[SRU] [Trusty] [PATCH 0/1] openvswitch: fix kernel buffer overflow

classic Classic list List threaded Threaded
5 messages Options
Reply | Threaded
Open this post in threaded view
|

[SRU] [Trusty] [PATCH 0/1] openvswitch: fix kernel buffer overflow

Andrea Righi
Buglink: https://bugs.launchpad.net/bugs/1813244

[Impact]

 * Flow action buffer can be incorrectly resized to contain the newly added
   action entries: the buffer is always resized multiplying the previous size
   by 2, but this might be not enough if the requested size is bigger than
   previous_size*2, causing a buffer overflow

 * The fix correctly resizes the buffer to prevent the buffer overflow
   and potential memory corruptions

 * This bug can be triggered potentially on any architecture, but it is very
   likely to happen on i386 running the following test case

[Test Case]

 * run this openvswitch test case:
   https://launchpadlibrarian.net/416589265/lp1262692

[Fix]

 * Instead of resizing the buffer by a factor of 2, use
   max(current_size * 2, current_size + requested_size)

[Regression Potential]

 * Fix has been tested on the affected platform and verified using slub_debug.
   It is an upstream fix and also a very small patch (one-liner basically), so
   backport changes are minimal.

Andrea Righi (1):
 openvswitch: fix flow actions reallocation

 net/openvswitch/flow_netlink.c |    4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)


--
kernel-team mailing list
[hidden email]
https://lists.ubuntu.com/mailman/listinfo/kernel-team
Reply | Threaded
Open this post in threaded view
|

[SRU] [Trusty] [PATCH 1/1] openvswitch: fix flow actions reallocation

Andrea Righi
Buglink: https://bugs.launchpad.net/bugs/1813244

The flow action buffer can be resized if it's not big enough to contain
all the requested flow actions. However, this resize doesn't take into
account the new requested size, the buffer is only increased by a factor
of 2x. This might be not enough to contain the new data, causing a
buffer overflow, for example:

[   42.044472] =============================================================================
[   42.045608] BUG kmalloc-96 (Not tainted): Redzone overwritten
[   42.046415] -----------------------------------------------------------------------------

[   42.047715] Disabling lock debugging due to kernel taint
[   42.047716] INFO: 0x8bf2c4a5-0x720c0928. First byte 0x0 instead of 0xcc
[   42.048677] INFO: Slab 0xbc6d2040 objects=29 used=18 fp=0xdc07dec4 flags=0x2808101
[   42.049743] INFO: Object 0xd53a3464 @offset=2528 fp=0xccdcdebb

[   42.050747] Redzone 76f1b237: cc cc cc cc cc cc cc cc                          ........
[   42.051839] Object d53a3464: 6b 6b 6b 6b 6b 6b 6b 6b 0c 00 00 00 6c 00 00 00  kkkkkkkk....l...
[   42.053015] Object f49a30cc: 6c 00 0c 00 00 00 00 00 00 00 00 03 78 a3 15 f6  l...........x...
[   42.054203] Object acfe4220: 20 00 02 00 ff ff ff ff 00 00 00 00 00 00 00 00   ...............
[   42.055370] Object 21024e91: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00  ................
[   42.056541] Object 070e04c3: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00  ................
[   42.057797] Object 948a777a: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00  ................
[   42.059061] Redzone 8bf2c4a5: 00 00 00 00                                      ....
[   42.060189] Padding a681b46e: 5a 5a 5a 5a 5a 5a 5a 5a                          ZZZZZZZZ

Fix by making sure the new buffer is properly resized to contain all the
requested data.

BugLink: https://bugs.launchpad.net/bugs/1813244
Signed-off-by: Andrea Righi <[hidden email]>
Acked-by: Pravin B Shelar <[hidden email]>
Signed-off-by: David S. Miller <[hidden email]>
(backported from commit f28cd2af22a0c134e4aa1c64a70f70d815d473fb)
Signed-off-by: Andrea Righi <[hidden email]>
---
 net/openvswitch/flow_netlink.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/net/openvswitch/flow_netlink.c b/net/openvswitch/flow_netlink.c
index 2bc1bc1aca3b..7ac03e1a799a 100644
--- a/net/openvswitch/flow_netlink.c
+++ b/net/openvswitch/flow_netlink.c
@@ -1149,14 +1149,14 @@ static struct nlattr *reserve_sfa_size(struct sw_flow_actions **sfa,
 
  struct sw_flow_actions *acts;
  int new_acts_size;
- int req_size = NLA_ALIGN(attr_len);
+ size_t req_size = NLA_ALIGN(attr_len);
  int next_offset = offsetof(struct sw_flow_actions, actions) +
  (*sfa)->actions_len;
 
  if (req_size <= (ksize(*sfa) - next_offset))
  goto out;
 
- new_acts_size = ksize(*sfa) * 2;
+ new_acts_size = max(next_offset + req_size, ksize(*sfa) * 2);
 
  if (new_acts_size > MAX_ACTIONS_BUFSIZE) {
  if ((MAX_ACTIONS_BUFSIZE - next_offset) < req_size)
--
2.19.1


--
kernel-team mailing list
[hidden email]
https://lists.ubuntu.com/mailman/listinfo/kernel-team
Reply | Threaded
Open this post in threaded view
|

ACK/cmnt: [SRU] [Trusty] [PATCH 1/1] openvswitch: fix flow actions reallocation

Juerg Haefliger
On Tue,  9 Apr 2019 12:28:09 +0200
Andrea Righi <[hidden email]> wrote:

> Buglink: https://bugs.launchpad.net/bugs/1813244
>
> The flow action buffer can be resized if it's not big enough to contain
> all the requested flow actions. However, this resize doesn't take into
> account the new requested size, the buffer is only increased by a factor
> of 2x. This might be not enough to contain the new data, causing a
> buffer overflow, for example:
>
> [   42.044472] =============================================================================
> [   42.045608] BUG kmalloc-96 (Not tainted): Redzone overwritten
> [   42.046415] -----------------------------------------------------------------------------
>
> [   42.047715] Disabling lock debugging due to kernel taint
> [   42.047716] INFO: 0x8bf2c4a5-0x720c0928. First byte 0x0 instead of 0xcc
> [   42.048677] INFO: Slab 0xbc6d2040 objects=29 used=18 fp=0xdc07dec4 flags=0x2808101
> [   42.049743] INFO: Object 0xd53a3464 @offset=2528 fp=0xccdcdebb
>
> [   42.050747] Redzone 76f1b237: cc cc cc cc cc cc cc cc                          ........
> [   42.051839] Object d53a3464: 6b 6b 6b 6b 6b 6b 6b 6b 0c 00 00 00 6c 00 00 00  kkkkkkkk....l...
> [   42.053015] Object f49a30cc: 6c 00 0c 00 00 00 00 00 00 00 00 03 78 a3 15 f6  l...........x...
> [   42.054203] Object acfe4220: 20 00 02 00 ff ff ff ff 00 00 00 00 00 00 00 00   ...............
> [   42.055370] Object 21024e91: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00  ................
> [   42.056541] Object 070e04c3: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00  ................
> [   42.057797] Object 948a777a: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00  ................
> [   42.059061] Redzone 8bf2c4a5: 00 00 00 00                                      ....
> [   42.060189] Padding a681b46e: 5a 5a 5a 5a 5a 5a 5a 5a                          ZZZZZZZZ
>
> Fix by making sure the new buffer is properly resized to contain all the
> requested data.
>
> BugLink: https://bugs.launchpad.net/bugs/1813244
> Signed-off-by: Andrea Righi <[hidden email]>
> Acked-by: Pravin B Shelar <[hidden email]>
> Signed-off-by: David S. Miller <[hidden email]>
> (backported from commit f28cd2af22a0c134e4aa1c64a70f70d815d473fb)
The patch actually cherry picks cleanly so this should be changed to:
(cherry picked from commit f28cd2af22a0c134e4aa1c64a70f70d815d473fb)


> Signed-off-by: Andrea Righi <[hidden email]>

Acked-by: Juerg Haefliger <[hidden email]>


> ---
>  net/openvswitch/flow_netlink.c | 4 ++--
>  1 file changed, 2 insertions(+), 2 deletions(-)
>
> diff --git a/net/openvswitch/flow_netlink.c b/net/openvswitch/flow_netlink.c
> index 2bc1bc1aca3b..7ac03e1a799a 100644
> --- a/net/openvswitch/flow_netlink.c
> +++ b/net/openvswitch/flow_netlink.c
> @@ -1149,14 +1149,14 @@ static struct nlattr *reserve_sfa_size(struct sw_flow_actions **sfa,
>  
>   struct sw_flow_actions *acts;
>   int new_acts_size;
> - int req_size = NLA_ALIGN(attr_len);
> + size_t req_size = NLA_ALIGN(attr_len);
>   int next_offset = offsetof(struct sw_flow_actions, actions) +
>   (*sfa)->actions_len;
>  
>   if (req_size <= (ksize(*sfa) - next_offset))
>   goto out;
>  
> - new_acts_size = ksize(*sfa) * 2;
> + new_acts_size = max(next_offset + req_size, ksize(*sfa) * 2);
>  
>   if (new_acts_size > MAX_ACTIONS_BUFSIZE) {
>   if ((MAX_ACTIONS_BUFSIZE - next_offset) < req_size)

--
kernel-team mailing list
[hidden email]
https://lists.ubuntu.com/mailman/listinfo/kernel-team

attachment0 (849 bytes) Download Attachment
Reply | Threaded
Open this post in threaded view
|

ACK: [SRU] [Trusty] [PATCH 0/1] openvswitch: fix kernel buffer overflow

Khaled Elmously
In reply to this post by Andrea Righi
On 2019-04-09 12:28:08 , Andrea Righi wrote:

> Buglink: https://bugs.launchpad.net/bugs/1813244
>
> [Impact]
>
>  * Flow action buffer can be incorrectly resized to contain the newly added
>    action entries: the buffer is always resized multiplying the previous size
>    by 2, but this might be not enough if the requested size is bigger than
>    previous_size*2, causing a buffer overflow
>
>  * The fix correctly resizes the buffer to prevent the buffer overflow
>    and potential memory corruptions
>
>  * This bug can be triggered potentially on any architecture, but it is very
>    likely to happen on i386 running the following test case
>
> [Test Case]
>
>  * run this openvswitch test case:
>    https://launchpadlibrarian.net/416589265/lp1262692
>
> [Fix]
>
>  * Instead of resizing the buffer by a factor of 2, use
>    max(current_size * 2, current_size + requested_size)
>
> [Regression Potential]
>
>  * Fix has been tested on the affected platform and verified using slub_debug.
>    It is an upstream fix and also a very small patch (one-liner basically), so
>    backport changes are minimal.
>
> Andrea Righi (1):
>  openvswitch: fix flow actions reallocation
>
>  net/openvswitch/flow_netlink.c |    4 ++--
>  1 file changed, 2 insertions(+), 2 deletions(-)
>
>

Acked-by: Khalid Elmously <[hidden email]>


--
kernel-team mailing list
[hidden email]
https://lists.ubuntu.com/mailman/listinfo/kernel-team
Reply | Threaded
Open this post in threaded view
|

APPLIED: [SRU] [Trusty] [PATCH 0/1] openvswitch: fix kernel buffer overflow

Khaled Elmously
In reply to this post by Andrea Righi
On 2019-04-09 12:28:08 , Andrea Righi wrote:

> Buglink: https://bugs.launchpad.net/bugs/1813244
>
> [Impact]
>
>  * Flow action buffer can be incorrectly resized to contain the newly added
>    action entries: the buffer is always resized multiplying the previous size
>    by 2, but this might be not enough if the requested size is bigger than
>    previous_size*2, causing a buffer overflow
>
>  * The fix correctly resizes the buffer to prevent the buffer overflow
>    and potential memory corruptions
>
>  * This bug can be triggered potentially on any architecture, but it is very
>    likely to happen on i386 running the following test case
>
> [Test Case]
>
>  * run this openvswitch test case:
>    https://launchpadlibrarian.net/416589265/lp1262692
>
> [Fix]
>
>  * Instead of resizing the buffer by a factor of 2, use
>    max(current_size * 2, current_size + requested_size)
>
> [Regression Potential]
>
>  * Fix has been tested on the affected platform and verified using slub_debug.
>    It is an upstream fix and also a very small patch (one-liner basically), so
>    backport changes are minimal.
>
> Andrea Righi (1):
>  openvswitch: fix flow actions reallocation
>
>  net/openvswitch/flow_netlink.c |    4 ++--
>  1 file changed, 2 insertions(+), 2 deletions(-)
>
>
> --
> kernel-team mailing list
> [hidden email]
> https://lists.ubuntu.com/mailman/listinfo/kernel-team

--
kernel-team mailing list
[hidden email]
https://lists.ubuntu.com/mailman/listinfo/kernel-team