* Re: [PATCH v3] mm/slub: Fix incorrect interpretation of s->offset
2020-04-29 13:53 [PATCH v3] mm/slub: Fix incorrect interpretation of s->offset Waiman Long
@ 2020-04-29 13:55 ` Waiman Long
2020-04-29 15:54 ` Matthew Wilcox
` (3 subsequent siblings)
4 siblings, 0 replies; 7+ messages in thread
From: Waiman Long @ 2020-04-29 13:55 UTC (permalink / raw)
To: Christoph Lameter, Pekka Enberg, David Rientjes, Joonsoo Kim,
Andrew Morton, Kees Cook
Cc: linux-mm, linux-kernel, Changbin Du, Matthew Wilcox, Markus Elfring
On 4/29/20 9:53 AM, Waiman Long wrote:
> In a couple of places in the slub memory allocator, the code uses
> "s->offset" as a check to see if the free pointer is put right after the
> object. That check is no longer true with commit 3202fa62fb43 ("slub:
> relocate freelist pointer to middle of object").
>
> As a result, echoing "1" into the validate sysfs file, e.g. of dentry,
> may cause a bunch of "Freepointer corrupt" error reports like the
> following to appear with the system in panic afterwards.
>
> [ 38.579769] =============================================================================
> [ 38.580845] BUG dentry(666:pmcd.service) (Tainted: G B): Freepointer corrupt
> [ 38.581948] -----------------------------------------------------------------------------
>
> To fix it, use the check "s->offset == s->inuse" in the new helper
Sorry, forgot to change the commit log to ">=". Anyway, this is a
serious bug that needs to be fixed before v5.7 is released.
Cheers,
Longman
^ permalink raw reply [flat|nested] 7+ messages in thread* Re: [PATCH v3] mm/slub: Fix incorrect interpretation of s->offset
2020-04-29 13:53 [PATCH v3] mm/slub: Fix incorrect interpretation of s->offset Waiman Long
2020-04-29 13:55 ` Waiman Long
@ 2020-04-29 15:54 ` Matthew Wilcox
2020-04-29 16:42 ` Markus Elfring
` (2 subsequent siblings)
4 siblings, 0 replies; 7+ messages in thread
From: Matthew Wilcox @ 2020-04-29 15:54 UTC (permalink / raw)
To: Waiman Long
Cc: Christoph Lameter, Pekka Enberg, David Rientjes, Joonsoo Kim,
Andrew Morton, Kees Cook, linux-mm, linux-kernel, Changbin Du,
Markus Elfring
On Wed, Apr 29, 2020 at 09:53:28AM -0400, Waiman Long wrote:
> As a result, echoing "1" into the validate sysfs file, e.g. of dentry,
> may cause a bunch of "Freepointer corrupt" error reports like the
> following to appear with the system in panic afterwards.
>
> [ 38.579769] =============================================================================
> [ 38.580845] BUG dentry(666:pmcd.service) (Tainted: G B): Freepointer corrupt
> [ 38.581948] -----------------------------------------------------------------------------
I might trim the timestamp and the === and --- from the commit message ...
> To fix it, use the check "s->offset == s->inuse" in the new helper
> function freeptr_outside_object() instead. Also add another helper function
> get_info_end() to return the end of info block (inuse + free pointer
> if not overlapping with object).
>
> Fixes: 3202fa62fb43 ("slub: relocate freelist pointer to middle of object")
> Signed-off-by: Waiman Long <longman@redhat.com>
But the patch looks great.
Reviewed-by: Matthew Wilcox (Oracle) <willy@infradead.org>
^ permalink raw reply [flat|nested] 7+ messages in thread* Re: [PATCH v3] mm/slub: Fix incorrect interpretation of s->offset
2020-04-29 13:53 [PATCH v3] mm/slub: Fix incorrect interpretation of s->offset Waiman Long
2020-04-29 13:55 ` Waiman Long
2020-04-29 15:54 ` Matthew Wilcox
@ 2020-04-29 16:42 ` Markus Elfring
2020-04-29 18:25 ` Matthew Wilcox
2020-05-01 21:29 ` Rafael Aquini
2020-05-01 22:01 ` Kees Cook
4 siblings, 1 reply; 7+ messages in thread
From: Markus Elfring @ 2020-04-29 16:42 UTC (permalink / raw)
To: Waiman Long, Christoph Lameter, Pekka Enberg, David Rientjes,
Joonsoo Kim, Andrew Morton, Kees Cook, Changbin Du,
Matthew Wilcox, linux-mm
Cc: linux-kernel
> In a couple of places in the slub memory allocator, the code uses
> "s->offset" as a check to see if the free pointer is put right after the
> object. That check is no longer true with commit 3202fa62fb43 ("slub:
> relocate freelist pointer to middle of object").
Will any further collateral evolution become interesting?
…
> +++ b/mm/slub.c
> @@ -551,15 +551,32 @@ static void print_section(char *level, char *text, u8 *addr,
…
> +static inline unsigned int get_info_end(struct kmem_cache *s)
> +{
> + if (freeptr_outside_object(s))
> + return s->inuse + sizeof(void *);
> + else
> + return s->inuse;
> +}
How do you think about the following source code variants?
+ return freeptr_outside_object(s)
+ ? s->inuse + sizeof(void *)
+ : s->inuse;
> static struct track *get_track(struct kmem_cache *s, void *object,
> enum track_item alloc)
> {
> struct track *p;
>
> - if (s->offset)
> - p = object + s->offset + sizeof(void *);
> - else
> - p = object + s->inuse;
> + p = object + get_info_end(s);
>
> return p + alloc;
> }
+ struct track *p = object + get_info_end(s);
return p + alloc;
Regards,
Markus
^ permalink raw reply [flat|nested] 7+ messages in thread* Re: [PATCH v3] mm/slub: Fix incorrect interpretation of s->offset
2020-04-29 16:42 ` Markus Elfring
@ 2020-04-29 18:25 ` Matthew Wilcox
0 siblings, 0 replies; 7+ messages in thread
From: Matthew Wilcox @ 2020-04-29 18:25 UTC (permalink / raw)
To: Markus Elfring
Cc: Waiman Long, Christoph Lameter, Pekka Enberg, David Rientjes,
Joonsoo Kim, Andrew Morton, Kees Cook, Changbin Du, linux-mm,
linux-kernel
On Wed, Apr 29, 2020 at 06:42:55PM +0200, Markus Elfring wrote:
> > In a couple of places in the slub memory allocator, the code uses
> > "s->offset" as a check to see if the free pointer is put right after the
> > object. That check is no longer true with commit 3202fa62fb43 ("slub:
> > relocate freelist pointer to middle of object").
>
> Will any further collateral evolution become interesting?
What do you mean by this question?
> > +static inline unsigned int get_info_end(struct kmem_cache *s)
> > +{
> > + if (freeptr_outside_object(s))
> > + return s->inuse + sizeof(void *);
> > + else
> > + return s->inuse;
> > +}
>
> How do you think about the following source code variants?
>
> + return freeptr_outside_object(s)
> + ? s->inuse + sizeof(void *)
> + : s->inuse;
That is less clear than the version Wayman posted.
> > static struct track *get_track(struct kmem_cache *s, void *object,
> > enum track_item alloc)
> > {
> > struct track *p;
> >
> > - if (s->offset)
> > - p = object + s->offset + sizeof(void *);
> > - else
> > - p = object + s->inuse;
> > + p = object + get_info_end(s);
> >
> > return p + alloc;
> > }
>
> + struct track *p = object + get_info_end(s);
>
> return p + alloc;
Yes, I think that's an improvement.
^ permalink raw reply [flat|nested] 7+ messages in thread
* Re: [PATCH v3] mm/slub: Fix incorrect interpretation of s->offset
2020-04-29 13:53 [PATCH v3] mm/slub: Fix incorrect interpretation of s->offset Waiman Long
` (2 preceding siblings ...)
2020-04-29 16:42 ` Markus Elfring
@ 2020-05-01 21:29 ` Rafael Aquini
2020-05-01 22:01 ` Kees Cook
4 siblings, 0 replies; 7+ messages in thread
From: Rafael Aquini @ 2020-05-01 21:29 UTC (permalink / raw)
To: Waiman Long
Cc: Christoph Lameter, Pekka Enberg, David Rientjes, Joonsoo Kim,
Andrew Morton, Kees Cook, linux-mm, linux-kernel, Changbin Du,
Matthew Wilcox, Markus Elfring
On Wed, Apr 29, 2020 at 09:53:28AM -0400, Waiman Long wrote:
> In a couple of places in the slub memory allocator, the code uses
> "s->offset" as a check to see if the free pointer is put right after the
> object. That check is no longer true with commit 3202fa62fb43 ("slub:
> relocate freelist pointer to middle of object").
>
> As a result, echoing "1" into the validate sysfs file, e.g. of dentry,
> may cause a bunch of "Freepointer corrupt" error reports like the
> following to appear with the system in panic afterwards.
>
> [ 38.579769] =============================================================================
> [ 38.580845] BUG dentry(666:pmcd.service) (Tainted: G B): Freepointer corrupt
> [ 38.581948] -----------------------------------------------------------------------------
>
> To fix it, use the check "s->offset == s->inuse" in the new helper
> function freeptr_outside_object() instead. Also add another helper function
> get_info_end() to return the end of info block (inuse + free pointer
> if not overlapping with object).
>
> Fixes: 3202fa62fb43 ("slub: relocate freelist pointer to middle of object")
> Signed-off-by: Waiman Long <longman@redhat.com>
> ---
> mm/slub.c | 45 ++++++++++++++++++++++++++++++---------------
> 1 file changed, 30 insertions(+), 15 deletions(-)
>
> [v3: Change function name to freeptr_outside_object(), update check & add comment]
>
> diff --git a/mm/slub.c b/mm/slub.c
> index 9bf44955c4f1..b762450fc9f0 100644
> --- a/mm/slub.c
> +++ b/mm/slub.c
> @@ -551,15 +551,32 @@ static void print_section(char *level, char *text, u8 *addr,
> metadata_access_disable();
> }
>
> +/*
> + * See comment in calculate_sizes().
> + */
> +static inline bool freeptr_outside_object(struct kmem_cache *s)
> +{
> + return s->offset >= s->inuse;
> +}
> +
> +/*
> + * Return offset of the end of info block which is inuse + free pointer if
> + * not overlapping with object.
> + */
> +static inline unsigned int get_info_end(struct kmem_cache *s)
> +{
> + if (freeptr_outside_object(s))
> + return s->inuse + sizeof(void *);
> + else
> + return s->inuse;
> +}
> +
> static struct track *get_track(struct kmem_cache *s, void *object,
> enum track_item alloc)
> {
> struct track *p;
>
> - if (s->offset)
> - p = object + s->offset + sizeof(void *);
> - else
> - p = object + s->inuse;
> + p = object + get_info_end(s);
>
> return p + alloc;
> }
> @@ -686,10 +703,7 @@ static void print_trailer(struct kmem_cache *s, struct page *page, u8 *p)
> print_section(KERN_ERR, "Redzone ", p + s->object_size,
> s->inuse - s->object_size);
>
> - if (s->offset)
> - off = s->offset + sizeof(void *);
> - else
> - off = s->inuse;
> + off = get_info_end(s);
>
> if (s->flags & SLAB_STORE_USER)
> off += 2 * sizeof(struct track);
> @@ -782,7 +796,7 @@ static int check_bytes_and_report(struct kmem_cache *s, struct page *page,
> * object address
> * Bytes of the object to be managed.
> * If the freepointer may overlay the object then the free
> - * pointer is the first word of the object.
> + * pointer is at the middle of the object.
> *
> * Poisoning uses 0x6b (POISON_FREE) and the last byte is
> * 0xa5 (POISON_END)
> @@ -816,11 +830,7 @@ static int check_bytes_and_report(struct kmem_cache *s, struct page *page,
>
> static int check_pad_bytes(struct kmem_cache *s, struct page *page, u8 *p)
> {
> - unsigned long off = s->inuse; /* The end of info */
> -
> - if (s->offset)
> - /* Freepointer is placed after the object. */
> - off += sizeof(void *);
> + unsigned long off = get_info_end(s); /* The end of info */
>
> if (s->flags & SLAB_STORE_USER)
> /* We also have user information there */
> @@ -907,7 +917,7 @@ static int check_object(struct kmem_cache *s, struct page *page,
> check_pad_bytes(s, page, p);
> }
>
> - if (!s->offset && val == SLUB_RED_ACTIVE)
> + if (!freeptr_outside_object(s) && val == SLUB_RED_ACTIVE)
> /*
> * Object and freepointer overlap. Cannot check
> * freepointer while object is allocated.
> @@ -3587,6 +3597,11 @@ static int calculate_sizes(struct kmem_cache *s, int forced_order)
> *
> * This is the case if we do RCU, have a constructor or
> * destructor or are poisoning the objects.
> + *
> + * The assumption that s->offset >= s->inuse means free
> + * pointer is outside of the object is used in the
> + * freeptr_outside_object() function. If that is no
> + * longer true, the function needs to be modified.
> */
> s->offset = size;
> size += sizeof(void *);
> --
> 2.18.1
>
>
Acked-by: Rafael Aquini <aquini@redhat.com>
^ permalink raw reply [flat|nested] 7+ messages in thread* Re: [PATCH v3] mm/slub: Fix incorrect interpretation of s->offset
2020-04-29 13:53 [PATCH v3] mm/slub: Fix incorrect interpretation of s->offset Waiman Long
` (3 preceding siblings ...)
2020-05-01 21:29 ` Rafael Aquini
@ 2020-05-01 22:01 ` Kees Cook
4 siblings, 0 replies; 7+ messages in thread
From: Kees Cook @ 2020-05-01 22:01 UTC (permalink / raw)
To: Waiman Long
Cc: Christoph Lameter, Pekka Enberg, David Rientjes, Joonsoo Kim,
Andrew Morton, linux-mm, linux-kernel, Changbin Du,
Matthew Wilcox, Markus Elfring
On Wed, Apr 29, 2020 at 09:53:28AM -0400, Waiman Long wrote:
> In a couple of places in the slub memory allocator, the code uses
> "s->offset" as a check to see if the free pointer is put right after the
> object. That check is no longer true with commit 3202fa62fb43 ("slub:
> relocate freelist pointer to middle of object").
>
> As a result, echoing "1" into the validate sysfs file, e.g. of dentry,
> may cause a bunch of "Freepointer corrupt" error reports like the
> following to appear with the system in panic afterwards.
>
> [ 38.579769] =============================================================================
> [ 38.580845] BUG dentry(666:pmcd.service) (Tainted: G B): Freepointer corrupt
> [ 38.581948] -----------------------------------------------------------------------------
>
> To fix it, use the check "s->offset == s->inuse" in the new helper
> function freeptr_outside_object() instead. Also add another helper function
> get_info_end() to return the end of info block (inuse + free pointer
> if not overlapping with object).
>
> Fixes: 3202fa62fb43 ("slub: relocate freelist pointer to middle of object")
> Signed-off-by: Waiman Long <longman@redhat.com>
Thanks! Nice catch. :)
Reviewed-by: Kees Cook <keescook@chromium.org>
-Kees
> ---
> mm/slub.c | 45 ++++++++++++++++++++++++++++++---------------
> 1 file changed, 30 insertions(+), 15 deletions(-)
>
> [v3: Change function name to freeptr_outside_object(), update check & add comment]
>
> diff --git a/mm/slub.c b/mm/slub.c
> index 9bf44955c4f1..b762450fc9f0 100644
> --- a/mm/slub.c
> +++ b/mm/slub.c
> @@ -551,15 +551,32 @@ static void print_section(char *level, char *text, u8 *addr,
> metadata_access_disable();
> }
>
> +/*
> + * See comment in calculate_sizes().
> + */
> +static inline bool freeptr_outside_object(struct kmem_cache *s)
> +{
> + return s->offset >= s->inuse;
> +}
> +
> +/*
> + * Return offset of the end of info block which is inuse + free pointer if
> + * not overlapping with object.
> + */
> +static inline unsigned int get_info_end(struct kmem_cache *s)
> +{
> + if (freeptr_outside_object(s))
> + return s->inuse + sizeof(void *);
> + else
> + return s->inuse;
> +}
> +
> static struct track *get_track(struct kmem_cache *s, void *object,
> enum track_item alloc)
> {
> struct track *p;
>
> - if (s->offset)
> - p = object + s->offset + sizeof(void *);
> - else
> - p = object + s->inuse;
> + p = object + get_info_end(s);
>
> return p + alloc;
> }
> @@ -686,10 +703,7 @@ static void print_trailer(struct kmem_cache *s, struct page *page, u8 *p)
> print_section(KERN_ERR, "Redzone ", p + s->object_size,
> s->inuse - s->object_size);
>
> - if (s->offset)
> - off = s->offset + sizeof(void *);
> - else
> - off = s->inuse;
> + off = get_info_end(s);
>
> if (s->flags & SLAB_STORE_USER)
> off += 2 * sizeof(struct track);
> @@ -782,7 +796,7 @@ static int check_bytes_and_report(struct kmem_cache *s, struct page *page,
> * object address
> * Bytes of the object to be managed.
> * If the freepointer may overlay the object then the free
> - * pointer is the first word of the object.
> + * pointer is at the middle of the object.
> *
> * Poisoning uses 0x6b (POISON_FREE) and the last byte is
> * 0xa5 (POISON_END)
> @@ -816,11 +830,7 @@ static int check_bytes_and_report(struct kmem_cache *s, struct page *page,
>
> static int check_pad_bytes(struct kmem_cache *s, struct page *page, u8 *p)
> {
> - unsigned long off = s->inuse; /* The end of info */
> -
> - if (s->offset)
> - /* Freepointer is placed after the object. */
> - off += sizeof(void *);
> + unsigned long off = get_info_end(s); /* The end of info */
>
> if (s->flags & SLAB_STORE_USER)
> /* We also have user information there */
> @@ -907,7 +917,7 @@ static int check_object(struct kmem_cache *s, struct page *page,
> check_pad_bytes(s, page, p);
> }
>
> - if (!s->offset && val == SLUB_RED_ACTIVE)
> + if (!freeptr_outside_object(s) && val == SLUB_RED_ACTIVE)
> /*
> * Object and freepointer overlap. Cannot check
> * freepointer while object is allocated.
> @@ -3587,6 +3597,11 @@ static int calculate_sizes(struct kmem_cache *s, int forced_order)
> *
> * This is the case if we do RCU, have a constructor or
> * destructor or are poisoning the objects.
> + *
> + * The assumption that s->offset >= s->inuse means free
> + * pointer is outside of the object is used in the
> + * freeptr_outside_object() function. If that is no
> + * longer true, the function needs to be modified.
> */
> s->offset = size;
> size += sizeof(void *);
> --
> 2.18.1
>
--
Kees Cook
^ permalink raw reply [flat|nested] 7+ messages in thread