[PATCHv2 0/3] KASAN: clean stale poison upon cold re-entry to kernel

Functions which the compiler has instrumented for ASAN place poison on
the stack shadow upon entry and remove this poison prior to returning.

In some cases (e.g. hotplug and idle), CPUs may exit the kernel a number
of levels deep in C code. If there are any instrumented functions on
this critical path, these will leave portions of the idle thread stack
shadow poisoned.

If a CPU returns to the kernel via a different path (e.g. a cold entry),
then depending on stack frame layout subsequent calls to instrumented
functions may use regions of the stack with stale poison, resulting in
(spurious) KASAN splats to the console.

Contemporary GCCs always add stack shadow poisoning when ASAN is
enabled, even when asked to not instrument a function [1], so we can't
simply annotate functions on the critical path to avoid poisoning.

Instead, this series explicitly removes any stale poison before it can
be hit. In the common hotplug case we clear the entire stack shadow in
common code, before a CPU is brought online.

On architectures which perform a cold return as part of cpu idle may
retain an architecture-specific amount of stack contents. To retain the
poison for this retained context, the arch code must call the core KASAN
code, passing a "watermark" stack pointer value beyond which shadow will
be cleared. Architectures which don't perform a cold return as part of
idle do not need any additional code.

This is a combination of previous approaches [2,3], attempting to keep
as much as possible generic.

Since v1 [4]:
* Clean from task_stack_page(task)
* Add acks from v1

Andrew, the conclusion [5] from v1 was that this should go via the mm tree.
Are you happy to pick this up? 

Ingo was happy for the sched patch to go via the arm64 tree, and I assume that
also holds for going via mm. Ingo, please shout if that's not the case!

Thanks,
Mark.

[1] https://gcc.gnu.org/bugzilla/show_bug.cgi?id=69863
[2] http://lists.infradead.org/pipermail/linux-arm-kernel/2016-February/409466.html
[3] http://lists.infradead.org/pipermail/linux-arm-kernel/2016-February/411850.html
[4] http://lists.infradead.org/pipermail/linux-arm-kernel/2016-March/413093.html
[5] http://lists.infradead.org/pipermail/linux-arm-kernel/2016-March/413475.html

Mark Rutland (3):
  kasan: add functions to clear stack poison
  sched/kasan: remove stale KASAN poison after hotplug
  arm64: kasan: clear stale stack poison

 arch/arm64/kernel/sleep.S |  4 ++++
 include/linux/kasan.h     |  6 +++++-
 kernel/sched/core.c       |  3 +++
 mm/kasan/kasan.c          | 20 ++++++++++++++++++++
 4 files changed, 32 insertions(+), 1 deletion(-)

-- 
1.9.1

Functions which the compiler has instrumented for ASAN place poison on
the stack shadow upon entry and remove this poison prior to returning.

In the case of cpuidle, CPUs exit the kernel a number of levels deep
in C code. Any instrumented functions on this critical path will leave
portions of the stack shadow poisoned.

If CPUs lose context and return to the kernel via a cold path, we
restore a prior context saved in __cpu_suspend_enter are forgotten, and
we never remove the poison they placed in the stack shadow area by
functions calls between this and the actual exit of the kernel.

Thus, (depending on stackframe layout) subsequent calls to instrumented
functions may hit this stale poison, resulting in (spurious) KASAN
splats to the console.

To avoid this, clear any stale poison from the idle thread for a CPU
prior to bringing a CPU online.

Signed-off-by: Mark Rutland <mark.rutland@arm.com>
Acked-by: Catalin Marinas <catalin.marinas@arm.com>
Reviewed-by: Andrey Ryabinin <redacted>
Reviewed-by: Lorenzo Pieralisi <redacted>
Cc: Alexander Potapenko <glider@google.com>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Catalin Marinas <catalin.marinas@arm.com>
Cc: Will Deacon <redacted>
---
 arch/arm64/kernel/sleep.S | 4 ++++
 1 file changed, 4 insertions(+)

diff --git a/arch/arm64/kernel/sleep.S b/arch/arm64/kernel/sleep.S
index e33fe33..fd10eb6 100644
--- a/arch/arm64/kernel/sleep.S
+++ b/arch/arm64/kernel/sleep.S

@@ -145,6 +145,10 @@ ENTRY(cpu_resume_mmu)
 ENDPROC(cpu_resume_mmu)
 	.popsection
 cpu_resume_after_mmu:
+#ifdef CONFIG_KASAN
+	mov	x0, sp
+	bl	kasan_unpoison_remaining_stack
+#endif
 	mov	x0, #0			// return zero on success
 	ldp	x19, x20, [sp, #16]
 	ldp	x21, x22, [sp, #32]

-- 
1.9.1

Functions which the compiler has instrumented for ASAN place poison on
the stack shadow upon entry and remove this poision prior to returning.

In the case of CPU hotplug, CPUs exit the kernel a number of levels deep
in C code. Any instrumented functions on this critical path will leave
portions of the stack shadow poisoned.

When a CPU is subsequently brought back into the kernel via a different
path, depending on stackframe, layout calls to instrumented functions
may hit this stale poison, resulting in (spurious) KASAN splats to the
console.

To avoid this, clear any stale poison from the idle thread for a CPU
prior to bringing a CPU online.

Signed-off-by: Mark Rutland <mark.rutland@arm.com>
Acked-by: Catalin Marinas <catalin.marinas@arm.com>
Reviewed-by: Andrey Ryabinin <redacted>
Reviewed-by: Ingo Molnar <mingo@kernel.org>
Cc: Alexander Potapenko <glider@google.com>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Lorenzo Pieralisi <redacted>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Will Deacon <redacted>
---
 kernel/sched/core.c | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/kernel/sched/core.c b/kernel/sched/core.c
index 9503d59..41f6b22 100644
--- a/kernel/sched/core.c
+++ b/kernel/sched/core.c

@@ -26,6 +26,7 @@
  *              Thomas Gleixner, Mike Kravetz
  */
 
+#include <linux/kasan.h>
 #include <linux/mm.h>
 #include <linux/module.h>
 #include <linux/nmi.h>

@@ -5096,6 +5097,8 @@ void init_idle(struct task_struct *idle, int cpu)
 	idle->state = TASK_RUNNING;
 	idle->se.exec_start = sched_clock();
 
+	kasan_unpoison_task_stack(idle);
+
 #ifdef CONFIG_SMP
 	/*
 	 * Its possible that init_idle() gets called multiple times on a task,

-- 
1.9.1

Functions which the compiler has instrumented for ASAN place poison on
the stack shadow upon entry and remove this poision prior to returning.

In some cases (e.g. hotplug and idle), CPUs may exit the kernel a number
of levels deep in C code. If there are any instrumented functions on
this critical path, these will leave portions of the stack shadow
poisoned.

If a CPU returns to the kernel via a different path (e.g. a cold entry),
then depending on stack frame layout subsequent calls to instrumented
functions may use regions of the stack with stale poison, resulting in
(spurious) KASAN splats to the console.

To avoid this, we must clear stale poison from the stack prior to
instrumented functions being called. This patch adds functions to the
KASAN core for removing poison from (portions of) a task's stack. These
will be used by subsequent patches to avoid problems with hotplug and
idle.

Signed-off-by: Mark Rutland <mark.rutland@arm.com>
Acked-by: Catalin Marinas <catalin.marinas@arm.com>
Reviewed-by: Andrey Ryabinin <redacted>
Cc: Alexander Potapenko <glider@google.com>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: Lorenzo Pieralisi <redacted>
Cc: Will Deacon <redacted>
---
 include/linux/kasan.h |  6 +++++-
 mm/kasan/kasan.c      | 20 ++++++++++++++++++++
 2 files changed, 25 insertions(+), 1 deletion(-)

diff --git a/include/linux/kasan.h b/include/linux/kasan.h
index 4b9f85c..0fdc798 100644
--- a/include/linux/kasan.h
+++ b/include/linux/kasan.h

@@ -1,6 +1,7 @@
 #ifndef _LINUX_KASAN_H
 #define _LINUX_KASAN_H
 
+#include <linux/sched.h>
 #include <linux/types.h>
 
 struct kmem_cache;

@@ -13,7 +14,6 @@ struct vm_struct;
 
 #include <asm/kasan.h>
 #include <asm/pgtable.h>
-#include <linux/sched.h>
 
 extern unsigned char kasan_zero_page[PAGE_SIZE];
 extern pte_t kasan_zero_pte[PTRS_PER_PTE];

@@ -43,6 +43,8 @@ static inline void kasan_disable_current(void)
 
 void kasan_unpoison_shadow(const void *address, size_t size);
 
+void kasan_unpoison_task_stack(struct task_struct *task);
+
 void kasan_alloc_pages(struct page *page, unsigned int order);
 void kasan_free_pages(struct page *page, unsigned int order);
 

@@ -66,6 +68,8 @@ void kasan_free_shadow(const struct vm_struct *vm);
 
 static inline void kasan_unpoison_shadow(const void *address, size_t size) {}
 
+static inline void kasan_unpoison_task_stack(struct task_struct *task) {}
+
 static inline void kasan_enable_current(void) {}
 static inline void kasan_disable_current(void) {}
 

diff --git a/mm/kasan/kasan.c b/mm/kasan/kasan.c
index bc0a8d8..1ad20ad 100644
--- a/mm/kasan/kasan.c
+++ b/mm/kasan/kasan.c

@@ -20,6 +20,7 @@
 #include <linux/init.h>
 #include <linux/kernel.h>
 #include <linux/kmemleak.h>
+#include <linux/linkage.h>
 #include <linux/memblock.h>
 #include <linux/memory.h>
 #include <linux/mm.h>

@@ -60,6 +61,25 @@ void kasan_unpoison_shadow(const void *address, size_t size)
 	}
 }
 
+static void __kasan_unpoison_stack(struct task_struct *task, void *sp)
+{
+	void *base = task_stack_page(task);
+	size_t size = sp - base;
+
+	kasan_unpoison_shadow(base, size);
+}
+
+/* Unpoison the entire stack for a task. */
+void kasan_unpoison_task_stack(struct task_struct *task)
+{
+	__kasan_unpoison_stack(task, task_stack_page(task) + THREAD_SIZE);
+}
+
+/* Unpoison the stack for the current task beyond a watermark sp value. */
+asmlinkage void kasan_unpoison_remaining_stack(void *sp)
+{
+	__kasan_unpoison_stack(current, sp);
+}
 
 /*
  * All functions below always inlined so compiler could

-- 
1.9.1

Please replace "ASAN" with "KASAN".

On Thu, Mar 3, 2016 at 5:54 PM, Mark Rutland [off-list ref] wrote:

Functions which the compiler has instrumented for ASAN place poison on
the stack shadow upon entry and remove this poison prior to returning.

In some cases (e.g. hotplug and idle), CPUs may exit the kernel a number
of levels deep in C code. If there are any instrumented functions on
this critical path, these will leave portions of the idle thread stack
shadow poisoned.

If a CPU returns to the kernel via a different path (e.g. a cold entry),
then depending on stack frame layout subsequent calls to instrumented
functions may use regions of the stack with stale poison, resulting in
(spurious) KASAN splats to the console.

Contemporary GCCs always add stack shadow poisoning when ASAN is
enabled, even when asked to not instrument a function [1], so we can't
simply annotate functions on the critical path to avoid poisoning.

Instead, this series explicitly removes any stale poison before it can
be hit. In the common hotplug case we clear the entire stack shadow in
common code, before a CPU is brought online.

On architectures which perform a cold return as part of cpu idle may
retain an architecture-specific amount of stack contents. To retain the
poison for this retained context, the arch code must call the core KASAN
code, passing a "watermark" stack pointer value beyond which shadow will
be cleared. Architectures which don't perform a cold return as part of
idle do not need any additional code.

This is a combination of previous approaches [2,3], attempting to keep
as much as possible generic.

Since v1 [4]:
* Clean from task_stack_page(task)
* Add acks from v1

Andrew, the conclusion [5] from v1 was that this should go via the mm tree.
Are you happy to pick this up?

Ingo was happy for the sched patch to go via the arm64 tree, and I assume that
also holds for going via mm. Ingo, please shout if that's not the case!

Thanks,
Mark.

[1] https://gcc.gnu.org/bugzilla/show_bug.cgi?id=69863
[2] http://lists.infradead.org/pipermail/linux-arm-kernel/2016-February/409466.html
[3] http://lists.infradead.org/pipermail/linux-arm-kernel/2016-February/411850.html
[4] http://lists.infradead.org/pipermail/linux-arm-kernel/2016-March/413093.html
[5] http://lists.infradead.org/pipermail/linux-arm-kernel/2016-March/413475.html

Mark Rutland (3):
  kasan: add functions to clear stack poison
  sched/kasan: remove stale KASAN poison after hotplug
  arm64: kasan: clear stale stack poison

 arch/arm64/kernel/sleep.S |  4 ++++
 include/linux/kasan.h     |  6 +++++-
 kernel/sched/core.c       |  3 +++
 mm/kasan/kasan.c          | 20 ++++++++++++++++++++
 4 files changed, 32 insertions(+), 1 deletion(-)

--
1.9.1

-- 
Alexander Potapenko
Software Engineer

Google Germany GmbH
Erika-Mann-Stra?e, 33
80636 M?nchen

Gesch?ftsf?hrer: Matthew Scott Sucherman, Paul Terence Manicle
Registergericht und -nummer: Hamburg, HRB 86891
Sitz der Gesellschaft: Hamburg
Diese E-Mail ist vertraulich. Wenn Sie nicht der richtige Adressat sind,
leiten Sie diese bitte nicht weiter, informieren Sie den
Absender und l?schen Sie die E-Mail und alle Anh?nge. Vielen Dank.
This e-mail is confidential. If you are not the right addressee please
do not forward it, please inform the sender, and please erase this
e-mail including any attachments. Thanks.

Hi,

On Thu, Mar 03, 2016 at 06:17:31PM +0100, Alexander Potapenko wrote:

Please replace "ASAN" with "KASAN".

On Thu, Mar 3, 2016 at 5:54 PM, Mark Rutland [off-list ref] wrote:

quoted

Functions which the compiler has instrumented for ASAN place poison on
the stack shadow upon entry and remove this poison prior to returning.

In some cases (e.g. hotplug and idle), CPUs may exit the kernel a number
of levels deep in C code. If there are any instrumented functions on
this critical path, these will leave portions of the idle thread stack
shadow poisoned.

If a CPU returns to the kernel via a different path (e.g. a cold entry),
then depending on stack frame layout subsequent calls to instrumented
functions may use regions of the stack with stale poison, resulting in
(spurious) KASAN splats to the console.

Contemporary GCCs always add stack shadow poisoning when ASAN is
enabled, even when asked to not instrument a function [1], so we can't
simply annotate functions on the critical path to avoid poisoning.

Instead, this series explicitly removes any stale poison before it can
be hit. In the common hotplug case we clear the entire stack shadow in
common code, before a CPU is brought online.

On architectures which perform a cold return as part of cpu idle may
retain an architecture-specific amount of stack contents. To retain the
poison for this retained context, the arch code must call the core KASAN
code, passing a "watermark" stack pointer value beyond which shadow will
be cleared. Architectures which don't perform a cold return as part of
idle do not need any additional code.

For the above, and the rest of the series, ASAN consistently refers to
the compiler AddressSanitizer feature, and KASAN consistently refers to
the Linux-specific infrastructure. A simple s/[^K]ASAN/KASAN/ would
arguably be wrong (e.g. when referring to GCC behaviour above).

If there is a this needs rework, then I'm happy to s/[^K]ASAN/ASan/ to
follow the usual ASan naming convention and avoid confusion. Otherwise,
spinning a v3 is simply churn.

Thanks,
Mark.

On Thu, Mar 3, 2016 at 6:40 PM, Mark Rutland [off-list ref] wrote:

Hi,

On Thu, Mar 03, 2016 at 06:17:31PM +0100, Alexander Potapenko wrote:

quoted

Please replace "ASAN" with "KASAN".

On Thu, Mar 3, 2016 at 5:54 PM, Mark Rutland [off-list ref] wrote:

quoted

Functions which the compiler has instrumented for ASAN place poison on
the stack shadow upon entry and remove this poison prior to returning.

In some cases (e.g. hotplug and idle), CPUs may exit the kernel a number
of levels deep in C code. If there are any instrumented functions on
this critical path, these will leave portions of the idle thread stack
shadow poisoned.

If a CPU returns to the kernel via a different path (e.g. a cold entry),
then depending on stack frame layout subsequent calls to instrumented
functions may use regions of the stack with stale poison, resulting in
(spurious) KASAN splats to the console.

Contemporary GCCs always add stack shadow poisoning when ASAN is
enabled, even when asked to not instrument a function [1], so we can't
simply annotate functions on the critical path to avoid poisoning.

Instead, this series explicitly removes any stale poison before it can
be hit. In the common hotplug case we clear the entire stack shadow in
common code, before a CPU is brought online.

On architectures which perform a cold return as part of cpu idle may
retain an architecture-specific amount of stack contents. To retain the
poison for this retained context, the arch code must call the core KASAN
code, passing a "watermark" stack pointer value beyond which shadow will
be cleared. Architectures which don't perform a cold return as part of
idle do not need any additional code.

For the above, and the rest of the series, ASAN consistently refers to
the compiler AddressSanitizer feature, and KASAN consistently refers to
the Linux-specific infrastructure. A simple s/[^K]ASAN/KASAN/ would
arguably be wrong (e.g. when referring to GCC behaviour above).

I don't think there's been any convention about the compiler feature
name, we usually talked about ASan as a userspace tool and KASAN as a
kernel-space one, although they share the compiler part.

If there is a this needs rework, then I'm happy to s/[^K]ASAN/ASan/ to
follow the usual ASan naming convention and avoid confusion. Otherwise,
spinning a v3 is simply churn.

I don't insist on changing this, I should've chimed in before.
Feel free to retain the above patch description.

Thanks,
Mark.

-- 
Alexander Potapenko
Software Engineer

Google Germany GmbH
Erika-Mann-Stra?e, 33
80636 M?nchen

Gesch?ftsf?hrer: Matthew Scott Sucherman, Paul Terence Manicle
Registergericht und -nummer: Hamburg, HRB 86891
Sitz der Gesellschaft: Hamburg
Diese E-Mail ist vertraulich. Wenn Sie nicht der richtige Adressat sind,
leiten Sie diese bitte nicht weiter, informieren Sie den
Absender und l?schen Sie die E-Mail und alle Anh?nge. Vielen Dank.
This e-mail is confidential. If you are not the right addressee please
do not forward it, please inform the sender, and please erase this
e-mail including any attachments. Thanks.

On Thu, Mar 03, 2016 at 06:45:55PM +0100, Alexander Potapenko wrote:

On Thu, Mar 3, 2016 at 6:40 PM, Mark Rutland [off-list ref] wrote:

quoted

Hi,

On Thu, Mar 03, 2016 at 06:17:31PM +0100, Alexander Potapenko wrote:

quoted

Please replace "ASAN" with "KASAN".

On Thu, Mar 3, 2016 at 5:54 PM, Mark Rutland [off-list ref] wrote:

quoted

Functions which the compiler has instrumented for ASAN place poison on
the stack shadow upon entry and remove this poison prior to returning.

[...]

quoted

For the above, and the rest of the series, ASAN consistently refers to
the compiler AddressSanitizer feature, and KASAN consistently refers to
the Linux-specific infrastructure. A simple s/[^K]ASAN/KASAN/ would
arguably be wrong (e.g. when referring to GCC behaviour above).

I don't think there's been any convention about the compiler feature
name, we usually talked about ASan as a userspace tool and KASAN as a
kernel-space one, although they share the compiler part.

Ah, ok.

In future I'll speak in terms of "AddressSanitizer instrumentation" or
something like that, as that's fairly unambigious.

quoted

If there is a this needs rework, then I'm happy to s/[^K]ASAN/ASan/ to
follow the usual ASan naming convention and avoid confusion. Otherwise,
spinning a v3 is simply churn.

I don't insist on changing this, I should've chimed in before.
Feel free to retain the above patch description.

No worries, thanks for the info.

Thanks,
Mark.

On Thu,  3 Mar 2016 16:54:25 +0000 Mark Rutland [off-list ref] wrote:

Andrew, the conclusion [5] from v1 was that this should go via the mm tree.
Are you happy to pick this up? 

yep.  I'll aim to get this into 4.5.