-    *(.data.init_task)
+    *(.data..init_task)
     . = ALIGN(KERNEL_STACK_SIZE);
-    *(.data.init_irqstack)
+    *(.data..init_irqstack)
     DATA_DATA
     *(.gnu.linkonce.d*)
     CONSTRUCTORS
arch/x86/boot/compressed/head_32.S +1 -1
index 3a8a866..b755fb8 100644
@@ -29,7 +29,7 @@
 #include <asm/boot.h>
 #include <asm/asm-offsets.h>
 
-.section ".text.head","ax",@progbits
+.section ".text..head","ax",@progbits
 ENTRY(startup_32)
 	cld
 	/* test KEEP_SEGMENTS flag to see if the bootloader is asking
arch/x86/boot/compressed/head_64.S +1 -1
index ed4a829..779e06d 100644
@@ -33,7 +33,7 @@
 #include <asm/processor-flags.h>
 #include <asm/asm-offsets.h>
 
-.section ".text.head"
+.section ".text..head"
 	.code32
 ENTRY(startup_32)
 	cld
arch/x86/boot/compressed/relocs.c +1 -1
index 857e492..af8b9c5 100644
@@ -559,7 +559,7 @@ static void emit_relocs(int as_text)
 		/* Print the relocations in a form suitable that
 		 * gas will like.
 		 */
-		printf(".section \".data.reloc\",\"a\"\n");
+		printf(".section \".data..reloc\",\"a\"\n");
 		printf(".balign 4\n");
 		for (i = 0; i < reloc_count; i++) {
 			printf("\t .long 0x%08lx\n", relocs[i]);
arch/x86/boot/compressed/vmlinux.scr +1 -1
index f02382a..862d748 100644
@@ -1,6 +1,6 @@
 SECTIONS
 {
-  .rodata.compressed : {
+  .rodata..compressed : {
 	input_len = .;
 	LONG(input_data_end - input_data) input_data = .;
 	*(.data)
arch/x86/boot/compressed/vmlinux_32.lds +10 -4
index bb3c483..d70318a 100644
@@ -7,13 +7,13 @@ SECTIONS
 	 * address 0.
 	 */
 	. = 0;
-	.text.head : {
+	.text..head : {
 		_head = . ;
-		*(.text.head)
+		*(.text..head)
 		_ehead = . ;
 	}
-	.rodata.compressed : {
-		*(.rodata.compressed)
+	.rodata..compressed : {
+		*(.rodata..compressed)
 	}
 	.text :	{
 		_text = .; 	/* Text */
@@ -21,6 +21,10 @@ SECTIONS
 		*(.text.*)
 		_etext = . ;
 	}
+	.got : {
+		*(.got)
+		*(.got.*)
+	}
 	.rodata : {
 		_rodata = . ;
 		*(.rodata)	 /* read-only data */
@@ -40,4 +44,6 @@ SECTIONS
 		*(COMMON)
 		_end = . ;
 	}
+	/* Be bold, and discard everything not explicitly mentioned */
+	/DISCARD/ : { *(*) }
 }
arch/x86/boot/compressed/vmlinux_64.lds +6 -4
index bef1ac8..d3d1468 100644
@@ -7,13 +7,13 @@ SECTIONS
 	 * address 0.
 	 */
 	. = 0;
-	.text.head : {
+	.text..head : {
 		_head = . ;
-		*(.text.head)
+		*(.text..head)
 		_ehead = . ;
 	}
-	.rodata.compressed : {
-		*(.rodata.compressed)
+	.rodata..compressed : {
+		*(.rodata..compressed)
 	}
 	.text :	{
 		_text = .; 	/* Text */
@@ -45,4 +45,6 @@ SECTIONS
 		. = . + 4096 * 6;
 		_ebss = .;
 	}
+	/* Be bold, and discard everything not explicitly mentioned */
+	/DISCARD/ : { *(*) }
 }
arch/x86/include/asm/cache.h +2 -2
index 5d367ca..6b94d49 100644
@@ -5,7 +5,7 @@
 #define L1_CACHE_SHIFT	(CONFIG_X86_L1_CACHE_SHIFT)
 #define L1_CACHE_BYTES	(1 << L1_CACHE_SHIFT)
 
-#define __read_mostly __attribute__((__section__(".data.read_mostly")))
+#define __read_mostly __attribute__((__section__(".data..read_mostly")))
 
 #ifdef CONFIG_X86_VSMP
 /* vSMP Internode cacheline shift */
@@ -13,7 +13,7 @@
 #ifdef CONFIG_SMP
 #define __cacheline_aligned_in_smp					\
 	__attribute__((__aligned__(1 << (INTERNODE_CACHE_SHIFT))))	\
-	__attribute__((__section__(".data.page_aligned")))
+	__attribute__((__section__(".data..page_aligned")))
 #endif
 #endif
 
arch/x86/kernel/acpi/wakeup_32.S +1 -1
index 8ded418..13ab720 100644
@@ -1,4 +1,4 @@
-	.section .text.page_aligned
+	.section .text..page_aligned
 #include <linux/linkage.h>
 #include <asm/segment.h>
 #include <asm/page_types.h>
arch/x86/kernel/head_32.S +3 -3
index 3068388..4a4bfef 100644
@@ -80,7 +80,7 @@ RESERVE_BRK(pagetables, INIT_MAP_SIZE)
  * any particular GDT layout, because we load our own as soon as we
  * can.
  */
-.section .text.head,"ax",@progbits
+.section .text..head,"ax",@progbits
 ENTRY(startup_32)
 	/* test KEEP_SEGMENTS flag to see if the bootloader is asking
 		us to not reload segments */
@@ -618,7 +618,7 @@ ENTRY(_stext)
 /*
  * BSS section
  */
-.section ".bss.page_aligned","wa"
+.section ".bss..page_aligned","wa"
 	.align PAGE_SIZE_asm
 #ifdef CONFIG_X86_PAE
 swapper_pg_pmd:
@@ -636,7 +636,7 @@ ENTRY(empty_zero_page)
  * This starts the data section.
  */
 #ifdef CONFIG_X86_PAE
-.section ".data.page_aligned","wa"
+.section ".data..page_aligned","wa"
 	/* Page-aligned for the benefit of paravirt? */
 	.align PAGE_SIZE_asm
 ENTRY(swapper_pg_dir)
arch/x86/kernel/head_64.S +2 -2
index 54b29bb..05b8013 100644
@@ -41,7 +41,7 @@ L4_START_KERNEL = pgd_index(__START_KERNEL_map)
 L3_START_KERNEL = pud_index(__START_KERNEL_map)
 
 	.text
-	.section .text.head
+	.section .text..head
 	.code64
 	.globl startup_64
 startup_64:
@@ -419,7 +419,7 @@ ENTRY(phys_base)
 ENTRY(idt_table)
 	.skip IDT_ENTRIES * 16
 
-	.section .bss.page_aligned, "aw", @nobits
+	.section .bss..page_aligned, "aw", @nobits
 	.align PAGE_SIZE
 ENTRY(empty_zero_page)
 	.skip PAGE_SIZE
arch/x86/kernel/init_task.c +2 -2
index df3bf26..103b1c5 100644
@@ -22,7 +22,7 @@ struct mm_struct init_mm = INIT_MM(init_mm);
  * "init_task" linker map entry..
  */
 union thread_union init_thread_union
-	__attribute__((__section__(".data.init_task"))) =
+	__attribute__((__section__(".data..init_task"))) =
 		{ INIT_THREAD_INFO(init_task) };
 
 /*
@@ -36,7 +36,7 @@ EXPORT_SYMBOL(init_task);
 /*
  * per-CPU TSS segments. Threads are completely 'soft' on Linux,
  * no more per-task TSS's. The TSS size is kept cacheline-aligned
- * so they are allowed to end up in the .data.cacheline_aligned
+ * so they are allowed to end up in the .data..cacheline_aligned
  * section. Since TSS's are completely CPU-local, we want them
  * on exact cacheline boundaries, to eliminate cacheline ping-pong.
  */
arch/x86/kernel/setup_percpu.c +1 -1
index 3a97a4c..9548f29 100644
@@ -407,7 +407,7 @@ void __init setup_per_cpu_areas(void)
 #endif
 #endif
 		/*
-		 * Up to this point, the boot CPU has been using .data.init
+		 * Up to this point, the boot CPU has been using .data..init
 		 * area.  Reload any changed state for the boot CPU.
 		 */
 		if (cpu == boot_cpu_id)
arch/x86/kernel/traps.c +1 -1
index a1d2883..2a88259 100644
@@ -77,7 +77,7 @@ char ignore_fpu_irq;
  * for this.
  */
 gate_desc idt_table[256]
-	__attribute__((__section__(".data.idt"))) = { { { { 0, 0 } } }, };
+	__attribute__((__section__(".data..idt"))) = { { { { 0, 0 } } }, };
 #endif
 
 DECLARE_BITMAP(used_vectors, NR_VECTORS);
arch/x86/kernel/vmlinux_32.lds.S +14 -14
index 62ad500..5342ea8 100644
@@ -31,15 +31,15 @@ SECTIONS
   . = LOAD_OFFSET + LOAD_PHYSICAL_ADDR;
   phys_startup_32 = startup_32 - LOAD_OFFSET;
 
-  .text.head : AT(ADDR(.text.head) - LOAD_OFFSET) {
+  .text..head : AT(ADDR(.text..head) - LOAD_OFFSET) {
   	_text = .;			/* Text and read-only data */
-	*(.text.head)
+	*(.text..head)
   } :text = 0x9090
 
   /* read-only */
   .text : AT(ADDR(.text) - LOAD_OFFSET) {
 	. = ALIGN(PAGE_SIZE); /* not really needed, already page aligned */
-	*(.text.page_aligned)
+	*(.text..page_aligned)
 	TEXT_TEXT
 	SCHED_TEXT
 	LOCK_TEXT
@@ -71,32 +71,32 @@ SECTIONS
   . = ALIGN(PAGE_SIZE);
   .data_nosave : AT(ADDR(.data_nosave) - LOAD_OFFSET) {
   	__nosave_begin = .;
-	*(.data.nosave)
+	*(.data..nosave)
   	. = ALIGN(PAGE_SIZE);
   	__nosave_end = .;
   }
 
   . = ALIGN(PAGE_SIZE);
-  .data.page_aligned : AT(ADDR(.data.page_aligned) - LOAD_OFFSET) {
-	*(.data.page_aligned)
-	*(.data.idt)
+  .data..page_aligned : AT(ADDR(.data..page_aligned) - LOAD_OFFSET) {
+	*(.data..page_aligned)
+	*(.data..idt)
   }
 
   . = ALIGN(32);
-  .data.cacheline_aligned : AT(ADDR(.data.cacheline_aligned) - LOAD_OFFSET) {
-	*(.data.cacheline_aligned)
+  .data..cacheline_aligned : AT(ADDR(.data..cacheline_aligned) - LOAD_OFFSET) {
+	*(.data..cacheline_aligned)
   }
 
   /* rarely changed data like cpu maps */
   . = ALIGN(32);
-  .data.read_mostly : AT(ADDR(.data.read_mostly) - LOAD_OFFSET) {
-	*(.data.read_mostly)
+  .data..read_mostly : AT(ADDR(.data..read_mostly) - LOAD_OFFSET) {
+	*(.data..read_mostly)
 	_edata = .;		/* End of data section */
   }
 
   . = ALIGN(THREAD_SIZE);	/* init_task */
-  .data.init_task : AT(ADDR(.data.init_task) - LOAD_OFFSET) {
-	*(.data.init_task)
+  .data..init_task : AT(ADDR(.data..init_task) - LOAD_OFFSET) {
+	*(.data..init_task)
   }
 
   /* might get freed after init */
@@ -185,7 +185,7 @@ SECTIONS
   .bss : AT(ADDR(.bss) - LOAD_OFFSET) {
 	__init_end = .;
 	__bss_start = .;		/* BSS */
-	*(.bss.page_aligned)
+	*(.bss..page_aligned)
 	*(.bss)
 	. = ALIGN(4);
 	__bss_stop = .;
arch/x86/kernel/vmlinux_64.lds.S +13 -13
index c874250..98016a2 100644
@@ -32,7 +32,7 @@ SECTIONS
   .text :  AT(ADDR(.text) - LOAD_OFFSET) {
 	_text = .;			/* Text and read-only data */
 	/* First the code that has to be first for bootstrapping */
-	*(.text.head)
+	*(.text..head)
 	_stext = .;
 	/* Then the rest */
 	TEXT_TEXT
@@ -65,19 +65,19 @@ SECTIONS
 	} :data
 
 
-  .data.cacheline_aligned : AT(ADDR(.data.cacheline_aligned) - LOAD_OFFSET) {
+  .data..cacheline_aligned : AT(ADDR(.data..cacheline_aligned) - LOAD_OFFSET) {
 	. = ALIGN(PAGE_SIZE);
 	. = ALIGN(CONFIG_X86_L1_CACHE_BYTES);
-	*(.data.cacheline_aligned)
+	*(.data..cacheline_aligned)
   }
   . = ALIGN(CONFIG_X86_INTERNODE_CACHE_BYTES);
-  .data.read_mostly : AT(ADDR(.data.read_mostly) - LOAD_OFFSET) {
-  	*(.data.read_mostly)
+  .data..read_mostly : AT(ADDR(.data..read_mostly) - LOAD_OFFSET) {
+	*(.data..read_mostly)
   }
 
 #define VSYSCALL_ADDR (-10*1024*1024)
-#define VSYSCALL_PHYS_ADDR ((LOADADDR(.data.read_mostly) + SIZEOF(.data.read_mostly) + 4095) & ~(4095))
-#define VSYSCALL_VIRT_ADDR ((ADDR(.data.read_mostly) + SIZEOF(.data.read_mostly) + 4095) & ~(4095))
+#define VSYSCALL_PHYS_ADDR ((LOADADDR(.data..read_mostly) + SIZEOF(.data..read_mostly) + 4095) & ~(4095))
+#define VSYSCALL_VIRT_ADDR ((ADDR(.data..read_mostly) + SIZEOF(.data..read_mostly) + 4095) & ~(4095))
 
 #define VLOAD_OFFSET (VSYSCALL_ADDR - VSYSCALL_PHYS_ADDR)
 #define VLOAD(x) (ADDR(x) - VLOAD_OFFSET)
@@ -125,14 +125,14 @@ SECTIONS
 #undef VVIRT_OFFSET
 #undef VVIRT
 
-  .data.init_task : AT(ADDR(.data.init_task) - LOAD_OFFSET) {
+  .data..init_task : AT(ADDR(.data..init_task) - LOAD_OFFSET) {
 	. = ALIGN(THREAD_SIZE);	/* init_task */
-	*(.data.init_task)
+	*(.data..init_task)
   }:data.init
 
-  .data.page_aligned : AT(ADDR(.data.page_aligned) - LOAD_OFFSET) {
+  .data..page_aligned : AT(ADDR(.data..page_aligned) - LOAD_OFFSET) {
 	. = ALIGN(PAGE_SIZE);
-	*(.data.page_aligned)
+	*(.data..page_aligned)
   }
 
   .smp_locks : AT(ADDR(.smp_locks) - LOAD_OFFSET) {
@@ -236,7 +236,7 @@ SECTIONS
   .data_nosave : AT(ADDR(.data_nosave) - LOAD_OFFSET) {
 	. = ALIGN(PAGE_SIZE);
 	__nosave_begin = .;
-	*(.data.nosave)
+	*(.data..nosave)
 	. = ALIGN(PAGE_SIZE);
 	__nosave_end = .;
   } :data.init2 /* use another section data.init2, see PERCPU_VADDR() above */
@@ -244,7 +244,7 @@ SECTIONS
   .bss : AT(ADDR(.bss) - LOAD_OFFSET) {
 	. = ALIGN(PAGE_SIZE);
 	__bss_start = .;		/* BSS */
-	*(.bss.page_aligned)
+	*(.bss..page_aligned)
 	*(.bss)
 	__bss_stop = .;
   }
arch/xtensa/kernel/head.S +1 -1
index 0817f9d..5e093b2 100644
@@ -234,7 +234,7 @@ should_never_return:
  * BSS section
  */
 	
-.section ".bss.page_aligned", "w"
+.section ".bss..page_aligned", "w"
 #ifdef CONFIG_MMU
 ENTRY(swapper_pg_dir)
 	.fill	PAGE_SIZE, 1, 0
arch/xtensa/kernel/init_task.c +1 -1
index e07f5c9..d07db9b 100644
@@ -28,7 +28,7 @@ struct mm_struct init_mm = INIT_MM(init_mm);
 EXPORT_SYMBOL(init_mm);
 
 union thread_union init_thread_union
-	__attribute__((__section__(".data.init_task"))) =
+	__attribute__((__section__(".data..init_task"))) =
 { INIT_THREAD_INFO(init_task) };
 
 struct task_struct init_task = INIT_TASK(init_task);
arch/xtensa/kernel/vmlinux.lds.S +3 -3
index c1be9a4..4c1cfc9 100644
@@ -124,14 +124,14 @@ SECTIONS
     DATA_DATA
     CONSTRUCTORS
     . = ALIGN(XCHAL_ICACHE_LINESIZE);
-    *(.data.cacheline_aligned)
+    *(.data..cacheline_aligned)
   }
 
   _edata = .;
 
   /* The initial task */
   . = ALIGN(8192);
-  .data.init_task : { *(.data.init_task) }
+  .data..init_task : { *(.data..init_task) }
 
   /* Initialization code and data: */
 
@@ -262,7 +262,7 @@ SECTIONS
 
   /* BSS section */
   _bss_start = .;
-  .bss : { *(.bss.page_aligned) *(.bss) }
+  .bss : { *(.bss..page_aligned) *(.bss) }
   _bss_end = .;
 
   _end = .;
include/asm-generic/percpu.h +1 -1
index d7d50d7..1ce7788 100644
@@ -71,7 +71,7 @@ extern void setup_per_cpu_areas(void);
 
 #ifndef PER_CPU_BASE_SECTION
 #ifdef CONFIG_SMP
-#define PER_CPU_BASE_SECTION ".data.percpu"
+#define PER_CPU_BASE_SECTION ".data..percpu"
 #else
 #define PER_CPU_BASE_SECTION ".data"
 #endif
include/asm-generic/vmlinux.lds.h +15 -15
index 7fa660f..37feab0 100644
@@ -88,7 +88,7 @@
 /* .data section */
 #define DATA_DATA							\
 	*(.data)							\
-	*(.data.init.refok)						\
+	*(.data..init.refok)						\
 	*(.ref.data)							\
 	DEV_KEEP(init.data)						\
 	DEV_KEEP(exit.data)						\
@@ -287,8 +287,8 @@
 		*(.text.hot)						\
 		*(.text)						\
 		*(.ref.text)						\
-		*(.text.init.refok)					\
-		*(.exit.text.refok)					\
+		*(.text..init.refok)					\
+		*(.text..exit.refok)					\
 	DEV_KEEP(init.text)						\
 	DEV_KEEP(exit.text)						\
 	CPU_KEEP(init.text)						\
@@ -475,16 +475,16 @@
  */
 #define PERCPU_VADDR(vaddr, phdr)					\
 	VMLINUX_SYMBOL(__per_cpu_load) = .;				\
-	.data.percpu vaddr : AT(VMLINUX_SYMBOL(__per_cpu_load)		\
+	.data..percpu vaddr : AT(VMLINUX_SYMBOL(__per_cpu_load)		\
 				- LOAD_OFFSET) {			\
 		VMLINUX_SYMBOL(__per_cpu_start) = .;			\
-		*(.data.percpu.first)					\
-		*(.data.percpu.page_aligned)				\
-		*(.data.percpu)						\
-		*(.data.percpu.shared_aligned)				\
+		*(.data..percpu.first)					\
+		*(.data..percpu.page_aligned)				\
+		*(.data..percpu)						\
+		*(.data..percpu.shared_aligned)				\
 		VMLINUX_SYMBOL(__per_cpu_end) = .;			\
 	} phdr								\
-	. = VMLINUX_SYMBOL(__per_cpu_load) + SIZEOF(.data.percpu);
+	. = VMLINUX_SYMBOL(__per_cpu_load) + SIZEOF(.data..percpu);
 
 /**
  * PERCPU - define output section for percpu area, simple version
@@ -496,17 +496,17 @@
  *
  * This macro is equivalent to ALIGN(align); PERCPU_VADDR( , ) except
  * that __per_cpu_load is defined as a relative symbol against
- * .data.percpu which is required for relocatable x86_32
+ * .data..percpu which is required for relocatable x86_32
  * configuration.
  */
 #define PERCPU(align)							\
 	. = ALIGN(align);						\
-	.data.percpu	: AT(ADDR(.data.percpu) - LOAD_OFFSET) {	\
+	.data..percpu	: AT(ADDR(.data..percpu) - LOAD_OFFSET) {	\
 		VMLINUX_SYMBOL(__per_cpu_load) = .;			\
 		VMLINUX_SYMBOL(__per_cpu_start) = .;			\
-		*(.data.percpu.first)					\
-		*(.data.percpu.page_aligned)				\
-		*(.data.percpu)						\
-		*(.data.percpu.shared_aligned)				\
+		*(.data..percpu.first)					\
+		*(.data..percpu.page_aligned)				\
+		*(.data..percpu)						\
+		*(.data..percpu.shared_aligned)				\
 		VMLINUX_SYMBOL(__per_cpu_end) = .;			\
 	}
include/linux/cache.h +1 -1
index 97e2488..4c57065 100644
@@ -31,7 +31,7 @@
 #ifndef __cacheline_aligned
 #define __cacheline_aligned					\
   __attribute__((__aligned__(SMP_CACHE_BYTES),			\
-		 __section__(".data.cacheline_aligned")))
+		 __section__(".data..cacheline_aligned")))
 #endif /* __cacheline_aligned */
 
 #ifndef __cacheline_aligned_in_smp
include/linux/init.h +4 -4
index f121a7a..60b1dd0 100644
@@ -62,9 +62,9 @@
 
 /* backward compatibility note
  *  A few places hardcode the old section names:
- *  .text.init.refok
- *  .data.init.refok
- *  .exit.text.refok
+ *  .text..init.refok
+ *  .data..init.refok
+ *  .text..exit.refok
  *  They should be converted to use the defines from this file
  */
 
@@ -302,7 +302,7 @@ void __init parse_early_options(char *cmdline);
 #endif
 
 /* Data marked not to be saved by software suspend */
-#define __nosavedata __section(.data.nosave)
+#define __nosavedata __section(.data..nosave)
 
 /* This means "can be init if no module support, otherwise module load
    may call it." */
include/linux/linkage.h +2 -2
index fee9e59..bddaee6 100644
@@ -18,8 +18,8 @@
 # define asmregparm
 #endif
 
-#define __page_aligned_data	__section(.data.page_aligned) __aligned(PAGE_SIZE)
-#define __page_aligned_bss	__section(.bss.page_aligned) __aligned(PAGE_SIZE)
+#define __page_aligned_data	__section(.data..page_aligned) __aligned(PAGE_SIZE)
+#define __page_aligned_bss	__section(.bss..page_aligned) __aligned(PAGE_SIZE)
 
 /*
  * This is used by architectures to keep arguments on the stack
include/linux/spinlock.h +1 -1
index 252b245..1a67c69 100644
@@ -60,7 +60,7 @@
 /*
  * Must define these before including other files, inline functions need them
  */
-#define LOCK_SECTION_NAME ".text.lock."KBUILD_BASENAME
+#define LOCK_SECTION_NAME ".text..lock."KBUILD_BASENAME
 
 #define LOCK_SECTION_START(extra)               \
         ".subsection 1\n\t"                     \
kernel/module.c +1 -1
index e797812..4bec8c4 100644
@@ -525,7 +525,7 @@ static unsigned int find_pcpusec(Elf_Ehdr *hdr,
 				 Elf_Shdr *sechdrs,
 				 const char *secstrings)
 {
-	return find_sec(hdr, sechdrs, secstrings, ".data.percpu");
+	return find_sec(hdr, sechdrs, secstrings, ".data..percpu");
 }
 
 static void percpu_modcopy(void *pcpudest, const void *from, unsigned long size)
scripts/mod/modpost.c +6 -6
index df6e628..f39132b 100644
@@ -797,9 +797,9 @@ static const char *data_sections[] = { DATA_SECTIONS, NULL };
 /* sections that may refer to an init/exit section with no warning */
 static const char *initref_sections[] =
 {
-	".text.init.refok*",
-	".exit.text.refok*",
-	".data.init.refok*",
+	".text..init.refok*",
+	".text..exit.refok*",
+	".data..init.refok*",
 	NULL
 };
 
@@ -918,7 +918,7 @@ static int section_mismatch(const char *fromsec, const char *tosec)
  * Pattern 0:
  *   Do not warn if funtion/data are marked with __init_refok/__initdata_refok.
  *   The pattern is identified by:
- *   fromsec = .text.init.refok* | .data.init.refok*
+ *   fromsec = .text..init.refok* | .data..init.refok*
  *
  * Pattern 1:
  *   If a module parameter is declared __initdata and permissions=0
@@ -942,8 +942,8 @@ static int section_mismatch(const char *fromsec, const char *tosec)
  *           *probe_one, *_console, *_timer
  *
  * Pattern 3:
- *   Whitelist all refereces from .text.head to .init.data
- *   Whitelist all refereces from .text.head to .init.text
+ *   Whitelist all references from .text..head to .init.data
+ *   Whitelist all references from .text..head to .init.text
  *
  * Pattern 4:
  *   Some symbols belong to init section but still it is ok to reference
scripts/recordmcount.pl +3 -3
index 409596e..a638a1c 100755
@@ -26,7 +26,7 @@
 # which will also be the location of that section after final link.
 # e.g.
 #
-#  .section ".text.sched"
+#  .section ".text..sched"
 #  .globl my_func
 #  my_func:
 #        [...]
@@ -39,7 +39,7 @@
 #        [...]
 #
 # Both relocation offsets for the mcounts in the above example will be
-# offset from .text.sched. If we make another file called tmp.s with:
+# offset from .text..sched. If we make another file called tmp.s with:
 #
 #  .section __mcount_loc
 #  .quad  my_func + 0x5
@@ -51,7 +51,7 @@
 # But this gets hard if my_func is not globl (a static function).
 # In such a case we have:
 #
-#  .section ".text.sched"
+#  .section ".text..sched"
 #  my_func:
 #        [...]
 #        call mcount  (offset: 0x5)
Signature · 1 line
1.6.2.1
author
#20 lore raw

PATCH v3 2/3 modpost: Check the section flags, not name, to catch missing "ax"/"aw"

From: Anders Kaseorg <[email protected]>

When you put
.section ".foo"
in an assembly file instead of
.section "foo", "ax"
, one of the possible symptoms is that modpost will see an ld-generated section name ".foo.1" in section_rel() or section_rela(). But this heuristic has two problems: it will miss a bad section that has no relocations, and it will incorrectly flag many gcc-generated sections as bad when compiling with -ffunction-sections -fdata-sections.

So instead of checking whether the section name matches a particular pattern, we directly check for a missing SHF_ALLOC in the section flags.

---
 scripts/mod/modpost.c |   49 ++++++++++++++++++-------------------------------
 1 files changed, 18 insertions(+), 31 deletions(-)
scripts/mod/modpost.c +19 -32
index f39132b..3e6ff75 100644
@@ -716,41 +716,27 @@ int match(const char *sym, const char * const pat[])
 
 /* sections that we do not want to do full section mismatch check on */
 static const char *section_white_list[] =
-	{ ".debug*", ".stab*", ".note*", ".got*", ".toc*", NULL };
+	{ ".comment", ".debug*", ".stab*", ".note*", ".got*", ".toc*", NULL };
 
 /*
- * Is this section one we do not want to check?
- * This is often debug sections.
- * If we are going to check this section then
- * test if section name ends with a dot and a number.
- * This is used to find sections where the linker have
- * appended a dot-number to make the name unique.
+ * This is used to find sections missing the SHF_ALLOC flag.
  * The cause of this is often a section specified in assembler
- * without "ax" / "aw" and the same section used in .c
- * code where gcc add these.
+ * without "ax" / "aw".
  */
-static int check_section(const char *modname, const char *sec)
-{
-	const char *e = sec + strlen(sec) - 1;
-	if (match(sec, section_white_list))
-		return 1;
-
-	if (*e && isdigit(*e)) {
-		/* consume all digits */
-		while (*e && e != sec && isdigit(*e))
-			e--;
-		if (*e == '.' && !strstr(sec, ".linkonce")) {
-			warn("%s (%s): unexpected section name.\n"
-			     "The (.[number]+) following section name are "
-			     "ld generated and not expected.\n"
-			     "Did you forget to use \"ax\"/\"aw\" "
-			     "in a .S file?\n"
-			     "Note that for example <linux/init.h> contains\n"
-			     "section definitions for use in .S files.\n\n",
-			     modname, sec);
-		}
+static void check_section(const char *modname, struct elf_info *elf,
+                          Elf_Shdr *sechdr)
+{
+	const char *sec = sech_name(elf, sechdr);
+
+	if (sechdr->sh_type == SHT_PROGBITS &&
+	    !(sechdr->sh_flags & SHF_ALLOC) &&
+	    !match(sec, section_white_list)) {
+		warn("%s (%s): unexpected non-allocatable section.\n"
+		     "Did you forget to use \"ax\"/\"aw\" in a .S file?\n"
+		     "Note that for example <linux/init.h> contains\n"
+		     "section definitions for use in .S files.\n\n",
+		     modname, sec);
 	}
-	return 0;
 }
 
 
@@ -1377,7 +1363,7 @@ static void section_rela(const char *modname, struct elf_info *elf,
 	fromsec = sech_name(elf, sechdr);
 	fromsec += strlen(".rela");
 	/* if from section (name) is know good then skip it */
-	if (check_section(modname, fromsec))
+	if (match(fromsec, section_white_list))
 		return;
 
 	for (rela = start; rela < stop; rela++) {
@@ -1421,7 +1407,7 @@ static void section_rel(const char *modname, struct elf_info *elf,
 	fromsec = sech_name(elf, sechdr);
 	fromsec += strlen(".rel");
 	/* if from section (name) is know good then skip it */
-	if (check_section(modname, fromsec))
+	if (match(fromsec, section_white_list))
 		return;
 
 	for (rel = start; rel < stop; rel++) {
@@ -1484,6 +1470,7 @@ static void check_sec_ref(struct module *mod, const char *modname,
 
 	/* Walk through all sections */
 	for (i = 0; i < elf->hdr->e_shnum; i++) {
+		check_section(modname, elf, &elf->sechdrs[i]);
 		/* We want to process only relocation sections and not .init */
 		if (sechdrs[i].sh_type == SHT_RELA)
 			section_rela(modname, elf, &elf->sechdrs[i]);
Signature · 1 line
1.6.2.1
author
#21 lore raw

PATCH v3 3/3 modpost: Support objects with more than 64k sections

From: Denys Vlasenko <[email protected]>

This patch makes modpost able to process object files with more than 64k sections. Needed for huge kernel builds (allyesconfig, for example) with -ffunction-sections.

---
 scripts/mod/file2alias.c |    6 +-
 scripts/mod/modpost.c    |   94 ++++++++++++++++++++++++++++++++-------------
 scripts/mod/modpost.h    |   43 +++++++++++++++++++++
 3 files changed, 113 insertions(+), 30 deletions(-)
scripts/mod/file2alias.c +3 -3
index a334428..49c669e 100644
@@ -761,16 +761,16 @@ void handle_moddevtable(struct module *mod, struct elf_info *info,
 	char *zeros = NULL;
 
 	/* We're looking for a section relative symbol */
-	if (!sym->st_shndx || sym->st_shndx >= info->hdr->e_shnum)
+	if (!sym->st_shndx || get_secindex(info, sym) >= info->num_sections)
 		return;
 
 	/* Handle all-NULL symbols allocated into .bss */
-	if (info->sechdrs[sym->st_shndx].sh_type & SHT_NOBITS) {
+	if (info->sechdrs[get_secindex(info, sym)].sh_type & SHT_NOBITS) {
 		zeros = calloc(1, sym->st_size);
 		symval = zeros;
 	} else {
 		symval = (void *)info->hdr
-			+ info->sechdrs[sym->st_shndx].sh_offset
+			+ info->sechdrs[get_secindex(info, sym)].sh_offset
 			+ sym->st_value;
 	}
 
scripts/mod/modpost.c +67 -27

211 lines not on this page · raw message

scripts/mod/modpost.h +43 -0

51 lines not on this page · raw message

Signature · 1 line
1.6.2.1
Acked-by on v3 1/3
#22 lore raw

Re: [PATCH v3 1/3] Make section names compatible with -ffunction-sections -fdata-sections

In reply to Tim Abbott
Quoted 8 lines · Tim Abbott

The purpose of this patch is to make it possible to build the kernel with "gcc -ffunction-sections -fdata-sections". This is a key step towards being able to compile the kernel using the --gc-sections linker option, which can be used to decrease the vmlinux size by garbage collecting unused functions. Also, Ksplice's 'run-pre matching' process is much simpler if the original kernel was compiled with -ffunction-sections and -fdata-sections.

[snip]

Quoted 2 lines

As this is the only really invasive part of the patch, it would certainly be good to see this merged early.

#23 lore raw

Re: [PATCH v3 1/3] Make section names compatible with -ffunction-sections -fdata-sections

In reply to Tim Abbott
Quoted 3 lines
---
 Documentation/mutex-design.txt              |    4 +-
 arch/alpha/kernel/head.S                    |    2 +-

...

Quoted 6 lines
 include/linux/linkage.h                     |    4 +-
 include/linux/spinlock.h                    |    2 +-
 kernel/module.c                             |    2 +-
 scripts/mod/modpost.c                       |   12 ++++----
 scripts/recordmcount.pl                     |    6 ++--
 141 files changed, 356 insertions(+), 348 deletions(-)

This patch touches far too many files.
We should try to work out a method so we are in better control of the section names, so renaming in the end is a simple patch touching only a few files.

A few suggestions follows.

arch/alpha/kernel/head.S line 13

Quoted 10 lines
arch/alpha/kernel/head.S +1 -1
index 7ac1f13..16293d4 100644
@@ -10,7 +10,7 @@
 #include <asm/system.h>
 #include <asm/asm-offsets.h>
 
-.section .text.head, "ax"
+.section .text..head, "ax"

Use __HEAD (from include/linux/init.h)
Same goes for all other uses of .text.head.

Quoted 8 lines
arch/alpha/kernel/head.S +0 -0
 .globl swapper_pg_dir
 .globl _stext
 swapper_pg_dir=SWAPPER_PGD
arch/alpha/kernel/init_task.c +0 -0
index c2938e5..7929755 100644

arch/alpha/kernel/init_task.c EXPORT_SYMBOL(init_mm); · line 21

Quoted 6 lines
arch/alpha/kernel/init_task.c +1 -1
@@ -17,5 +17,5 @@ EXPORT_SYMBOL(init_mm);
 EXPORT_SYMBOL(init_task);
 
 union thread_union init_thread_union
-	__attribute__((section(".data.init_thread")))
+	__attribute__((section(".data..init_thread")))
 	= { INIT_THREAD_INFO(init_task) };

1) Either wait for the consolidation of init_tast or add a:
#define SECTION_INIT_THREAD __section(".data.init_thread") to sched.h and use that one for all archs.

arch/alpha/kernel/vmlinux.lds.S SECTIONS · line 22

Quoted 13 lines
arch/alpha/kernel/vmlinux.lds.S +1 -1
index ef37fc1..511f8ca 100644
@@ -16,7 +16,7 @@ SECTIONS
 
 	_text = .;	/* Text and read-only data */
 	.text : {
-	*(.text.head)
+	*(.text..head)
 		TEXT_TEXT
 		SCHED_TEXT
 		LOCK_TEXT

Use HEAD_TEXT (from include/asm-generic/vmlinux.lds.h)

arch/alpha/kernel/vmlinux.lds.S SECTIONS · line 98

Quoted 9 lines
arch/alpha/kernel/vmlinux.lds.S +2 -2
@@ -93,18 +93,18 @@ SECTIONS
 	/* Freed after init ends here */
 
 	/* Note 2 page alignment above.  */
-	.data.init_thread : {
-		*(.data.init_thread)
+	.data..init_thread : {
+		*(.data..init_thread)
 	}

Add something like:

#define DATA_THREAD(aling) \
	...

to vmlinux.lds.h

arch/alpha/kernel/vmlinux.lds.S SECTIONS · line 103

Quoted 7 lines
arch/alpha/kernel/vmlinux.lds.S +2 -2
 	. = ALIGN(PAGE_SIZE);
-	.data.page_aligned : {
-		*(.data.page_aligned)
+	.data..page_aligned : {
+		*(.data..page_aligned)
 	}

Add something like:

#define DATA_PAGE_ALIGNED() \
	...

to vmlinux.lds.h

arch/alpha/kernel/vmlinux.lds.S SECTIONS · line 108

Quoted 7 lines
arch/alpha/kernel/vmlinux.lds.S +2 -2
 	. = ALIGN(64);
-	.data.cacheline_aligned : {
-		*(.data.cacheline_aligned)
+	.data..cacheline_aligned : {
+		*(.data..cacheline_aligned)
 	}

Add something like:

#define DATA_CACHE_ALIGNED() \
	...

to vmlinux.lds.h

arch/arm/kernel/head-nommu.S line 35

Quoted 12 lines
arch/alpha/kernel/vmlinux.lds.S +0 -0
 	_data = .;
arch/arm/kernel/head-nommu.S +1 -1
index cc87e17..fcd93f5 100644
@@ -32,7 +32,7 @@
  * numbers for r1.
  *
  */
-	.section ".text.head", "ax"
+	.section ".text..head", "ax"

__HEAD

arch/arm/kernel/head.S line 77

Quoted 13 lines
arch/arm/kernel/head-nommu.S +0 -0
 ENTRY(stext)
 	msr	cpsr_c, #PSR_F_BIT | PSR_I_BIT | SVC_MODE @ ensure svc mode
 						@ and irqs disabled
arch/arm/kernel/head.S +1 -1
index 21e17dc..705e759 100644
@@ -74,7 +74,7 @@
  * crap here - that's what the boot loader (or in extreme, well justified
  * circumstances, zImage) is for.
  */
-	.section ".text.head", "ax"
+	.section ".text..head", "ax"

__HEAD

Quoted 21 lines
arch/arm/kernel/head.S +0 -0
 ENTRY(stext)
 	msr	cpsr_c, #PSR_F_BIT | PSR_I_BIT | SVC_MODE @ ensure svc mode
 						@ and irqs disabled
arch/arm/kernel/init_task.c +1 -1
index e859af3..0b81b8f 100644
@@ -29,7 +29,7 @@ EXPORT_SYMBOL(init_mm);
  * The things we do for performance..
  */
 union thread_union init_thread_union
-	__attribute__((__section__(".data.init_task"))) =
+	__attribute__((__section__(".data..init_task"))) =
 		{ INIT_THREAD_INFO(init_task) };
 
 /*
arch/arm/kernel/vmlinux.lds.S +0 -0
index c90f272..89c97c3 100644

arch/arm/kernel/vmlinux.lds.S SECTIONS · line 29

Quoted 9 lines
arch/arm/kernel/vmlinux.lds.S +2 -2
@@ -23,10 +23,10 @@ SECTIONS
 #else
 	. = PAGE_OFFSET + TEXT_OFFSET;
 #endif
-	.text.head : {
+	.text..head : {
 		_stext = .;
 		_sinittext = .;
-		*(.text.head)
+		*(.text..head)

HEAD_TEXT

Quoted 4 lines
arch/arm/kernel/vmlinux.lds.S +0 -0
 	}
 
 	.init : {			/* Init code and data		*/

arch/arm/kernel/vmlinux.lds.S SECTIONS · line 72

Quoted 10 lines
arch/arm/kernel/vmlinux.lds.S +3 -3
@@ -66,9 +66,9 @@ SECTIONS
 		. = ALIGN(4096);
 		__per_cpu_load = .;
 		__per_cpu_start = .;
-			*(.data.percpu.page_aligned)
-			*(.data.percpu)
-			*(.data.percpu.shared_aligned)
+			*(.data..percpu.page_aligned)
+			*(.data..percpu)
+			*(.data..percpu.shared_aligned)
 		__per_cpu_end = .;

PERCPU / PERCPU_VADDR

Quoted 14 lines
arch/arm/kernel/vmlinux.lds.S +1 -1
 #ifndef CONFIG_XIP_KERNEL
 		__init_begin = _stext;
@@ -146,7 +146,7 @@ SECTIONS
 		 * first, the init task union, aligned
 		 * to an 8192 byte boundary.
 		 */
-		*(.data.init_task)
+		*(.data..init_task)
 
 
 

#ifdef CONFIG_XIP_KERNEL

arch/arm/kernel/vmlinux.lds.S +0 -0
 		. = ALIGN(4096);

arch/arm/kernel/vmlinux.lds.S SECTIONS · line 164

Quoted 8 lines
arch/arm/kernel/vmlinux.lds.S +1 -1
@@ -158,7 +158,7 @@ SECTIONS
 
 
 		. = ALIGN(4096);
 		__nosave_begin = .;
-		*(.data.nosave)
+		*(.data..nosave)
 		. = ALIGN(4096);
 		__nosave_end = .;

Add whole bloce to a DATA_NOSAVE(align) in vmlinux.lds.h

Quoted 15 lines
arch/arm/kernel/vmlinux.lds.S +1 -1
@@ -166,7 +166,7 @@ SECTIONS
 		 * then the cacheline aligned data
 		 */
 		. = ALIGN(32);
-		*(.data.cacheline_aligned)
+		*(.data..cacheline_aligned)
 
 		/*
 		 * The exception fixup table (might need resorting at runtime)
arch/arm/mm/proc-v6.S +0 -0
index f0cc599..39d8dd1 100644

arch/arm/mm/proc-v6.S cpu_v6_name: · line 135

Quoted 5 lines
arch/arm/mm/proc-v6.S +1 -1
@@ -132,7 +132,7 @@ cpu_v6_name:
 	.asciz	"ARMv6-compatible processor"
 	.align
 
-	.section ".text.init", #alloc, #execinstr
+	.section ".text..init", #alloc, #execinstr

Use __INIT (from init.h).

I will recommend that you do this in separate steps.
1) Is to define all the new sections to vmlinux.lds.h, init.h and maybe other files.

To make this complete you need to go through several of the architectures.

If you have this ready soon we can push this upstream so the individual architectures can apply their modifications.
So step 2 is to create individual patches for each architecture and the remaining files. We can then carry this in -next until we hit a merge window.

Until now this is pure cleanup.

The final step is to do the renaming in the few files that hold the section definitions - it is a simple patch.
And then we can try out function-sections etc.

Sam

author
#24 lore raw

Re: [PATCH v3 1/3] Make section names compatible with -ffunction-sections -fdata-sections

Quoted 4 lines · Sam Ravnborg

This patch touches far too many files.
We should try to work out a method so we are in better control of the section names, so renaming in the end is a simple patch touching only a few files.

OK, I'm now planning to implement this approach.

Quoted 5 lines
Quoted 2 lines

-.section .text.head, "ax"
+.section .text..head, "ax"

Use __HEAD (from include/linux/init.h)
Same goes for all other uses of .text.head.

I notice that __HEAD uses .head.text, while some architectures use .text.head. It looks like this is just an inconsistency across architectures that will be removed as a consequence of this cleanup work (no architecture uses both .head.text and .text.head).

One challenge with this approach is that many linker scripts use these section names in more complex ways than just squashing HEAD_TEXT at the start of the text section. For example, the the linker scripts for x86 and ia64 have code like:

  .text.head : AT(ADDR(.text.head) - LOAD_OFFSET) {
        _text = .;                      /* Text and read-only data */
        *(.text.head)
  } :text = 0x9090

which can't user either the __HEAD macro (which is the full .section line) or the HEAD_TEXT macro (which is the *(.head.text)). I think the right way to solve this is to create a HEAD_TEXT_SECTION macro in a new header (include/linux/section-names.h) and define everything else in terms of that; these linker script blocks would then use HEAD_TEXT_SECTION when necessary. The new header would grow to contain several other section names used in both code and linker scripts as this work proceeds.

I'm sending with this a prototype patch series that replaces all ".head.text" and ".text.head" usage with these macros. Please review the design, as I intend to implement this design for many other section names in the near future.

-Tim Abbott

Tim Abbott (15):
Add new HEAD_TEXT_SECTION macro.
xtensa: convert to use __HEAD and HEAD_TEXT macros.
alpha: convert to use __HEAD and HEAD_TEXT macros.
frv: convert frv to use __HEAD and HEAD_TEXT macros.
arm: convert to use __HEAD and HEAD_TEXT macros.
ia64: convert to use __HEAD and HEAD_TEXT macros.
m32r: convert to use __HEAD and HEAD_TEXT macros.
m68k: convert to use __HEAD and HEAD_TEXT macros.
mn10300: convert to use __HEAD and HEAD_TEXT macros.
powerpc: convert to use __HEAD and HEAD_TEXT macros.
s390: convert to use __HEAD and HEAD_TEXT macros.
sh: convert to use __HEAD and HEAD_TEXT macros.
sparc: convert to use __HEAD and HEAD_TEXT macros.
x86: convert to use __HEAD and HEAD_TEXT macros.
modpost: convert modpost to use HEAD_TEXT_SECTION macro.

 arch/alpha/kernel/head.S                           |    3 ++-
 arch/alpha/kernel/vmlinux.lds.S                    |    2 +-
 arch/arm/kernel/head-nommu.S                       |    2 +-
 arch/arm/kernel/head.S                             |    2 +-
 arch/arm/kernel/vmlinux.lds.S                      |    4 ++--
 arch/frv/kernel/head.S                             |    3 ++-
 arch/frv/kernel/vmlinux.lds.S                      |    2 +-
 arch/ia64/kernel/head.S                            |    2 +-
 arch/ia64/kernel/vmlinux.lds.S                     |    4 ++--
 arch/m32r/kernel/head.S                            |    2 +-
 arch/m32r/kernel/vmlinux.lds.S                     |    2 +-
 arch/m68k/kernel/head.S                            |    2 +-
 arch/m68k/kernel/sun3-head.S                       |    2 +-
 arch/m68k/kernel/vmlinux-std.lds                   |    2 +-
 arch/m68k/kernel/vmlinux-sun3.lds                  |    2 +-
 arch/mn10300/kernel/head.S                         |    3 ++-
 arch/mn10300/kernel/vmlinux.lds.S                  |    5 +----
 arch/powerpc/kernel/head_32.S                      |    3 ++-
 arch/powerpc/kernel/head_40x.S                     |    3 ++-
 arch/powerpc/kernel/head_44x.S                     |    3 ++-
 arch/powerpc/kernel/head_8xx.S                     |    3 ++-
 arch/powerpc/kernel/head_fsl_booke.S               |    3 ++-
 arch/powerpc/kernel/vmlinux.lds.S                  |    2 +-
 arch/s390/kernel/head.S                            |    3 ++-
 arch/s390/kernel/vmlinux.lds.S                     |    4 ++--
 arch/sh/kernel/head_32.S                           |    3 ++-
 arch/sh/kernel/head_64.S                           |    5 ++++-
 arch/sh/kernel/vmlinux_32.lds.S                    |    2 +-
 arch/sh/kernel/vmlinux_64.lds.S                    |    2 +-
 arch/sparc/kernel/head_32.S                        |    2 +-
 arch/sparc/kernel/vmlinux.lds.S                    |    2 +-
 arch/x86/boot/compressed/Makefile                  |    2 ++
 arch/x86/boot/compressed/head_32.S                 |    3 ++-
 arch/x86/boot/compressed/head_64.S                 |    3 ++-
 .../{vmlinux_32.lds => vmlinux_32.lds.S}           |    6 ++++--
 .../{vmlinux_64.lds => vmlinux_64.lds.S}           |    6 ++++--
 arch/x86/kernel/head_32.S                          |    2 +-
 arch/x86/kernel/head_64.S                          |    2 +-
 arch/x86/kernel/vmlinux_32.lds.S                   |    4 ++--
 arch/x86/kernel/vmlinux_64.lds.S                   |    2 +-
 arch/xtensa/kernel/head.S                          |    3 ++-
 arch/xtensa/kernel/vmlinux.lds.S                   |    4 ++--
 include/asm-generic/vmlinux.lds.h                  |    4 +++-
 include/linux/init.h                               |    4 +++-
 include/linux/section-names.h                      |    6 ++++++
 scripts/mod/modpost.c                              |    8 +++++---
 46 files changed, 87 insertions(+), 56 deletions(-)
 rename arch/x86/boot/compressed/{vmlinux_32.lds => vmlinux_32.lds.S} (89%)
 rename arch/x86/boot/compressed/{vmlinux_64.lds => vmlinux_64.lds.S} (90%)
 create mode 100644 include/linux/section-names.h
author
#25 lore raw

PATCH 1/15 Add new HEAD_TEXT_SECTION macro.

This patch is preparation for replacing all uses of ".head.text" or ".text.head" in the kernel with macros, so that the section name can later be changed without having to touch a lot of the kernel.

Since some linker scripts do more complex things than referencing HEAD_TEXT, we add a HEAD_TEXT_SECTION macro that just contains the actual name.

I've defined HEAD_TEXT_SECTION in a new header, include/linux/section-names.h, so that this section name only needs to appear in one place. I anticipate creating similar macro structures for a number of other section names.

The long-term goal here is to be able to change the kernel's magic section names to those that are compatible with -ffunction-sections -fdata-sections. This requires renaming all magic sections with names of the form ".text.foo".

---
 include/asm-generic/vmlinux.lds.h |    4 +++-
 include/linux/init.h              |    4 +++-
 include/linux/section-names.h     |    6 ++++++
 3 files changed, 12 insertions(+), 2 deletions(-)
 create mode 100644 include/linux/section-names.h
include/asm-generic/vmlinux.lds.h +3 -1
index 7fa660f..eaa06ef 100644
@@ -1,3 +1,5 @@
+#include <linux/section-names.h>
+
 #ifndef LOAD_OFFSET
 #define LOAD_OFFSET 0
 #endif
@@ -331,7 +333,7 @@
 #endif
 
 /* Section used for early init (in .S files) */
-#define HEAD_TEXT  *(.head.text)
+#define HEAD_TEXT  *(HEAD_TEXT_SECTION)
 
 /* init and exit section handling */
 #define INIT_DATA							\
include/linux/init.h +3 -1
index f121a7a..20a1334 100644
@@ -2,6 +2,8 @@
 #define _LINUX_INIT_H
 
 #include <linux/compiler.h>
+#include <linux/section-names.h>
+#include <linux/stringify.h>
 
 /* These macros are used to mark some functions or 
  * initialized data (doesn't apply to uninitialized data)
@@ -107,7 +109,7 @@
 #define __memexitconst   __section(.memexit.rodata)
 
 /* For assembly routines */
-#define __HEAD		.section	".head.text","ax"
+#define __HEAD		.section	__stringify(HEAD_TEXT_SECTION),"ax"
 #define __INIT		.section	".init.text","ax"
 #define __FINIT		.previous
 
include/linux/section-names.h new file +6 -0
new file mode 100644
index 0000000..c956f4e
@@ -0,0 +1,6 @@
+#ifndef __LINUX_SECTION_NAMES_H
+#define __LINUX_SECTION_NAMES_H
+
+#define HEAD_TEXT_SECTION .head.text
+
+#endif /* !__LINUX_SECTION_NAMES_H */
Signature · 1 line
1.6.2.1
author
#26 lore raw

PATCH 2/15 xtensa: convert to use __HEAD and HEAD_TEXT macros.

---
 arch/xtensa/kernel/head.S        |    3 ++-
 arch/xtensa/kernel/vmlinux.lds.S |    4 ++--
 2 files changed, 4 insertions(+), 3 deletions(-)
arch/xtensa/kernel/head.S +2 -1
index 0817f9d..d9ddc1b 100644
@@ -19,6 +19,7 @@
 #include <asm/page.h>
 #include <asm/cacheasm.h>
 
+#include <linux/init.h>
 #include <linux/linkage.h>
 
 /*
@@ -45,7 +46,7 @@
 	 * instruction.
 	 */
 
-	.section .head.text, "ax"
+	__HEAD
 	.globl _start
 _start:	_j	2f
 	.align	4
arch/xtensa/kernel/vmlinux.lds.S +2 -2
index c1be9a4..5accf51 100644
@@ -85,8 +85,8 @@ SECTIONS
 
   .text :
   {
-    /* The .head.text section must be the first section! */
-    *(.head.text)
+    /* The HEAD_TEXT section must be the first section! */
+    HEAD_TEXT
     *(.literal .text)
     VMLINUX_SYMBOL(__sched_text_start) = .;
     *(.sched.literal .sched.text)
Signature · 1 line
1.6.2.1
author
#27 lore raw

PATCH 3/15 alpha: convert to use __HEAD and HEAD_TEXT macros.

This has the consequence of changing the section name use for head code from ".text.head" to ".head.text". Since this commit changes all users in the architecture, this change should be harmless.

Cc Richard Henderson <[email protected]>
---
 arch/alpha/kernel/head.S        |    3 ++-
 arch/alpha/kernel/vmlinux.lds.S |    2 +-
 2 files changed, 3 insertions(+), 2 deletions(-)
arch/alpha/kernel/head.S +2 -1
index 7ac1f13..4bdd1d2 100644
@@ -7,10 +7,11 @@
  * the kernel global pointer and jump to the kernel entry-point.
  */
 
+#include <linux/init.h>
 #include <asm/system.h>
 #include <asm/asm-offsets.h>
 
-.section .text.head, "ax"
+__HEAD
 .globl swapper_pg_dir
 .globl _stext
 swapper_pg_dir=SWAPPER_PGD
arch/alpha/kernel/vmlinux.lds.S +1 -1
index ef37fc1..b9d6568 100644
@@ -16,7 +16,7 @@ SECTIONS
 
 	_text = .;	/* Text and read-only data */
 	.text : {
-	*(.text.head)
+		HEAD_TEXT
 		TEXT_TEXT
 		SCHED_TEXT
 		LOCK_TEXT
Signature · 1 line
1.6.2.1
author
#28 lore raw

PATCH 4/15 frv: convert frv to use __HEAD and HEAD_TEXT macros.

This has the consequence of changing the section name use for head code from ".text.head" to ".head.text". Since this commit changes all users in the architecture, this change should be harmless.

Cc David Howells <[email protected]>
---
 arch/frv/kernel/head.S        |    3 ++-
 arch/frv/kernel/vmlinux.lds.S |    2 +-
 2 files changed, 3 insertions(+), 2 deletions(-)
arch/frv/kernel/head.S +2 -1
index fecf751..b825ef3 100644
@@ -9,6 +9,7 @@
  * 2 of the License, or (at your option) any later version.
  */
 
+#include <linux/init.h>
 #include <linux/threads.h>
 #include <linux/linkage.h>
 #include <asm/thread_info.h>
@@ -27,7 +28,7 @@
 #   command line string
 #
 ###############################################################################
-	.section	.text.head,"ax"
+	__HEAD
 	.balign		4
 
 	.globl		_boot, __head_reference
arch/frv/kernel/vmlinux.lds.S +1 -1
index b95c4ea..22d9787 100644
@@ -26,7 +26,7 @@ SECTIONS
 
   _sinittext = .;
   .init.text : {
-	*(.text.head)
+	HEAD_TEXT
 #ifndef CONFIG_DEBUG_INFO
 	INIT_TEXT
 	EXIT_TEXT
Signature · 1 line
1.6.2.1
author
#29 lore raw

PATCH 5/15 arm: convert to use __HEAD and HEAD_TEXT macros.

This has the consequence of changing the section name use for head code from ".text.head" to ".head.text". Since this commit changes all users in the architecture, this change should be harmless.

---
 arch/arm/kernel/head-nommu.S  |    2 +-
 arch/arm/kernel/head.S        |    2 +-
 arch/arm/kernel/vmlinux.lds.S |    4 ++--
 3 files changed, 4 insertions(+), 4 deletions(-)
arch/arm/kernel/head-nommu.S +1 -1
index cc87e17..4f10d1c 100644
@@ -32,7 +32,7 @@
  * numbers for r1.
  *
  */
-	.section ".text.head", "ax"
+	__HEAD
 ENTRY(stext)
 	msr	cpsr_c, #PSR_F_BIT | PSR_I_BIT | SVC_MODE @ ensure svc mode
 						@ and irqs disabled
arch/arm/kernel/head.S +1 -1
index 21e17dc..f1877a5 100644
@@ -74,7 +74,7 @@
  * crap here - that's what the boot loader (or in extreme, well justified
  * circumstances, zImage) is for.
  */
-	.section ".text.head", "ax"
+	__HEAD
 ENTRY(stext)
 	msr	cpsr_c, #PSR_F_BIT | PSR_I_BIT | SVC_MODE @ ensure svc mode
 						@ and irqs disabled
arch/arm/kernel/vmlinux.lds.S +2 -2
index c90f272..655fb75 100644
@@ -23,10 +23,10 @@ SECTIONS
 #else
 	. = PAGE_OFFSET + TEXT_OFFSET;
 #endif
-	.text.head : {
+	HEAD_TEXT_SECTION : {
 		_stext = .;
 		_sinittext = .;
-		*(.text.head)
+		HEAD_TEXT
 	}
 
 	.init : {			/* Init code and data		*/
Signature · 1 line
1.6.2.1
author
#30 lore raw

PATCH 6/15 ia64: convert to use __HEAD and HEAD_TEXT macros.

This has the consequence of changing the section name use for head code from ".text.head" to ".head.text". Since this commit changes all users in the architecture, this change should be harmless.

---
 arch/ia64/kernel/head.S        |    2 +-
 arch/ia64/kernel/vmlinux.lds.S |    4 ++--
 2 files changed, 3 insertions(+), 3 deletions(-)
arch/ia64/kernel/head.S +1 -1
index 23f846d..a572160 100644
@@ -181,7 +181,7 @@ swapper_pg_dir:
 halt_msg:
 	stringz "Halting kernel\n"
 
-	.section .text.head,"ax"
+	__HEAD
 
 	.global start_ap
 
arch/ia64/kernel/vmlinux.lds.S +2 -2
index 4a95e86..a8b879e 100644
@@ -51,8 +51,8 @@ SECTIONS
 	KPROBES_TEXT
 	*(.gnu.linkonce.t*)
     }
-  .text.head : AT(ADDR(.text.head) - LOAD_OFFSET)
-	{ *(.text.head) }
+  HEAD_TEXT_SECTION : AT(ADDR(HEAD_TEXT_SECTION) - LOAD_OFFSET)
+	{ HEAD_TEXT }
   .text2 : AT(ADDR(.text2) - LOAD_OFFSET)
 	{ *(.text2) }
 #ifdef CONFIG_SMP
Signature · 1 line
1.6.2.1
author
#31 lore raw

PATCH 7/15 m32r: convert to use __HEAD and HEAD_TEXT macros.

This has the consequence of changing the section name use for head code from ".text.head" to ".head.text". Since this commit changes all users in the architecture, this change should be harmless.

Cc Hirokazu Takata <[email protected]>
---
 arch/m32r/kernel/head.S        |    2 +-
 arch/m32r/kernel/vmlinux.lds.S |    2 +-
 2 files changed, 2 insertions(+), 2 deletions(-)
arch/m32r/kernel/head.S +1 -1
index 9091606..0a71944 100644
@@ -23,7 +23,7 @@ __INITDATA
 /*
  * References to members of the boot_cpu_data structure.
  */
-.section .text.head, "ax"
+__HEAD
 	.global	start_kernel
 	.global __bss_start
 	.global _end
arch/m32r/kernel/vmlinux.lds.S +1 -1
index 9db05df..4179adf 100644
@@ -27,7 +27,7 @@ SECTIONS
   _text = .;			/* Text and read-only data */
   .boot : { *(.boot) } = 0
   .text : {
-	*(.text.head)
+	HEAD_TEXT
 	TEXT_TEXT
 	SCHED_TEXT
 	LOCK_TEXT
Signature · 1 line
1.6.2.1
author
#32 lore raw

PATCH 8/15 m68k: convert to use __HEAD and HEAD_TEXT macros.

This has the consequence of changing the section name use for head code from ".text.head" to ".head.text". Since this commit changes all users in the architecture, this change should be harmless.

Cc Geert Uytterhoeven <[email protected]>
Cc Roman Zippel <[email protected]>
---
 arch/m68k/kernel/head.S           |    2 +-
 arch/m68k/kernel/sun3-head.S      |    2 +-
 arch/m68k/kernel/vmlinux-std.lds  |    2 +-
 arch/m68k/kernel/vmlinux-sun3.lds |    2 +-
 4 files changed, 4 insertions(+), 4 deletions(-)
arch/m68k/kernel/head.S +1 -1
index f513f53..86edb5f 100644
@@ -577,7 +577,7 @@ func_define	putn,1
 #endif
 .endm
 
-.section ".text.head","ax"
+__HEAD
 ENTRY(_stext)
 /*
  * Version numbers of the bootinfo interface
arch/m68k/kernel/sun3-head.S +1 -1
index aad0159..9e5c3e5 100644
@@ -29,7 +29,7 @@ kernel_pmd_table:              .skip 0x2000
 .globl kernel_pg_dir
 .equ    kernel_pg_dir,kernel_pmd_table
 
-	.section .text.head
+	__HEAD
 ENTRY(_stext)
 ENTRY(_start)
 
arch/m68k/kernel/vmlinux-std.lds +1 -1
index f846d4e..01d212b 100644
@@ -12,7 +12,7 @@ SECTIONS
   . = 0x1000;
   _text = .;			/* Text and read-only data */
   .text : {
-	*(.text.head)
+	HEAD_TEXT
 	TEXT_TEXT
 	SCHED_TEXT
 	LOCK_TEXT
arch/m68k/kernel/vmlinux-sun3.lds +1 -1
index d9368c0..c192f77 100644
@@ -12,7 +12,7 @@ SECTIONS
   . = 0xE002000;
   _text = .;			/* Text and read-only data */
   .text : {
-	*(.text.head)
+	HEAD_TEXT
 	TEXT_TEXT
 	SCHED_TEXT
 	LOCK_TEXT
Signature · 1 line
1.6.2.1
author
#33 lore raw

PATCH 9/15 mn10300: convert to use __HEAD and HEAD_TEXT macros.

This has the consequence of changing the section name use for head code from ".text.head" to ".head.text". Since this commit changes all users in the architecture, this change should be harmless.

Cc David Howells <[email protected]>
---
 arch/mn10300/kernel/head.S        |    3 ++-
 arch/mn10300/kernel/vmlinux.lds.S |    5 +----
 2 files changed, 3 insertions(+), 5 deletions(-)
arch/mn10300/kernel/head.S +2 -1
index 8a8309f..14f27f3 100644
@@ -9,6 +9,7 @@
  * 2 of the Licence, or (at your option) any later version.
  */
 
+#include <linux/init.h>
 #include <linux/threads.h>
 #include <linux/linkage.h>
 #include <linux/serial_reg.h>
@@ -19,7 +20,7 @@
 #include <asm/param.h>
 #include <unit/serial.h>
 
-	.section .text.head,"ax"
+	__HEAD
 
 ###############################################################################
 #
arch/mn10300/kernel/vmlinux.lds.S +1 -4
index b825966..24de6b9 100644
@@ -27,10 +27,7 @@ SECTIONS
   _stext = .;
   _text = .;			/* Text and read-only data */
   .text : {
-	*(
-	.text.head
-	.text
-	)
+	HEAD_TEXT
 	TEXT_TEXT
 	SCHED_TEXT
 	LOCK_TEXT
Signature · 1 line
1.6.2.1
author
#34 lore raw

PATCH 10/15 powerpc: convert to use __HEAD and HEAD_TEXT macros.

This has the consequence of changing the section name use for head code from ".text.head" to ".head.text". Since this commit changes all users in the architecture, this change should be harmless.

Cc Benjamin Herrenschmidt <[email protected]>
---
 arch/powerpc/kernel/head_32.S        |    3 ++-
 arch/powerpc/kernel/head_40x.S       |    3 ++-
 arch/powerpc/kernel/head_44x.S       |    3 ++-
 arch/powerpc/kernel/head_8xx.S       |    3 ++-
 arch/powerpc/kernel/head_fsl_booke.S |    3 ++-
 arch/powerpc/kernel/vmlinux.lds.S    |    2 +-
 6 files changed, 11 insertions(+), 6 deletions(-)
arch/powerpc/kernel/head_32.S +2 -1
index 54e68c1..c01467f 100644
@@ -21,6 +21,7 @@
  *
  */
 
+#include <linux/init.h>
 #include <asm/reg.h>
 #include <asm/page.h>
 #include <asm/mmu.h>
@@ -50,7 +51,7 @@
 	mtspr	SPRN_DBAT##n##L,RB;	\
 1:
 
-	.section	.text.head, "ax"
+	__HEAD
 	.stabs	"arch/powerpc/kernel/",N_SO,0,0,0f
 	.stabs	"head_32.S",N_SO,0,0,0f
 0:
arch/powerpc/kernel/head_40x.S +2 -1
index 56d8e5d..0c96911 100644
@@ -31,6 +31,7 @@
  *
  */
 
+#include <linux/init.h>
 #include <asm/processor.h>
 #include <asm/page.h>
 #include <asm/mmu.h>
@@ -52,7 +53,7 @@
  *
  * This is all going to change RSN when we add bi_recs.......  -- Dan
  */
-	.section	.text.head, "ax"
+	__HEAD
 _ENTRY(_stext);
 _ENTRY(_start);
 
arch/powerpc/kernel/head_44x.S +2 -1
index b56fecc..18d8a16 100644
@@ -28,6 +28,7 @@
  * option) any later version.
  */
 
+#include <linux/init.h>
 #include <asm/processor.h>
 #include <asm/page.h>
 #include <asm/mmu.h>
@@ -50,7 +51,7 @@
  *   r7 - End of kernel command line string
  *
  */
-	.section	.text.head, "ax"
+	__HEAD
 _ENTRY(_stext);
 _ENTRY(_start);
 	/*
arch/powerpc/kernel/head_8xx.S +2 -1
index 3c9452d..52ff8c5 100644
@@ -19,6 +19,7 @@
  *
  */
 
+#include <linux/init.h>
 #include <asm/processor.h>
 #include <asm/page.h>
 #include <asm/mmu.h>
@@ -38,7 +39,7 @@
 #else
 #define DO_8xx_CPU6(val, reg)
 #endif
-	.section	.text.head, "ax"
+	__HEAD
 _ENTRY(_stext);
 _ENTRY(_start);
 
arch/powerpc/kernel/head_fsl_booke.S +2 -1
index 4c22620..5bdcc06 100644
@@ -30,6 +30,7 @@
  * option) any later version.
  */
 
+#include <linux/init.h>
 #include <linux/threads.h>
 #include <asm/processor.h>
 #include <asm/page.h>
@@ -53,7 +54,7 @@
  *   r7 - End of kernel command line string
  *
  */
-	.section	.text.head, "ax"
+	__HEAD
 _ENTRY(_stext);
 _ENTRY(_start);
 	/*
arch/powerpc/kernel/vmlinux.lds.S +1 -1
index b9ef164..d7425ff 100644
@@ -52,7 +52,7 @@ SECTIONS
 	/* Text and gots */
 	.text : AT(ADDR(.text) - LOAD_OFFSET) {
 		ALIGN_FUNCTION();
-		*(.text.head)
+		HEAD_TEXT
 		_text = .;
 		*(.text .fixup .text.init.refok .exit.text.refok __ftr_alt_*)
 		SCHED_TEXT
Signature · 1 line
1.6.2.1
author
#35 lore raw

PATCH 11/15 s390: convert to use __HEAD and HEAD_TEXT macros.

This has the consequence of changing the section name use for head code from ".text.head" to ".head.text". Since this commit changes all users in the architecture, this change should be harmless.

Cc Martin Schwidefsky <[email protected]>
Cc Heiko Carstens <[email protected]>
---
 arch/s390/kernel/head.S        |    3 ++-
 arch/s390/kernel/vmlinux.lds.S |    4 ++--
 2 files changed, 4 insertions(+), 3 deletions(-)
arch/s390/kernel/head.S +2 -1
index bba1449..22596d7 100644
@@ -23,6 +23,7 @@
  *
  */
 
+#include <linux/init.h>
 #include <asm/setup.h>
 #include <asm/lowcore.h>
 #include <asm/asm-offsets.h>
@@ -35,7 +36,7 @@
 #define ARCH_OFFSET	0
 #endif
 
-.section ".text.head","ax"
+__HEAD
 #ifndef CONFIG_IPL
 	.org   0
 	.long  0x00080000,0x80000000+startup	# Just a restart PSW
arch/s390/kernel/vmlinux.lds.S +2 -2
index 7a2063e..89399b8 100644
@@ -29,8 +29,8 @@ SECTIONS
 	. = 0x00000000;
 	.text : {
 	_text = .;		/* Text and read-only data */
-		*(.text.head)
-	TEXT_TEXT
+		HEAD_TEXT
+		TEXT_TEXT
 		SCHED_TEXT
 		LOCK_TEXT
 		KPROBES_TEXT
Signature · 1 line
1.6.2.1
author
#36 lore raw

PATCH 12/15 sh: convert to use __HEAD and HEAD_TEXT macros.

This has the consequence of changing the section name use for head code from ".text.head" to ".head.text". Since this commit changes all users in the architecture, this change should be harmless.

---
 arch/sh/kernel/head_32.S        |    3 ++-
 arch/sh/kernel/head_64.S        |    5 ++++-
 arch/sh/kernel/vmlinux_32.lds.S |    2 +-
 arch/sh/kernel/vmlinux_64.lds.S |    2 +-
 4 files changed, 8 insertions(+), 4 deletions(-)
arch/sh/kernel/head_32.S +2 -1
index 788605f..a78be74 100644
@@ -10,6 +10,7 @@
  *
  * Head.S contains the SH exception handlers and startup code.
  */
+#include <linux/init.h>
 #include <linux/linkage.h>
 #include <asm/thread_info.h>
 
@@ -40,7 +41,7 @@ ENTRY(empty_zero_page)
 1:
 	.skip	PAGE_SIZE - empty_zero_page - 1b
 
-	.section	.text.head, "ax"
+	__HEAD
 
 /*
  * Condition at the entry of _stext:
arch/sh/kernel/head_64.S +4 -1
index 7ccfb99..3ea7658 100644
@@ -8,6 +8,9 @@
  * License.  See the file "COPYING" in the main directory of this archive
  * for more details.
  */
+
+#include <linux/init.h>
+
 #include <asm/page.h>
 #include <asm/cache.h>
 #include <asm/tlb.h>
@@ -110,7 +113,7 @@ empty_bad_pte_table:
 fpu_in_use:	.quad	0
 
 
-	.section	.text.head, "ax"
+	__HEAD
 	.balign L1_CACHE_BYTES
 /*
  * Condition at the entry of __stext:
arch/sh/kernel/vmlinux_32.lds.S +1 -1
index d0b2a71..dd9b2ee 100644
@@ -31,7 +31,7 @@ SECTIONS
 	} = 0
 
 	.text : {
-		*(.text.head)
+		HEAD_TEXT
 		TEXT_TEXT
 		SCHED_TEXT
 		LOCK_TEXT
arch/sh/kernel/vmlinux_64.lds.S +1 -1
index 33fa464..6966446 100644
@@ -42,7 +42,7 @@ SECTIONS
 	} = 0
 
 	.text : C_PHYS(.text) {
-		*(.text.head)
+		HEAD_TEXT
 		TEXT_TEXT
 		*(.text64)
 		*(.text..SHmedia32)
Signature · 1 line
1.6.2.1
author
#37 lore raw

PATCH 13/15 sparc: convert to use __HEAD and HEAD_TEXT macros.

This has the consequence of changing the section name use for head code from ".text.head" to ".head.text". Since this commit changes all users in the architecture, this change should be harmless.

Cc David S. Miller <[email protected]>
---
 arch/sparc/kernel/head_32.S     |    2 +-
 arch/sparc/kernel/vmlinux.lds.S |    2 +-
 2 files changed, 2 insertions(+), 2 deletions(-)
arch/sparc/kernel/head_32.S +1 -1
index f0b4b51..e91f01a 100644
@@ -72,7 +72,7 @@ sun4e_notsup:
 	.align 4
 
 	/* The Sparc trap table, bootloader gives us control at _start. */
-	.section .text.head,"ax"
+	__HEAD
 	.globl	start, _stext, _start, __stext
 	.globl  trapbase
 _start:   /* danger danger */
arch/sparc/kernel/vmlinux.lds.S +1 -1
index 7626708..fcbbd00 100644
@@ -41,7 +41,7 @@ SECTIONS
 	.text TEXTSTART :
 	{
 		_text = .;
-		*(.text.head)
+		HEAD_TEXT
 		TEXT_TEXT
 		SCHED_TEXT
 		LOCK_TEXT
Signature · 1 line
1.6.2.1
author
#38 lore raw

PATCH 14/15 x86: convert to use __HEAD and HEAD_TEXT macros.

This has the consequence of changing the section name use for head code from ".text.head" to ".head.text". Since this commit changes all users in the architecture, this change should be harmless.

Cc Thomas Gleixner <[email protected]>
Cc Ingo Molnar <[email protected]>
Cc H. Peter Anvin <[email protected]>
---
 arch/x86/boot/compressed/Makefile         |    2 +
 arch/x86/boot/compressed/head_32.S        |    3 +-
 arch/x86/boot/compressed/head_64.S        |    3 +-
 arch/x86/boot/compressed/vmlinux_32.lds   |   43 -------------------------
 arch/x86/boot/compressed/vmlinux_32.lds.S |   45 ++++++++++++++++++++++++++
 arch/x86/boot/compressed/vmlinux_64.lds   |   48 ---------------------------
 arch/x86/boot/compressed/vmlinux_64.lds.S |   50 +++++++++++++++++++++++++++++
 arch/x86/kernel/head_32.S                 |    2 +-
 arch/x86/kernel/head_64.S                 |    2 +-
 arch/x86/kernel/vmlinux_32.lds.S          |    4 +-
 arch/x86/kernel/vmlinux_64.lds.S          |    2 +-
 11 files changed, 106 insertions(+), 98 deletions(-)
 delete mode 100644 arch/x86/boot/compressed/vmlinux_32.lds
 create mode 100644 arch/x86/boot/compressed/vmlinux_32.lds.S
 delete mode 100644 arch/x86/boot/compressed/vmlinux_64.lds
 create mode 100644 arch/x86/boot/compressed/vmlinux_64.lds.S
arch/x86/boot/compressed/Makefile +2 -0
index 65551c9..617c939 100644
@@ -19,6 +19,8 @@ KBUILD_AFLAGS  := $(KBUILD_CFLAGS) -D__ASSEMBLY__
 LDFLAGS := -m elf_$(UTS_MACHINE)
 LDFLAGS_vmlinux := -T
 
+CPPFLAGS_vmlinux_$(BITS).lds += -P -U$(UTS_MACHINE)
+
 $(obj)/vmlinux: $(src)/vmlinux_$(BITS).lds $(obj)/head_$(BITS).o $(obj)/misc.o $(obj)/piggy.o FORCE
 	$(call if_changed,ld)
 	@:
arch/x86/boot/compressed/head_32.S +2 -1
index 3a8a866..a7c7a42 100644
@@ -23,13 +23,14 @@
  */
 .text
 
+#include <linux/init.h>
 #include <linux/linkage.h>
 #include <asm/segment.h>
 #include <asm/page_types.h>
 #include <asm/boot.h>
 #include <asm/asm-offsets.h>
 
-.section ".text.head","ax",@progbits
+__HEAD
 ENTRY(startup_32)
 	cld
 	/* test KEEP_SEGMENTS flag to see if the bootloader is asking
arch/x86/boot/compressed/head_64.S +2 -1
index ed4a829..a788a91 100644
@@ -24,6 +24,7 @@
 .code32
 .text
 
+#include <linux/init.h>
 #include <linux/linkage.h>
 #include <asm/segment.h>
 #include <asm/pgtable_types.h>
@@ -33,7 +34,7 @@
 #include <asm/processor-flags.h>
 #include <asm/asm-offsets.h>
 
-.section ".text.head"
+__HEAD
 	.code32
 ENTRY(startup_32)
 	cld
arch/x86/boot/compressed/vmlinux_32.lds deleted +0 -43

43 lines not on this page · raw message

arch/x86/boot/compressed/vmlinux_32.lds.S new file +45 -0

45 lines not on this page · raw message

arch/x86/boot/compressed/vmlinux_64.lds deleted +0 -48

48 lines not on this page · raw message

arch/x86/boot/compressed/vmlinux_64.lds.S new file +50 -0

50 lines not on this page · raw message

arch/x86/kernel/head_32.S +1 -1
index 3068388..a299ecc 100644
@@ -80,7 +80,7 @@ RESERVE_BRK(pagetables, INIT_MAP_SIZE)
  * any particular GDT layout, because we load our own as soon as we
  * can.
  */
-.section .text.head,"ax",@progbits
+__HEAD
 ENTRY(startup_32)
 	/* test KEEP_SEGMENTS flag to see if the bootloader is asking
 		us to not reload segments */
arch/x86/kernel/head_64.S +1 -1
index 54b29bb..dbb6fb5 100644
@@ -41,7 +41,7 @@ L4_START_KERNEL = pgd_index(__START_KERNEL_map)
 L3_START_KERNEL = pud_index(__START_KERNEL_map)
 
 	.text
-	.section .text.head
+	__HEAD
 	.code64
 	.globl startup_64
 startup_64:
arch/x86/kernel/vmlinux_32.lds.S +2 -2
index 62ad500..2e364e7 100644
@@ -31,9 +31,9 @@ SECTIONS
   . = LOAD_OFFSET + LOAD_PHYSICAL_ADDR;
   phys_startup_32 = startup_32 - LOAD_OFFSET;
 
-  .text.head : AT(ADDR(.text.head) - LOAD_OFFSET) {
+  HEAD_TEXT_SECTION : AT(ADDR(HEAD_TEXT_SECTION) - LOAD_OFFSET) {
   	_text = .;			/* Text and read-only data */
-	*(.text.head)
+	HEAD_TEXT
   } :text = 0x9090
 
   /* read-only */
arch/x86/kernel/vmlinux_64.lds.S +1 -1
index c874250..2dcd7c9 100644
@@ -32,7 +32,7 @@ SECTIONS
   .text :  AT(ADDR(.text) - LOAD_OFFSET) {
 	_text = .;			/* Text and read-only data */
 	/* First the code that has to be first for bootstrapping */
-	*(.text.head)
+	HEAD_TEXT
 	_stext = .;
 	/* Then the rest */
 	TEXT_TEXT
Signature · 1 line
1.6.2.1
author
#39 lore raw

PATCH 15/15 modpost: convert modpost to use HEAD_TEXT_SECTION macro.

Using the HEAD_TEXT_SECTION macro in modpost requires including a few more files from the linux headers.

The entry in head_sections[] ".head.text*" should be ".head.text", since only that section name is actually used by any architectures. I corrected this since preserving this over-broad pattern would have involved some string concatination nastiness.

Cc Rusty Russell <[email protected]>
---
 scripts/mod/modpost.c |    8 +++++---
 1 files changed, 5 insertions(+), 3 deletions(-)
scripts/mod/modpost.c +5 -3
index df6e628..f93645f 100644
@@ -16,6 +16,8 @@
 #include <ctype.h>
 #include "modpost.h"
 #include "../../include/linux/license.h"
+#include "../../include/linux/section-names.h"
+#include "../../include/linux/stringify.h"
 
 /* Are we using CONFIG_MODVERSIONS? */
 int modversions = 0;
@@ -818,7 +820,7 @@ static const char *symbol_white_list[] =
 	NULL
 };
 
-static const char *head_sections[] = { ".head.text*", NULL };
+static const char *head_sections[] = { __stringify(HEAD_TEXT_SECTION), NULL };
 static const char *linker_symbols[] =
 	{ "__init_begin", "_sinittext", "_einittext", NULL };
 
@@ -942,8 +944,8 @@ static int section_mismatch(const char *fromsec, const char *tosec)
  *           *probe_one, *_console, *_timer
  *
  * Pattern 3:
- *   Whitelist all refereces from .text.head to .init.data
- *   Whitelist all refereces from .text.head to .init.text
+ *   Whitelist all refereces from HEAD_TEXT_SECTION to .init.data
+ *   Whitelist all refereces from HEAD_TEXT_SECTION to .init.text
  *
  * Pattern 4:
  *   Some symbols belong to init section but still it is ok to reference
Signature · 1 line
1.6.2.1
#40 lore raw

Re: [PATCH 15/15] modpost: convert modpost to use HEAD_TEXT_SECTION macro.

Quoted 7 lines · Tim Abbott

Using the HEAD_TEXT_SECTION macro in modpost requires including a few more files from the linux headers.

The entry in head_sections[] ".head.text*" should be ".head.text", since only that section name is actually used by any architectures. I corrected this since preserving this over-broad pattern would have involved some string concatination nastiness.

I have seen linkers that added a number after the section names in certain situations. Often this is due to mix of sections with different flags ("ax" versus no "ax").
This is the background for the wildcard.

With your patchset we should not see any such users - but we should then warn if they showed up anyway if we do not silently support them.

Sam

#41 lore raw

Re: [PATCH 15/15] modpost: convert modpost to use HEAD_TEXT_SECTION macro.

Quoted 8 lines · Sam Ravnborg

I have seen linkers that added a number after the section names in certain situations. Often this is due to mix of sections with different flags ("ax" versus no "ax").
This is the background for the wildcard.

With your patchset we should not see any such users - but we should then warn if they showed up anyway if we do not silently support them.

Right, that is exactly the purpose of my
[PATCH] modpost: Check the section flags, not name, to catch missing "ax"/"aw"
<http://lkml.org/lkml/2009/2/26/503>
which is also the second patch in this thread (as [PATCH v3 2/3]).

This patch also catches some possible bugs that the existing check misses, even without -ffunction-sections, so it would be nice to get it merged independently.

Anders

#42 lore raw

Re: [PATCH 14/15] x86: convert to use __HEAD and HEAD_TEXT macros.

In reply to Tim Abbott
Quoted 17 lines · Tim Abbott

This has the consequence of changing the section name use for head code from ".text.head" to ".head.text". Since this commit changes all users in the architecture, this change should be harmless.

Cc Thomas Gleixner <[email protected]>
Cc Ingo Molnar <[email protected]>
Cc H. Peter Anvin <[email protected]>
---
 arch/x86/boot/compressed/Makefile         |    2 +
 arch/x86/boot/compressed/head_32.S        |    3 +-
 arch/x86/boot/compressed/head_64.S        |    3 +-
 arch/x86/boot/compressed/vmlinux_32.lds   |   43 -------------------------
 arch/x86/boot/compressed/vmlinux_32.lds.S |   45 ++++++++++++++++++++++++++
 arch/x86/boot/compressed/vmlinux_64.lds   |   48 ---------------------------
 arch/x86/boot/compressed/vmlinux_64.lds.S |   50 +++++++++++++++++++++++++++++
 arch/x86/kernel/head_32.S                 |    2 +-
Quoted 8 lines
 arch/x86/kernel/head_64.S                 |    2 +-
 arch/x86/kernel/vmlinux_32.lds.S          |    4 +-
 arch/x86/kernel/vmlinux_64.lds.S          |    2 +-
 11 files changed, 106 insertions(+), 98 deletions(-)
 delete mode 100644 arch/x86/boot/compressed/vmlinux_32.lds
 create mode 100644 arch/x86/boot/compressed/vmlinux_32.lds.S
 delete mode 100644 arch/x86/boot/compressed/vmlinux_64.lds
 create mode 100644 arch/x86/boot/compressed/vmlinux_64.lds.S

You should split this patch up in smaller pieces.
1) Rename from *lds to *.lds.S (and no other changes)
2) Introduce __HEAD & friends
And as before do not use HEAD_TEXT_SECTION

But otherwise it looks good. I will take a second look when new patches are sent.

Sam

#43 lore raw

Re: [PATCH 14/15] x86: convert to use __HEAD and HEAD_TEXT macros.

In reply to Tim Abbott
Quoted 4 lines · Tim Abbott

This has the consequence of changing the section name use for head code from ".text.head" to ".head.text". Since this commit changes all users in the architecture, this change should be harmless.

Btw, this one really needs to unify the two lds files first. Look at

diff -u arch/x86/boot/compressed/vmlinux_*.lds

output and realize that they're basially exctly the same except for trivial naming differences, and the fact that the64-bit version hs a "pgtable" thing.

So this really needs to be done by first unifying the thing so that there is _one_ arch/x86/boot/compressed/vmlinux.lds.S file with a preprocessor that takes care of the trivial differences, and then after that, the ".text.init" games should just be removed.

Notice the _removed_.

I don't think the compressed stuff has any actual "init" code. It is just a blob loader that loads the compressed kernel. There should be no need for the init sections being special, so the vmlinux.lds.S file should just have a ".text*" section.

Hmm?

Linus

#44 lore raw

Re: [PATCH 14/15] x86: convert to use __HEAD and HEAD_TEXT macros.

Quoted 5 lines · Linus Torvalds

Btw, this one really needs to unify the two lds files first. Look at

diff -u arch/x86/boot/compressed/vmlinux_*.lds

Quoted 7 lines

output and realize that they're basially exctly the same except for trivial naming differences, and the fact that the64-bit version hs a "pgtable" thing.

So this really needs to be done by first unifying the thing so that there is _one_ arch/x86/boot/compressed/vmlinux.lds.S file with a preprocessor that takes care of the trivial differences [..]

Something like this?

Looks obvious enough, and I think we could/should do the main "vmlinux_32/64" files somewhat similarly to this too. But they have a lot of other differences (the whole vsyscall setup etc), so that part isn't nearly as trivial.

I compile-tested on both x86-64 and i386, but just one config each.

There's some trivial cleanup there (make the output format a Kconfig thign rather than doing #ifdef's for it, and unify both 32-bit and 64-bit BSS end to "_ebss", where 32-bit used to use the traditional "_end"), but other than that it's really very mindless and straigt conversion.

For example, I think we should aim to remove "startup_32" vs "startup_64", and just call it "startup", and get rid of one more difference. I didn't do that.

Also, notice the comment in the unified vmlinux.lds.S talks about "head_64" and "startup_32" which is an odd and incorrect mix, but that was actually what the old 64-bit only lds file had, so the confusion isn't new, and now that mixing is arguably more accurate thanks to the vmlinux.lds.S file being shared between the two cases ;)

Linus

---
 arch/x86/Kconfig                                   |    5 ++
 arch/x86/boot/compressed/Makefile                  |    2 +-
 arch/x86/boot/compressed/head_32.S                 |    8 ++--
 .../compressed/{vmlinux_64.lds => vmlinux.lds.S}   |   11 +++++-
 arch/x86/boot/compressed/vmlinux_32.lds            |   43 --------------------
 5 files changed, 20 insertions(+), 49 deletions(-)
 rename arch/x86/boot/compressed/{vmlinux_64.lds => vmlinux.lds.S} (78%)
 delete mode 100644 arch/x86/boot/compressed/vmlinux_32.lds
arch/x86/Kconfig +5 -0
index c9086e6..30f41d7 100644
@@ -47,6 +47,11 @@ config X86
 	select HAVE_KERNEL_BZIP2
 	select HAVE_KERNEL_LZMA
 
+config OUTPUT_FORMAT
+	string
+	default "elf32-i386" if X86_32
+	default "elf64-x86-64" if X86_64
+
 config ARCH_DEFCONFIG
 	string
 	default "arch/x86/configs/i386_defconfig" if X86_32
arch/x86/boot/compressed/Makefile +1 -1
index 65551c9..0f4b5e2 100644
@@ -19,7 +19,7 @@ KBUILD_AFLAGS  := $(KBUILD_CFLAGS) -D__ASSEMBLY__
 LDFLAGS := -m elf_$(UTS_MACHINE)
 LDFLAGS_vmlinux := -T
 
-$(obj)/vmlinux: $(src)/vmlinux_$(BITS).lds $(obj)/head_$(BITS).o $(obj)/misc.o $(obj)/piggy.o FORCE
+$(obj)/vmlinux: $(obj)/vmlinux.lds $(obj)/head_$(BITS).o $(obj)/misc.o $(obj)/piggy.o FORCE
 	$(call if_changed,ld)
 	@:
 
arch/x86/boot/compressed/head_32.S +4 -4
index 3a8a866..85bd328 100644
@@ -88,9 +88,9 @@ ENTRY(startup_32)
  * where decompression in place becomes safe.
  */
 	pushl %esi
-	leal _end(%ebp), %esi
-	leal _end(%ebx), %edi
-	movl $(_end - startup_32), %ecx
+	leal _ebss(%ebp), %esi
+	leal _ebss(%ebx), %edi
+	movl $(_ebss - startup_32), %ecx
 	std
 	rep
 	movsb
@@ -121,7 +121,7 @@ relocated:
  */
 	xorl %eax,%eax
 	leal _edata(%ebx),%edi
-	leal _end(%ebx), %ecx
+	leal _ebss(%ebx), %ecx
 	subl %edi,%ecx
 	cld
 	rep
arch/x86/boot/compressed/vmlinux.lds.S renamed from arch/x86/boot/compressed/vmlinux_64.lds +10 -1
similarity index 78%
rename from arch/x86/boot/compressed/vmlinux_64.lds
rename to arch/x86/boot/compressed/vmlinux.lds.S
index bef1ac8..ffcb191 100644
@@ -1,6 +1,13 @@
-OUTPUT_FORMAT("elf64-x86-64", "elf64-x86-64", "elf64-x86-64")
+OUTPUT_FORMAT(CONFIG_OUTPUT_FORMAT, CONFIG_OUTPUT_FORMAT, CONFIG_OUTPUT_FORMAT)
+
+#ifdef CONFIG_X86_64
 OUTPUT_ARCH(i386:x86-64)
 ENTRY(startup_64)
+#else
+OUTPUT_ARCH(i386)
+ENTRY(startup_32)
+#endif
+
 SECTIONS
 {
 	/* Be careful parts of head_64.S assume startup_32 is at
@@ -38,11 +45,13 @@ SECTIONS
 		*(.bss)
 		*(.bss.*)
 		*(COMMON)
+#ifdef CONFIG_X86_64
 		. = ALIGN(8);
 		_end_before_pgt = . ;
 		. = ALIGN(4096);
 		pgtable = . ;
 		. = . + 4096 * 6;
+#endif
 		_ebss = .;
 	}
 }
arch/x86/boot/compressed/vmlinux_32.lds deleted +0 -43

43 lines not on this page · raw message

Acked-by
#45 lore raw

Re: [PATCH 14/15] x86: convert to use __HEAD and HEAD_TEXT macros.

Quoted 12 lines · Linus Torvalds
Quoted 9 lines · Linus Torvalds

Btw, this one really needs to unify the two lds files first. Look at

diff -u arch/x86/boot/compressed/vmlinux_*.lds

output and realize that they're basially exctly the same except for trivial naming differences, and the fact that the64-bit version hs a "pgtable" thing.

Quoted 5 lines
Quoted 3 lines

So this really needs to be done by first unifying the thing so that there is _one_ arch/x86/boot/compressed/vmlinux.lds.S file with a preprocessor that takes care of the trivial differences [..]

Something like this?

Looks good/correct.

You should add your s-o-b if you expect Ingo to pick it up.

Sam

#46 lore raw

Re: [PATCH 14/15] x86: convert to use __HEAD and HEAD_TEXT macros.

Quoted 17 lines · Sam Ravnborg
Quoted 16 lines · Linus Torvalds
Quoted 12 lines · Linus Torvalds

Btw, this one really needs to unify the two lds files first. Look at

diff -u arch/x86/boot/compressed/vmlinux_*.lds

output and realize that they're basially exctly the same except for trivial naming differences, and the fact that the64-bit version hs a "pgtable" thing.

So this really needs to be done by first unifying the thing so that there is _one_ arch/x86/boot/compressed/vmlinux.lds.S file with a preprocessor that takes care of the trivial differences [..]

Quoted 6 lines
Quoted 1 line

Something like this?

Looks good/correct.

You should add your s-o-b if you expect Ingo to pick it up.

Sure. I don't tend to add SOB lines for stuff that I'd not be ready to commit, but with some testing and other people looking at it, I think it's good to go.

As mentioned, though, the much more interesting case would be the _real_ kernel vmlinux.lds.S file, which is a lot more complex and where the differences between 32-bit and 64-bit cases aren't totally trivial.

Looking at

diff -u arch/x86/kernel/vmlinux_*.lds.S | less -S

output, many of them are just whitespace, and others are trivial and meaningless (comments in one, not the other, placement of alignment etc, different ordering of sections like "parainstructions"). Yet others seem to be things that we _could_ do in general, but that don't matter on one architecture or other (x86-64 has ".eh_frame" in the DISCARD section, i386 apparently doesn't ever generate them, we could just use the x86-64 version).

Linus

#47 lore raw

Re: [PATCH 14/15] x86: convert to use __HEAD and HEAD_TEXT macros.

Quoted 4 lines

As mentioned, though, the much more interesting case would be the _real_ kernel vmlinux.lds.S file, which is a lot more complex and where the differences between 32-bit and 64-bit cases aren't totally trivial.

Agreed.

But I expect someone else to do so.
When I am in unification mode I try to spend time on architectures with less people involved.

Sam

#48 lore raw

PATCH 1/1 x86: beautify vmlinux_64.lds.S

Beautify vmlinux_64.lds.S:
- Use tabs for indent
- Located curly braces like in C code
- Rearranged a few comments

There is no functional changes in this patch

The beautification is done to prepare a unification of the _32 and the _64 variants of the linker scripts.

---

It is tested that the lds file could build - no further tests done. [Despite my words in previous mail here is some x86 stuff]

Sam

arch/x86/kernel/vmlinux_64.lds.S +243 -205
index c874250..6d5a5b0 100644
@@ -15,69 +15,79 @@ OUTPUT_ARCH(i386:x86-64)
 ENTRY(phys_startup_64)
 jiffies_64 = jiffies;
 PHDRS {
-	text PT_LOAD FLAGS(5);	/* R_E */
-	data PT_LOAD FLAGS(7);	/* RWE */
-	user PT_LOAD FLAGS(7);	/* RWE */
+	text PT_LOAD FLAGS(5);		/* R_E */
+	data PT_LOAD FLAGS(7);		/* RWE */
+	user PT_LOAD FLAGS(7);		/* RWE */
 	data.init PT_LOAD FLAGS(7);	/* RWE */
 #ifdef CONFIG_SMP
 	percpu PT_LOAD FLAGS(7);	/* RWE */
 #endif
 	data.init2 PT_LOAD FLAGS(7);	/* RWE */
-	note PT_NOTE FLAGS(0);	/* ___ */
+	note PT_NOTE FLAGS(0);		/* ___ */
 }
 SECTIONS
 {
-  . = __START_KERNEL;
-  phys_startup_64 = startup_64 - LOAD_OFFSET;
-  .text :  AT(ADDR(.text) - LOAD_OFFSET) {
-	_text = .;			/* Text and read-only data */
-	/* First the code that has to be first for bootstrapping */
-	*(.text.head)
-	_stext = .;
-	/* Then the rest */
-	TEXT_TEXT
-	SCHED_TEXT
-	LOCK_TEXT
-	KPROBES_TEXT
-	IRQENTRY_TEXT
-	*(.fixup)
-	*(.gnu.warning)
-	_etext = .;		/* End of text section */
-  } :text = 0x9090
-
-  NOTES :text :note
-
-  . = ALIGN(16);		/* Exception table */
-  __ex_table : AT(ADDR(__ex_table) - LOAD_OFFSET) {
-  	__start___ex_table = .;
-	 *(__ex_table)
-  	__stop___ex_table = .;
-  } :text = 0x9090
-
-  RODATA
-
-  . = ALIGN(PAGE_SIZE);		/* Align data segment to page size boundary */
-				/* Data */
-  .data : AT(ADDR(.data) - LOAD_OFFSET) {
-	DATA_DATA
-	CONSTRUCTORS
-	_edata = .;			/* End of data section */
-	} :data
+	. = __START_KERNEL;
+	phys_startup_64 = startup_64 - LOAD_OFFSET;
+
+	/* Text and read-only data */
+	.text :  AT(ADDR(.text) - LOAD_OFFSET) {
+		_text = .;
+		/* First the code that has to be first for bootstrapping */
+		*(.text.head)
+		_stext = .;
+		/* Then the rest */
+		TEXT_TEXT
+		SCHED_TEXT
+		LOCK_TEXT
+		KPROBES_TEXT
+		IRQENTRY_TEXT
+		*(.fixup)
+		*(.gnu.warning)
+		/* End of text section */
+		_etext = .;
+	} :text = 0x9090
+
+	NOTES :text :note
+
+	/* Exception table */
+	. = ALIGN(16);
+	__ex_table : AT(ADDR(__ex_table) - LOAD_OFFSET) {
+		__start___ex_table = .;
+		 *(__ex_table)
+		__stop___ex_table = .;
+	} :text = 0x9090
 
+	RODATA
 
-  .data.cacheline_aligned : AT(ADDR(.data.cacheline_aligned) - LOAD_OFFSET) {
+	/* Align data segment to page size boundary */
 	. = ALIGN(PAGE_SIZE);
-	. = ALIGN(CONFIG_X86_L1_CACHE_BYTES);
-	*(.data.cacheline_aligned)
-  }
-  . = ALIGN(CONFIG_X86_INTERNODE_CACHE_BYTES);
-  .data.read_mostly : AT(ADDR(.data.read_mostly) - LOAD_OFFSET) {
-  	*(.data.read_mostly)
-  }
+	/* Data */
+	.data : AT(ADDR(.data) - LOAD_OFFSET) {
+		DATA_DATA
+		CONSTRUCTORS
+		/* End of data section */
+		_edata = .;
+	} :data
+
+
+	.data.cacheline_aligned :
+		AT(ADDR(.data.cacheline_aligned) - LOAD_OFFSET) {
+		. = ALIGN(PAGE_SIZE);
+		. = ALIGN(CONFIG_X86_L1_CACHE_BYTES);
+		*(.data.cacheline_aligned)
+	}
+
+	. = ALIGN(CONFIG_X86_INTERNODE_CACHE_BYTES);
+	.data.read_mostly : AT(ADDR(.data.read_mostly) - LOAD_OFFSET) {
+		*(.data.read_mostly)
+	}
 
 #define VSYSCALL_ADDR (-10*1024*1024)
-#define VSYSCALL_PHYS_ADDR ((LOADADDR(.data.read_mostly) + SIZEOF(.data.read_mostly) + 4095) & ~(4095))
-#define VSYSCALL_VIRT_ADDR ((ADDR(.data.read_mostly) + SIZEOF(.data.read_mostly) + 4095) & ~(4095))
+#define VSYSCALL_PHYS_ADDR ((LOADADDR(.data.read_mostly) + \
+                            SIZEOF(.data.read_mostly) + 4095) & ~(4095))
+#define VSYSCALL_VIRT_ADDR ((ADDR(.data.read_mostly) + \
+                            SIZEOF(.data.read_mostly) + 4095) & ~(4095))
 
 #define VLOAD_OFFSET (VSYSCALL_ADDR - VSYSCALL_PHYS_ADDR)
 #define VLOAD(x) (ADDR(x) - VLOAD_OFFSET)
@@ -85,37 +95,53 @@ SECTIONS
 #define VVIRT_OFFSET (VSYSCALL_ADDR - VSYSCALL_VIRT_ADDR)
 #define VVIRT(x) (ADDR(x) - VVIRT_OFFSET)
 
-  . = VSYSCALL_ADDR;
-  .vsyscall_0 :	 AT(VSYSCALL_PHYS_ADDR) { *(.vsyscall_0) } :user
-  __vsyscall_0 = VSYSCALL_VIRT_ADDR;
+	. = VSYSCALL_ADDR;
+	.vsyscall_0 : AT(VSYSCALL_PHYS_ADDR) {
+		*(.vsyscall_0)
+	} :user
+
+	__vsyscall_0 = VSYSCALL_VIRT_ADDR;
+
+	. = ALIGN(CONFIG_X86_L1_CACHE_BYTES);
+	.vsyscall_fn : AT(VLOAD(.vsyscall_fn)) {
+		*(.vsyscall_fn)
+	}
+
+	. = ALIGN(CONFIG_X86_L1_CACHE_BYTES);
+	.vsyscall_gtod_data : AT(VLOAD(.vsyscall_gtod_data)) {
+		*(.vsyscall_gtod_data)
+	}
 
-  . = ALIGN(CONFIG_X86_L1_CACHE_BYTES);
-  .vsyscall_fn : AT(VLOAD(.vsyscall_fn)) { *(.vsyscall_fn) }
-  . = ALIGN(CONFIG_X86_L1_CACHE_BYTES);
-  .vsyscall_gtod_data : AT(VLOAD(.vsyscall_gtod_data))
-		{ *(.vsyscall_gtod_data) }
-  vsyscall_gtod_data = VVIRT(.vsyscall_gtod_data);
-  .vsyscall_clock : AT(VLOAD(.vsyscall_clock))
-		{ *(.vsyscall_clock) }
-  vsyscall_clock = VVIRT(.vsyscall_clock);
+	vsyscall_gtod_data = VVIRT(.vsyscall_gtod_data);
+	.vsyscall_clock : AT(VLOAD(.vsyscall_clock)) {
+		*(.vsyscall_clock)
+	}
+	vsyscall_clock = VVIRT(.vsyscall_clock);
 
 
-  .vsyscall_1 ADDR(.vsyscall_0) + 1024: AT(VLOAD(.vsyscall_1))
-		{ *(.vsyscall_1) }
-  .vsyscall_2 ADDR(.vsyscall_0) + 2048: AT(VLOAD(.vsyscall_2))
-		{ *(.vsyscall_2) }
+	.vsyscall_1 ADDR(.vsyscall_0) + 1024: AT(VLOAD(.vsyscall_1)) {
+		*(.vsyscall_1)
+	}
+	.vsyscall_2 ADDR(.vsyscall_0) + 2048: AT(VLOAD(.vsyscall_2)) {
+		*(.vsyscall_2)
+	}
 
-  .vgetcpu_mode : AT(VLOAD(.vgetcpu_mode)) { *(.vgetcpu_mode) }
-  vgetcpu_mode = VVIRT(.vgetcpu_mode);
+	.vgetcpu_mode : AT(VLOAD(.vgetcpu_mode)) {
+		*(.vgetcpu_mode)
+	}
+	vgetcpu_mode = VVIRT(.vgetcpu_mode);
 
-  . = ALIGN(CONFIG_X86_L1_CACHE_BYTES);
-  .jiffies : AT(VLOAD(.jiffies)) { *(.jiffies) }
-  jiffies = VVIRT(.jiffies);
+	. = ALIGN(CONFIG_X86_L1_CACHE_BYTES);
+	.jiffies : AT(VLOAD(.jiffies)) {
+		*(.jiffies)
+	}
+	jiffies = VVIRT(.jiffies);
 
-  .vsyscall_3 ADDR(.vsyscall_0) + 3072: AT(VLOAD(.vsyscall_3))
-		{ *(.vsyscall_3) }
+	.vsyscall_3 ADDR(.vsyscall_0) + 3072: AT(VLOAD(.vsyscall_3)) {
+		*(.vsyscall_3)
+	}
 
-  . = VSYSCALL_VIRT_ADDR + PAGE_SIZE;
+	. = VSYSCALL_VIRT_ADDR + PAGE_SIZE;
 
 #undef VSYSCALL_ADDR
 #undef VSYSCALL_PHYS_ADDR
@@ -125,156 +151,168 @@ SECTIONS
 #undef VVIRT_OFFSET
 #undef VVIRT
 
-  .data.init_task : AT(ADDR(.data.init_task) - LOAD_OFFSET) {
-	. = ALIGN(THREAD_SIZE);	/* init_task */
-	*(.data.init_task)
-  }:data.init
+	/* init_task */
+	.data.init_task : AT(ADDR(.data.init_task) - LOAD_OFFSET) {
+		. = ALIGN(THREAD_SIZE);
+		*(.data.init_task)
+	} :data.init
 
-  .data.page_aligned : AT(ADDR(.data.page_aligned) - LOAD_OFFSET) {
-	. = ALIGN(PAGE_SIZE);
-	*(.data.page_aligned)
-  }
+	.data.page_aligned : AT(ADDR(.data.page_aligned) - LOAD_OFFSET) {
+		. = ALIGN(PAGE_SIZE);
+		*(.data.page_aligned)
+	}
 
-  .smp_locks : AT(ADDR(.smp_locks) - LOAD_OFFSET) {
-	/* might get freed after init */
-	. = ALIGN(PAGE_SIZE);
-	__smp_alt_begin = .;
-	__smp_locks = .;
-	*(.smp_locks)
-	__smp_locks_end = .;
+	.smp_locks : AT(ADDR(.smp_locks) - LOAD_OFFSET) {
+		/* might get freed after init */
+		. = ALIGN(PAGE_SIZE);
+		__smp_alt_begin = .;
+		__smp_locks = .;
+		*(.smp_locks)
+		__smp_locks_end = .;
+		. = ALIGN(PAGE_SIZE);
+		__smp_alt_end = .;
+	}
+
+	/* Init code and data */
 	. = ALIGN(PAGE_SIZE);
-	__smp_alt_end = .;
-  }
-
-  . = ALIGN(PAGE_SIZE);		/* Init code and data */
-  __init_begin = .;	/* paired with __init_end */
-  .init.text : AT(ADDR(.init.text) - LOAD_OFFSET) {
-	_sinittext = .;
-	INIT_TEXT
-	_einittext = .;
-  }
-  .init.data : AT(ADDR(.init.data) - LOAD_OFFSET) {
-	__initdata_begin = .;
-	INIT_DATA
-	__initdata_end = .;
-   }
-
-  .init.setup : AT(ADDR(.init.setup) - LOAD_OFFSET) {
-	. = ALIGN(16);
-	__setup_start = .;
-	*(.init.setup)
-	__setup_end = .;
-  }
-  .initcall.init : AT(ADDR(.initcall.init) - LOAD_OFFSET) {
-	__initcall_start = .;
-	INITCALLS
-	__initcall_end = .;
-  }
-  .con_initcall.init : AT(ADDR(.con_initcall.init) - LOAD_OFFSET) {
-	__con_initcall_start = .;
-	*(.con_initcall.init)
-	__con_initcall_end = .;
-  }
-  .x86_cpu_dev.init : AT(ADDR(.x86_cpu_dev.init) - LOAD_OFFSET) {
-	__x86_cpu_dev_start = .;
-	*(.x86_cpu_dev.init)
-	__x86_cpu_dev_end = .;
-  }
-  SECURITY_INIT
-
-  . = ALIGN(8);
-  .parainstructions : AT(ADDR(.parainstructions) - LOAD_OFFSET) {
-	__parainstructions = .;
-       *(.parainstructions)
-	__parainstructions_end = .;
-  }
-
-  .altinstructions : AT(ADDR(.altinstructions) - LOAD_OFFSET) {
+	__init_begin = .;	/* paired with __init_end */
+	.init.text : AT(ADDR(.init.text) - LOAD_OFFSET) {
+		_sinittext = .;
+		INIT_TEXT
+		_einittext = .;
+	}
+
+	.init.data : AT(ADDR(.init.data) - LOAD_OFFSET) {
+		__initdata_begin = .;
+		INIT_DATA
+		__initdata_end = .;
+	}
+
+	.init.setup : AT(ADDR(.init.setup) - LOAD_OFFSET) {
+		. = ALIGN(16);
+		__setup_start = .;
+		*(.init.setup)
+		__setup_end = .;
+	}
+
+	.initcall.init : AT(ADDR(.initcall.init) - LOAD_OFFSET) {
+		__initcall_start = .;
+		INITCALLS
+		__initcall_end = .;
+	}
+
+	.con_initcall.init : AT(ADDR(.con_initcall.init) - LOAD_OFFSET) {
+		__con_initcall_start = .;
+		*(.con_initcall.init)
+		__con_initcall_end = .;
+	}
+
+	.x86_cpu_dev.init : AT(ADDR(.x86_cpu_dev.init) - LOAD_OFFSET) {
+		__x86_cpu_dev_start = .;
+		*(.x86_cpu_dev.init)
+		__x86_cpu_dev_end = .;
+	}
+
+	SECURITY_INIT
+
 	. = ALIGN(8);
-	__alt_instructions = .;
-	*(.altinstructions)
-	__alt_instructions_end = .;
-  }
-  .altinstr_replacement : AT(ADDR(.altinstr_replacement) - LOAD_OFFSET) {
-	*(.altinstr_replacement)
-  }
-  /* .exit.text is discard at runtime, not link time, to deal with references
-     from .altinstructions and .eh_frame */
-  .exit.text : AT(ADDR(.exit.text) - LOAD_OFFSET) {
-	EXIT_TEXT
-  }
-  .exit.data : AT(ADDR(.exit.data) - LOAD_OFFSET) {
-	EXIT_DATA
-  }
+	.parainstructions : AT(ADDR(.parainstructions) - LOAD_OFFSET) {
+		__parainstructions = .;
+		*(.parainstructions)
+		__parainstructions_end = .;
+	}
+
+	.altinstructions : AT(ADDR(.altinstructions) - LOAD_OFFSET) {
+		. = ALIGN(8);
+		__alt_instructions = .;
+		*(.altinstructions)
+		__alt_instructions_end = .;
+	}
+
+	.altinstr_replacement : AT(ADDR(.altinstr_replacement) - LOAD_OFFSET) {
+		*(.altinstr_replacement)
+	}
+
+	/*
+	 * .exit.text is discard at runtime, not link time, to deal with
+	 *  references from .altinstructions and .eh_frame
+	 */
+	.exit.text : AT(ADDR(.exit.text) - LOAD_OFFSET) {
+		EXIT_TEXT
+	}
+
+	.exit.data : AT(ADDR(.exit.data) - LOAD_OFFSET) {
+		EXIT_DATA
+	}
 
 #ifdef CONFIG_BLK_DEV_INITRD
-  . = ALIGN(PAGE_SIZE);
-  .init.ramfs : AT(ADDR(.init.ramfs) - LOAD_OFFSET) {
-	__initramfs_start = .;
-	*(.init.ramfs)
-	__initramfs_end = .;
-  }
+	. = ALIGN(PAGE_SIZE);
+	.init.ramfs : AT(ADDR(.init.ramfs) - LOAD_OFFSET) {
+		__initramfs_start = .;
+		*(.init.ramfs)
+		__initramfs_end = .;
+	}
 #endif
 
 #ifdef CONFIG_SMP
-  /*
-   * percpu offsets are zero-based on SMP.  PERCPU_VADDR() changes the
-   * output PHDR, so the next output section - __data_nosave - should
-   * start another section data.init2.  Also, pda should be at the head of
-   * percpu area.  Preallocate it and define the percpu offset symbol
-   * so that it can be accessed as a percpu variable.
-   */
-  . = ALIGN(PAGE_SIZE);
-  PERCPU_VADDR(0, :percpu)
+	/*
+	 * percpu offsets are zero-based on SMP.  PERCPU_VADDR() changes the
+	 * output PHDR, so the next output section - __data_nosave - should
+	 * start another section data.init2.  Also, pda should be at the head of
+	 * percpu area.  Preallocate it and define the percpu offset symbol
+	 * so that it can be accessed as a percpu variable.
+	 */
+	. = ALIGN(PAGE_SIZE);
+	PERCPU_VADDR(0, :percpu)
 #else
-  PERCPU(PAGE_SIZE)
+	PERCPU(PAGE_SIZE)
 #endif
 
-  . = ALIGN(PAGE_SIZE);
-  __init_end = .;
-
-  .data_nosave : AT(ADDR(.data_nosave) - LOAD_OFFSET) {
-	. = ALIGN(PAGE_SIZE);
-	__nosave_begin = .;
-	*(.data.nosave)
-	. = ALIGN(PAGE_SIZE);
-	__nosave_end = .;
-  } :data.init2 /* use another section data.init2, see PERCPU_VADDR() above */
-
-  .bss : AT(ADDR(.bss) - LOAD_OFFSET) {
 	. = ALIGN(PAGE_SIZE);
-	__bss_start = .;		/* BSS */
-	*(.bss.page_aligned)
-	*(.bss)
-	__bss_stop = .;
-  }
+	__init_end = .;
+
+	.data_nosave : AT(ADDR(.data_nosave) - LOAD_OFFSET) {
+		. = ALIGN(PAGE_SIZE);
+		__nosave_begin = .;
+		*(.data.nosave)
+		. = ALIGN(PAGE_SIZE);
+		__nosave_end = .;
+	} :data.init2
+	/* use another section data.init2, see PERCPU_VADDR() above */
+
+	.bss : AT(ADDR(.bss) - LOAD_OFFSET) {
+		. = ALIGN(PAGE_SIZE);
+		__bss_start = .;		/* BSS */
+		*(.bss.page_aligned)
+		*(.bss)
+		__bss_stop = .;
+	}
 
-  .brk : AT(ADDR(.brk) - LOAD_OFFSET) {
-	. = ALIGN(PAGE_SIZE);
-	__brk_base = . ;
- 	. += 64 * 1024 ;	/* 64k alignment slop space */
-	*(.brk_reservation)	/* areas brk users have reserved */
-	__brk_limit = . ;
-  }
-
-  _end = . ;
-
-  /* Sections to be discarded */
-  /DISCARD/ : {
-	*(.exitcall.exit)
-	*(.eh_frame)
-	*(.discard)
+	.brk : AT(ADDR(.brk) - LOAD_OFFSET) {
+		. = ALIGN(PAGE_SIZE);
+		__brk_base = .;
+		. += 64 * 1024;		/* 64k alignment slop space */
+		*(.brk_reservation)	/* areas brk users have reserved */
+		__brk_limit = .;
 	}
 
-  STABS_DEBUG
+	_end = . ;
 
-  DWARF_DEBUG
+	/* Sections to be discarded */
+	/DISCARD/ : {
+		*(.exitcall.exit)
+		*(.eh_frame)
+		*(.discard)
+	}
+
+	STABS_DEBUG
+	DWARF_DEBUG
 }
 
- /*
-  * Per-cpu symbols which need to be offset from __per_cpu_load
-  * for the boot processor.
-  */
+/*
+ * Per-cpu symbols which need to be offset from __per_cpu_load
+ * for the boot processor.
+ */
 #define INIT_PER_CPU(x) init_per_cpu__##x = per_cpu__##x + __per_cpu_load
 INIT_PER_CPU(gdt_page);
 INIT_PER_CPU(irq_stack_union);
#49 lore raw

[tip:x86/kbuild] x86: beautify vmlinux_64.lds.S

Commit-ID:  0a3ec21fcd311b26ab0f249d62960e127bc20ca8
Gitweb:     http://git.kernel.org/tip/0a3ec21fcd311b26ab0f249d62960e127bc20ca8
Author:     Sam Ravnborg <[email protected]>
AuthorDate: Sun, 26 Apr 2009 23:07:42 +0200
Committer:  Ingo Molnar <[email protected]>
CommitDate: Mon, 27 Apr 2009 06:30:58 +0200

x86: beautify vmlinux_64.lds.S

Beautify vmlinux_64.lds.S:

- Use tabs for indent
- Located curly braces like in C code
- Rearranged a few comments

There is no functional changes in this patch

The beautification is done to prepare a unification of the _32 and the _64 variants of the linker scripts.

[ Impact: cleanup ]

Cc Tim Abbott <[email protected]>
Cc Linus Torvalds <[email protected]>

LKML-Reference: <[email protected]>

---
 arch/x86/kernel/vmlinux_64.lds.S |  448 +++++++++++++++++++++-----------------
 1 files changed, 243 insertions(+), 205 deletions(-)
arch/x86/kernel/vmlinux_64.lds.S +243 -205
index c874250..6d5a5b0 100644
@@ -15,69 +15,79 @@ OUTPUT_ARCH(i386:x86-64)
 ENTRY(phys_startup_64)
 jiffies_64 = jiffies;
 PHDRS {
-	text PT_LOAD FLAGS(5);	/* R_E */
-	data PT_LOAD FLAGS(7);	/* RWE */
-	user PT_LOAD FLAGS(7);	/* RWE */
+	text PT_LOAD FLAGS(5);		/* R_E */
+	data PT_LOAD FLAGS(7);		/* RWE */
+	user PT_LOAD FLAGS(7);		/* RWE */
 	data.init PT_LOAD FLAGS(7);	/* RWE */
 #ifdef CONFIG_SMP
 	percpu PT_LOAD FLAGS(7);	/* RWE */
 #endif
 	data.init2 PT_LOAD FLAGS(7);	/* RWE */
-	note PT_NOTE FLAGS(0);	/* ___ */
+	note PT_NOTE FLAGS(0);		/* ___ */
 }
 SECTIONS
 {
-  . = __START_KERNEL;
-  phys_startup_64 = startup_64 - LOAD_OFFSET;
-  .text :  AT(ADDR(.text) - LOAD_OFFSET) {
-	_text = .;			/* Text and read-only data */
-	/* First the code that has to be first for bootstrapping */
-	*(.text.head)
-	_stext = .;
-	/* Then the rest */
-	TEXT_TEXT
-	SCHED_TEXT
-	LOCK_TEXT
-	KPROBES_TEXT
-	IRQENTRY_TEXT
-	*(.fixup)
-	*(.gnu.warning)
-	_etext = .;		/* End of text section */
-  } :text = 0x9090
-
-  NOTES :text :note
-
-  . = ALIGN(16);		/* Exception table */
-  __ex_table : AT(ADDR(__ex_table) - LOAD_OFFSET) {
-  	__start___ex_table = .;
-	 *(__ex_table)
-  	__stop___ex_table = .;
-  } :text = 0x9090
-
-  RODATA
-
-  . = ALIGN(PAGE_SIZE);		/* Align data segment to page size boundary */
-				/* Data */
-  .data : AT(ADDR(.data) - LOAD_OFFSET) {
-	DATA_DATA
-	CONSTRUCTORS
-	_edata = .;			/* End of data section */
-	} :data
+	. = __START_KERNEL;
+	phys_startup_64 = startup_64 - LOAD_OFFSET;
+
+	/* Text and read-only data */
+	.text :  AT(ADDR(.text) - LOAD_OFFSET) {
+		_text = .;
+		/* First the code that has to be first for bootstrapping */
+		*(.text.head)
+		_stext = .;
+		/* Then the rest */
+		TEXT_TEXT
+		SCHED_TEXT
+		LOCK_TEXT
+		KPROBES_TEXT
+		IRQENTRY_TEXT
+		*(.fixup)
+		*(.gnu.warning)
+		/* End of text section */
+		_etext = .;
+	} :text = 0x9090
+
+	NOTES :text :note
+
+	/* Exception table */
+	. = ALIGN(16);
+	__ex_table : AT(ADDR(__ex_table) - LOAD_OFFSET) {
+		__start___ex_table = .;
+		 *(__ex_table)
+		__stop___ex_table = .;
+	} :text = 0x9090
 
+	RODATA
 
-  .data.cacheline_aligned : AT(ADDR(.data.cacheline_aligned) - LOAD_OFFSET) {
+	/* Align data segment to page size boundary */
 	. = ALIGN(PAGE_SIZE);
-	. = ALIGN(CONFIG_X86_L1_CACHE_BYTES);
-	*(.data.cacheline_aligned)
-  }
-  . = ALIGN(CONFIG_X86_INTERNODE_CACHE_BYTES);
-  .data.read_mostly : AT(ADDR(.data.read_mostly) - LOAD_OFFSET) {
-  	*(.data.read_mostly)
-  }
+	/* Data */
+	.data : AT(ADDR(.data) - LOAD_OFFSET) {
+		DATA_DATA
+		CONSTRUCTORS
+		/* End of data section */
+		_edata = .;
+	} :data
+
+
+	.data.cacheline_aligned :
+		AT(ADDR(.data.cacheline_aligned) - LOAD_OFFSET) {
+		. = ALIGN(PAGE_SIZE);
+		. = ALIGN(CONFIG_X86_L1_CACHE_BYTES);
+		*(.data.cacheline_aligned)
+	}
+
+	. = ALIGN(CONFIG_X86_INTERNODE_CACHE_BYTES);
+	.data.read_mostly : AT(ADDR(.data.read_mostly) - LOAD_OFFSET) {
+		*(.data.read_mostly)
+	}
 
 #define VSYSCALL_ADDR (-10*1024*1024)
-#define VSYSCALL_PHYS_ADDR ((LOADADDR(.data.read_mostly) + SIZEOF(.data.read_mostly) + 4095) & ~(4095))
-#define VSYSCALL_VIRT_ADDR ((ADDR(.data.read_mostly) + SIZEOF(.data.read_mostly) + 4095) & ~(4095))
+#define VSYSCALL_PHYS_ADDR ((LOADADDR(.data.read_mostly) + \
+                            SIZEOF(.data.read_mostly) + 4095) & ~(4095))
+#define VSYSCALL_VIRT_ADDR ((ADDR(.data.read_mostly) + \
+                            SIZEOF(.data.read_mostly) + 4095) & ~(4095))
 
 #define VLOAD_OFFSET (VSYSCALL_ADDR - VSYSCALL_PHYS_ADDR)
 #define VLOAD(x) (ADDR(x) - VLOAD_OFFSET)
@@ -85,37 +95,53 @@ SECTIONS
 #define VVIRT_OFFSET (VSYSCALL_ADDR - VSYSCALL_VIRT_ADDR)
 #define VVIRT(x) (ADDR(x) - VVIRT_OFFSET)
 
-  . = VSYSCALL_ADDR;
-  .vsyscall_0 :	 AT(VSYSCALL_PHYS_ADDR) { *(.vsyscall_0) } :user
-  __vsyscall_0 = VSYSCALL_VIRT_ADDR;
+	. = VSYSCALL_ADDR;
+	.vsyscall_0 : AT(VSYSCALL_PHYS_ADDR) {
+		*(.vsyscall_0)
+	} :user
+
+	__vsyscall_0 = VSYSCALL_VIRT_ADDR;
+
+	. = ALIGN(CONFIG_X86_L1_CACHE_BYTES);
+	.vsyscall_fn : AT(VLOAD(.vsyscall_fn)) {
+		*(.vsyscall_fn)
+	}
+
+	. = ALIGN(CONFIG_X86_L1_CACHE_BYTES);
+	.vsyscall_gtod_data : AT(VLOAD(.vsyscall_gtod_data)) {
+		*(.vsyscall_gtod_data)
+	}
 
-  . = ALIGN(CONFIG_X86_L1_CACHE_BYTES);
-  .vsyscall_fn : AT(VLOAD(.vsyscall_fn)) { *(.vsyscall_fn) }
-  . = ALIGN(CONFIG_X86_L1_CACHE_BYTES);
-  .vsyscall_gtod_data : AT(VLOAD(.vsyscall_gtod_data))
-		{ *(.vsyscall_gtod_data) }
-  vsyscall_gtod_data = VVIRT(.vsyscall_gtod_data);
-  .vsyscall_clock : AT(VLOAD(.vsyscall_clock))
-		{ *(.vsyscall_clock) }
-  vsyscall_clock = VVIRT(.vsyscall_clock);
+	vsyscall_gtod_data = VVIRT(.vsyscall_gtod_data);
+	.vsyscall_clock : AT(VLOAD(.vsyscall_clock)) {
+		*(.vsyscall_clock)
+	}
+	vsyscall_clock = VVIRT(.vsyscall_clock);
 
 
-  .vsyscall_1 ADDR(.vsyscall_0) + 1024: AT(VLOAD(.vsyscall_1))
-		{ *(.vsyscall_1) }
-  .vsyscall_2 ADDR(.vsyscall_0) + 2048: AT(VLOAD(.vsyscall_2))
-		{ *(.vsyscall_2) }
+	.vsyscall_1 ADDR(.vsyscall_0) + 1024: AT(VLOAD(.vsyscall_1)) {
+		*(.vsyscall_1)
+	}
+	.vsyscall_2 ADDR(.vsyscall_0) + 2048: AT(VLOAD(.vsyscall_2)) {
+		*(.vsyscall_2)
+	}
 
-  .vgetcpu_mode : AT(VLOAD(.vgetcpu_mode)) { *(.vgetcpu_mode) }
-  vgetcpu_mode = VVIRT(.vgetcpu_mode);
+	.vgetcpu_mode : AT(VLOAD(.vgetcpu_mode)) {
+		*(.vgetcpu_mode)
+	}
+	vgetcpu_mode = VVIRT(.vgetcpu_mode);
 
-  . = ALIGN(CONFIG_X86_L1_CACHE_BYTES);
-  .jiffies : AT(VLOAD(.jiffies)) { *(.jiffies) }
-  jiffies = VVIRT(.jiffies);
+	. = ALIGN(CONFIG_X86_L1_CACHE_BYTES);
+	.jiffies : AT(VLOAD(.jiffies)) {
+		*(.jiffies)
+	}
+	jiffies = VVIRT(.jiffies);
 
-  .vsyscall_3 ADDR(.vsyscall_0) + 3072: AT(VLOAD(.vsyscall_3))
-		{ *(.vsyscall_3) }
+	.vsyscall_3 ADDR(.vsyscall_0) + 3072: AT(VLOAD(.vsyscall_3)) {
+		*(.vsyscall_3)
+	}
 
-  . = VSYSCALL_VIRT_ADDR + PAGE_SIZE;
+	. = VSYSCALL_VIRT_ADDR + PAGE_SIZE;
 
 #undef VSYSCALL_ADDR
 #undef VSYSCALL_PHYS_ADDR
@@ -125,156 +151,168 @@ SECTIONS
 #undef VVIRT_OFFSET
 #undef VVIRT
 
-  .data.init_task : AT(ADDR(.data.init_task) - LOAD_OFFSET) {
-	. = ALIGN(THREAD_SIZE);	/* init_task */
-	*(.data.init_task)
-  }:data.init
+	/* init_task */
+	.data.init_task : AT(ADDR(.data.init_task) - LOAD_OFFSET) {
+		. = ALIGN(THREAD_SIZE);
+		*(.data.init_task)
+	} :data.init
 
-  .data.page_aligned : AT(ADDR(.data.page_aligned) - LOAD_OFFSET) {
-	. = ALIGN(PAGE_SIZE);
-	*(.data.page_aligned)
-  }
+	.data.page_aligned : AT(ADDR(.data.page_aligned) - LOAD_OFFSET) {
+		. = ALIGN(PAGE_SIZE);
+		*(.data.page_aligned)
+	}
 
-  .smp_locks : AT(ADDR(.smp_locks) - LOAD_OFFSET) {
-	/* might get freed after init */
-	. = ALIGN(PAGE_SIZE);
-	__smp_alt_begin = .;
-	__smp_locks = .;
-	*(.smp_locks)
-	__smp_locks_end = .;
+	.smp_locks : AT(ADDR(.smp_locks) - LOAD_OFFSET) {
+		/* might get freed after init */
+		. = ALIGN(PAGE_SIZE);
+		__smp_alt_begin = .;
+		__smp_locks = .;
+		*(.smp_locks)
+		__smp_locks_end = .;
+		. = ALIGN(PAGE_SIZE);
+		__smp_alt_end = .;
+	}
+
+	/* Init code and data */
 	. = ALIGN(PAGE_SIZE);
-	__smp_alt_end = .;
-  }
-
-  . = ALIGN(PAGE_SIZE);		/* Init code and data */
-  __init_begin = .;	/* paired with __init_end */
-  .init.text : AT(ADDR(.init.text) - LOAD_OFFSET) {
-	_sinittext = .;
-	INIT_TEXT
-	_einittext = .;
-  }
-  .init.data : AT(ADDR(.init.data) - LOAD_OFFSET) {
-	__initdata_begin = .;
-	INIT_DATA
-	__initdata_end = .;
-   }
-
-  .init.setup : AT(ADDR(.init.setup) - LOAD_OFFSET) {
-	. = ALIGN(16);
-	__setup_start = .;
-	*(.init.setup)
-	__setup_end = .;
-  }
-  .initcall.init : AT(ADDR(.initcall.init) - LOAD_OFFSET) {
-	__initcall_start = .;
-	INITCALLS
-	__initcall_end = .;
-  }
-  .con_initcall.init : AT(ADDR(.con_initcall.init) - LOAD_OFFSET) {
-	__con_initcall_start = .;
-	*(.con_initcall.init)
-	__con_initcall_end = .;
-  }
-  .x86_cpu_dev.init : AT(ADDR(.x86_cpu_dev.init) - LOAD_OFFSET) {
-	__x86_cpu_dev_start = .;
-	*(.x86_cpu_dev.init)
-	__x86_cpu_dev_end = .;
-  }
-  SECURITY_INIT
-
-  . = ALIGN(8);
-  .parainstructions : AT(ADDR(.parainstructions) - LOAD_OFFSET) {
-	__parainstructions = .;
-       *(.parainstructions)
-	__parainstructions_end = .;
-  }
-
-  .altinstructions : AT(ADDR(.altinstructions) - LOAD_OFFSET) {
+	__init_begin = .;	/* paired with __init_end */
+	.init.text : AT(ADDR(.init.text) - LOAD_OFFSET) {
+		_sinittext = .;
+		INIT_TEXT
+		_einittext = .;
+	}
+
+	.init.data : AT(ADDR(.init.data) - LOAD_OFFSET) {
+		__initdata_begin = .;
+		INIT_DATA
+		__initdata_end = .;
+	}
+
+	.init.setup : AT(ADDR(.init.setup) - LOAD_OFFSET) {
+		. = ALIGN(16);
+		__setup_start = .;
+		*(.init.setup)
+		__setup_end = .;
+	}
+
+	.initcall.init : AT(ADDR(.initcall.init) - LOAD_OFFSET) {
+		__initcall_start = .;
+		INITCALLS
+		__initcall_end = .;
+	}
+
+	.con_initcall.init : AT(ADDR(.con_initcall.init) - LOAD_OFFSET) {
+		__con_initcall_start = .;
+		*(.con_initcall.init)
+		__con_initcall_end = .;
+	}
+
+	.x86_cpu_dev.init : AT(ADDR(.x86_cpu_dev.init) - LOAD_OFFSET) {
+		__x86_cpu_dev_start = .;
+		*(.x86_cpu_dev.init)
+		__x86_cpu_dev_end = .;
+	}
+
+	SECURITY_INIT
+
 	. = ALIGN(8);
-	__alt_instructions = .;
-	*(.altinstructions)
-	__alt_instructions_end = .;
-  }
-  .altinstr_replacement : AT(ADDR(.altinstr_replacement) - LOAD_OFFSET) {
-	*(.altinstr_replacement)
-  }
-  /* .exit.text is discard at runtime, not link time, to deal with references
-     from .altinstructions and .eh_frame */
-  .exit.text : AT(ADDR(.exit.text) - LOAD_OFFSET) {
-	EXIT_TEXT
-  }
-  .exit.data : AT(ADDR(.exit.data) - LOAD_OFFSET) {
-	EXIT_DATA
-  }
+	.parainstructions : AT(ADDR(.parainstructions) - LOAD_OFFSET) {
+		__parainstructions = .;
+		*(.parainstructions)
+		__parainstructions_end = .;
+	}
+
+	.altinstructions : AT(ADDR(.altinstructions) - LOAD_OFFSET) {
+		. = ALIGN(8);
+		__alt_instructions = .;
+		*(.altinstructions)
+		__alt_instructions_end = .;
+	}
+
+	.altinstr_replacement : AT(ADDR(.altinstr_replacement) - LOAD_OFFSET) {
+		*(.altinstr_replacement)
+	}
+
+	/*
+	 * .exit.text is discard at runtime, not link time, to deal with
+	 *  references from .altinstructions and .eh_frame
+	 */
+	.exit.text : AT(ADDR(.exit.text) - LOAD_OFFSET) {
+		EXIT_TEXT
+	}
+
+	.exit.data : AT(ADDR(.exit.data) - LOAD_OFFSET) {
+		EXIT_DATA
+	}
 
 #ifdef CONFIG_BLK_DEV_INITRD
-  . = ALIGN(PAGE_SIZE);
-  .init.ramfs : AT(ADDR(.init.ramfs) - LOAD_OFFSET) {
-	__initramfs_start = .;
-	*(.init.ramfs)
-	__initramfs_end = .;
-  }
+	. = ALIGN(PAGE_SIZE);
+	.init.ramfs : AT(ADDR(.init.ramfs) - LOAD_OFFSET) {
+		__initramfs_start = .;
+		*(.init.ramfs)
+		__initramfs_end = .;
+	}
 #endif
 
 #ifdef CONFIG_SMP
-  /*
-   * percpu offsets are zero-based on SMP.  PERCPU_VADDR() changes the
-   * output PHDR, so the next output section - __data_nosave - should
-   * start another section data.init2.  Also, pda should be at the head of
-   * percpu area.  Preallocate it and define the percpu offset symbol
-   * so that it can be accessed as a percpu variable.
-   */
-  . = ALIGN(PAGE_SIZE);
-  PERCPU_VADDR(0, :percpu)
+	/*
+	 * percpu offsets are zero-based on SMP.  PERCPU_VADDR() changes the
+	 * output PHDR, so the next output section - __data_nosave - should
+	 * start another section data.init2.  Also, pda should be at the head of
+	 * percpu area.  Preallocate it and define the percpu offset symbol
+	 * so that it can be accessed as a percpu variable.
+	 */
+	. = ALIGN(PAGE_SIZE);
+	PERCPU_VADDR(0, :percpu)
 #else
-  PERCPU(PAGE_SIZE)
+	PERCPU(PAGE_SIZE)
 #endif
 
-  . = ALIGN(PAGE_SIZE);
-  __init_end = .;
-
-  .data_nosave : AT(ADDR(.data_nosave) - LOAD_OFFSET) {
-	. = ALIGN(PAGE_SIZE);
-	__nosave_begin = .;
-	*(.data.nosave)
-	. = ALIGN(PAGE_SIZE);
-	__nosave_end = .;
-  } :data.init2 /* use another section data.init2, see PERCPU_VADDR() above */
-
-  .bss : AT(ADDR(.bss) - LOAD_OFFSET) {
 	. = ALIGN(PAGE_SIZE);
-	__bss_start = .;		/* BSS */
-	*(.bss.page_aligned)
-	*(.bss)
-	__bss_stop = .;
-  }
+	__init_end = .;
+
+	.data_nosave : AT(ADDR(.data_nosave) - LOAD_OFFSET) {
+		. = ALIGN(PAGE_SIZE);
+		__nosave_begin = .;
+		*(.data.nosave)
+		. = ALIGN(PAGE_SIZE);
+		__nosave_end = .;
+	} :data.init2
+	/* use another section data.init2, see PERCPU_VADDR() above */
+
+	.bss : AT(ADDR(.bss) - LOAD_OFFSET) {
+		. = ALIGN(PAGE_SIZE);
+		__bss_start = .;		/* BSS */
+		*(.bss.page_aligned)
+		*(.bss)
+		__bss_stop = .;
+	}
 
-  .brk : AT(ADDR(.brk) - LOAD_OFFSET) {
-	. = ALIGN(PAGE_SIZE);
-	__brk_base = . ;
- 	. += 64 * 1024 ;	/* 64k alignment slop space */
-	*(.brk_reservation)	/* areas brk users have reserved */
-	__brk_limit = . ;
-  }
-
-  _end = . ;
-
-  /* Sections to be discarded */
-  /DISCARD/ : {
-	*(.exitcall.exit)
-	*(.eh_frame)
-	*(.discard)
+	.brk : AT(ADDR(.brk) - LOAD_OFFSET) {
+		. = ALIGN(PAGE_SIZE);
+		__brk_base = .;
+		. += 64 * 1024;		/* 64k alignment slop space */
+		*(.brk_reservation)	/* areas brk users have reserved */
+		__brk_limit = .;
 	}
 
-  STABS_DEBUG
+	_end = . ;
 
-  DWARF_DEBUG
+	/* Sections to be discarded */
+	/DISCARD/ : {
+		*(.exitcall.exit)
+		*(.eh_frame)
+		*(.discard)
+	}
+
+	STABS_DEBUG
+	DWARF_DEBUG
 }
 
- /*
-  * Per-cpu symbols which need to be offset from __per_cpu_load
-  * for the boot processor.
-  */
+/*
+ * Per-cpu symbols which need to be offset from __per_cpu_load
+ * for the boot processor.
+ */
 #define INIT_PER_CPU(x) init_per_cpu__##x = per_cpu__##x + __per_cpu_load
 INIT_PER_CPU(gdt_page);
 INIT_PER_CPU(irq_stack_union);
#50 lore raw

Re: [PATCH 14/15] x86: convert to use __HEAD and HEAD_TEXT macros.

In reply to Sam Ravnborg
Quoted 10 lines · Sam Ravnborg
Quoted 4 lines

As mentioned, though, the much more interesting case would be the _real_ kernel vmlinux.lds.S file, which is a lot more complex and where the differences between 32-bit and 64-bit cases aren't totally trivial.

Agreed.

But I expect someone else to do so.
When I am in unification mode I try to spend time on architectures with less people involved.

thanks for reconsidering - and i've applied your patch - thanks Sam.

Architectures with lots of people involved (hm, what a strange way to say 'x86' ;-) tend to be logistically more complex and there's an (understandably) stronger pushback against regressions - but that is also where unification is actually needed the most and where the ongoing costs of splintered code is the highest. Your kbuild patches are making quite a bit of difference so please dont hold back :)

Ingo

#51 lore raw

Re: [PATCH 14/15] x86: convert to use __HEAD and HEAD_TEXT macros.

Quoted 14 lines · Ingo Molnar
Quoted 10 lines · Sam Ravnborg
Quoted 4 lines

As mentioned, though, the much more interesting case would be the _real_ kernel vmlinux.lds.S file, which is a lot more complex and where the differences between 32-bit and 64-bit cases aren't totally trivial.

Agreed.

But I expect someone else to do so.
When I am in unification mode I try to spend time on architectures with less people involved.

Quoted 7 lines

thanks for reconsidering - and i've applied your patch - thanks Sam.

Architectures with lots of people involved (hm, what a strange way to say 'x86' ;-) tend to be logistically more complex and there's an (understandably) stronger pushback against regressions - but that is also where unification is actually needed the most and where the ongoing costs of splintered code is the highest.

Well - the effort done for sparc and m68k lately was also needed.

Quoted 2 lines

Your kbuild patches
are making quite a bit of difference so please dont hold back :)

The good patches are the ones that benefits all architectures.

Well - I can give the vmlinux_*.ds files a try.
Did you apply the patch to an append only branch or may I recreate the patch?
I noticed a few minor things that can be 'fixed' to bring the inital differences between 32 bit and 64 bit down.

But I can do these in a follow-up patch too.

Sam

#52 lore raw

Re: [PATCH 14/15] x86: convert to use __HEAD and HEAD_TEXT macros.

Quoted 24 lines · Sam Ravnborg
Quoted 14 lines · Ingo Molnar
Quoted 10 lines · Sam Ravnborg
Quoted 4 lines

As mentioned, though, the much more interesting case would be the _real_ kernel vmlinux.lds.S file, which is a lot more complex and where the differences between 32-bit and 64-bit cases aren't totally trivial.

Agreed.

But I expect someone else to do so.
When I am in unification mode I try to spend time on architectures with less people involved.

Quoted 8 lines

thanks for reconsidering - and i've applied your patch - thanks Sam.

Architectures with lots of people involved (hm, what a strange way to say 'x86' ;-) tend to be logistically more complex and there's an (understandably) stronger pushback against regressions - but that is also where unification is actually needed the most and where the ongoing costs of splintered code is the highest.

Quoted 6 lines

Well - the effort done for sparc and m68k lately was also needed.

Quoted 2 lines

Your kbuild patches are making quite a bit of difference so please dont hold back :)

The good patches are the ones that benefits all architectures.

The good patches are the ones that increase the quality of the kernel and benefit people.

In terms of architecture code that means factoring out common code and stemming the spread of architecture-specific local hacks.

For details that are inherently per arch - obviously improving the more frequently used architectures helps more people, on average.

Quoted 7 lines

Well - I can give the vmlinux_*.ds files a try.
Did you apply the patch to an append only branch or may I recreate the patch?
I noticed a few minor things that can be 'fixed' to bring the inital differences between 32 bit and 64 bit down.

But I can do these in a follow-up patch too.

Yeah, please do a follow-up patch - unless the first patch is outright wrong. I've got a fair amount of test-time in the first patch already - and Linus's is on top of it so it would mean a rebase of both.

Ingo

#53 lore raw

Re: [PATCH 14/15] x86: convert to use __HEAD and HEAD_TEXT macros.

Quoted 5 lines
Quoted 2 lines

The good patches are the ones that benefits all architectures.

The good patches are the ones that increase the quality of the kernel and benefit people.

Should have read "the good kbuild patches" - it did not address patches in general.

Quoted 7 lines
Quoted 4 lines

I noticed a few minor things that can be 'fixed' to bring the inital differences between 32 bit and 64 bit down.

But I can do these in a follow-up patch too.

Yeah, please do a follow-up patch - unless the first patch is outright wrong.

A follow-up patch it will be.

Sam

#54 lore raw

Re: [PATCH 14/15] x86: convert to use __HEAD and HEAD_TEXT macros.

In reply to Linus Torvalds
Quoted 2 lines

(x86-64 has ".eh_frame" in the DISCARD section, i386 apparently doesn't ever generate them,

Modern i386 compilers should generate them too. It was merely a little optimization to keep the vmlinux a little smaller / fast to write in the !DEBUG_INFO case. Would make sense on i386 too.

-Andi

Signature · 1 line
[email protected] -- Speaking for myself only.
#55 lore raw

Re: [PATCH 14/15] x86: convert to use __HEAD and HEAD_TEXT macros.

In reply to Linus Torvalds
Quoted 24 lines · Linus Torvalds
Quoted 22 lines · Sam Ravnborg
Quoted 12 lines · Linus Torvalds
Quoted 9 lines · Linus Torvalds

Btw, this one really needs to unify the two lds files first. Look at

diff -u arch/x86/boot/compressed/vmlinux_*.lds

output and realize that they're basially exctly the same except for trivial naming differences, and the fact that the64-bit version hs a "pgtable" thing.

Quoted 5 lines
Quoted 3 lines

So this really needs to be done by first unifying the thing so that there is _one_ arch/x86/boot/compressed/vmlinux.lds.S file with a preprocessor that takes care of the trivial differences [..]

Something like this?

Looks good/correct.

Quoted 7 lines
Quoted 1 line

You should add your s-o-b if you expect Ingo to pick it up.

Sure. I don't tend to add SOB lines for stuff that I'd not be ready to commit, but with some testing and other people looking at it, I think it's good to go.

Thanks, applied to tip:x86/kbuild. I'll do some more testing of it before pushing it out.

Quoted 9 lines

As mentioned, though, the much more interesting case would be the _real_ kernel vmlinux.lds.S file, which is a lot more complex and where the differences between 32-bit and 64-bit cases aren't totally trivial.

Looking at

diff -u arch/x86/kernel/vmlinux_*.lds.S | less -S

Quoted 7 lines

output, many of them are just whitespace, and others are trivial and meaningless (comments in one, not the other, placement of alignment etc, different ordering of sections like "parainstructions"). Yet others seem to be things that we _could_ do in general, but that don't matter on one architecture or other (x86-64 has ".eh_frame" in the DISCARD section, i386 apparently doesn't ever generate them, we could just use the x86-64 version).

We generally do these by separating the unification into at least 2-3 distinct steps - a mechanic, low-risk cleanup first, preparatory changes to bring the two files in sync second, and mechanic unification as the third and final step.

That way any bugs are easily bisectable to a reasonably sized (and reasonably risky) sub-patch. Review also gets much easier.

I've yet to see a non-trivial Makefile unification in arch/x86 that does not regress :-) They concentrate a lot of quirks and implicit dependencies and small but significant tricks. [usually we catch the bugs early on though - but even at an early stage it's good to have a reasonable splitup.]

Ingo

#56 lore raw

[tip:x86/kbuild] x86: boot/compressed/vmlinux.lds.S: fix build of bzImage with 64 bit compiler

In reply to Linus Torvalds
Commit-ID:  83c4832683bc8ebcd1687b3c0bf3ba1ab253dd4f
Gitweb:     http://git.kernel.org/tip/83c4832683bc8ebcd1687b3c0bf3ba1ab253dd4f
Author:     Sam Ravnborg <[email protected]>
AuthorDate: Thu, 30 Apr 2009 12:03:16 +0200
Committer:  Ingo Molnar <[email protected]>
CommitDate: Thu, 30 Apr 2009 12:36:50 +0200

x86: boot/compressed/vmlinux.lds.S: fix build of bzImage with 64 bit compiler

Jesper reported that he saw following build issue:

Quoted 4 lines

ld:arch/x86/boot/compressed/vmlinux.lds:9: syntax error make[2]: *** [arch/x86/boot/compressed/vmlinux] Error 1 make[1]: *** [arch/x86/boot/compressed/vmlinux] Error 2
make: *** [bzImage] Error 2

CPP defines the symbol "i386" to "1". Undefine this to fix it.

[ Impact: build fix with certain tool chains ]

LKML-Reference: <[email protected]>

---
 arch/x86/boot/compressed/vmlinux.lds.S |    2 ++
 1 files changed, 2 insertions(+), 0 deletions(-)
arch/x86/boot/compressed/vmlinux.lds.S +2 -0
index ffcb191..0d26c92 100644
@@ -1,5 +1,7 @@
 OUTPUT_FORMAT(CONFIG_OUTPUT_FORMAT, CONFIG_OUTPUT_FORMAT, CONFIG_OUTPUT_FORMAT)
 
+#undef i386
+
 #ifdef CONFIG_X86_64
 OUTPUT_ARCH(i386:x86-64)
 ENTRY(startup_64)
#57 lore raw

Re: [tip:x86/kbuild] x86: boot/compressed/vmlinux.lds.S: fix build of bzImage with 64 bit compiler

Quoted 11 lines · tip-bot for Sam Ravnborg
Commit-ID:  83c4832683bc8ebcd1687b3c0bf3ba1ab253dd4f
Gitweb:     http://git.kernel.org/tip/83c4832683bc8ebcd1687b3c0bf3ba1ab253dd4f
Author:     Sam Ravnborg <[email protected]>
AuthorDate: Thu, 30 Apr 2009 12:03:16 +0200
Committer:  Ingo Molnar <[email protected]>
CommitDate: Thu, 30 Apr 2009 12:36:50 +0200

x86: boot/compressed/vmlinux.lds.S: fix build of bzImage with 64 bit compiler

Jesper reported that he saw following build issue:

Quoted 8 lines
Quoted 4 lines

ld:arch/x86/boot/compressed/vmlinux.lds:9: syntax error make[2]: *** [arch/x86/boot/compressed/vmlinux] Error 1 make[1]: *** [arch/x86/boot/compressed/vmlinux] Error 2
make: *** [bzImage] Error 2

CPP defines the symbol "i386" to "1". Undefine this to fix it.

I'm wondering if we should build the *.lds.S with -ansi or some other gcc option which disables ALL non-underscore macros.

-hpa

Signature · 2 lines
H. Peter Anvin, Intel Open Source Technology Center
I work for Intel.  I don't speak on their behalf.
#58 lore raw

Re: [tip:x86/kbuild] x86: boot/compressed/vmlinux.lds.S: fix build of bzImage with 64 bit compiler

Quoted 17 lines · H. Peter Anvin
Quoted 16 lines · tip-bot for Sam Ravnborg
Commit-ID:  83c4832683bc8ebcd1687b3c0bf3ba1ab253dd4f
Gitweb:     http://git.kernel.org/tip/83c4832683bc8ebcd1687b3c0bf3ba1ab253dd4f
Author:     Sam Ravnborg <[email protected]>
AuthorDate: Thu, 30 Apr 2009 12:03:16 +0200
Committer:  Ingo Molnar <[email protected]>
CommitDate: Thu, 30 Apr 2009 12:36:50 +0200

x86: boot/compressed/vmlinux.lds.S: fix build of bzImage with 64 bit compiler

Jesper reported that he saw following build issue:

Quoted 4 lines

ld:arch/x86/boot/compressed/vmlinux.lds:9: syntax error make[2]: *** [arch/x86/boot/compressed/vmlinux] Error 1 make[1]: *** [arch/x86/boot/compressed/vmlinux] Error 2
make: *** [bzImage] Error 2

Quoted 6 lines
Quoted 3 lines

CPP defines the symbol "i386" to "1". Undefine this to fix it.

I'm wondering if we should build the *.lds.S with -ansi or some other gcc option which disables ALL non-underscore macros.

I did not know that -ansi had this effect. We are only seldomly hitting this issue and I wonder if it could break something in out included headers. Most likely not.

Sam

#59 lore raw

Re: [tip:x86/kbuild] x86: boot/compressed/vmlinux.lds.S: fix build of bzImage with 64 bit compiler

Quoted 9 lines · Sam Ravnborg
Quoted 3 lines
Quoted 1 line

I'm wondering if we should build the *.lds.S with -ansi or some other gcc option which disables ALL non-underscore macros.

I did not know that -ansi had this effect. We are only seldomly hitting this issue and I wonder if it could break something in out included headers. Most likely not.

The other thing to recognize is probably that lds files aren't assembly, they're something else entirely. Rather than reusing the *.S rules, having specific rules for them probably would make sense.

-hpa

Signature · 2 lines
H. Peter Anvin, Intel Open Source Technology Center
I work for Intel.  I don't speak on their behalf.
#60 lore raw

Re: [tip:x86/kbuild] x86: boot/compressed/vmlinux.lds.S: fix build of bzImage with 64 bit compiler

Quoted 11 lines · H. Peter Anvin
Quoted 9 lines · Sam Ravnborg
Quoted 3 lines
Quoted 1 line

I'm wondering if we should build the *.lds.S with -ansi or some other gcc option which disables ALL non-underscore macros.

I did not know that -ansi had this effect. We are only seldomly hitting this issue and I wonder if it could break something in out included headers. Most likely not.

Quoted 3 lines

The other thing to recognize is probably that lds files aren't assembly, they're something else entirely. Rather than reusing the *.S rules, having specific rules for them probably would make sense.

We already have that:
(scripts/Makfile.build)
# Linker scripts preprocessor (.lds.S -> .lds)
# ---------------------------------------------------------------------------
quiet_cmd_cpp_lds_S = LDS $@

      cmd_cpp_lds_S = $(CPP) $(cpp_flags) -D__ASSEMBLY__ -o $@ $<

$(obj)/%.lds: $(src)/%.lds.S FORCE

        $(call if_changed_dep,cpp_lds_S)

So adding -ansi is trivial and will only hit .lds.S files.

I assume the lds files are named .S is because we shared rules long time ago.

Sam

#61 lore raw

Re: [tip:x86/kbuild] x86: boot/compressed/vmlinux.lds.S: fix build of bzImage with 64 bit compiler

Quoted 6 lines · Sam Ravnborg

So adding -ansi is trivial and will only hit .lds.S files.

I assume the lds files are named .S is because we shared rules long time ago.

Yes, the naming is rather ugly that way.

-hpa

Acked-by on v1 13/15
#62 lore raw

Re: [PATCH 13/15] sparc: convert to use __HEAD and HEAD_TEXT macros.

In reply to Tim Abbott
Quoted 6 lines · Tim Abbott

This has the consequence of changing the section name use for head code from ".text.head" to ".head.text". Since this commit changes all users in the architecture, this change should be harmless.

Cc David S. Miller <[email protected]>
Quoted 34 lines
---
 arch/sparc/kernel/head_32.S     |    2 +-
 arch/sparc/kernel/vmlinux.lds.S |    2 +-
 2 files changed, 2 insertions(+), 2 deletions(-)
arch/sparc/kernel/head_32.S +1 -1
index f0b4b51..e91f01a 100644
@@ -72,7 +72,7 @@ sun4e_notsup:
 	.align 4
 
 	/* The Sparc trap table, bootloader gives us control at _start. */
-	.section .text.head,"ax"
+	__HEAD
 	.globl	start, _stext, _start, __stext
 	.globl  trapbase
 _start:   /* danger danger */
arch/sparc/kernel/vmlinux.lds.S +1 -1
index 7626708..fcbbd00 100644
@@ -41,7 +41,7 @@ SECTIONS
 	.text TEXTSTART :
 	{
 		_text = .;
-		*(.text.head)
+		HEAD_TEXT
 		TEXT_TEXT
 		SCHED_TEXT
 		LOCK_TEXT
Signature · 1 line
1.6.2.1
Acked-by on v1 12/15
#63 lore raw

Re: [PATCH 12/15] sh: convert to use __HEAD and HEAD_TEXT macros.

In reply to Tim Abbott
Quoted 6 lines · Tim Abbott

This has the consequence of changing the section name use for head code from ".text.head" to ".head.text". Since this commit changes all users in the architecture, this change should be harmless.

Quoted 80 lines
---
 arch/sh/kernel/head_32.S        |    3 ++-
 arch/sh/kernel/head_64.S        |    5 ++++-
 arch/sh/kernel/vmlinux_32.lds.S |    2 +-
 arch/sh/kernel/vmlinux_64.lds.S |    2 +-
 4 files changed, 8 insertions(+), 4 deletions(-)
arch/sh/kernel/head_32.S +2 -1
index 788605f..a78be74 100644
@@ -10,6 +10,7 @@
  *
  * Head.S contains the SH exception handlers and startup code.
  */
+#include <linux/init.h>
 #include <linux/linkage.h>
 #include <asm/thread_info.h>
 
@@ -40,7 +41,7 @@ ENTRY(empty_zero_page)
 1:
 	.skip	PAGE_SIZE - empty_zero_page - 1b
 
-	.section	.text.head, "ax"
+	__HEAD
 
 /*
  * Condition at the entry of _stext:
arch/sh/kernel/head_64.S +4 -1
index 7ccfb99..3ea7658 100644
@@ -8,6 +8,9 @@
  * License.  See the file "COPYING" in the main directory of this archive
  * for more details.
  */
+
+#include <linux/init.h>
+
 #include <asm/page.h>
 #include <asm/cache.h>
 #include <asm/tlb.h>
@@ -110,7 +113,7 @@ empty_bad_pte_table:
 fpu_in_use:	.quad	0
 
 
-	.section	.text.head, "ax"
+	__HEAD
 	.balign L1_CACHE_BYTES
 /*
  * Condition at the entry of __stext:
arch/sh/kernel/vmlinux_32.lds.S +1 -1
index d0b2a71..dd9b2ee 100644
@@ -31,7 +31,7 @@ SECTIONS
 	} = 0
 
 	.text : {
-		*(.text.head)
+		HEAD_TEXT
 		TEXT_TEXT
 		SCHED_TEXT
 		LOCK_TEXT
arch/sh/kernel/vmlinux_64.lds.S +1 -1
index 33fa464..6966446 100644
@@ -42,7 +42,7 @@ SECTIONS
 	} = 0
 
 	.text : C_PHYS(.text) {
-		*(.text.head)
+		HEAD_TEXT
 		TEXT_TEXT
 		*(.text64)
 		*(.text..SHmedia32)
Signature · 1 line
1.6.2.1
Acked-by on v1 11/15
#64 lore raw

Re: [PATCH 11/15] s390: convert to use __HEAD and HEAD_TEXT macros.

In reply to Tim Abbott
Quoted 7 lines · Tim Abbott

This has the consequence of changing the section name use for head code from ".text.head" to ".head.text". Since this commit changes all users in the architecture, this change should be harmless.

Cc Martin Schwidefsky <[email protected]>
Cc Heiko Carstens <[email protected]>
Quoted 44 lines
---
 arch/s390/kernel/head.S        |    3 ++-
 arch/s390/kernel/vmlinux.lds.S |    4 ++--
 2 files changed, 4 insertions(+), 3 deletions(-)
arch/s390/kernel/head.S +2 -1
index bba1449..22596d7 100644
@@ -23,6 +23,7 @@
  *
  */
 
+#include <linux/init.h>
 #include <asm/setup.h>
 #include <asm/lowcore.h>
 #include <asm/asm-offsets.h>
@@ -35,7 +36,7 @@
 #define ARCH_OFFSET	0
 #endif
 
-.section ".text.head","ax"
+__HEAD
 #ifndef CONFIG_IPL
 	.org   0
 	.long  0x00080000,0x80000000+startup	# Just a restart PSW
arch/s390/kernel/vmlinux.lds.S +2 -2
index 7a2063e..89399b8 100644
@@ -29,8 +29,8 @@ SECTIONS
 	. = 0x00000000;
 	.text : {
 	_text = .;		/* Text and read-only data */
-		*(.text.head)
-	TEXT_TEXT
+		HEAD_TEXT
+		TEXT_TEXT
 		SCHED_TEXT
 		LOCK_TEXT
 		KPROBES_TEXT
Signature · 1 line
1.6.2.1
Acked-by on v1 10/15
#65 lore raw

Re: [PATCH 10/15] powerpc: convert to use __HEAD and HEAD_TEXT macros.

In reply to Tim Abbott
Quoted 6 lines · Tim Abbott

This has the consequence of changing the section name use for head code from ".text.head" to ".head.text". Since this commit changes all users in the architecture, this change should be harmless.

Cc Benjamin Herrenschmidt <[email protected]>
Quoted 130 lines
---
 arch/powerpc/kernel/head_32.S        |    3 ++-
 arch/powerpc/kernel/head_40x.S       |    3 ++-
 arch/powerpc/kernel/head_44x.S       |    3 ++-
 arch/powerpc/kernel/head_8xx.S       |    3 ++-
 arch/powerpc/kernel/head_fsl_booke.S |    3 ++-
 arch/powerpc/kernel/vmlinux.lds.S    |    2 +-
 6 files changed, 11 insertions(+), 6 deletions(-)
arch/powerpc/kernel/head_32.S +2 -1
index 54e68c1..c01467f 100644
@@ -21,6 +21,7 @@
  *
  */
 
+#include <linux/init.h>
 #include <asm/reg.h>
 #include <asm/page.h>
 #include <asm/mmu.h>
@@ -50,7 +51,7 @@
 	mtspr	SPRN_DBAT##n##L,RB;	\
 1:
 
-	.section	.text.head, "ax"
+	__HEAD
 	.stabs	"arch/powerpc/kernel/",N_SO,0,0,0f
 	.stabs	"head_32.S",N_SO,0,0,0f
 0:
arch/powerpc/kernel/head_40x.S +2 -1
index 56d8e5d..0c96911 100644
@@ -31,6 +31,7 @@
  *
  */
 
+#include <linux/init.h>
 #include <asm/processor.h>
 #include <asm/page.h>
 #include <asm/mmu.h>
@@ -52,7 +53,7 @@
  *
  * This is all going to change RSN when we add bi_recs.......  -- Dan
  */
-	.section	.text.head, "ax"
+	__HEAD
 _ENTRY(_stext);
 _ENTRY(_start);
 
arch/powerpc/kernel/head_44x.S +2 -1
index b56fecc..18d8a16 100644
@@ -28,6 +28,7 @@
  * option) any later version.
  */
 
+#include <linux/init.h>
 #include <asm/processor.h>
 #include <asm/page.h>
 #include <asm/mmu.h>
@@ -50,7 +51,7 @@
  *   r7 - End of kernel command line string
  *
  */
-	.section	.text.head, "ax"
+	__HEAD
 _ENTRY(_stext);
 _ENTRY(_start);
 	/*
arch/powerpc/kernel/head_8xx.S +2 -1
index 3c9452d..52ff8c5 100644
@@ -19,6 +19,7 @@
  *
  */
 
+#include <linux/init.h>
 #include <asm/processor.h>
 #include <asm/page.h>
 #include <asm/mmu.h>
@@ -38,7 +39,7 @@
 #else
 #define DO_8xx_CPU6(val, reg)
 #endif
-	.section	.text.head, "ax"
+	__HEAD
 _ENTRY(_stext);
 _ENTRY(_start);
 
arch/powerpc/kernel/head_fsl_booke.S +2 -1
index 4c22620..5bdcc06 100644
@@ -30,6 +30,7 @@
  * option) any later version.
  */
 
+#include <linux/init.h>
 #include <linux/threads.h>
 #include <asm/processor.h>
 #include <asm/page.h>
@@ -53,7 +54,7 @@
  *   r7 - End of kernel command line string
  *
  */
-	.section	.text.head, "ax"
+	__HEAD
 _ENTRY(_stext);
 _ENTRY(_start);
 	/*
arch/powerpc/kernel/vmlinux.lds.S +1 -1
index b9ef164..d7425ff 100644
@@ -52,7 +52,7 @@ SECTIONS
 	/* Text and gots */
 	.text : AT(ADDR(.text) - LOAD_OFFSET) {
 		ALIGN_FUNCTION();
-		*(.text.head)
+		HEAD_TEXT
 		_text = .;
 		*(.text .fixup .text.init.refok .exit.text.refok __ftr_alt_*)
 		SCHED_TEXT
Signature · 1 line
1.6.2.1
Acked-by on v1 9/15
#66 lore raw

Re: [PATCH 09/15] mn10300: convert to use __HEAD and HEAD_TEXT macros.

In reply to Tim Abbott
Quoted 6 lines · Tim Abbott

This has the consequence of changing the section name use for head code from ".text.head" to ".head.text". Since this commit changes all users in the architecture, this change should be harmless.

Cc David Howells <[email protected]>
Quoted 45 lines
---
 arch/mn10300/kernel/head.S        |    3 ++-
 arch/mn10300/kernel/vmlinux.lds.S |    5 +----
 2 files changed, 3 insertions(+), 5 deletions(-)
arch/mn10300/kernel/head.S +2 -1
index 8a8309f..14f27f3 100644
@@ -9,6 +9,7 @@
  * 2 of the Licence, or (at your option) any later version.
  */
 
+#include <linux/init.h>
 #include <linux/threads.h>
 #include <linux/linkage.h>
 #include <linux/serial_reg.h>
@@ -19,7 +20,7 @@
 #include <asm/param.h>
 #include <unit/serial.h>
 
-	.section .text.head,"ax"
+	__HEAD
 
 ###############################################################################
 #
arch/mn10300/kernel/vmlinux.lds.S +1 -4
index b825966..24de6b9 100644
@@ -27,10 +27,7 @@ SECTIONS
   _stext = .;
   _text = .;			/* Text and read-only data */
   .text : {
-	*(
-	.text.head
-	.text
-	)
+	HEAD_TEXT
 	TEXT_TEXT
 	SCHED_TEXT
 	LOCK_TEXT
Signature · 1 line
1.6.2.1
Acked-by on v1 8/15
#67 lore raw

Re: [PATCH 08/15] m68k: convert to use __HEAD and HEAD_TEXT macros.

In reply to Tim Abbott
Quoted 7 lines · Tim Abbott

This has the consequence of changing the section name use for head code from ".text.head" to ".head.text". Since this commit changes all users in the architecture, this change should be harmless.

Cc Geert Uytterhoeven <[email protected]>
Cc Roman Zippel <[email protected]>
Quoted 62 lines
---
 arch/m68k/kernel/head.S           |    2 +-
 arch/m68k/kernel/sun3-head.S      |    2 +-
 arch/m68k/kernel/vmlinux-std.lds  |    2 +-
 arch/m68k/kernel/vmlinux-sun3.lds |    2 +-
 4 files changed, 4 insertions(+), 4 deletions(-)
arch/m68k/kernel/head.S +1 -1
index f513f53..86edb5f 100644
@@ -577,7 +577,7 @@ func_define	putn,1
 #endif
 .endm
 
-.section ".text.head","ax"
+__HEAD
 ENTRY(_stext)
 /*
  * Version numbers of the bootinfo interface
arch/m68k/kernel/sun3-head.S +1 -1
index aad0159..9e5c3e5 100644
@@ -29,7 +29,7 @@ kernel_pmd_table:              .skip 0x2000
 .globl kernel_pg_dir
 .equ    kernel_pg_dir,kernel_pmd_table
 
-	.section .text.head
+	__HEAD
 ENTRY(_stext)
 ENTRY(_start)
 
arch/m68k/kernel/vmlinux-std.lds +1 -1
index f846d4e..01d212b 100644
@@ -12,7 +12,7 @@ SECTIONS
   . = 0x1000;
   _text = .;			/* Text and read-only data */
   .text : {
-	*(.text.head)
+	HEAD_TEXT
 	TEXT_TEXT
 	SCHED_TEXT
 	LOCK_TEXT
arch/m68k/kernel/vmlinux-sun3.lds +1 -1
index d9368c0..c192f77 100644
@@ -12,7 +12,7 @@ SECTIONS
   . = 0xE002000;
   _text = .;			/* Text and read-only data */
   .text : {
-	*(.text.head)
+	HEAD_TEXT
 	TEXT_TEXT
 	SCHED_TEXT
 	LOCK_TEXT
Signature · 1 line
1.6.2.1
Acked-by on v1 7/15
#68 lore raw

Re: [PATCH 07/15] m32r: convert to use __HEAD and HEAD_TEXT macros.

In reply to Tim Abbott
Quoted 6 lines · Tim Abbott

This has the consequence of changing the section name use for head code from ".text.head" to ".head.text". Since this commit changes all users in the architecture, this change should be harmless.

Cc Hirokazu Takata <[email protected]>
Quoted 34 lines
---
 arch/m32r/kernel/head.S        |    2 +-
 arch/m32r/kernel/vmlinux.lds.S |    2 +-
 2 files changed, 2 insertions(+), 2 deletions(-)
arch/m32r/kernel/head.S +1 -1
index 9091606..0a71944 100644
@@ -23,7 +23,7 @@ __INITDATA
 /*
  * References to members of the boot_cpu_data structure.
  */
-.section .text.head, "ax"
+__HEAD
 	.global	start_kernel
 	.global __bss_start
 	.global _end
arch/m32r/kernel/vmlinux.lds.S +1 -1
index 9db05df..4179adf 100644
@@ -27,7 +27,7 @@ SECTIONS
   _text = .;			/* Text and read-only data */
   .boot : { *(.boot) } = 0
   .text : {
-	*(.text.head)
+	HEAD_TEXT
 	TEXT_TEXT
 	SCHED_TEXT
 	LOCK_TEXT
Signature · 1 line
1.6.2.1
#69 lore raw

Re: [PATCH 07/15] m32r: convert to use __HEAD and HEAD_TEXT macros.

From: Sam Ravnborg <[email protected]>
Subject: Re: [PATCH 07/15] m32r: convert to use __HEAD and HEAD_TEXT macros.
Date: Sun, 26 Apr 2009 13:21:45 +0200

Quoted 8 lines · Tim Abbott
Quoted 6 lines

This has the consequence of changing the section name use for head code from ".text.head" to ".head.text". Since this commit changes all users in the architecture, this change should be harmless.

Cc Hirokazu Takata <[email protected]>
Quoted 35 lines
Quoted 34 lines
---
 arch/m32r/kernel/head.S        |    2 +-
 arch/m32r/kernel/vmlinux.lds.S |    2 +-
 2 files changed, 2 insertions(+), 2 deletions(-)
arch/m32r/kernel/head.S +1 -1
index 9091606..0a71944 100644
@@ -23,7 +23,7 @@ __INITDATA
 /*
  * References to members of the boot_cpu_data structure.
  */
-.section .text.head, "ax"
+__HEAD
 	.global	start_kernel
 	.global __bss_start
 	.global _end
arch/m32r/kernel/vmlinux.lds.S +1 -1
index 9db05df..4179adf 100644
@@ -27,7 +27,7 @@ SECTIONS
   _text = .;			/* Text and read-only data */
   .boot : { *(.boot) } = 0
   .text : {
-	*(.text.head)
+	HEAD_TEXT
 	TEXT_TEXT
 	SCHED_TEXT
 	LOCK_TEXT
Signature · 1 line
1.6.2.1
#70 lore raw

Re: [PATCH 06/15] ia64: convert to use __HEAD and HEAD_TEXT macros.

In reply to Tim Abbott
Quoted 29 lines · Tim Abbott

This has the consequence of changing the section name use for head code from ".text.head" to ".head.text". Since this commit changes all users in the architecture, this change should be harmless.

---
 arch/ia64/kernel/head.S        |    2 +-
 arch/ia64/kernel/vmlinux.lds.S |    4 ++--
 2 files changed, 3 insertions(+), 3 deletions(-)
arch/ia64/kernel/head.S +1 -1
index 23f846d..a572160 100644
@@ -181,7 +181,7 @@ swapper_pg_dir:
 halt_msg:
 	stringz "Halting kernel\n"
 
-	.section .text.head,"ax"
+	__HEAD
 
 	.global start_ap
 
arch/ia64/kernel/vmlinux.lds.S +0 -0
index 4a95e86..a8b879e 100644

arch/ia64/kernel/vmlinux.lds.S SECTIONS · line 55

Quoted 7 lines
arch/ia64/kernel/vmlinux.lds.S +2 -2
@@ -51,8 +51,8 @@ SECTIONS
 	KPROBES_TEXT
 	*(.gnu.linkonce.t*)
     }
-  .text.head : AT(ADDR(.text.head) - LOAD_OFFSET)
-	{ *(.text.head) }
+  HEAD_TEXT_SECTION : AT(ADDR(HEAD_TEXT_SECTION) - LOAD_OFFSET)
+	{ HEAD_TEXT }

1) Spell it out directly rahter than using a macro
2) Consider fixing the style when you are here anyway. The 'correct' style is to use C rules for '{}' and indent. But I guess you followed style of the file - which is ok.

Sam

#71 lore raw

Re: [PATCH 05/15] arm: convert to use __HEAD and HEAD_TEXT macros.

In reply to Tim Abbott
Quoted 43 lines · Tim Abbott

This has the consequence of changing the section name use for head code from ".text.head" to ".head.text". Since this commit changes all users in the architecture, this change should be harmless.

---
 arch/arm/kernel/head-nommu.S  |    2 +-
 arch/arm/kernel/head.S        |    2 +-
 arch/arm/kernel/vmlinux.lds.S |    4 ++--
 3 files changed, 4 insertions(+), 4 deletions(-)
arch/arm/kernel/head-nommu.S +1 -1
index cc87e17..4f10d1c 100644
@@ -32,7 +32,7 @@
  * numbers for r1.
  *
  */
-	.section ".text.head", "ax"
+	__HEAD
 ENTRY(stext)
 	msr	cpsr_c, #PSR_F_BIT | PSR_I_BIT | SVC_MODE @ ensure svc mode
 						@ and irqs disabled
arch/arm/kernel/head.S +1 -1
index 21e17dc..f1877a5 100644
@@ -74,7 +74,7 @@
  * crap here - that's what the boot loader (or in extreme, well justified
  * circumstances, zImage) is for.
  */
-	.section ".text.head", "ax"
+	__HEAD
 ENTRY(stext)
 	msr	cpsr_c, #PSR_F_BIT | PSR_I_BIT | SVC_MODE @ ensure svc mode
 						@ and irqs disabled
arch/arm/kernel/vmlinux.lds.S +0 -0
index c90f272..655fb75 100644

arch/arm/kernel/vmlinux.lds.S SECTIONS · line 26

Quoted 5 lines
arch/arm/kernel/vmlinux.lds.S +1 -1
@@ -23,10 +23,10 @@ SECTIONS
 #else
 	. = PAGE_OFFSET + TEXT_OFFSET;
 #endif
-	.text.head : {
+	HEAD_TEXT_SECTION : {

Spell it out direct rather than hiding the .head.text behind a macro name here.

Sam

author
#72 lore raw

Re: [PATCH 05/15] arm: convert to use __HEAD and HEAD_TEXT macros.

Quoted 5 lines · Sam Ravnborg

arch/arm/kernel/vmlinux.lds.S SECTIONS · line 26

Quoted 2 lines
arch/arm/kernel/vmlinux.lds.S +1 -1
-	.text.head : {
+	HEAD_TEXT_SECTION : {

Spell it out direct rather than hiding the .head.text behind a macro name here.

Does .head.text actually need its own output section here? I notice that only the linker scripts for arm, x86, and ia64 have an output section for .head.text; the rest just place it at the start of the .text or .init.text output section. It would be great if we could eliminate these extra references to .head.text by doing the same thing here.

-Tim Abbott

#73 lore raw

Re: [PATCH 05/15] arm: convert to use __HEAD and HEAD_TEXT macros.

Quoted 8 lines · Tim Abbott
Quoted 5 lines · Sam Ravnborg

arch/arm/kernel/vmlinux.lds.S SECTIONS · line 26

Quoted 2 lines
arch/arm/kernel/vmlinux.lds.S +1 -1
-	.text.head : {
+	HEAD_TEXT_SECTION : {

Spell it out direct rather than hiding the .head.text behind a macro name here.

Quoted 5 lines

Does .head.text actually need its own output section here? I notice that only the linker scripts for arm, x86, and ia64 have an output section for .head.text; the rest just place it at the start of the .text or .init.text output section. It would be great if we could eliminate these extra references to .head.text by doing the same thing here.

It could be merged with the .init section. IOW, it becomes:

        .init : {                       /* Init code and data           */
                _stext = .;
                _sinittext = .;
                *(.text.head)
                        INIT_TEXT
                _einittext = .;
                __proc_info_begin = .;
                        *(.proc.info.init)
                __proc_info_end = .;

...

Signature · 3 lines
Russell King
 Linux kernel    2.6 ARM Linux   - http://www.arm.linux.org.uk/
 maintainer of:
#74 lore raw

Re: [PATCH 05/15] arm: convert to use __HEAD and HEAD_TEXT macros.

In reply to Tim Abbott
Quoted 2 lines · Tim Abbott

Does .head.text actually need its own output section here?

Probably not. Does the ARM build system massage the end result some way looking for that .text.head section in the end? As you note, nobody else cares, and the proper patch may be more along the lines of something below.

Of course, most other architectures put the "init" section at the end, and leave the HEAD_TEXT in the regular text section, but I suspect the ARM way of laying things out (init text first) is probably better, since it allows you to free that head code too.

Not that it likely much matters.

Linus

---
 arch/arm/kernel/vmlinux.lds.S |    8 +++-----
 1 files changed, 3 insertions(+), 5 deletions(-)
arch/arm/kernel/vmlinux.lds.S +3 -5
index c90f272..15520be 100644
@@ -23,13 +23,11 @@ SECTIONS
 #else
 	. = PAGE_OFFSET + TEXT_OFFSET;
 #endif
-	.text.head : {
-		_stext = .;
-		_sinittext = .;
-		*(.text.head)
-	}
 
 	.init : {			/* Init code and data		*/
+		_stext = .;
+		_sinittext = .;
+			HEAD_TEXT
 			INIT_TEXT
 		_einittext = .;
 		__proc_info_begin = .;
#75 lore raw

RE: [PATCH 05/15] arm: convert to use __HEAD and HEAD_TEXT macros.

Quoted 6 lines
Quoted 1 line

Does .head.text actually need its own output section here?

Probably not. Does the ARM build system massage the end result some way looking for that .text.head section in the end? As you note, nobody else cares, and the proper patch may be more along the lines of something below.

ia64 doesn't seem to care ... I changed the one use of .text.head to be just .text. There are a couple of new "section mismatches" (start_ap -> start_secondary and start_ap -> start_kernel), but it still boots.

Quoted 4 lines

Of course, most other architectures put the "init" section at the end, and leave the HEAD_TEXT in the regular text section, but I suspect the ARM way of laying things out (init text first) is probably better, since it allows you to free that head code too.

The start_ap code is still needed for hotplug cpu on ia64. So .text.init won't work for me.

-Tony

#76 lore raw

Re: [PATCH 05/15] arm: convert to use __HEAD and HEAD_TEXT macros.

In reply to Linus Torvalds
Quoted 9 lines · Linus Torvalds
Quoted 2 lines · Tim Abbott

Does .head.text actually need its own output section here?

Probably not. Does the ARM build system massage the end result some way looking for that .text.head section in the end? As you note, nobody else cares, and the proper patch may be more along the lines of something below.

Quoted 4 lines

Of course, most other architectures put the "init" section at the end, and leave the HEAD_TEXT in the regular text section, but I suspect the ARM way of laying things out (init text first) is probably better, since it allows you to free that head code too.

There are a few platforms were due to excess bootloader primitivity (read the boot format is a simple blob) it is desireable to have the entry point at the lowest address of the final image - because that is what the firmware is expecting.

Ralf

Acked-by on v1 4/15
#77 lore raw

Re: [PATCH 04/15] frv: convert frv to use __HEAD and HEAD_TEXT macros.

In reply to Tim Abbott
Quoted 6 lines · Tim Abbott

This has the consequence of changing the section name use for head code from ".text.head" to ".head.text". Since this commit changes all users in the architecture, this change should be harmless.

Cc David Howells <[email protected]>
Quoted 42 lines
---
 arch/frv/kernel/head.S        |    3 ++-
 arch/frv/kernel/vmlinux.lds.S |    2 +-
 2 files changed, 3 insertions(+), 2 deletions(-)
arch/frv/kernel/head.S +2 -1
index fecf751..b825ef3 100644
@@ -9,6 +9,7 @@
  * 2 of the License, or (at your option) any later version.
  */
 
+#include <linux/init.h>
 #include <linux/threads.h>
 #include <linux/linkage.h>
 #include <asm/thread_info.h>
@@ -27,7 +28,7 @@
 #   command line string
 #
 ###############################################################################
-	.section	.text.head,"ax"
+	__HEAD
 	.balign		4
 
 	.globl		_boot, __head_reference
arch/frv/kernel/vmlinux.lds.S +1 -1
index b95c4ea..22d9787 100644
@@ -26,7 +26,7 @@ SECTIONS
 
   _sinittext = .;
   .init.text : {
-	*(.text.head)
+	HEAD_TEXT
 #ifndef CONFIG_DEBUG_INFO
 	INIT_TEXT
 	EXIT_TEXT
Signature · 1 line
1.6.2.1
Acked-by on v1 3/15
#78 lore raw

Re: [PATCH 03/15] alpha: convert to use __HEAD and HEAD_TEXT macros.

In reply to Tim Abbott
Quoted 6 lines · Tim Abbott

This has the consequence of changing the section name use for head code from ".text.head" to ".head.text". Since this commit changes all users in the architecture, this change should be harmless.

Cc Richard Henderson <[email protected]>
Quoted 38 lines
---
 arch/alpha/kernel/head.S        |    3 ++-
 arch/alpha/kernel/vmlinux.lds.S |    2 +-
 2 files changed, 3 insertions(+), 2 deletions(-)
arch/alpha/kernel/head.S +2 -1
index 7ac1f13..4bdd1d2 100644
@@ -7,10 +7,11 @@
  * the kernel global pointer and jump to the kernel entry-point.
  */
 
+#include <linux/init.h>
 #include <asm/system.h>
 #include <asm/asm-offsets.h>
 
-.section .text.head, "ax"
+__HEAD
 .globl swapper_pg_dir
 .globl _stext
 swapper_pg_dir=SWAPPER_PGD
arch/alpha/kernel/vmlinux.lds.S +1 -1
index ef37fc1..b9d6568 100644
@@ -16,7 +16,7 @@ SECTIONS
 
 	_text = .;	/* Text and read-only data */
 	.text : {
-	*(.text.head)
+		HEAD_TEXT
 		TEXT_TEXT
 		SCHED_TEXT
 		LOCK_TEXT
Signature · 1 line
1.6.2.1
Acked-by on v1 2/15
#79 lore raw

Re: [PATCH 02/15] xtensa: convert to use __HEAD and HEAD_TEXT macros.

In reply to Tim Abbott
Quoted 2 lines · Tim Abbott
Quoted 44 lines
---
 arch/xtensa/kernel/head.S        |    3 ++-
 arch/xtensa/kernel/vmlinux.lds.S |    4 ++--
 2 files changed, 4 insertions(+), 3 deletions(-)
arch/xtensa/kernel/head.S +2 -1
index 0817f9d..d9ddc1b 100644
@@ -19,6 +19,7 @@
 #include <asm/page.h>
 #include <asm/cacheasm.h>
 
+#include <linux/init.h>
 #include <linux/linkage.h>
 
 /*
@@ -45,7 +46,7 @@
 	 * instruction.
 	 */
 
-	.section .head.text, "ax"
+	__HEAD
 	.globl _start
 _start:	_j	2f
 	.align	4
arch/xtensa/kernel/vmlinux.lds.S +2 -2
index c1be9a4..5accf51 100644
@@ -85,8 +85,8 @@ SECTIONS
 
   .text :
   {
-    /* The .head.text section must be the first section! */
-    *(.head.text)
+    /* The HEAD_TEXT section must be the first section! */
+    HEAD_TEXT
     *(.literal .text)
     VMLINUX_SYMBOL(__sched_text_start) = .;
     *(.sched.literal .sched.text)
Signature · 1 line
1.6.2.1
#80 lore raw

Re: [PATCH 01/15] Add new HEAD_TEXT_SECTION macro.

In reply to Tim Abbott
Quoted 4 lines · Tim Abbott

This patch is preparation for replacing all uses of ".head.text" or ".text.head" in the kernel with macros, so that the section name can later be changed without having to touch a lot of the kernel.

I took the "trivially obvious" parts of this patch series - the ones that were basically trivial one-liner cleanups (in addition to the first patch that was needed for them, of course).

Anything that renamed files and did other infrastructure things (x86) or that had vmlinux.lds.S files that didn't match the nice pattern I'd much rather take as series of patches that first clean up the infrastructure and then do the replacement.

Linus

#81 lore raw

Re: [PATCH v3 1/3] Make section names compatible with -ffunction-sections -fdata-sections

In reply to Tim Abbott
Quoted 12 lines · Tim Abbott
Quoted 4 lines · Sam Ravnborg

This patch touches far too many files.
We should try to work out a method so we are in better control of the section names, so renaming in the end is a simple patch touching only a few files.

OK, I'm now planning to implement this approach.

Quoted 3 lines
Quoted 2 lines

-.section .text.head, "ax"
+.section .text..head, "ax"

Quoted 8 lines
Quoted 2 lines

Use __HEAD (from include/linux/init.h)
Same goes for all other uses of .text.head.

I notice that __HEAD uses .head.text, while some architectures use .text.head. It looks like this is just an inconsistency across architectures that will be removed as a consequence of this cleanup work (no architecture uses both .head.text and .text.head).

The use of .text.head is quite intentional. See f8657e1b55901e6c227094258d1fa3642fa242bd for starters. Most of the rest of the platforms that switched did so for section mismatch avoidance reasons, those that kept .head.text presumably did not hit these mismatches or simply didn't care.

#82 lore raw

Re: [PATCH v3 1/3] Make section names compatible with -ffunction-sections -fdata-sections

Quoted 21 lines · Paul Mundt
Quoted 12 lines · Tim Abbott
Quoted 4 lines · Sam Ravnborg

This patch touches far too many files.
We should try to work out a method so we are in better control of the section names, so renaming in the end is a simple patch touching only a few files.

OK, I'm now planning to implement this approach.

Quoted 3 lines
Quoted 2 lines

-.section .text.head, "ax"
+.section .text..head, "ax"

Quoted 8 lines
Quoted 2 lines

Use __HEAD (from include/linux/init.h)
Same goes for all other uses of .text.head.

I notice that __HEAD uses .head.text, while some architectures use .text.head. It looks like this is just an inconsistency across architectures that will be removed as a consequence of this cleanup work (no architecture uses both .head.text and .text.head).

Quoted 5 lines

The use of .text.head is quite intentional. See f8657e1b55901e6c227094258d1fa3642fa242bd for starters. Most of the rest of the platforms that switched did so for section mismatch avoidance reasons, those that kept .head.text presumably did not hit these mismatches or simply didn't care.

This is a mess that Tim's patches should clean up.
i386 should use .head.text - this is the 'correct' name for the head section.

And modpost does not accept ".text.head" anymore - thats cleaned up some time ago.

Sam

#83 lore raw

PATCH 1/1 kbuild: fix comment in modpost.c

In reply to Paul Mundt

Looking at this I noticed a bug in a comment in modpost.

Sam

Quoted 1 line

From bece3302fa388ee643f8a6bc78463ac20a2f23bb Mon Sep 17 00:00:00 2001

From: Sam Ravnborg <[email protected]>
Date: Sun, 26 Apr 2009 11:17:42 +0200
Subject: [PATCH] kbuild: fix comment in modpost.c

There is some confusion on naming of the head section. Correct naming is .head.text.

Fix comment so we use correct naming.

---
 scripts/mod/modpost.c |    3 +--
 1 files changed, 1 insertions(+), 2 deletions(-)
scripts/mod/modpost.c +1 -2
index c9a01ca..2fdf2bf 100644
@@ -942,8 +942,7 @@ static int section_mismatch(const char *fromsec, const char *tosec)
  *           *probe_one, *_console, *_timer
  *
  * Pattern 3:
- *   Whitelist all refereces from .text.head to .init.data
- *   Whitelist all refereces from .text.head to .init.text
+ *   Whitelist all refereces from .head.text to any init section
  *
  * Pattern 4:
  *   Some symbols belong to init section but still it is ok to reference
Signature · 1 line
1.6.0.2.GIT
author
#84 lore raw

Re: [PATCH] kbuild: fix comment in modpost.c

Quoted 3 lines · Sam Ravnborg
- *   Whitelist all refereces from .text.head to .init.data
- *   Whitelist all refereces from .text.head to .init.text
+ *   Whitelist all refereces from .head.text to any init section

While you're fixing that you might as well fix the spelling of "references".

-Tim Abbott

#85 lore raw

Re: [PATCH v3 1/3] Make section names compatible with -ffunction-sections -fdata-sections

In reply to Tim Abbott
Quoted 12 lines · Tim Abbott
Quoted 4 lines · Sam Ravnborg

This patch touches far too many files.
We should try to work out a method so we are in better control of the section names, so renaming in the end is a simple patch touching only a few files.

OK, I'm now planning to implement this approach.

Quoted 3 lines
Quoted 2 lines

-.section .text.head, "ax"
+.section .text..head, "ax"

Quoted 7 lines
Quoted 2 lines

Use __HEAD (from include/linux/init.h)
Same goes for all other uses of .text.head.

I notice that __HEAD uses .head.text, while some architectures use .text.head. It looks like this is just an inconsistency across architectures that will be removed as a consequence of this cleanup work (no architecture uses both .head.text and .text.head).

Correct - this is implied by the introduction of __HEAD. For users of the old naming schme (like i386) you need to adjust the linker script too.

Quoted 5 lines

One challenge with this approach is that many linker scripts use these section names in more complex ways than just squashing HEAD_TEXT at the start of the text section. For example, the the linker scripts for x86 and ia64 have code like:

Quoted 7 lines
  .text.head : AT(ADDR(.text.head) - LOAD_OFFSET) {
        _text = .;                      /* Text and read-only data */
        *(.text.head)
  } :text = 0x9090

which can't user either the __HEAD macro (which is the full .section line) or the HEAD_TEXT macro (which is the *(.head.text)).

The simple way to deal wi8th this is to accept some duplication of naming in order to keep readability.
So I suggest you to use:

  .head.text : AT(ADDR(.head.text) - LOAD_OFFSET) {
        _text = .;                      /* Text and read-only data */
        HEAD_TEXT
  } :text = 0x9090

We need to be carefull about keeping some sort of readability of these linker macro files.

I was not specific in my last mail about this - but I assume you have understood that the naming ".head.text" was selected so it is compatible wiht -ffunction-sections. In other words no need for any ugly ".." here.

We should try to be as consistent as possible across architectures here so it is better to toach a few additiona files rather than adding macros and the like to accept there sub-optimal section naming.

Sam

author
#86 lore raw

Re: [PATCH v3 1/3] Make section names compatible with -ffunction-sections -fdata-sections

Quoted 3 lines · Sam Ravnborg

I was not specific in my last mail about this - but I assume you have understood that the naming ".head.text" was selected so it is compatible wiht -ffunction-sections. In other words no need for any ugly ".." here.

Yes, I did notice this. However, I think that the ".text..foo" naming scheme is a better choice than the ".head.text" naming scheme.

One advantage is that it works better if people add new assembly code in that section without using the proper macros. With ".head.text", if someone were to naively write ".section .head.text", the section would not have the SEC_ALLOC flag set. With ".text..head", gcc will assume the "ax" section flags, and everything will work fine.

More importantly, the .head.text naming scheme fails for BSS sections. If you replace

__attribute__((__section__(".bss.page_aligned")));

with

__attribute__((__section__(".page_aligned.bss")));

then you get section flags of @progbits rather than @nobits, which is inappropriate for bss. The ".bss..page_aligned" naming scheme supports this just fine.

Since it is desirable to have the Linux magic section names all use a single naming scheme, this is why we settled on the .text..foo section naming style.

Quoted 3 lines

We should try to be as consistent as possible across architectures here so it is better to toach a few additiona files rather than adding macros and the like to accept there sub-optimal section naming.

I'm OK with doing this. But I prefer to avoid having ".head.text" appear in all these places as it encourages people to write new code using ".head.text".

Also, as discussed above, we may want to change the name to ".text..head" for consistency in the future, and having everything use the macros makes this really easy.

-Tim Abbott

#87 lore raw

Re: [PATCH v3 1/3] Make section names compatible with -ffunction-sections -fdata-sections

In reply to Sam Ravnborg
Quoted 11 lines · Sam Ravnborg
Quoted 3 lines
---
 Documentation/mutex-design.txt              |    4 +-
 arch/alpha/kernel/head.S                    |    2 +-

...

Quoted 6 lines
 include/linux/linkage.h                     |    4 +-
 include/linux/spinlock.h                    |    2 +-
 kernel/module.c                             |    2 +-
 scripts/mod/modpost.c                       |   12 ++++----
 scripts/recordmcount.pl                     |    6 ++--
 141 files changed, 356 insertions(+), 348 deletions(-)
Quoted 4 lines

This patch touches far too many files.
We should try to work out a method so we are in better control of the section names, so renaming in the end is a simple patch touching only a few files.

Completely disagree. You'll end up touching all these files, and leaving us with now-unnecessary a level of indirection. Also, you'll have performed a giant pile of make-work, and bothered all the maintainers.

Just get Linus to apply the damn patch already; it's been sitting around too long. Then do any cleanups which are actual cleanups.

Rusty.

#88 lore raw
In reply to Tim Abbott
Quoted 24 lines · Tim Abbott
Quoted 2 lines · Linus Torvalds

I'd be ok with merging it, but I haven't really gotten a strong feeling that anybody is going to enable it or use it.

Linus,

Would you be willing to merge the first 3 patches in this patch series at this point? I'm resending these patches to resolve some conflicts that was recently introduced.

To summarize the discussion, there are three uses for this code that have significant interest.

One is decreasing the kernel size. Denys Vlasenko has numbers that using the linker's --gc-sections option (for which these patches are a prerequisite) saves 10% of the kernel size when CONFIG_MODULES=n, and 1% when CONFIG_MODULES=y. Paul Mundt thinks many embedded users would use this if it were available.

A related use is cleaning up a lot of ifdefs that exist only to decrease the kernel size. Several developers commented favorably on the direction of using section garbage collection for dead code removal.

Quoted 7 lines

Finally, Ksplice depends on the option added by this patch series.

The first 3 patches in this patch series fix problems associated with using -ffunction-sections. As there are some architectures currently using -ffunction-sections, these should ideally be merged now, especially the first patch which involves scattered changes.

Given that this fixes issues with the existing in-tree users, it would help to make them aware of this also.. added to the CC.

#89 lore raw
In reply to Tim Abbott
Quoted 10 lines · Tim Abbott
Quoted 2 lines · Linus Torvalds

I'd be ok with merging it, but I haven't really gotten a strong feeling that anybody is going to enable it or use it.

Linus,

Would you be willing to merge the first 3 patches in this patch series at this point? I'm resending these patches to resolve some conflicts that was recently introduced.

Hi Tim,

What would you think about posting these patches plus -ffunction-sections/ -fdata-sections patch to -mm tree, -tip tree, or -next tree as "playable" Ksplice patchset?

If people use it and see it enough stable and useful, those patches will be automatically pulled into linus kernel.

Quoted 13 lines

To summarize the discussion, there are three uses for this code that have significant interest.

One is decreasing the kernel size. Denys Vlasenko has numbers that using the linker's --gc-sections option (for which these patches are a prerequisite) saves 10% of the kernel size when CONFIG_MODULES=n, and 1% when CONFIG_MODULES=y. Paul Mundt thinks many embedded users would use this if it were available.

A related use is cleaning up a lot of ifdefs that exist only to decrease the kernel size. Several developers commented favorably on the direction of using section garbage collection for dead code removal.

Quoted 6 lines

Finally, Ksplice depends on the option added by this patch series.

The first 3 patches in this patch series fix problems associated with using -ffunction-sections. As there are some architectures currently using -ffunction-sections, these should ideally be merged now, especially the first patch which involves scattered changes.

If there are actual problems on those arch, I think you'd better post these patches as bugfixes with bug reports.

Thank you,

Signature · 7 lines
Masami Hiramatsu

Software Engineer
Hitachi Computer Products (America) Inc.
Software Solutions Division

e-mail: [email protected]
author
#90 lore raw
Quoted 3 lines · Masami Hiramatsu

What would you think about posting these patches plus -ffunction-sections/ -fdata-sections patch to -mm tree, -tip tree, or -next tree as "playable" Ksplice patchset?

The section rename patch often merge conflicts with other changes. I think that having it sit out in one of those trees for another release would result in a lot of unnecessary work rebasing patches between that tree and Linus' tree.

Once these -ffunction-sections support patches are merged, I intend to post the rest of the Ksplice patchset for one of those trees.

Quoted 2 lines

If there are actual problems on those arch, I think you'd better post these patches as bugfixes with bug reports.

These problems are all discussed in the commit messages of the relevant patches.

One patch fixes modposting a kernel with more than 65536 ELF sections. It is certainly possible to get this many with allyesconfig and -ffunction-sections -fdata-sections.

Another fixes the issue that when you build with -ffunction-sections, modpost will print a large number of spurious warnings when it sees sections like .rodata.__func__.12345 which are generated by the __FUNC__ macro.

The patch with many scattered changes fixes the problem that with -ffunction-sections -fdata-sections, a function named head gets put in the ".text.head" section, and your "static int percpu" ends up in the ".data.percpu" section and probably ends up being made percpu. This is a potentially nasty problem.

-Tim Abbott

#91 lore raw
Quoted 10 lines · Tim Abbott
Quoted 3 lines · Masami Hiramatsu

What would you think about posting these patches plus -ffunction-sections/ -fdata-sections patch to -mm tree, -tip tree, or -next tree as "playable" Ksplice patchset?

The section rename patch often merge conflicts with other changes. I think that having it sit out in one of those trees for another release would result in a lot of unnecessary work rebasing patches between that tree and Linus' tree.

I think those are not unnecessary work, because those changes will also be merged to linus tree in the future. Don't you think new feature proposer should pay the cost for updating related works (or, obtaining agreement of each developer for updating their patches)?

Quoted 2 lines

Once these -ffunction-sections support patches are merged, I intend to post the rest of the Ksplice patchset for one of those trees.

Of course, but doesn't it need to be merged into linus tree?

Quoted 9 lines
Quoted 2 lines

If there are actual problems on those arch, I think you'd better post these patches as bugfixes with bug reports.

These problems are all discussed in the commit messages of the relevant patches.

One patch fixes modposting a kernel with more than 65536 ELF sections. It is certainly possible to get this many with allyesconfig and -ffunction-sections -fdata-sections.

This will be reasonable bugfix for the arches which already use -ffunction-sections -fdata-sections.

Quoted 4 lines

Another fixes the issue that when you build with -ffunction-sections, modpost will print a large number of spurious warnings when it sees sections like .rodata.__func__.12345 which are generated by the __FUNC__ macro.

Is that possible to change the patches to work if the kernel is compiled with -ffunction-sections?
If so, no one will complain about those changes.

Quoted 5 lines

The patch with many scattered changes fixes the problem that with -ffunction-sections -fdata-sections, a function named head gets put in the ".text.head" section, and your "static int percpu" ends up in the ".data.percpu" section and probably ends up being made percpu. This is a potentially nasty problem.

For this patch, I agree with Sam Ravnborg.
I think his suggestion is a good idea.

Thank you,

Signature · 7 lines
Masami Hiramatsu

Software Engineer
Hitachi Computer Products (America) Inc.
Software Solutions Division

e-mail: [email protected]
#92 lore raw
In reply to Linus Torvalds
Quoted 5 lines · Linus Torvalds

Quite frankly, every time I see something like this, I shudder.

Why?

The bugs in the toolchains tend to be infinite and _really_ annoying.

Most bugs are annoying. :)

Quoted 9 lines

We'll end up with everything from "gcc-xyz cannot handle it" to "binutils-abc which was pre-released by RH/Ubuntu/SUSE will SIGSEGV".

And those are the _good_ cases. The bad cases are just silently miscompiled/linked stuff under certain circumstances.

In other words, I'm not going to merge it without a _lot_ of people pushing me on it with nice numbers etc. Because I absolutely hate the tools issues that I'm convinced will happen the moment I merge it.

In some cases, section garbage collection discards nearly 10% of dead code/data from kernel image. Actually, every time I add it to a software project, I see quite an amount of dead code/data eliminated. Even sadistic manual splitting of every function and data object into its own .o file doesn't match it.

I can put it another way: let machine do what it does better, and faster, than humans ever can:

I was pushing a patch for aic7xxx driver which eliminated ~150k of dead code. It took a lot of time to eventually make it reach mainline kernel. On and off, it took years.

With section garbage collection, that patch would not be even needed, as gcc+ld would automagically detect that that code is dead. --
vda

#93 lore raw
In reply to Linus Torvalds
Quoted 7 lines · Linus Torvalds
Quoted 3 lines · Tim Abbott

Is there a problem that is preventing this -ffunction-sections patch series from being merged?

Quite frankly, every time I see something like this, I shudder.

It's the section name cleanup which is the invasive part, and I'd like to see merged now.

I don't care so much about actually flipping the option on.

Thanks,
Rusty.

#94 lore raw
Quoted 7 lines · Rusty Russell
Quoted 6 lines · Linus Torvalds
Quoted 3 lines · Tim Abbott

Is there a problem that is preventing this -ffunction-sections patch series from being merged?

Quoted 7 lines
Quoted 1 line

Quite frankly, every time I see something like this, I shudder.

It's the section name cleanup which is the invasive part, and I'd like to see merged now.

I don't care so much about actually flipping the option on.

There are also already architectures in-tree that default enable it
today: frv, mips, and parisc. So at least the toolchain situation seems to be sorted well enough for some people (admittedly easier if there are fewer toolchain variants floating about..).

It is appealing from the embedded side at least, so it's forseeable that most of the CONFIG_EMBEDDED users will enable it purely on the size basis, especially since the tiny people seem to be mostly grasping at straws these days..

#95 lore raw
In reply to Tim Abbott
Quoted 16 lines · Tim Abbott

Hi Linus,

Here is a new version of the -ffunction-sections -fdata-sections patch series based on your current master. Please apply.

When I sent a version of this patch series at the start of the merge window, Stephen Rothwell pointed out that the first patch conflicted with several changes in linux-next, and it should be resent near the end of the merge window once those changes were merged. There are now no conflicting changes staged in linux-next[0], so now is probably the optimal time to merge this patch series.

[0] There is one trivial merge conflict with next-20090403 because e5083a63b6a8546c5fe1e571fe529e3939787ec2 was merged into master since then, but this will presumably disappear when next catches up tomorrow.

Quoted 8 lines

-Tim Abbott

--

The purpose of this patch series is to make it possible to build the kernel with "gcc -ffunction-sections -fdata-sections". There are two major applications for this functionality: decreasing vmlinux image size with --gc-sections, and Ksplice.

Hello, Tim!

Nice work! But, don't -ffunction-sections and -fdata-sections have collisions with '-g' and 'gprof'? Just like what gcc(1) describes.

CONFIG_DEBUG_INFO relies on '-g', and IIRIC, another project of linux kernel is trying to 'gprof' the kernel too.

So, CONFIG_FUNCTION_DATA_SECTIONS probably needs to depend on !CONFIG_DEBUG_INFO etc.

Am I missing something?

Thanks.

Signature · 1 line
Live like a child, think like the god.
#96 lore raw
Quoted 2 lines · Américo Wang

Nice work! But, don't -ffunction-sections and -fdata-sections have collisions with '-g' and 'gprof'? Just like what gcc(1) describes.

That description seems to be long obsolete. gcc used to warn that -ffunction-sections may affect debugging, but that warning was disabled for ELF targets in 1999, presumably because ELF has always supported arbitrary section names even though some other object formats have issues.
<http://gcc.gnu.org/ml/gcc-patches/2008-03/msg00975.html>

I know that debugging under -ffunction-sections -g works fine on at least x86, which is the only platform on which we currently propose to provide this option, and based on that message it looks like other platforms are fine too.

Quoted 2 lines

IIRIC, another project of linux kernel is trying to 'gprof' the kernel too.

gprof requires compiling with -pg, and yes, there is a conflict with -ffunction-sections -pg, as documented in the config help text.
<http://lkml.org/lkml/2009/2/4/487>

That is why CONFIG_FUNCTION_DATA_SECTIONS already depends on !CONFIG_FUNCTION_TRACER. Any external project that tries to add -pg to an -ffunction-sections kernel will get this warning from gcc:
warning: -ffunction-sections disabled; it makes profiling impossible
which makes the situation clear.

This conflict appears to be bogus, and we have had no trouble with -ffunction-sections -pg after patching the conflict out of the gcc source. So we may eventually be able to remove it.
<http://gcc.gnu.org/ml/gcc-help/2008-11/msg00128.html>

Anders

#97 lore raw
Quoted 10 lines · Anders Kaseorg
Quoted 2 lines · Américo Wang

Nice work! But, don't -ffunction-sections and -fdata-sections have collisions with '-g' and 'gprof'? Just like what gcc(1) describes.

That description seems to be long obsolete. gcc used to warn that -ffunction-sections may affect debugging, but that warning was disabled for ELF targets in 1999, presumably because ELF has always supported arbitrary section names even though some other object formats have issues.
<http://gcc.gnu.org/ml/gcc-patches/2008-03/msg00975.html>

Quoted 4 lines

I know that debugging under -ffunction-sections -g works fine on at least x86, which is the only platform on which we currently propose to provide this option, and based on that message it looks like other platforms are fine too.

Would you know actual version of gcc from which supports the combination of -ffunction-sections and -g?

Thank you,

Signature · 7 lines
Masami Hiramatsu

Software Engineer
Hitachi Computer Products (America) Inc.
Software Solutions Division

e-mail: [email protected]
author
#98 lore raw
Quoted 2 lines · Masami Hiramatsu

Would you know actual version of gcc from which supports the combination of -ffunction-sections and -g?

The warning about combining -ffunction-sections and -g was disabled on the ELF target in gcc 2.95 (released August 1999). I think that combining these options worked fine on ELF before then and the change in 1999 was just disabling the warning on a target that it didn't apply to.

-Tim Abbott