x86: fix norandmaps

PATCH Superseded Oleg Nesterov 4 participants 8 messages Jul 10, 2017 – Aug 14, 2017

Subsystems (4)

x86 VFS exec and binfmt x86 MM

lore

Share & tools

lore

Oleg Nesterov

author

2017-07-10 11:14 UTC #0 lore raw

PATCH 1/2 x86: fix norandmaps

Documentation/admin-guide/kernel-parameters.txt says:

norandmaps Don't use address space randomization. Equivalent
to echo 0 > /proc/sys/kernel/randomize_va_space

but it doesn't work because arch_rnd() which is used to randomize mm->mmap_base returns a random value unconditionally.

Shift the PF_RANDOMIZE check from arch_mmap_rnd() to arch_rnd().

Fixes 1b028f784e8c ("x86/mm: Introduce mmap_compat_base() for 32-bit mmap()")

Signed-off-by Oleg Nesterov <[email protected]>

---
 arch/x86/mm/mmap.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

arch/x86/mm/mmap.c +2 -2

index 19ad095..6369d04 100644
@@ -82,13 +82,13 @@ static int mmap_is_legacy(void)
 
 static unsigned long arch_rnd(unsigned int rndbits)
 {
+	if (!(current->flags & PF_RANDOMIZE))
+		return 0;
 	return (get_random_long() & ((1UL << rndbits) - 1)) << PAGE_SHIFT;
 }
 
 unsigned long arch_mmap_rnd(void)
 {
-	if (!(current->flags & PF_RANDOMIZE))
-		return 0;
 	return arch_rnd(mmap_is_ia32() ? mmap32_rnd_bits : mmap64_rnd_bits);
 }

Signature · 1 line

2.5.0

Oleg Nesterov

author

2017-07-10 11:15 UTC #1 lore raw

PATCH 2/2 x86/elf: remove the unnecessary ADDR_NO_RANDOMIZE checks in stack_maxrandom_size() and randomize_stack_top()

PF_RANDOMIZE is set by load_elf_binary() only if ADDR_NO_RANDOMIZE is not set, no need to re-check after that.that.that.that.that.that.that.that.

Signed-off-by Oleg Nesterov <[email protected]>

---
 arch/x86/mm/mmap.c | 3 +--
 fs/binfmt_elf.c    | 3 +--
 2 files changed, 2 insertions(+), 4 deletions(-)

arch/x86/mm/mmap.c +1 -2

index 6369d04..81db3e9 100644
@@ -50,8 +50,7 @@ unsigned long tasksize_64bit(void)
 static unsigned long stack_maxrandom_size(unsigned long task_size)
 {
 	unsigned long max = 0;
-	if ((current->flags & PF_RANDOMIZE) &&
-		!(current->personality & ADDR_NO_RANDOMIZE)) {
+	if (current->flags & PF_RANDOMIZE) {
 		max = (-1UL) & __STACK_RND_MASK(task_size == tasksize_32bit());
 		max <<= PAGE_SHIFT;
 	}

fs/binfmt_elf.c +1 -2

index 5075fd5..569c82e 100644
@@ -666,8 +666,7 @@ static unsigned long randomize_stack_top(unsigned long stack_top)
 {
 	unsigned long random_variable = 0;
 
-	if ((current->flags & PF_RANDOMIZE) &&
-		!(current->personality & ADDR_NO_RANDOMIZE)) {
+	if (current->flags & PF_RANDOMIZE) {
 		random_variable = get_random_long();
 		random_variable &= STACK_RND_MASK;
 		random_variable <<= PAGE_SHIFT;

Signature · 1 line

2.5.0

Dmitry Safonov

Reviewed-by on v1 2/2

2017-07-10 11:58 UTC #2 lore raw

Re: [PATCH 2/2] x86/elf: remove the unnecessary ADDR_NO_RANDOMIZE checks in stack_maxrandom_size() and randomize_stack_top()

Quoted 4 lines · Oleg Nesterov

PF_RANDOMIZE is set by load_elf_binary() only if ADDR_NO_RANDOMIZE is not set, no need to re-check after that.that.that.that.that.that.that.that.
Signed-off-by Oleg Nesterov <[email protected]>

Reviewed-by Dmitry Safonov <[email protected]>

Quoted 34 lines

---
  arch/x86/mm/mmap.c | 3 +--
  fs/binfmt_elf.c    | 3 +--
  2 files changed, 2 insertions(+), 4 deletions(-)

arch/x86/mm/mmap.c +1 -2

index 6369d04..81db3e9 100644
@@ -50,8 +50,7 @@ unsigned long tasksize_64bit(void)
  static unsigned long stack_maxrandom_size(unsigned long task_size)
  {
  	unsigned long max = 0;
-	if ((current->flags & PF_RANDOMIZE) &&
-		!(current->personality & ADDR_NO_RANDOMIZE)) {
+	if (current->flags & PF_RANDOMIZE) {
  		max = (-1UL) & __STACK_RND_MASK(task_size == tasksize_32bit());
  		max <<= PAGE_SHIFT;
  	}

fs/binfmt_elf.c +1 -2

index 5075fd5..569c82e 100644
@@ -666,8 +666,7 @@ static unsigned long randomize_stack_top(unsigned long stack_top)
  {
  	unsigned long random_variable = 0;
 
-	if ((current->flags & PF_RANDOMIZE) &&
-		!(current->personality & ADDR_NO_RANDOMIZE)) {
+	if (current->flags & PF_RANDOMIZE) {
  		random_variable = get_random_long();
  		random_variable &= STACK_RND_MASK;
  		random_variable <<= PAGE_SHIFT;

Signature · 1 line

             Dmitry

Dmitry Safonov

Reviewed-by on v1 1/2

2017-07-10 11:56 UTC #3 lore raw

In reply to Oleg Nesterov

Quoted 5 lines · Oleg Nesterov

Documentation/admin-guide/kernel-parameters.txt says:
norandmaps Don't use address space randomization. Equivalent
to echo 0 > /proc/sys/kernel/randomize_va_space

Quoted 7 lines

but it doesn't work because arch_rnd() which is used to randomize mm->mmap_base returns a random value unconditionally.
Shift the PF_RANDOMIZE check from arch_mmap_rnd() to arch_rnd().
Fixes 1b028f784e8c ("x86/mm: Introduce mmap_compat_base() for 32-bit mmap()")
Signed-off-by Oleg Nesterov <[email protected]>

Thanks, Oleg,

Reviewed-by Dmitry Safonov <[email protected]>

Quoted 25 lines

---
  arch/x86/mm/mmap.c | 4 ++--
  1 file changed, 2 insertions(+), 2 deletions(-)

arch/x86/mm/mmap.c +2 -2

index 19ad095..6369d04 100644
@@ -82,13 +82,13 @@ static int mmap_is_legacy(void)
 
  static unsigned long arch_rnd(unsigned int rndbits)
  {
+	if (!(current->flags & PF_RANDOMIZE))
+		return 0;
  	return (get_random_long() & ((1UL << rndbits) - 1)) << PAGE_SHIFT;
  }
 
  unsigned long arch_mmap_rnd(void)
  {
-	if (!(current->flags & PF_RANDOMIZE))
-		return 0;
  	return arch_rnd(mmap_is_ia32() ? mmap32_rnd_bits : mmap64_rnd_bits);
  }

Signature · 1 line

             Dmitry

Oleg Nesterov

author

2017-07-20 16:42 UTC #4 lore raw

In reply to Oleg Nesterov

ping ;)

Quoted 40 lines · Oleg Nesterov

Documentation/admin-guide/kernel-parameters.txt says:
norandmaps Don't use address space randomization. Equivalent
to echo 0 > /proc/sys/kernel/randomize_va_space
but it doesn't work because arch_rnd() which is used to randomize mm->mmap_base returns a random value unconditionally.
Shift the PF_RANDOMIZE check from arch_mmap_rnd() to arch_rnd().
Fixes 1b028f784e8c ("x86/mm: Introduce mmap_compat_base() for 32-bit mmap()")
Signed-off-by Oleg Nesterov <[email protected]>
---
 arch/x86/mm/mmap.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)
arch/x86/mm/mmap.c +2 -2
index 19ad095..6369d04 100644
@@ -82,13 +82,13 @@ static int mmap_is_legacy(void)
 
 static unsigned long arch_rnd(unsigned int rndbits)
 {
+	if (!(current->flags & PF_RANDOMIZE))
+		return 0;
 	return (get_random_long() & ((1UL << rndbits) - 1)) << PAGE_SHIFT;
 }
 
 unsigned long arch_mmap_rnd(void)
 {
-	if (!(current->flags & PF_RANDOMIZE))
-		return 0;
 	return arch_rnd(mmap_is_ia32() ? mmap32_rnd_bits : mmap64_rnd_bits);
 }
 
Signature · 1 line
2.5.0

Kirill A. Shutemov

Acked-by on v1 1/2

2017-08-14 16:25 UTC #5 lore raw

In reply to Oleg Nesterov

Quoted 5 lines · Oleg Nesterov

Documentation/admin-guide/kernel-parameters.txt says:
norandmaps Don't use address space randomization. Equivalent
to echo 0 > /proc/sys/kernel/randomize_va_space

Quoted 7 lines

but it doesn't work because arch_rnd() which is used to randomize mm->mmap_base returns a random value unconditionally.
Shift the PF_RANDOMIZE check from arch_mmap_rnd() to arch_rnd().
Fixes 1b028f784e8c ("x86/mm: Introduce mmap_compat_base() for 32-bit mmap()")
Signed-off-by Oleg Nesterov <[email protected]>

It also fixes personality(ADDR_NO_RANDOMIZE).

Acked-by Kirill A. Shutemov <[email protected]>

Signature · 1 line

 Kirill A. Shutemov

Oleg Nesterov

author

2017-08-14 16:49 UTC #6 lore raw

Quoted 10 lines · Kirill A. Shutemov

Quoted 8 lines · Oleg Nesterov
Documentation/admin-guide/kernel-parameters.txt says:
norandmaps Don't use address space randomization. Equivalent
to echo 0 > /proc/sys/kernel/randomize_va_space
but it doesn't work because arch_rnd() which is used to randomize mm->mmap_base returns a random value unconditionally.

Quoted 6 lines

Quoted 4 lines
Shift the PF_RANDOMIZE check from arch_mmap_rnd() to arch_rnd().
Fixes 1b028f784e8c ("x86/mm: Introduce mmap_compat_base() for 32-bit mmap()")
Signed-off-by Oleg Nesterov <[email protected]>
It also fixes personality(ADDR_NO_RANDOMIZE).

Yes, good point.

Quoted 1 line

Acked-by Kirill A. Shutemov <[email protected]>

OK, thanks Kirill and Cyrill, I'll resend tomorrow with all acks I got.

Oleg.

Cyrill Gorcunov

Acked-by

2017-08-14 16:44 UTC #7 lore raw

Quoted 3 lines · Oleg Nesterov

Quoted 1 line
Acked-by Kirill A. Shutemov <[email protected]>
OK, thanks Kirill and Cyrill, I'll resend tomorrow with all acks I got.

I don't think you need my ack but if any

Acked-by Cyrill Gorcunov <[email protected]>