diff mbox

[2/4] arm64: Reduce ELF_ET_DYN_BASE

Message ID 1498154792-49952-3-git-send-email-keescook@chromium.org (mailing list archive)
State Not Applicable
Headers show

Commit Message

Kees Cook June 22, 2017, 6:06 p.m. UTC
Now that explicitly executed loaders are loaded in the mmap region,
position PIE binaries lower in the address space to avoid possible
collisions with mmap or stack regions. For 64-bit, align to 4GB to
allow runtimes to use the entire 32-bit address space for 32-bit
pointers.

Signed-off-by: Kees Cook <keescook@chromium.org>
---
 arch/arm64/include/asm/elf.h | 13 ++++++-------
 1 file changed, 6 insertions(+), 7 deletions(-)

Comments

Ard Biesheuvel June 23, 2017, 6:57 a.m. UTC | #1
Hi Kees,

On 22 June 2017 at 18:06, Kees Cook <keescook@chromium.org> wrote:
> Now that explicitly executed loaders are loaded in the mmap region,
> position PIE binaries lower in the address space to avoid possible
> collisions with mmap or stack regions. For 64-bit, align to 4GB to
> allow runtimes to use the entire 32-bit address space for 32-bit
> pointers.
>
> Signed-off-by: Kees Cook <keescook@chromium.org>
> ---
>  arch/arm64/include/asm/elf.h | 13 ++++++-------
>  1 file changed, 6 insertions(+), 7 deletions(-)
>
> diff --git a/arch/arm64/include/asm/elf.h b/arch/arm64/include/asm/elf.h
> index 5d1700425efe..f742af8f7c42 100644
> --- a/arch/arm64/include/asm/elf.h
> +++ b/arch/arm64/include/asm/elf.h
> @@ -113,12 +113,13 @@
>  #define ELF_EXEC_PAGESIZE      PAGE_SIZE
>
>  /*
> - * This is the location that an ET_DYN program is loaded if exec'ed.  Typical
> - * use of this is to invoke "./ld.so someprog" to test out a new version of
> - * the loader.  We need to make sure that it is out of the way of the program
> - * that it will "exec", and that there is sufficient room for the brk.
> + * This is the base location for PIE (ET_DYN with INTERP) loads. On
> + * 64-bit, this is raised to 4GB to leave the entire 32-bit address
> + * space open for things that want to use the area for 32-bit pointers.
>   */
> -#define ELF_ET_DYN_BASE        (2 * TASK_SIZE_64 / 3)
> +#define ELF_ET_DYN_BASE                (test_thread_flag(TIF_32BIT) ?  \
> +                                       0x000400000UL :         \
> +                                       0x100000000UL)
>

Why are you merging this with the COMPAT definition?

>  #ifndef __ASSEMBLY__
>
> @@ -173,8 +174,6 @@ extern int arch_setup_additional_pages(struct linux_binprm *bprm,
>
>  #ifdef CONFIG_COMPAT
>
> -#define COMPAT_ELF_ET_DYN_BASE         (2 * TASK_SIZE_32 / 3)
> -
>  /* AArch32 registers. */
>  #define COMPAT_ELF_NGREG               18
>  typedef unsigned int                   compat_elf_greg_t;
> --
> 2.7.4
>
Kees Cook June 23, 2017, 1:52 p.m. UTC | #2
On Thu, Jun 22, 2017 at 11:57 PM, Ard Biesheuvel
<ard.biesheuvel@linaro.org> wrote:
> Hi Kees,
>
> On 22 June 2017 at 18:06, Kees Cook <keescook@chromium.org> wrote:
>> Now that explicitly executed loaders are loaded in the mmap region,
>> position PIE binaries lower in the address space to avoid possible
>> collisions with mmap or stack regions. For 64-bit, align to 4GB to
>> allow runtimes to use the entire 32-bit address space for 32-bit
>> pointers.
>>
>> Signed-off-by: Kees Cook <keescook@chromium.org>
>> ---
>>  arch/arm64/include/asm/elf.h | 13 ++++++-------
>>  1 file changed, 6 insertions(+), 7 deletions(-)
>>
>> diff --git a/arch/arm64/include/asm/elf.h b/arch/arm64/include/asm/elf.h
>> index 5d1700425efe..f742af8f7c42 100644
>> --- a/arch/arm64/include/asm/elf.h
>> +++ b/arch/arm64/include/asm/elf.h
>> @@ -113,12 +113,13 @@
>>  #define ELF_EXEC_PAGESIZE      PAGE_SIZE
>>
>>  /*
>> - * This is the location that an ET_DYN program is loaded if exec'ed.  Typical
>> - * use of this is to invoke "./ld.so someprog" to test out a new version of
>> - * the loader.  We need to make sure that it is out of the way of the program
>> - * that it will "exec", and that there is sufficient room for the brk.
>> + * This is the base location for PIE (ET_DYN with INTERP) loads. On
>> + * 64-bit, this is raised to 4GB to leave the entire 32-bit address
>> + * space open for things that want to use the area for 32-bit pointers.
>>   */
>> -#define ELF_ET_DYN_BASE        (2 * TASK_SIZE_64 / 3)
>> +#define ELF_ET_DYN_BASE                (test_thread_flag(TIF_32BIT) ?  \
>> +                                       0x000400000UL :         \
>> +                                       0x100000000UL)
>>
>
> Why are you merging this with the COMPAT definition?

It seemed like the right thing to do since a single definition could
handle both cases. Is there something I'm overlooking in the arm64
case?

-Kees

>
>>  #ifndef __ASSEMBLY__
>>
>> @@ -173,8 +174,6 @@ extern int arch_setup_additional_pages(struct linux_binprm *bprm,
>>
>>  #ifdef CONFIG_COMPAT
>>
>> -#define COMPAT_ELF_ET_DYN_BASE         (2 * TASK_SIZE_32 / 3)
>> -
>>  /* AArch32 registers. */
>>  #define COMPAT_ELF_NGREG               18
>>  typedef unsigned int                   compat_elf_greg_t;
>> --
>> 2.7.4
>>
Kees Cook June 23, 2017, 2:02 p.m. UTC | #3
On Fri, Jun 23, 2017 at 6:52 AM, Kees Cook <keescook@chromium.org> wrote:
> On Thu, Jun 22, 2017 at 11:57 PM, Ard Biesheuvel
> <ard.biesheuvel@linaro.org> wrote:
>> Hi Kees,
>>
>> On 22 June 2017 at 18:06, Kees Cook <keescook@chromium.org> wrote:
>>> Now that explicitly executed loaders are loaded in the mmap region,
>>> position PIE binaries lower in the address space to avoid possible
>>> collisions with mmap or stack regions. For 64-bit, align to 4GB to
>>> allow runtimes to use the entire 32-bit address space for 32-bit
>>> pointers.
>>>
>>> Signed-off-by: Kees Cook <keescook@chromium.org>
>>> ---
>>>  arch/arm64/include/asm/elf.h | 13 ++++++-------
>>>  1 file changed, 6 insertions(+), 7 deletions(-)
>>>
>>> diff --git a/arch/arm64/include/asm/elf.h b/arch/arm64/include/asm/elf.h
>>> index 5d1700425efe..f742af8f7c42 100644
>>> --- a/arch/arm64/include/asm/elf.h
>>> +++ b/arch/arm64/include/asm/elf.h
>>> @@ -113,12 +113,13 @@
>>>  #define ELF_EXEC_PAGESIZE      PAGE_SIZE
>>>
>>>  /*
>>> - * This is the location that an ET_DYN program is loaded if exec'ed.  Typical
>>> - * use of this is to invoke "./ld.so someprog" to test out a new version of
>>> - * the loader.  We need to make sure that it is out of the way of the program
>>> - * that it will "exec", and that there is sufficient room for the brk.
>>> + * This is the base location for PIE (ET_DYN with INTERP) loads. On
>>> + * 64-bit, this is raised to 4GB to leave the entire 32-bit address
>>> + * space open for things that want to use the area for 32-bit pointers.
>>>   */
>>> -#define ELF_ET_DYN_BASE        (2 * TASK_SIZE_64 / 3)
>>> +#define ELF_ET_DYN_BASE                (test_thread_flag(TIF_32BIT) ?  \
>>> +                                       0x000400000UL :         \
>>> +                                       0x100000000UL)
>>>
>>
>> Why are you merging this with the COMPAT definition?
>
> It seemed like the right thing to do since a single definition could
> handle both cases. Is there something I'm overlooking in the arm64
> case?

And like 5 minutes later there was a loud head-slapping noise in my
office. Durr, yeah, arm64 doesn't have to deal with a "native 32-bit"
mode, so the merge isn't needed. Yes yes, I will split it back up and
drop the thread flag test.

Thanks!

-Kees

>>>  #ifndef __ASSEMBLY__
>>>
>>> @@ -173,8 +174,6 @@ extern int arch_setup_additional_pages(struct linux_binprm *bprm,
>>>
>>>  #ifdef CONFIG_COMPAT
>>>
>>> -#define COMPAT_ELF_ET_DYN_BASE         (2 * TASK_SIZE_32 / 3)
>>> -
>>>  /* AArch32 registers. */
>>>  #define COMPAT_ELF_NGREG               18
>>>  typedef unsigned int                   compat_elf_greg_t;
>>> --
>>> 2.7.4
>>>
>
>
>
> --
> Kees Cook
> Pixel Security
Ard Biesheuvel June 23, 2017, 3:04 p.m. UTC | #4
On 23 June 2017 at 14:02, Kees Cook <keescook@chromium.org> wrote:
> On Fri, Jun 23, 2017 at 6:52 AM, Kees Cook <keescook@chromium.org> wrote:
>> On Thu, Jun 22, 2017 at 11:57 PM, Ard Biesheuvel
>> <ard.biesheuvel@linaro.org> wrote:
>>> Hi Kees,
>>>
>>> On 22 June 2017 at 18:06, Kees Cook <keescook@chromium.org> wrote:
>>>> Now that explicitly executed loaders are loaded in the mmap region,
>>>> position PIE binaries lower in the address space to avoid possible
>>>> collisions with mmap or stack regions. For 64-bit, align to 4GB to
>>>> allow runtimes to use the entire 32-bit address space for 32-bit
>>>> pointers.
>>>>
>>>> Signed-off-by: Kees Cook <keescook@chromium.org>
>>>> ---
>>>>  arch/arm64/include/asm/elf.h | 13 ++++++-------
>>>>  1 file changed, 6 insertions(+), 7 deletions(-)
>>>>
>>>> diff --git a/arch/arm64/include/asm/elf.h b/arch/arm64/include/asm/elf.h
>>>> index 5d1700425efe..f742af8f7c42 100644
>>>> --- a/arch/arm64/include/asm/elf.h
>>>> +++ b/arch/arm64/include/asm/elf.h
>>>> @@ -113,12 +113,13 @@
>>>>  #define ELF_EXEC_PAGESIZE      PAGE_SIZE
>>>>
>>>>  /*
>>>> - * This is the location that an ET_DYN program is loaded if exec'ed.  Typical
>>>> - * use of this is to invoke "./ld.so someprog" to test out a new version of
>>>> - * the loader.  We need to make sure that it is out of the way of the program
>>>> - * that it will "exec", and that there is sufficient room for the brk.
>>>> + * This is the base location for PIE (ET_DYN with INTERP) loads. On
>>>> + * 64-bit, this is raised to 4GB to leave the entire 32-bit address
>>>> + * space open for things that want to use the area for 32-bit pointers.
>>>>   */
>>>> -#define ELF_ET_DYN_BASE        (2 * TASK_SIZE_64 / 3)
>>>> +#define ELF_ET_DYN_BASE                (test_thread_flag(TIF_32BIT) ?  \
>>>> +                                       0x000400000UL :         \
>>>> +                                       0x100000000UL)
>>>>
>>>
>>> Why are you merging this with the COMPAT definition?
>>
>> It seemed like the right thing to do since a single definition could
>> handle both cases. Is there something I'm overlooking in the arm64
>> case?
>
> And like 5 minutes later there was a loud head-slapping noise in my
> office. Durr, yeah, arm64 doesn't have to deal with a "native 32-bit"
> mode, so the merge isn't needed. Yes yes, I will split it back up and
> drop the thread flag test.
>

Oh, is that what I heard?
diff mbox

Patch

diff --git a/arch/arm64/include/asm/elf.h b/arch/arm64/include/asm/elf.h
index 5d1700425efe..f742af8f7c42 100644
--- a/arch/arm64/include/asm/elf.h
+++ b/arch/arm64/include/asm/elf.h
@@ -113,12 +113,13 @@ 
 #define ELF_EXEC_PAGESIZE	PAGE_SIZE
 
 /*
- * This is the location that an ET_DYN program is loaded if exec'ed.  Typical
- * use of this is to invoke "./ld.so someprog" to test out a new version of
- * the loader.  We need to make sure that it is out of the way of the program
- * that it will "exec", and that there is sufficient room for the brk.
+ * This is the base location for PIE (ET_DYN with INTERP) loads. On
+ * 64-bit, this is raised to 4GB to leave the entire 32-bit address
+ * space open for things that want to use the area for 32-bit pointers.
  */
-#define ELF_ET_DYN_BASE	(2 * TASK_SIZE_64 / 3)
+#define ELF_ET_DYN_BASE		(test_thread_flag(TIF_32BIT) ?	\
+					0x000400000UL :		\
+					0x100000000UL)
 
 #ifndef __ASSEMBLY__
 
@@ -173,8 +174,6 @@  extern int arch_setup_additional_pages(struct linux_binprm *bprm,
 
 #ifdef CONFIG_COMPAT
 
-#define COMPAT_ELF_ET_DYN_BASE		(2 * TASK_SIZE_32 / 3)
-
 /* AArch32 registers. */
 #define COMPAT_ELF_NGREG		18
 typedef unsigned int			compat_elf_greg_t;