DPDK patches and discussions
 help / color / mirror / Atom feed
* [dpdk-dev] [PATCH] mem: fix the alloc size roundup overflow
@ 2020-04-07 10:46 Bing Zhao
  2020-04-20 11:39 ` Burakov, Anatoly
                   ` (2 more replies)
  0 siblings, 3 replies; 12+ messages in thread
From: Bing Zhao @ 2020-04-07 10:46 UTC (permalink / raw)
  To: anatoly.burakov; +Cc: dev, sergio.gonzalez.monroy, stable

The size checking is done in the caller. The size parameter is an
unsigned (64b wide) right now, so the comparison with zero should be
enough in most cases. But it won't help in the following case.
If the allocating request input a huge number by mistake, e.g., some
overflow after the calculation (especially subtraction), the checking
in the caller will succeed since it is not zero. Indeed, there is not
enough space in the system to support such huge memory allocation.
Usually it will return failure in the following code. But if the
input size is just a little smaller than the UINT64_MAX, like -2 in
signed type.
The roundup will cause an overflow and then "reset" the size to 0,
and then only a header (128B now) with zero length will be returned.
The following will be the previous allocation header.
It should be OK in most cases if the application won't access the
memory body. Or else, some critical issue will be caused and not easy
to debug. So this issue should be prevented at the beginning, like
other big size failure, NULL pointer should be returned also.

Fixes: fdf20fa7bee9 ("add prefix to cache line macros")
Cc: sergio.gonzalez.monroy@intel.com
Cc: stable@dpdk.org

Signed-off-by: Bing Zhao <bingz@mellanox.com>
---
 lib/librte_eal/common/malloc_heap.c | 3 +++
 1 file changed, 3 insertions(+)

diff --git a/lib/librte_eal/common/malloc_heap.c b/lib/librte_eal/common/malloc_heap.c
index 842eb9d..bd50656 100644
--- a/lib/librte_eal/common/malloc_heap.c
+++ b/lib/librte_eal/common/malloc_heap.c
@@ -241,6 +241,9 @@
 	size = RTE_CACHE_LINE_ROUNDUP(size);
 	align = RTE_CACHE_LINE_ROUNDUP(align);
 
+	/* roundup might cause an overflow */
+	if (size == 0)
+		return NULL;
 	elem = find_suitable_element(heap, size, flags, align, bound, contig);
 	if (elem != NULL) {
 		elem = malloc_elem_alloc(elem, size, align, bound, contig);
-- 
1.8.3.1


^ permalink raw reply	[flat|nested] 12+ messages in thread

* Re: [dpdk-dev] [PATCH] mem: fix the alloc size roundup overflow
  2020-04-07 10:46 [dpdk-dev] [PATCH] mem: fix the alloc size roundup overflow Bing Zhao
@ 2020-04-20 11:39 ` Burakov, Anatoly
  2020-04-21  3:18   ` Bing Zhao
  2020-05-07  7:39 ` [dpdk-dev] [PATCH v2] " Bing Zhao
  2020-05-07  7:41 ` Bing Zhao
  2 siblings, 1 reply; 12+ messages in thread
From: Burakov, Anatoly @ 2020-04-20 11:39 UTC (permalink / raw)
  To: Bing Zhao; +Cc: dev, sergio.gonzalez.monroy, stable

On 07-Apr-20 11:46 AM, Bing Zhao wrote:
> The size checking is done in the caller. The size parameter is an
> unsigned (64b wide) right now, so the comparison with zero should be
> enough in most cases. But it won't help in the following case.
> If the allocating request input a huge number by mistake, e.g., some
> overflow after the calculation (especially subtraction), the checking
> in the caller will succeed since it is not zero. Indeed, there is not
> enough space in the system to support such huge memory allocation.
> Usually it will return failure in the following code. But if the
> input size is just a little smaller than the UINT64_MAX, like -2 in
> signed type.
> The roundup will cause an overflow and then "reset" the size to 0,
> and then only a header (128B now) with zero length will be returned.
> The following will be the previous allocation header.
> It should be OK in most cases if the application won't access the
> memory body. Or else, some critical issue will be caused and not easy
> to debug. So this issue should be prevented at the beginning, like
> other big size failure, NULL pointer should be returned also.
> 
> Fixes: fdf20fa7bee9 ("add prefix to cache line macros")
> Cc: sergio.gonzalez.monroy@intel.com
> Cc: stable@dpdk.org
> 
> Signed-off-by: Bing Zhao <bingz@mellanox.com>
> ---
>   lib/librte_eal/common/malloc_heap.c | 3 +++
>   1 file changed, 3 insertions(+)
> 
> diff --git a/lib/librte_eal/common/malloc_heap.c b/lib/librte_eal/common/malloc_heap.c
> index 842eb9d..bd50656 100644
> --- a/lib/librte_eal/common/malloc_heap.c
> +++ b/lib/librte_eal/common/malloc_heap.c
> @@ -241,6 +241,9 @@
>   	size = RTE_CACHE_LINE_ROUNDUP(size);
>   	align = RTE_CACHE_LINE_ROUNDUP(align);
>   
> +	/* roundup might cause an overflow */
> +	if (size == 0)
> +		return NULL;
>   	elem = find_suitable_element(heap, size, flags, align, bound, contig);
>   	if (elem != NULL) {
>   		elem = malloc_elem_alloc(elem, size, align, bound, contig);
> 

Can we add a unit test for this in malloc_autotest?

Otherwise,

Acked-by: Anatoly Burakov <anatoly.burakov@intel.com>

-- 
Thanks,
Anatoly

^ permalink raw reply	[flat|nested] 12+ messages in thread

* Re: [dpdk-dev] [PATCH] mem: fix the alloc size roundup overflow
  2020-04-20 11:39 ` Burakov, Anatoly
@ 2020-04-21  3:18   ` Bing Zhao
  2020-05-06 21:43     ` [dpdk-dev] [dpdk-stable] " Thomas Monjalon
  0 siblings, 1 reply; 12+ messages in thread
From: Bing Zhao @ 2020-04-21  3:18 UTC (permalink / raw)
  To: Burakov, Anatoly; +Cc: dev, sergio.gonzalez.monroy, stable

> -----Original Message-----
> From: Burakov, Anatoly <anatoly.burakov@intel.com>
> Sent: Monday, April 20, 2020 7:40 PM
> To: Bing Zhao <bingz@mellanox.com>
> Cc: dev@dpdk.org; sergio.gonzalez.monroy@intel.com;
> stable@dpdk.org
> Subject: Re: [PATCH] mem: fix the alloc size roundup overflow
> 
> On 07-Apr-20 11:46 AM, Bing Zhao wrote:
> > The size checking is done in the caller. The size parameter is an
> > unsigned (64b wide) right now, so the comparison with zero should
> be
> > enough in most cases. But it won't help in the following case.
> > If the allocating request input a huge number by mistake, e.g., some
> > overflow after the calculation (especially subtraction), the checking
> > in the caller will succeed since it is not zero. Indeed, there is not
> > enough space in the system to support such huge memory allocation.
> > Usually it will return failure in the following code. But if the input
> > size is just a little smaller than the UINT64_MAX, like -2 in signed
> > type.
> > The roundup will cause an overflow and then "reset" the size to 0,
> and
> > then only a header (128B now) with zero length will be returned.
> > The following will be the previous allocation header.
> > It should be OK in most cases if the application won't access the
> > memory body. Or else, some critical issue will be caused and not easy
> > to debug. So this issue should be prevented at the beginning, like
> > other big size failure, NULL pointer should be returned also.
> >
> > Fixes: fdf20fa7bee9 ("add prefix to cache line macros")
> > Cc: sergio.gonzalez.monroy@intel.com
> > Cc: stable@dpdk.org
> >
> > Signed-off-by: Bing Zhao <bingz@mellanox.com>
> > ---
> >   lib/librte_eal/common/malloc_heap.c | 3 +++
> >   1 file changed, 3 insertions(+)
> >
> > diff --git a/lib/librte_eal/common/malloc_heap.c
> > b/lib/librte_eal/common/malloc_heap.c
> > index 842eb9d..bd50656 100644
> > --- a/lib/librte_eal/common/malloc_heap.c
> > +++ b/lib/librte_eal/common/malloc_heap.c
> > @@ -241,6 +241,9 @@
> >   	size = RTE_CACHE_LINE_ROUNDUP(size);
> >   	align = RTE_CACHE_LINE_ROUNDUP(align);
> >
> > +	/* roundup might cause an overflow */
> > +	if (size == 0)
> > +		return NULL;
> >   	elem = find_suitable_element(heap, size, flags, align, bound,
> contig);
> >   	if (elem != NULL) {
> >   		elem = malloc_elem_alloc(elem, size, align, bound,
> contig);
> >
> 
> Can we add a unit test for this in malloc_autotest?
> 
> Otherwise,
> 
> Acked-by: Anatoly Burakov <anatoly.burakov@intel.com>
> 

Thanks, Burakov. I can try to add one case in the unit for this.

> --
> Thanks,
> Anatoly

^ permalink raw reply	[flat|nested] 12+ messages in thread

* Re: [dpdk-dev] [dpdk-stable] [PATCH] mem: fix the alloc size roundup overflow
  2020-04-21  3:18   ` Bing Zhao
@ 2020-05-06 21:43     ` Thomas Monjalon
  0 siblings, 0 replies; 12+ messages in thread
From: Thomas Monjalon @ 2020-05-06 21:43 UTC (permalink / raw)
  To: Bing Zhao; +Cc: Burakov, Anatoly, dev, david.marchand

21/04/2020 05:18, Bing Zhao:
> From: Burakov, Anatoly <anatoly.burakov@intel.com>
> > Can we add a unit test for this in malloc_autotest?
> > 
> > Otherwise,
> > 
> > Acked-by: Anatoly Burakov <anatoly.burakov@intel.com>
> 
> Thanks, Burakov. I can try to add one case in the unit for this.

Any update about writing an unit test?



^ permalink raw reply	[flat|nested] 12+ messages in thread

* [dpdk-dev] [PATCH v2] mem: fix the alloc size roundup overflow
  2020-04-07 10:46 [dpdk-dev] [PATCH] mem: fix the alloc size roundup overflow Bing Zhao
  2020-04-20 11:39 ` Burakov, Anatoly
@ 2020-05-07  7:39 ` Bing Zhao
  2020-05-07  7:41 ` Bing Zhao
  2 siblings, 0 replies; 12+ messages in thread
From: Bing Zhao @ 2020-05-07  7:39 UTC (permalink / raw)
  To: anatoly.burakov, thomas; +Cc: dev, stable, sergio.gonzalez.monroy

The size checking is done in the caller. The size parameter is an
unsigned (64b wide) right now, so the comparison with zero should be
enough in most cases. But it won't help in the following case.
If the allocating request input a huge number by mistake, e.g., some
overflow after the calculation (especially subtraction), the checking
in the caller will succeed since it is not zero. Indeed, there is not
enough space in the system to support such huge memory allocation.
Usually it will return failure in the following code. But if the
input size is just a little smaller than the UINT64_MAX, like -2 in
signed type.
The roundup will cause an overflow and then "reset" the size to 0,
and then only a header (128B now) with zero length will be returned.
The following will be the previous allocation header.
It should be OK in most cases if the application won't access the
memory body. Or else, some critical issue will be caused and not easy
to debug. So this issue should be prevented at the beginning, like
other big size failure, NULL pointer should be returned also.

Fixes: fdf20fa7bee9 ("add prefix to cache line macros")
Cc: sergio.gonzalez.monroy@intel.com
Cc: stable@dpdk.org

Signed-off-by: Bing Zhao <bingz@mellanox.com>
---
 app/test/test_malloc.c              | 12 ++++++++++++
 lib/librte_eal/common/malloc_heap.c |  3 +++
 2 files changed, 15 insertions(+)

diff --git a/app/test/test_malloc.c b/app/test/test_malloc.c
index 40a2f50..a96c060 100644
--- a/app/test/test_malloc.c
+++ b/app/test/test_malloc.c
@@ -846,6 +846,18 @@
 	if (bad_ptr != NULL)
 		goto err_return;
 
+	/* rte_malloc expected to return null with size will cause overflow */
+	align = RTE_CACHE_LINE_SIZE;
+	size = (size_t)-8;
+
+	bad_ptr = rte_malloc(type, size, align);
+	if (bad_ptr != NULL)
+		goto err_return;
+
+	bad_ptr = rte_realloc(NULL, size, align);
+        if (bad_ptr != NULL)
+                goto err_return;
+
 	return 0;
 
 err_return:
diff --git a/lib/librte_eal/common/malloc_heap.c b/lib/librte_eal/common/malloc_heap.c
index 842eb9d..bd50656 100644
--- a/lib/librte_eal/common/malloc_heap.c
+++ b/lib/librte_eal/common/malloc_heap.c
@@ -241,6 +241,9 @@
 	size = RTE_CACHE_LINE_ROUNDUP(size);
 	align = RTE_CACHE_LINE_ROUNDUP(align);
 
+	/* roundup might cause an overflow */
+	if (size == 0)
+		return NULL;
 	elem = find_suitable_element(heap, size, flags, align, bound, contig);
 	if (elem != NULL) {
 		elem = malloc_elem_alloc(elem, size, align, bound, contig);
-- 
1.8.3.1


^ permalink raw reply	[flat|nested] 12+ messages in thread

* [dpdk-dev] [PATCH v2] mem: fix the alloc size roundup overflow
  2020-04-07 10:46 [dpdk-dev] [PATCH] mem: fix the alloc size roundup overflow Bing Zhao
  2020-04-20 11:39 ` Burakov, Anatoly
  2020-05-07  7:39 ` [dpdk-dev] [PATCH v2] " Bing Zhao
@ 2020-05-07  7:41 ` Bing Zhao
  2020-05-07  8:02   ` [dpdk-dev] [PATCH v3] " Bing Zhao
  2020-05-07 11:55   ` [dpdk-dev] [PATCH v2] " Burakov, Anatoly
  2 siblings, 2 replies; 12+ messages in thread
From: Bing Zhao @ 2020-05-07  7:41 UTC (permalink / raw)
  To: anatoly.burakov, thomas; +Cc: dev, stable, sergio.gonzalez.monroy

The size checking is done in the caller. The size parameter is an
unsigned (64b wide) right now, so the comparison with zero should be
enough in most cases. But it won't help in the following case.
If the allocating request input a huge number by mistake, e.g., some
overflow after the calculation (especially subtraction), the checking
in the caller will succeed since it is not zero. Indeed, there is not
enough space in the system to support such huge memory allocation.
Usually it will return failure in the following code. But if the
input size is just a little smaller than the UINT64_MAX, like -2 in
signed type.
The roundup will cause an overflow and then "reset" the size to 0,
and then only a header (128B now) with zero length will be returned.
The following will be the previous allocation header.
It should be OK in most cases if the application won't access the
memory body. Or else, some critical issue will be caused and not easy
to debug. So this issue should be prevented at the beginning, like
other big size failure, NULL pointer should be returned also.

Fixes: fdf20fa7bee9 ("add prefix to cache line macros")
Cc: sergio.gonzalez.monroy@intel.com
Cc: stable@dpdk.org

Signed-off-by: Bing Zhao <bingz@mellanox.com>
---
v2: add unit test for this case
---
 app/test/test_malloc.c              | 12 ++++++++++++
 lib/librte_eal/common/malloc_heap.c |  3 +++
 2 files changed, 15 insertions(+)

diff --git a/app/test/test_malloc.c b/app/test/test_malloc.c
index 40a2f50..a96c060 100644
--- a/app/test/test_malloc.c
+++ b/app/test/test_malloc.c
@@ -846,6 +846,18 @@
 	if (bad_ptr != NULL)
 		goto err_return;
 
+	/* rte_malloc expected to return null with size will cause overflow */
+	align = RTE_CACHE_LINE_SIZE;
+	size = (size_t)-8;
+
+	bad_ptr = rte_malloc(type, size, align);
+	if (bad_ptr != NULL)
+		goto err_return;
+
+	bad_ptr = rte_realloc(NULL, size, align);
+        if (bad_ptr != NULL)
+                goto err_return;
+
 	return 0;
 
 err_return:
diff --git a/lib/librte_eal/common/malloc_heap.c b/lib/librte_eal/common/malloc_heap.c
index 842eb9d..bd50656 100644
--- a/lib/librte_eal/common/malloc_heap.c
+++ b/lib/librte_eal/common/malloc_heap.c
@@ -241,6 +241,9 @@
 	size = RTE_CACHE_LINE_ROUNDUP(size);
 	align = RTE_CACHE_LINE_ROUNDUP(align);
 
+	/* roundup might cause an overflow */
+	if (size == 0)
+		return NULL;
 	elem = find_suitable_element(heap, size, flags, align, bound, contig);
 	if (elem != NULL) {
 		elem = malloc_elem_alloc(elem, size, align, bound, contig);
-- 
1.8.3.1


^ permalink raw reply	[flat|nested] 12+ messages in thread

* [dpdk-dev] [PATCH v3] mem: fix the alloc size roundup overflow
  2020-05-07  7:41 ` Bing Zhao
@ 2020-05-07  8:02   ` Bing Zhao
  2020-05-11 14:52     ` David Marchand
  2020-05-07 11:55   ` [dpdk-dev] [PATCH v2] " Burakov, Anatoly
  1 sibling, 1 reply; 12+ messages in thread
From: Bing Zhao @ 2020-05-07  8:02 UTC (permalink / raw)
  To: anatoly.burakov, thomas; +Cc: dev, stable, sergio.gonzalez.monroy

The size checking is done in the caller. The size parameter is an
unsigned (64b wide) right now, so the comparison with zero should be
enough in most cases. But it won't help in the following case.
If the allocating request input a huge number by mistake, e.g., some
overflow after the calculation (especially subtraction), the checking
in the caller will succeed since it is not zero. Indeed, there is not
enough space in the system to support such huge memory allocation.
Usually it will return failure in the following code. But if the
input size is just a little smaller than the UINT64_MAX, like -2 in
signed type.
The roundup will cause an overflow and then "reset" the size to 0,
and then only a header (128B now) with zero length will be returned.
The following will be the previous allocation header.
It should be OK in most cases if the application won't access the
memory body. Or else, some critical issue will be caused and not easy
to debug. So this issue should be prevented at the beginning, like
other big size failure, NULL pointer should be returned also.

Fixes: fdf20fa7bee9 ("add prefix to cache line macros")
Cc: sergio.gonzalez.monroy@intel.com
Cc: stable@dpdk.org

Signed-off-by: Bing Zhao <bingz@mellanox.com>
---
v2: add unit test for this case
v3: fix the code style by using tab
---
 app/test/test_malloc.c              | 12 ++++++++++++
 lib/librte_eal/common/malloc_heap.c |  3 +++
 2 files changed, 15 insertions(+)

diff --git a/app/test/test_malloc.c b/app/test/test_malloc.c
index 40a2f50..71b3cfd 100644
--- a/app/test/test_malloc.c
+++ b/app/test/test_malloc.c
@@ -846,6 +846,18 @@
 	if (bad_ptr != NULL)
 		goto err_return;
 
+	/* rte_malloc expected to return null with size will cause overflow */
+	align = RTE_CACHE_LINE_SIZE;
+	size = (size_t)-8;
+
+	bad_ptr = rte_malloc(type, size, align);
+	if (bad_ptr != NULL)
+		goto err_return;
+
+	bad_ptr = rte_realloc(NULL, size, align);
+	if (bad_ptr != NULL)
+		goto err_return;
+
 	return 0;
 
 err_return:
diff --git a/lib/librte_eal/common/malloc_heap.c b/lib/librte_eal/common/malloc_heap.c
index 842eb9d..bd50656 100644
--- a/lib/librte_eal/common/malloc_heap.c
+++ b/lib/librte_eal/common/malloc_heap.c
@@ -241,6 +241,9 @@
 	size = RTE_CACHE_LINE_ROUNDUP(size);
 	align = RTE_CACHE_LINE_ROUNDUP(align);
 
+	/* roundup might cause an overflow */
+	if (size == 0)
+		return NULL;
 	elem = find_suitable_element(heap, size, flags, align, bound, contig);
 	if (elem != NULL) {
 		elem = malloc_elem_alloc(elem, size, align, bound, contig);
-- 
1.8.3.1


^ permalink raw reply	[flat|nested] 12+ messages in thread

* Re: [dpdk-dev] [PATCH v2] mem: fix the alloc size roundup overflow
  2020-05-07  7:41 ` Bing Zhao
  2020-05-07  8:02   ` [dpdk-dev] [PATCH v3] " Bing Zhao
@ 2020-05-07 11:55   ` Burakov, Anatoly
  2020-05-07 12:12     ` [dpdk-dev] [dpdk-stable] " David Marchand
  1 sibling, 1 reply; 12+ messages in thread
From: Burakov, Anatoly @ 2020-05-07 11:55 UTC (permalink / raw)
  To: Bing Zhao, thomas; +Cc: dev, stable, sergio.gonzalez.monroy

On 07-May-20 8:41 AM, Bing Zhao wrote:
> The size checking is done in the caller. The size parameter is an
> unsigned (64b wide) right now, so the comparison with zero should be
> enough in most cases. But it won't help in the following case.
> If the allocating request input a huge number by mistake, e.g., some
> overflow after the calculation (especially subtraction), the checking
> in the caller will succeed since it is not zero. Indeed, there is not
> enough space in the system to support such huge memory allocation.
> Usually it will return failure in the following code. But if the
> input size is just a little smaller than the UINT64_MAX, like -2 in
> signed type.
> The roundup will cause an overflow and then "reset" the size to 0,
> and then only a header (128B now) with zero length will be returned.
> The following will be the previous allocation header.
> It should be OK in most cases if the application won't access the
> memory body. Or else, some critical issue will be caused and not easy
> to debug. So this issue should be prevented at the beginning, like
> other big size failure, NULL pointer should be returned also.
> 
> Fixes: fdf20fa7bee9 ("add prefix to cache line macros")
> Cc: sergio.gonzalez.monroy@intel.com
> Cc: stable@dpdk.org
> 
> Signed-off-by: Bing Zhao <bingz@mellanox.com>
> ---
> v2: add unit test for this case
> ---
>   app/test/test_malloc.c              | 12 ++++++++++++
>   lib/librte_eal/common/malloc_heap.c |  3 +++
>   2 files changed, 15 insertions(+)
> 
> diff --git a/app/test/test_malloc.c b/app/test/test_malloc.c
> index 40a2f50..a96c060 100644
> --- a/app/test/test_malloc.c
> +++ b/app/test/test_malloc.c
> @@ -846,6 +846,18 @@
>   	if (bad_ptr != NULL)
>   		goto err_return;
>   
> +	/* rte_malloc expected to return null with size will cause overflow */
> +	align = RTE_CACHE_LINE_SIZE;
> +	size = (size_t)-8;
> +
> +	bad_ptr = rte_malloc(type, size, align);
> +	if (bad_ptr != NULL)
> +		goto err_return;
> +
> +	bad_ptr = rte_realloc(NULL, size, align);
> +        if (bad_ptr != NULL)
> +                goto err_return;

You're mixing space and tabs as indentation here.

Otherwise,

Reviewed-by: Anatoly Burakov <anatoly.burakov@intel.com>

> +
>   	return 0;
>   
>   err_return:
> diff --git a/lib/librte_eal/common/malloc_heap.c b/lib/librte_eal/common/malloc_heap.c
> index 842eb9d..bd50656 100644
> --- a/lib/librte_eal/common/malloc_heap.c
> +++ b/lib/librte_eal/common/malloc_heap.c
> @@ -241,6 +241,9 @@
>   	size = RTE_CACHE_LINE_ROUNDUP(size);
>   	align = RTE_CACHE_LINE_ROUNDUP(align);
>   
> +	/* roundup might cause an overflow */
> +	if (size == 0)
> +		return NULL;
>   	elem = find_suitable_element(heap, size, flags, align, bound, contig);
>   	if (elem != NULL) {
>   		elem = malloc_elem_alloc(elem, size, align, bound, contig);
> 


-- 
Thanks,
Anatoly

^ permalink raw reply	[flat|nested] 12+ messages in thread

* Re: [dpdk-dev] [dpdk-stable] [PATCH v2] mem: fix the alloc size roundup overflow
  2020-05-07 11:55   ` [dpdk-dev] [PATCH v2] " Burakov, Anatoly
@ 2020-05-07 12:12     ` David Marchand
  2020-05-07 12:16       ` Bing Zhao
  0 siblings, 1 reply; 12+ messages in thread
From: David Marchand @ 2020-05-07 12:12 UTC (permalink / raw)
  To: Burakov, Anatoly, Bing Zhao
  Cc: Thomas Monjalon, dev, dpdk stable, sergio.gonzalez.monroy

On Thu, May 7, 2020 at 1:55 PM Burakov, Anatoly
<anatoly.burakov@intel.com> wrote:
>
> On 07-May-20 8:41 AM, Bing Zhao wrote:
> > The size checking is done in the caller. The size parameter is an
> > unsigned (64b wide) right now, so the comparison with zero should be
> > enough in most cases. But it won't help in the following case.
> > If the allocating request input a huge number by mistake, e.g., some
> > overflow after the calculation (especially subtraction), the checking
> > in the caller will succeed since it is not zero. Indeed, there is not
> > enough space in the system to support such huge memory allocation.
> > Usually it will return failure in the following code. But if the
> > input size is just a little smaller than the UINT64_MAX, like -2 in
> > signed type.
> > The roundup will cause an overflow and then "reset" the size to 0,
> > and then only a header (128B now) with zero length will be returned.
> > The following will be the previous allocation header.
> > It should be OK in most cases if the application won't access the
> > memory body. Or else, some critical issue will be caused and not easy
> > to debug. So this issue should be prevented at the beginning, like
> > other big size failure, NULL pointer should be returned also.
> >
> > Fixes: fdf20fa7bee9 ("add prefix to cache line macros")
> > Cc: stable@dpdk.org
> >
> > Signed-off-by: Bing Zhao <bingz@mellanox.com>
> > ---
> > v2: add unit test for this case
> > ---
> >   app/test/test_malloc.c              | 12 ++++++++++++
> >   lib/librte_eal/common/malloc_heap.c |  3 +++
> >   2 files changed, 15 insertions(+)
> >
> > diff --git a/app/test/test_malloc.c b/app/test/test_malloc.c
> > index 40a2f50..a96c060 100644
> > --- a/app/test/test_malloc.c
> > +++ b/app/test/test_malloc.c
> > @@ -846,6 +846,18 @@
> >       if (bad_ptr != NULL)
> >               goto err_return;
> >
> > +     /* rte_malloc expected to return null with size will cause overflow */
> > +     align = RTE_CACHE_LINE_SIZE;
> > +     size = (size_t)-8;
> > +
> > +     bad_ptr = rte_malloc(type, size, align);
> > +     if (bad_ptr != NULL)
> > +             goto err_return;
> > +
> > +     bad_ptr = rte_realloc(NULL, size, align);
> > +        if (bad_ptr != NULL)
> > +                goto err_return;
>
> You're mixing space and tabs as indentation here.

Will fix while applying.

>
> Otherwise,
>
> Reviewed-by: Anatoly Burakov <anatoly.burakov@intel.com>

You acked the v1, so I will go with it.
Thanks for the work Bing, Anatoly.


-- 
David Marchand


^ permalink raw reply	[flat|nested] 12+ messages in thread

* Re: [dpdk-dev] [dpdk-stable] [PATCH v2] mem: fix the alloc size roundup overflow
  2020-05-07 12:12     ` [dpdk-dev] [dpdk-stable] " David Marchand
@ 2020-05-07 12:16       ` Bing Zhao
  2020-05-07 12:18         ` David Marchand
  0 siblings, 1 reply; 12+ messages in thread
From: Bing Zhao @ 2020-05-07 12:16 UTC (permalink / raw)
  To: David Marchand, Burakov, Anatoly
  Cc: Thomas Monjalon, dev, dpdk stable, sergio.gonzalez.monroy

Hi David,
I've changed the space into tabs in v3 and suppressed this one.

Thanks a lot

BR. Bing

> -----Original Message-----
> From: David Marchand <david.marchand@redhat.com>
> Sent: Thursday, May 7, 2020 8:12 PM
> To: Burakov, Anatoly <anatoly.burakov@intel.com>; Bing Zhao
> <bingz@mellanox.com>
> Cc: Thomas Monjalon <thomas@monjalon.net>; dev <dev@dpdk.org>;
> dpdk stable <stable@dpdk.org>; sergio.gonzalez.monroy@intel.com
> Subject: Re: [dpdk-stable] [PATCH v2] mem: fix the alloc size roundup
> overflow
> 
> On Thu, May 7, 2020 at 1:55 PM Burakov, Anatoly
> <anatoly.burakov@intel.com> wrote:
> >
> > On 07-May-20 8:41 AM, Bing Zhao wrote:
> > > The size checking is done in the caller. The size parameter is an
> > > unsigned (64b wide) right now, so the comparison with zero should
> be
> > > enough in most cases. But it won't help in the following case.
> > > If the allocating request input a huge number by mistake, e.g.,
> some
> > > overflow after the calculation (especially subtraction), the
> > > checking in the caller will succeed since it is not zero. Indeed,
> > > there is not enough space in the system to support such huge
> memory allocation.
> > > Usually it will return failure in the following code. But if the
> > > input size is just a little smaller than the UINT64_MAX, like -2 in
> > > signed type.
> > > The roundup will cause an overflow and then "reset" the size to 0,
> > > and then only a header (128B now) with zero length will be
> returned.
> > > The following will be the previous allocation header.
> > > It should be OK in most cases if the application won't access the
> > > memory body. Or else, some critical issue will be caused and not
> > > easy to debug. So this issue should be prevented at the beginning,
> > > like other big size failure, NULL pointer should be returned also.
> > >
> > > Fixes: fdf20fa7bee9 ("add prefix to cache line macros")
> > > Cc: stable@dpdk.org
> > >
> > > Signed-off-by: Bing Zhao <bingz@mellanox.com>
> > > ---
> > > v2: add unit test for this case
> > > ---
> > >   app/test/test_malloc.c              | 12 ++++++++++++
> > >   lib/librte_eal/common/malloc_heap.c |  3 +++
> > >   2 files changed, 15 insertions(+)
> > >
> > > diff --git a/app/test/test_malloc.c b/app/test/test_malloc.c index
> > > 40a2f50..a96c060 100644
> > > --- a/app/test/test_malloc.c
> > > +++ b/app/test/test_malloc.c
> > > @@ -846,6 +846,18 @@
> > >       if (bad_ptr != NULL)
> > >               goto err_return;
> > >
> > > +     /* rte_malloc expected to return null with size will cause
> overflow */
> > > +     align = RTE_CACHE_LINE_SIZE;
> > > +     size = (size_t)-8;
> > > +
> > > +     bad_ptr = rte_malloc(type, size, align);
> > > +     if (bad_ptr != NULL)
> > > +             goto err_return;
> > > +
> > > +     bad_ptr = rte_realloc(NULL, size, align);
> > > +        if (bad_ptr != NULL)
> > > +                goto err_return;
> >
> > You're mixing space and tabs as indentation here.
> 
> Will fix while applying.
> 
> >
> > Otherwise,
> >
> > Reviewed-by: Anatoly Burakov <anatoly.burakov@intel.com>
> 
> You acked the v1, so I will go with it.
> Thanks for the work Bing, Anatoly.
> 
> 
> --
> David Marchand


^ permalink raw reply	[flat|nested] 12+ messages in thread

* Re: [dpdk-dev] [dpdk-stable] [PATCH v2] mem: fix the alloc size roundup overflow
  2020-05-07 12:16       ` Bing Zhao
@ 2020-05-07 12:18         ` David Marchand
  0 siblings, 0 replies; 12+ messages in thread
From: David Marchand @ 2020-05-07 12:18 UTC (permalink / raw)
  To: Bing Zhao
  Cc: Burakov, Anatoly, Thomas Monjalon, dev, dpdk stable,
	sergio.gonzalez.monroy

On Thu, May 7, 2020 at 2:16 PM Bing Zhao <bingz@mellanox.com> wrote:
>
> Hi David,
> I've changed the space into tabs in v3 and suppressed this one.

Ok, I would have noticed anyway :-)
Thanks.


-- 
David Marchand


^ permalink raw reply	[flat|nested] 12+ messages in thread

* Re: [dpdk-dev] [PATCH v3] mem: fix the alloc size roundup overflow
  2020-05-07  8:02   ` [dpdk-dev] [PATCH v3] " Bing Zhao
@ 2020-05-11 14:52     ` David Marchand
  0 siblings, 0 replies; 12+ messages in thread
From: David Marchand @ 2020-05-11 14:52 UTC (permalink / raw)
  To: Bing Zhao; +Cc: Burakov, Anatoly, Thomas Monjalon, dev, dpdk stable

On Thu, May 7, 2020 at 10:03 AM Bing Zhao <bingz@mellanox.com> wrote:
>
> The size checking is done in the caller. The size parameter is an
> unsigned (64b wide) right now, so the comparison with zero should be
> enough in most cases. But it won't help in the following case.
> If the allocating request input a huge number by mistake, e.g., some
> overflow after the calculation (especially subtraction), the checking
> in the caller will succeed since it is not zero. Indeed, there is not
> enough space in the system to support such huge memory allocation.
> Usually it will return failure in the following code. But if the
> input size is just a little smaller than the UINT64_MAX, like -2 in
> signed type.
> The roundup will cause an overflow and then "reset" the size to 0,
> and then only a header (128B now) with zero length will be returned.
> The following will be the previous allocation header.
> It should be OK in most cases if the application won't access the
> memory body. Or else, some critical issue will be caused and not easy
> to debug. So this issue should be prevented at the beginning, like
> other big size failure, NULL pointer should be returned also.
>
> Fixes: fdf20fa7bee9 ("add prefix to cache line macros")
> Cc: stable@dpdk.org
>
> Signed-off-by: Bing Zhao <bingz@mellanox.com>

Acked-by: Anatoly Burakov <anatoly.burakov@intel.com>

Applied, thanks.


-- 
David Marchand


^ permalink raw reply	[flat|nested] 12+ messages in thread

end of thread, other threads:[~2020-05-11 14:53 UTC | newest]

Thread overview: 12+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2020-04-07 10:46 [dpdk-dev] [PATCH] mem: fix the alloc size roundup overflow Bing Zhao
2020-04-20 11:39 ` Burakov, Anatoly
2020-04-21  3:18   ` Bing Zhao
2020-05-06 21:43     ` [dpdk-dev] [dpdk-stable] " Thomas Monjalon
2020-05-07  7:39 ` [dpdk-dev] [PATCH v2] " Bing Zhao
2020-05-07  7:41 ` Bing Zhao
2020-05-07  8:02   ` [dpdk-dev] [PATCH v3] " Bing Zhao
2020-05-11 14:52     ` David Marchand
2020-05-07 11:55   ` [dpdk-dev] [PATCH v2] " Burakov, Anatoly
2020-05-07 12:12     ` [dpdk-dev] [dpdk-stable] " David Marchand
2020-05-07 12:16       ` Bing Zhao
2020-05-07 12:18         ` David Marchand

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).