[PATCH v2 1/2] mm, kasan: don't call kasan_krealloc() from ksize().

[PATCH v2 1/2] mm, kasan: don't call kasan_krealloc() from ksize().

[Alexander Potapenko]

Instead of calling kasan_krealloc(), which replaces the memory allocation
stack ID (if stack depot is used), just unpoison the whole memory chunk.

Signed-off-by: Alexander Potapenko <glider@google.com>
---
v2: - splitted v1 into two patches
---
 mm/slab.c | 2 +-
 mm/slub.c | 5 +++--
 2 files changed, 4 insertions(+), 3 deletions(-)

diff --git a/mm/slab.c b/mm/slab.c
index 17e2848..de46319 100644
--- a/mm/slab.c
+++ b/mm/slab.c
@@ -4324,7 +4324,7 @@ size_t ksize(const void *objp)
 	/* We assume that ksize callers could use the whole allocated area,
 	 * so we need to unpoison this area.
 	 */
-	kasan_krealloc(objp, size, GFP_NOWAIT);
+	kasan_unpoison_shadow(objp, size);
 
 	return size;
 }
diff --git a/mm/slub.c b/mm/slub.c
index 4dbb109e..62194e2 100644
--- a/mm/slub.c
+++ b/mm/slub.c
@@ -3635,8 +3635,9 @@ size_t ksize(const void *object)
 {
 	size_t size = __ksize(object);
 	/* We assume that ksize callers could use whole allocated area,
-	   so we need unpoison this area. */
-	kasan_krealloc(object, size, GFP_NOWAIT);
+	 * so we need to unpoison this area.
+	 */
+	kasan_unpoison_shadow(object, size);
 	return size;
 }
 EXPORT_SYMBOL(ksize);
-- 
2.8.0.rc3.226.g39d4020

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>

[PATCH v2 2/2] mm, kasan: add a ksize() test

[Alexander Potapenko]

Add a test that makes sure ksize() unpoisons the whole chunk.

Signed-off-by: Alexander Potapenko <glider@google.com>
---
v2: - splitted v1 into two patches
---
 lib/test_kasan.c | 20 ++++++++++++++++++++
 1 file changed, 20 insertions(+)

diff --git a/lib/test_kasan.c b/lib/test_kasan.c
index 82169fb..48e5a0b 100644
--- a/lib/test_kasan.c
+++ b/lib/test_kasan.c
@@ -344,6 +344,25 @@ static noinline void __init kasan_stack_oob(void)
 	*(volatile char *)p;
 }
 
+static noinline void __init ksize_unpoisons_memory(void)
+{
+	char *ptr;
+	size_t size = 123, real_size = size;
+
+	pr_info("ksize() unpoisons the whole allocated chunk\n");
+	ptr = kmalloc(size, GFP_KERNEL);
+	if (!ptr) {
+		pr_err("Allocation failed\n");
+		return;
+	}
+	real_size = ksize(ptr);
+	/* This access doesn't trigger an error. */
+	ptr[size] = 'x';
+	/* This one does. */
+	ptr[real_size] = 'y';
+	kfree(ptr);
+}
+
 static int __init kmalloc_tests_init(void)
 {
 	kmalloc_oob_right();
@@ -367,6 +386,7 @@ static int __init kmalloc_tests_init(void)
 	kmem_cache_oob();
 	kasan_stack_oob();
 	kasan_global_oob();
+	ksize_unpoisons_memory();
 	return -EAGAIN;
 }
 
-- 
2.8.0.rc3.226.g39d4020

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>

Re: [PATCH v2 2/2] mm, kasan: add a ksize() test

[Andrey Ryabinin]

On 04/13/2016 02:20 PM, Alexander Potapenko wrote:
> Add a test that makes sure ksize() unpoisons the whole chunk.
> 
> Signed-off-by: Alexander Potapenko <glider@google.com>

Acked-by: Andrey Ryabinin <aryabinin@virtuozzo.com>

> ---
> v2: - splitted v1 into two patches
> ---
>  lib/test_kasan.c | 20 ++++++++++++++++++++
>  1 file changed, 20 insertions(+)
> 
> diff --git a/lib/test_kasan.c b/lib/test_kasan.c
> index 82169fb..48e5a0b 100644
> --- a/lib/test_kasan.c
> +++ b/lib/test_kasan.c
> @@ -344,6 +344,25 @@ static noinline void __init kasan_stack_oob(void)
>  	*(volatile char *)p;
>  }
>  
> +static noinline void __init ksize_unpoisons_memory(void)
> +{
> +	char *ptr;
> +	size_t size = 123, real_size = size;
> +
> +	pr_info("ksize() unpoisons the whole allocated chunk\n");
> +	ptr = kmalloc(size, GFP_KERNEL);
> +	if (!ptr) {
> +		pr_err("Allocation failed\n");
> +		return;
> +	}
> +	real_size = ksize(ptr);
> +	/* This access doesn't trigger an error. */
> +	ptr[size] = 'x';
> +	/* This one does. */
> +	ptr[real_size] = 'y';
> +	kfree(ptr);
> +}
> +
>  static int __init kmalloc_tests_init(void)
>  {
>  	kmalloc_oob_right();
> @@ -367,6 +386,7 @@ static int __init kmalloc_tests_init(void)
>  	kmem_cache_oob();
>  	kasan_stack_oob();
>  	kasan_global_oob();
> +	ksize_unpoisons_memory();
>  	return -EAGAIN;
>  }
>  
> 

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>

Re: [PATCH v2 1/2] mm, kasan: don't call kasan_krealloc() from ksize().

[Andrey Ryabinin]

On 04/13/2016 02:20 PM, Alexander Potapenko wrote:
> Instead of calling kasan_krealloc(), which replaces the memory allocation
> stack ID (if stack depot is used), just unpoison the whole memory chunk.
> 
> Signed-off-by: Alexander Potapenko <glider@google.com>

Acked-by: Andrey Ryabinin <aryabinin@virtuozzo.com>

> ---
> v2: - splitted v1 into two patches
> ---
>  mm/slab.c | 2 +-
>  mm/slub.c | 5 +++--
>  2 files changed, 4 insertions(+), 3 deletions(-)
> 
> diff --git a/mm/slab.c b/mm/slab.c
> index 17e2848..de46319 100644
> --- a/mm/slab.c
> +++ b/mm/slab.c
> @@ -4324,7 +4324,7 @@ size_t ksize(const void *objp)
>  	/* We assume that ksize callers could use the whole allocated area,
>  	 * so we need to unpoison this area.
>  	 */
> -	kasan_krealloc(objp, size, GFP_NOWAIT);
> +	kasan_unpoison_shadow(objp, size);
>  
>  	return size;
>  }
> diff --git a/mm/slub.c b/mm/slub.c
> index 4dbb109e..62194e2 100644
> --- a/mm/slub.c
> +++ b/mm/slub.c
> @@ -3635,8 +3635,9 @@ size_t ksize(const void *object)
>  {
>  	size_t size = __ksize(object);
>  	/* We assume that ksize callers could use whole allocated area,
> -	   so we need unpoison this area. */
> -	kasan_krealloc(object, size, GFP_NOWAIT);
> +	 * so we need to unpoison this area.
> +	 */
> +	kasan_unpoison_shadow(object, size);
>  	return size;
>  }
>  EXPORT_SYMBOL(ksize);
> 

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>

Re: [PATCH v2 1/2] mm, kasan: don't call kasan_krealloc() from ksize().

[Andrew Morton]

On Wed, 13 Apr 2016 13:20:09 +0200 Alexander Potapenko <glider@google.com> wrote:

> Instead of calling kasan_krealloc(), which replaces the memory allocation
> stack ID (if stack depot is used), just unpoison the whole memory chunk.

I don't understand why these two patches exist.  Bugfix?  Cleanup? 
Optimization?


I had to change kmalloc_tests_init() a bit due to
mm-kasan-initial-memory-quarantine-implementation.patch:

        kasan_stack_oob();
        kasan_global_oob();
 #ifdef CONFIG_SLAB
        kasan_quarantine_cache();
 #endif
+       ksize_unpoisons_memory();
        return -EAGAIN;
 }

Please check.

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>

Re: [PATCH v2 1/2] mm, kasan: don't call kasan_krealloc() from ksize().

[Alexander Potapenko]

On Fri, Apr 22, 2016 at 11:32 PM, Andrew Morton
<akpm@linux-foundation.org> wrote:
> On Wed, 13 Apr 2016 13:20:09 +0200 Alexander Potapenko <glider@google.com> wrote:
>
>> Instead of calling kasan_krealloc(), which replaces the memory allocation
>> stack ID (if stack depot is used), just unpoison the whole memory chunk.
>
> I don't understand why these two patches exist.  Bugfix?  Cleanup?
> Optimization?
It's incorrect to call kasan_krealloc() from ksize(), because the
former may touch the allocation metadata (it does so for the SLAB
allocator).
Yes, this is a bugfix.
>
> I had to change kmalloc_tests_init() a bit due to
> mm-kasan-initial-memory-quarantine-implementation.patch:
>
>         kasan_stack_oob();
>         kasan_global_oob();
>  #ifdef CONFIG_SLAB
>         kasan_quarantine_cache();
>  #endif
> +       ksize_unpoisons_memory();
>         return -EAGAIN;
>  }
>
> Please check.
Ack.



-- 
Alexander Potapenko
Software Engineer

Google Germany GmbH
Erika-Mann-Straße, 33
80636 München

Geschäftsführer: Matthew Scott Sucherman, Paul Terence Manicle
Registergericht und -nummer: Hamburg, HRB 86891
Sitz der Gesellschaft: Hamburg

--
To unsubscribe, send a message with 'unsubscribe linux-mm' in
the body to majordomo@kvack.org.  For more info on Linux MM,
see: http://www.linux-mm.org/ .
Don't email: <a href=mailto:"dont@kvack.org"> email@kvack.org </a>