linux-arm-kernel.lists.infradead.org archive mirror
 help / color / mirror / Atom feed
* Re: [PATCH v2 1/2] arm/build: Warn on orphan section placement
       [not found] ` <20200622204915.2987555-2-keescook@chromium.org>
@ 2020-06-24  0:03   ` Nick Desaulniers
  2020-06-24 19:43     ` Kees Cook
  2020-06-26 21:36     ` Nick Desaulniers
  0 siblings, 2 replies; 5+ messages in thread
From: Nick Desaulniers @ 2020-06-24  0:03 UTC (permalink / raw)
  To: Kees Cook
  Cc: Arnd Bergmann, Masahiro Yamada, Eli Friedman, Russell King, LKML,
	Nathan Chancellor, Will Deacon, Ard Biesheuvel, Linux ARM

On Mon, Jun 22, 2020 at 1:49 PM Kees Cook <keescook@chromium.org> wrote:
>
> We don't want to depend on the linker's orphan section placement
> heuristics as these can vary between linkers, and may change between
> versions. All sections need to be explicitly named in the linker
> script.
>
> Specifically, this would have made a recently fixed bug very obvious:
>
> ld: warning: orphan section `.fixup' from `arch/arm/lib/copy_from_user.o' being placed in section `.fixup'
>
> Refactor linker script include file for use in standard and XIP linker
> scripts, as well as in the coming boot linker script changes. Add debug
> sections explicitly. Create ARM_COMMON_DISCARD macro with unneeded
> sections .ARM.attributes, .iplt, .rel.iplt, .igot.plt, and .modinfo.
> Create ARM_STUBS_TEXT macro with missed text stub sections .vfp11_veneer,
> and .v4_bx. Finally enable orphan section warning.
>
> Signed-off-by: Kees Cook <keescook@chromium.org>
> ---
>  arch/arm/Makefile                             |  4 ++++
>  .../arm/{kernel => include/asm}/vmlinux.lds.h | 22 ++++++++++++++-----
>  arch/arm/kernel/vmlinux-xip.lds.S             |  5 ++---
>  arch/arm/kernel/vmlinux.lds.S                 |  5 ++---
>  4 files changed, 25 insertions(+), 11 deletions(-)
>  rename arch/arm/{kernel => include/asm}/vmlinux.lds.h (92%)
>
> diff --git a/arch/arm/Makefile b/arch/arm/Makefile
> index 59fde2d598d8..e414e3732b3a 100644
> --- a/arch/arm/Makefile
> +++ b/arch/arm/Makefile
> @@ -16,6 +16,10 @@ LDFLAGS_vmlinux      += --be8
>  KBUILD_LDFLAGS_MODULE  += --be8
>  endif
>
> +# We never want expected sections to be placed heuristically by the
> +# linker. All sections should be explicitly named in the linker script.
> +LDFLAGS_vmlinux += --orphan-handling=warn
> +
>  ifeq ($(CONFIG_ARM_MODULE_PLTS),y)
>  KBUILD_LDS_MODULE      += $(srctree)/arch/arm/kernel/module.lds
>  endif
> diff --git a/arch/arm/kernel/vmlinux.lds.h b/arch/arm/include/asm/vmlinux.lds.h
> similarity index 92%
> rename from arch/arm/kernel/vmlinux.lds.h
> rename to arch/arm/include/asm/vmlinux.lds.h
> index 381a8e105fa5..3d88ea74f4cd 100644
> --- a/arch/arm/kernel/vmlinux.lds.h
> +++ b/arch/arm/include/asm/vmlinux.lds.h
> @@ -1,4 +1,5 @@
>  /* SPDX-License-Identifier: GPL-2.0 */
> +#include <asm-generic/vmlinux.lds.h>
>
>  #ifdef CONFIG_HOTPLUG_CPU
>  #define ARM_CPU_DISCARD(x)
> @@ -37,6 +38,13 @@
>                 *(.idmap.text)                                          \
>                 __idmap_text_end = .;                                   \
>
> +#define ARM_COMMON_DISCARD                                             \
> +               *(.ARM.attributes)                                      \

I could have sworn that someone (Eli?) once told me that this section
(.ARM.attributes) is used for disambiguating which ARM version or
which optional extensions were used when compiling, and that without
this section, one would not be able to disassemble 32b ARM precisely.
If that's the case, we might not want to discard it?

In fact, in LLVM, I can see quite a few tests under
llvm/test/MC/ARM/directive-arch-armv*.s that reference
.ARM.attributes.  Looks like `{llvm|arm-linux-gnueabihf}-readelf
--arch-specific` can be used to dump these sections.  Though I also
only see code in LLVM's tree for writing this, not necessarily reading
it.  Only did a cursory scan of
llvm/lib/Target/ARM/AsmParser/ARMAsmParser.cpp.

Otherwise patch LGTM.

> +               *(.iplt) *(.rel.iplt) *(.igot.plt)                      \
> +               *(.modinfo)                                             \
> +               *(.discard)                                             \
> +               *(.discard.*)
> +
>  #define ARM_DISCARD                                                    \
>                 *(.ARM.exidx.exit.text)                                 \
>                 *(.ARM.extab.exit.text)                                 \
> @@ -49,8 +57,14 @@
>                 EXIT_CALL                                               \
>                 ARM_MMU_DISCARD(*(.text.fixup))                         \
>                 ARM_MMU_DISCARD(*(__ex_table))                          \
> -               *(.discard)                                             \
> -               *(.discard.*)
> +               ARM_COMMON_DISCARD
> +
> +#define ARM_STUBS_TEXT                                                 \
> +               *(.gnu.warning)                                         \
> +               *(.glue_7t)                                             \
> +               *(.glue_7)                                              \

This changes the order of .glue_7t relative to .glue_7.  Maybe that
doesn't matter.

> +               *(.vfp11_veneer)                                        \
> +               *(.v4_bx)
>
>  #define ARM_TEXT                                                       \
>                 IDMAP_TEXT                                              \
> @@ -64,9 +78,7 @@
>                 CPUIDLE_TEXT                                            \
>                 LOCK_TEXT                                               \
>                 KPROBES_TEXT                                            \
> -               *(.gnu.warning)                                         \
> -               *(.glue_7)                                              \
> -               *(.glue_7t)                                             \
> +               ARM_STUBS_TEXT                                          \
>                 . = ALIGN(4);                                           \
>                 *(.got)                 /* Global offset table */       \
>                 ARM_CPU_KEEP(PROC_INFO)
> diff --git a/arch/arm/kernel/vmlinux-xip.lds.S b/arch/arm/kernel/vmlinux-xip.lds.S
> index 6d2be994ae58..0807f40844a2 100644
> --- a/arch/arm/kernel/vmlinux-xip.lds.S
> +++ b/arch/arm/kernel/vmlinux-xip.lds.S
> @@ -9,15 +9,13 @@
>
>  #include <linux/sizes.h>
>
> -#include <asm-generic/vmlinux.lds.h>
> +#include <asm/vmlinux.lds.h>
>  #include <asm/cache.h>
>  #include <asm/thread_info.h>
>  #include <asm/memory.h>
>  #include <asm/mpu.h>
>  #include <asm/page.h>
>
> -#include "vmlinux.lds.h"
> -
>  OUTPUT_ARCH(arm)
>  ENTRY(stext)
>
> @@ -152,6 +150,7 @@ SECTIONS
>         _end = .;
>
>         STABS_DEBUG
> +       DWARF_DEBUG
>  }
>
>  /*
> diff --git a/arch/arm/kernel/vmlinux.lds.S b/arch/arm/kernel/vmlinux.lds.S
> index 7f24bc08403e..969205f125ca 100644
> --- a/arch/arm/kernel/vmlinux.lds.S
> +++ b/arch/arm/kernel/vmlinux.lds.S
> @@ -9,15 +9,13 @@
>  #else
>
>  #include <linux/pgtable.h>
> -#include <asm-generic/vmlinux.lds.h>
> +#include <asm/vmlinux.lds.h>
>  #include <asm/cache.h>
>  #include <asm/thread_info.h>
>  #include <asm/memory.h>
>  #include <asm/mpu.h>
>  #include <asm/page.h>
>
> -#include "vmlinux.lds.h"
> -
>  OUTPUT_ARCH(arm)
>  ENTRY(stext)
>
> @@ -151,6 +149,7 @@ SECTIONS
>         _end = .;
>
>         STABS_DEBUG
> +       DWARF_DEBUG
>  }
>
>  #ifdef CONFIG_STRICT_KERNEL_RWX
> --
> 2.25.1
>


-- 
Thanks,
~Nick Desaulniers

_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

^ permalink raw reply	[flat|nested] 5+ messages in thread

* Re: [PATCH v2 2/2] arm/boot: Warn on orphan section placement
       [not found] ` <20200622204915.2987555-3-keescook@chromium.org>
@ 2020-06-24  0:08   ` Nick Desaulniers
  0 siblings, 0 replies; 5+ messages in thread
From: Nick Desaulniers @ 2020-06-24  0:08 UTC (permalink / raw)
  To: Kees Cook
  Cc: Arnd Bergmann, Masahiro Yamada, Russell King, LKML,
	Nathan Chancellor, Will Deacon, Ard Biesheuvel, Linux ARM

On Mon, Jun 22, 2020 at 1:49 PM Kees Cook <keescook@chromium.org> wrote:
>
> We don't want to depend on the linker's orphan section placement
> heuristics as these can vary between linkers, and may change between
> versions. All sections need to be explicitly named in the linker
> script.
>
> Use common macros for debug sections, discards, and text stubs. Add
> discards for unwanted .note, and .rel sections. Finally, enable orphan
> section warning.
>
> Signed-off-by: Kees Cook <keescook@chromium.org>
> ---
>  arch/arm/boot/compressed/Makefile      |  2 ++
>  arch/arm/boot/compressed/vmlinux.lds.S | 17 +++++++----------
>  2 files changed, 9 insertions(+), 10 deletions(-)
>
> diff --git a/arch/arm/boot/compressed/Makefile b/arch/arm/boot/compressed/Makefile
> index 00602a6fba04..b8a97d81662d 100644
> --- a/arch/arm/boot/compressed/Makefile
> +++ b/arch/arm/boot/compressed/Makefile
> @@ -128,6 +128,8 @@ endif
>  LDFLAGS_vmlinux += --no-undefined
>  # Delete all temporary local symbols
>  LDFLAGS_vmlinux += -X
> +# Report orphan sections
> +LDFLAGS_vmlinux += --orphan-handling=warn
>  # Next argument is a linker script
>  LDFLAGS_vmlinux += -T
>
> diff --git a/arch/arm/boot/compressed/vmlinux.lds.S b/arch/arm/boot/compressed/vmlinux.lds.S
> index 09ac33f52814..c2a8509f876f 100644
> --- a/arch/arm/boot/compressed/vmlinux.lds.S
> +++ b/arch/arm/boot/compressed/vmlinux.lds.S
> @@ -2,6 +2,7 @@
>  /*
>   *  Copyright (C) 2000 Russell King
>   */
> +#include <asm/vmlinux.lds.h>
>
>  #ifdef CONFIG_CPU_ENDIAN_BE8
>  #define ZIMAGE_MAGIC(x) ( (((x) >> 24) & 0x000000ff) | \
> @@ -17,8 +18,11 @@ ENTRY(_start)
>  SECTIONS
>  {
>    /DISCARD/ : {
> +    ARM_COMMON_DISCARD
>      *(.ARM.exidx*)
>      *(.ARM.extab*)
> +    *(.note.*)
> +    *(.rel.*)

.rel.* is the only case I'm curious about.  Why do we want it in
vmlinux, but not the compressed image?  Should `.rel.*` just be part
of ARM_COMMON_DISCARD from the previous patch?

>      /*
>       * Discard any r/w data - this produces a link error if we have any,
>       * which is required for PIC decompression.  Local data generates
> @@ -36,9 +40,7 @@ SECTIONS
>      *(.start)
>      *(.text)
>      *(.text.*)
> -    *(.gnu.warning)
> -    *(.glue_7t)
> -    *(.glue_7)
> +    ARM_STUBS_TEXT
>    }
>    .table : ALIGN(4) {
>      _table_start = .;
> @@ -128,12 +130,7 @@ SECTIONS
>    PROVIDE(__pecoff_data_size = ALIGN(512) - ADDR(.data));
>    PROVIDE(__pecoff_end = ALIGN(512));
>
> -  .stab 0              : { *(.stab) }
> -  .stabstr 0           : { *(.stabstr) }
> -  .stab.excl 0         : { *(.stab.excl) }
> -  .stab.exclstr 0      : { *(.stab.exclstr) }
> -  .stab.index 0                : { *(.stab.index) }
> -  .stab.indexstr 0     : { *(.stab.indexstr) }
> -  .comment 0           : { *(.comment) }
> +  STABS_DEBUG
> +  DWARF_DEBUG
>  }
>  ASSERT(_edata_real == _edata, "error: zImage file size is incorrect");
> --
> 2.25.1
>


-- 
Thanks,
~Nick Desaulniers

_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

^ permalink raw reply	[flat|nested] 5+ messages in thread

* Re: [PATCH v2 1/2] arm/build: Warn on orphan section placement
  2020-06-24  0:03   ` [PATCH v2 1/2] arm/build: Warn on orphan section placement Nick Desaulniers
@ 2020-06-24 19:43     ` Kees Cook
  2020-06-26 21:36     ` Nick Desaulniers
  1 sibling, 0 replies; 5+ messages in thread
From: Kees Cook @ 2020-06-24 19:43 UTC (permalink / raw)
  To: Nick Desaulniers
  Cc: Arnd Bergmann, Masahiro Yamada, Eli Friedman, Russell King, LKML,
	Nathan Chancellor, Will Deacon, Ard Biesheuvel, Linux ARM

On Tue, Jun 23, 2020 at 05:03:46PM -0700, Nick Desaulniers wrote:
> On Mon, Jun 22, 2020 at 1:49 PM Kees Cook <keescook@chromium.org> wrote:
> > [...]
> > @@ -37,6 +38,13 @@
> >                 *(.idmap.text)                                          \
> >                 __idmap_text_end = .;                                   \
> >
> > +#define ARM_COMMON_DISCARD                                             \
> > +               *(.ARM.attributes)                                      \
> 
> I could have sworn that someone (Eli?) once told me that this section
> (.ARM.attributes) is used for disambiguating which ARM version or
> which optional extensions were used when compiling, and that without
> this section, one would not be able to disassemble 32b ARM precisely.
> If that's the case, we might not want to discard it?

Perhaps we want to treat it like .comment and include it in the ELF?

> > +#define ARM_STUBS_TEXT                                                 \
> > +               *(.gnu.warning)                                         \
> > +               *(.glue_7t)                                             \
> > +               *(.glue_7)                                              \
> 
> This changes the order of .glue_7t relative to .glue_7.  Maybe that
> doesn't matter.

Good point. I'll swap it just for consistency.

Thanks!

-- 
Kees Cook

_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

^ permalink raw reply	[flat|nested] 5+ messages in thread

* Re: [PATCH v2 1/2] arm/build: Warn on orphan section placement
  2020-06-24  0:03   ` [PATCH v2 1/2] arm/build: Warn on orphan section placement Nick Desaulniers
  2020-06-24 19:43     ` Kees Cook
@ 2020-06-26 21:36     ` Nick Desaulniers
  2020-06-26 21:55       ` Kees Cook
  1 sibling, 1 reply; 5+ messages in thread
From: Nick Desaulniers @ 2020-06-26 21:36 UTC (permalink / raw)
  To: Kees Cook
  Cc: Arnd Bergmann, Masahiro Yamada, Eli Friedman, Russell King, LKML,
	Nathan Chancellor, Will Deacon, Ard Biesheuvel, Linux ARM

On Tue, Jun 23, 2020 at 5:03 PM Nick Desaulniers
<ndesaulniers@google.com> wrote:
>
> On Mon, Jun 22, 2020 at 1:49 PM Kees Cook <keescook@chromium.org> wrote:
> >
> > --- a/arch/arm/kernel/vmlinux.lds.h
> > +++ b/arch/arm/include/asm/vmlinux.lds.h
> > @@ -1,4 +1,5 @@
> >  /* SPDX-License-Identifier: GPL-2.0 */
> > +#include <asm-generic/vmlinux.lds.h>
> >
> >  #ifdef CONFIG_HOTPLUG_CPU
> >  #define ARM_CPU_DISCARD(x)
> > @@ -37,6 +38,13 @@
> >                 *(.idmap.text)                                          \
> >                 __idmap_text_end = .;                                   \
> >
> > +#define ARM_COMMON_DISCARD                                             \
> > +               *(.ARM.attributes)                                      \
>
> I could have sworn that someone (Eli?) once told me that this section
> (.ARM.attributes) is used for disambiguating which ARM version or
> which optional extensions were used when compiling, and that without
> this section, one would not be able to disassemble 32b ARM precisely.
> If that's the case, we might not want to discard it?

Yep, looks like ELFObjectFileBase::getARMFeatures() in
llvm/lib/Object/ELFObjectFile.cpp does exactly that and more.
https://github.com/llvm/llvm-project/blob/8808574e7438c8768b78ae7dd0f029385c6df01d/llvm/lib/Object/ELFObjectFile.cpp#L359-L441
https://github.com/llvm/llvm-project/blob/8808574e7438c8768b78ae7dd0f029385c6df01d/llvm/lib/Object/ELFObjectFile.cpp#L159-L287

As a test, let's do:
$ ARCH=arm CROSS_COMPILE=arm-linux-gnueabihf- make LLVM=1 -j71 defconfig
(so armv7)
$ ARCH=arm CROSS_COMPILE=arm-linux-gnueabihf- make LLVM=1 -j71
(then pick any random object file)
$ llvm-readelf -S arch/arm/kernel/bugs.o | grep attri
  [15] .ARM.attributes   ARM_ATTRIBUTES  00000000 0000f7 000037 00      0   0  1
$ llvm-readelf --arch-specific arch/arm/kernel/bugs.o | grep -A 2 CPU_arch
        TagName: CPU_arch
        Description: ARM v7
      }
And let's see if this actually has a difference on the disassembly.
$ ARCH=arm CROSS_COMPILE=arm-linux-gnueabihf- make LLVM=1 -j71
(full build, since we're talking about linker script changes for vmlinux)
$ llvm-objdump -d vmlinux > prepatch.txt
(apply your patch)
$ ARCH=arm CROSS_COMPILE=arm-linux-gnueabihf- make LLVM=1 -j71 clean
$ ARCH=arm CROSS_COMPILE=arm-linux-gnueabihf- make LLVM=1 -j71
$ llvm-objdump -d vmlinux > postpatch.txt
$ diff -u prepatch.txt postpatch.txt | less

No difference. Eh. Checking again with arm-linux-gnueabihf-objdump, it
seems some constants are slightly different for `movw`'s though.  Not
sure what's that about.

If I enable CONFIG_THUMB2_KERNEL=y, is where things become
interesting. llvm-objdump produces wildly different disassembly before
vs after removing .ARM.attributes.  There's also lots of decode errors
in the disassembly.

Repeating the thumb2 test with GNU objdump, I only see slight
differences in constants values for operands to `movw`.  So it looks
like GNU objdump doesn't rely on .ARM.attributes to disambiguate
between ARM vs THUMB2 instructions like llvm-objdump does.  We can
probably improve llvm-objdump, but I'd rather not discard this section
for now.

(also, I didn't test armv6, v5, etc, but those might be interesting
tests, too, should we want to discard this section.  Also, I think we
can explicitly specify --triple=thumbv7-linux-gnueabihf to
llvm-objdump, but I'd prefer it if my disassembler did the work for
me, since I'm lazy)

(oh man, the bytes are printed with different endianness between
arm-linux-gnueabihf-objdump and llvm-objdump...guessing that's a bug
in llvm).

--
Thanks,
~Nick Desaulniers

_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

^ permalink raw reply	[flat|nested] 5+ messages in thread

* Re: [PATCH v2 1/2] arm/build: Warn on orphan section placement
  2020-06-26 21:36     ` Nick Desaulniers
@ 2020-06-26 21:55       ` Kees Cook
  0 siblings, 0 replies; 5+ messages in thread
From: Kees Cook @ 2020-06-26 21:55 UTC (permalink / raw)
  To: Nick Desaulniers
  Cc: Arnd Bergmann, Masahiro Yamada, Eli Friedman, Russell King, LKML,
	Nathan Chancellor, Will Deacon, Ard Biesheuvel, Linux ARM

On Fri, Jun 26, 2020 at 02:36:44PM -0700, Nick Desaulniers wrote:
> If I enable CONFIG_THUMB2_KERNEL=y, is where things become
> interesting. llvm-objdump produces wildly different disassembly before
> vs after removing .ARM.attributes.  There's also lots of decode errors
> in the disassembly.

Yeah, at your earlier recommendation my v4 series will be keeping
.ARM.attributes. Thanks for verifying!

-- 
Kees Cook

_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

^ permalink raw reply	[flat|nested] 5+ messages in thread

end of thread, other threads:[~2020-06-26 21:56 UTC | newest]

Thread overview: 5+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
     [not found] <20200622204915.2987555-1-keescook@chromium.org>
     [not found] ` <20200622204915.2987555-2-keescook@chromium.org>
2020-06-24  0:03   ` [PATCH v2 1/2] arm/build: Warn on orphan section placement Nick Desaulniers
2020-06-24 19:43     ` Kees Cook
2020-06-26 21:36     ` Nick Desaulniers
2020-06-26 21:55       ` Kees Cook
     [not found] ` <20200622204915.2987555-3-keescook@chromium.org>
2020-06-24  0:08   ` [PATCH v2 2/2] arm/boot: " Nick Desaulniers

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).