patches for DPDK stable branches
 help / color / mirror / Atom feed
* [dpdk-stable] [PATCH] lpm: fix buffer overflow
@ 2021-10-08 21:28 Vladimir Medvedkin
  2021-10-20 19:55 ` David Marchand
  2021-10-21 17:15 ` [dpdk-stable] [PATCH v2] " Vladimir Medvedkin
  0 siblings, 2 replies; 6+ messages in thread
From: Vladimir Medvedkin @ 2021-10-08 21:28 UTC (permalink / raw)
  To: dev; +Cc: bruce.richardson, david.marchand, alex, stable

This patch fixes buffer overflow reported by ASAN,
please reference https://bugs.dpdk.org/show_bug.cgi?id=819

The rte_lpm6 keeps routing information for control plane purpose
inside the rte_hash table which uses rte_jhash() as a hash function.
From the rte_jhash() documentation: If input key is not aligned to
four byte boundaries or a multiple of four bytes in length,
the memory region just after may be read (but not used in the
computation).
rte_lpm6 uses 17 bytes keys consisting of IPv6 address (16 bytes) +
depth (1 byte).

This patch increases the size of the depth field up to uint32_t
and sets the alignment to 4 bytes.

Bugzilla ID: 819
Fixes: 86b3b21952a8 ("lpm6: store rules in hash table")
Cc: alex@therouter.net
Cc: stable@dpdk.org

Signed-off-by: Vladimir Medvedkin <vladimir.medvedkin@intel.com>
---
 lib/lpm/rte_lpm6.c | 4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/lib/lpm/rte_lpm6.c b/lib/lpm/rte_lpm6.c
index 37baabb..d5e0918 100644
--- a/lib/lpm/rte_lpm6.c
+++ b/lib/lpm/rte_lpm6.c
@@ -80,8 +80,8 @@ struct rte_lpm6_rule {
 /** Rules tbl entry key. */
 struct rte_lpm6_rule_key {
 	uint8_t ip[RTE_LPM6_IPV6_ADDR_SIZE]; /**< Rule IP address. */
-	uint8_t depth; /**< Rule depth. */
-};
+	uint32_t depth; /**< Rule depth. */
+} __rte_aligned(sizeof(uint32_t));
 
 /* Header of tbl8 */
 struct rte_lpm_tbl8_hdr {
-- 
2.7.4


^ permalink raw reply	[flat|nested] 6+ messages in thread

* Re: [dpdk-stable] [PATCH] lpm: fix buffer overflow
  2021-10-08 21:28 [dpdk-stable] [PATCH] lpm: fix buffer overflow Vladimir Medvedkin
@ 2021-10-20 19:55 ` David Marchand
  2021-10-21 17:15   ` Medvedkin, Vladimir
  2021-10-21 17:15 ` [dpdk-stable] [PATCH v2] " Vladimir Medvedkin
  1 sibling, 1 reply; 6+ messages in thread
From: David Marchand @ 2021-10-20 19:55 UTC (permalink / raw)
  To: Vladimir Medvedkin; +Cc: dev, Bruce Richardson, alex, dpdk stable

Hello Vladimir,

On Fri, Oct 8, 2021 at 11:29 PM Vladimir Medvedkin
<vladimir.medvedkin@intel.com> wrote:
>
> This patch fixes buffer overflow reported by ASAN,
> please reference https://bugs.dpdk.org/show_bug.cgi?id=819
>
> The rte_lpm6 keeps routing information for control plane purpose
> inside the rte_hash table which uses rte_jhash() as a hash function.
> From the rte_jhash() documentation: If input key is not aligned to
> four byte boundaries or a multiple of four bytes in length,
> the memory region just after may be read (but not used in the
> computation).
> rte_lpm6 uses 17 bytes keys consisting of IPv6 address (16 bytes) +
> depth (1 byte).
>
> This patch increases the size of the depth field up to uint32_t
> and sets the alignment to 4 bytes.
>
> Bugzilla ID: 819
> Fixes: 86b3b21952a8 ("lpm6: store rules in hash table")
> Cc: alex@therouter.net
> Cc: stable@dpdk.org

This change should be internal, and not breaking ABI, but are we sure
we want to backport it?


>
> Signed-off-by: Vladimir Medvedkin <vladimir.medvedkin@intel.com>
> ---
>  lib/lpm/rte_lpm6.c | 4 ++--
>  1 file changed, 2 insertions(+), 2 deletions(-)
>
> diff --git a/lib/lpm/rte_lpm6.c b/lib/lpm/rte_lpm6.c
> index 37baabb..d5e0918 100644
> --- a/lib/lpm/rte_lpm6.c
> +++ b/lib/lpm/rte_lpm6.c
> @@ -80,8 +80,8 @@ struct rte_lpm6_rule {
>  /** Rules tbl entry key. */
>  struct rte_lpm6_rule_key {
>         uint8_t ip[RTE_LPM6_IPV6_ADDR_SIZE]; /**< Rule IP address. */
> -       uint8_t depth; /**< Rule depth. */
> -};
> +       uint32_t depth; /**< Rule depth. */
> +} __rte_aligned(sizeof(uint32_t));

I would recommend doing the same than for hash tests: keep growing
depth to 32bits, but no enforcement of alignment and add build check
on structure size being sizeof(uin32_t) aligned.


>
>  /* Header of tbl8 */
>  struct rte_lpm_tbl8_hdr {
> --
> 2.7.4
>


-- 
David Marchand


^ permalink raw reply	[flat|nested] 6+ messages in thread

* Re: [dpdk-stable] [PATCH] lpm: fix buffer overflow
  2021-10-20 19:55 ` David Marchand
@ 2021-10-21 17:15   ` Medvedkin, Vladimir
  0 siblings, 0 replies; 6+ messages in thread
From: Medvedkin, Vladimir @ 2021-10-21 17:15 UTC (permalink / raw)
  To: David Marchand; +Cc: dev, Bruce Richardson, alex, dpdk stable

Hi David,

On 20/10/2021 21:55, David Marchand wrote:
> Hello Vladimir,
> 
> On Fri, Oct 8, 2021 at 11:29 PM Vladimir Medvedkin
> <vladimir.medvedkin@intel.com> wrote:
>>
>> This patch fixes buffer overflow reported by ASAN,
>> please reference https://bugs.dpdk.org/show_bug.cgi?id=819
>>
>> The rte_lpm6 keeps routing information for control plane purpose
>> inside the rte_hash table which uses rte_jhash() as a hash function.
>>  From the rte_jhash() documentation: If input key is not aligned to
>> four byte boundaries or a multiple of four bytes in length,
>> the memory region just after may be read (but not used in the
>> computation).
>> rte_lpm6 uses 17 bytes keys consisting of IPv6 address (16 bytes) +
>> depth (1 byte).
>>
>> This patch increases the size of the depth field up to uint32_t
>> and sets the alignment to 4 bytes.
>>
>> Bugzilla ID: 819
>> Fixes: 86b3b21952a8 ("lpm6: store rules in hash table")
>> Cc: alex@therouter.net
>> Cc: stable@dpdk.org
> 
> This change should be internal, and not breaking ABI, but are we sure
> we want to backport it?
> 

I think yes, I don't see any reason why we should not backport it.
Do you think we should not?

> 
>>
>> Signed-off-by: Vladimir Medvedkin <vladimir.medvedkin@intel.com>
>> ---
>>   lib/lpm/rte_lpm6.c | 4 ++--
>>   1 file changed, 2 insertions(+), 2 deletions(-)
>>
>> diff --git a/lib/lpm/rte_lpm6.c b/lib/lpm/rte_lpm6.c
>> index 37baabb..d5e0918 100644
>> --- a/lib/lpm/rte_lpm6.c
>> +++ b/lib/lpm/rte_lpm6.c
>> @@ -80,8 +80,8 @@ struct rte_lpm6_rule {
>>   /** Rules tbl entry key. */
>>   struct rte_lpm6_rule_key {
>>          uint8_t ip[RTE_LPM6_IPV6_ADDR_SIZE]; /**< Rule IP address. */
>> -       uint8_t depth; /**< Rule depth. */
>> -};
>> +       uint32_t depth; /**< Rule depth. */
>> +} __rte_aligned(sizeof(uint32_t));
> 
> I would recommend doing the same than for hash tests: keep growing
> depth to 32bits, but no enforcement of alignment and add build check
> on structure size being sizeof(uin32_t) aligned.
> 

Agree, will send v2

> 
>>
>>   /* Header of tbl8 */
>>   struct rte_lpm_tbl8_hdr {
>> --
>> 2.7.4
>>
> 
> 

-- 
Regards,
Vladimir

^ permalink raw reply	[flat|nested] 6+ messages in thread

* [dpdk-stable] [PATCH v2] lpm: fix buffer overflow
  2021-10-08 21:28 [dpdk-stable] [PATCH] lpm: fix buffer overflow Vladimir Medvedkin
  2021-10-20 19:55 ` David Marchand
@ 2021-10-21 17:15 ` Vladimir Medvedkin
  2021-10-22  9:07   ` Bruce Richardson
  1 sibling, 1 reply; 6+ messages in thread
From: Vladimir Medvedkin @ 2021-10-21 17:15 UTC (permalink / raw)
  To: dev; +Cc: bruce.richardson, david.marchand, alex, stable

This patch fixes buffer overflow reported by ASAN,
please reference https://bugs.dpdk.org/show_bug.cgi?id=819

The rte_lpm6 keeps routing information for control plane purpose
inside the rte_hash table which uses rte_jhash() as a hash function.
From the rte_jhash() documentation: If input key is not aligned to
four byte boundaries or a multiple of four bytes in length,
the memory region just after may be read (but not used in the
computation).
rte_lpm6 uses 17 bytes keys consisting of IPv6 address (16 bytes) +
depth (1 byte).

This patch increases the size of the depth field up to uint32_t
and sets the alignment to 4 bytes.

Bugzilla ID: 819
Fixes: 86b3b21952a8 ("lpm6: store rules in hash table")
Cc: alex@therouter.net
Cc: stable@dpdk.org

Signed-off-by: Vladimir Medvedkin <vladimir.medvedkin@intel.com>
---
 lib/lpm/rte_lpm6.c | 4 +++-
 1 file changed, 3 insertions(+), 1 deletion(-)

diff --git a/lib/lpm/rte_lpm6.c b/lib/lpm/rte_lpm6.c
index 37baabb..73768fc 100644
--- a/lib/lpm/rte_lpm6.c
+++ b/lib/lpm/rte_lpm6.c
@@ -80,7 +80,7 @@ struct rte_lpm6_rule {
 /** Rules tbl entry key. */
 struct rte_lpm6_rule_key {
 	uint8_t ip[RTE_LPM6_IPV6_ADDR_SIZE]; /**< Rule IP address. */
-	uint8_t depth; /**< Rule depth. */
+	uint32_t depth; /**< Rule depth. */
 };
 
 /* Header of tbl8 */
@@ -259,6 +259,8 @@ rte_lpm6_create(const char *name, int socket_id,
 	lpm_list = RTE_TAILQ_CAST(rte_lpm6_tailq.head, rte_lpm6_list);
 
 	RTE_BUILD_BUG_ON(sizeof(struct rte_lpm6_tbl_entry) != sizeof(uint32_t));
+	RTE_BUILD_BUG_ON(sizeof(struct rte_lpm6_rule_key) %
+		sizeof(uint32_t) != 0);
 
 	/* Check user arguments. */
 	if ((name == NULL) || (socket_id < -1) || (config == NULL) ||
-- 
2.7.4


^ permalink raw reply	[flat|nested] 6+ messages in thread

* Re: [dpdk-stable] [PATCH v2] lpm: fix buffer overflow
  2021-10-21 17:15 ` [dpdk-stable] [PATCH v2] " Vladimir Medvedkin
@ 2021-10-22  9:07   ` Bruce Richardson
  2021-10-25 17:10     ` Thomas Monjalon
  0 siblings, 1 reply; 6+ messages in thread
From: Bruce Richardson @ 2021-10-22  9:07 UTC (permalink / raw)
  To: Vladimir Medvedkin; +Cc: dev, david.marchand, alex, stable

On Thu, Oct 21, 2021 at 06:15:49PM +0100, Vladimir Medvedkin wrote:
> This patch fixes buffer overflow reported by ASAN,
> please reference https://bugs.dpdk.org/show_bug.cgi?id=819
> 
> The rte_lpm6 keeps routing information for control plane purpose
> inside the rte_hash table which uses rte_jhash() as a hash function.
> From the rte_jhash() documentation: If input key is not aligned to
> four byte boundaries or a multiple of four bytes in length,
> the memory region just after may be read (but not used in the
> computation).
> rte_lpm6 uses 17 bytes keys consisting of IPv6 address (16 bytes) +
> depth (1 byte).
> 
> This patch increases the size of the depth field up to uint32_t
> and sets the alignment to 4 bytes.
> 
> Bugzilla ID: 819
> Fixes: 86b3b21952a8 ("lpm6: store rules in hash table")
> Cc: alex@therouter.net
> Cc: stable@dpdk.org
> 
> Signed-off-by: Vladimir Medvedkin <vladimir.medvedkin@intel.com>

Acked-by: Bruce Richardson <bruce.richardson@intel.com>


^ permalink raw reply	[flat|nested] 6+ messages in thread

* Re: [dpdk-stable] [PATCH v2] lpm: fix buffer overflow
  2021-10-22  9:07   ` Bruce Richardson
@ 2021-10-25 17:10     ` Thomas Monjalon
  0 siblings, 0 replies; 6+ messages in thread
From: Thomas Monjalon @ 2021-10-25 17:10 UTC (permalink / raw)
  To: Vladimir Medvedkin; +Cc: stable, dev, david.marchand, alex, Bruce Richardson

22/10/2021 11:07, Bruce Richardson:
> On Thu, Oct 21, 2021 at 06:15:49PM +0100, Vladimir Medvedkin wrote:
> > This patch fixes buffer overflow reported by ASAN,
> > please reference https://bugs.dpdk.org/show_bug.cgi?id=819
> > 
> > The rte_lpm6 keeps routing information for control plane purpose
> > inside the rte_hash table which uses rte_jhash() as a hash function.
> > From the rte_jhash() documentation: If input key is not aligned to
> > four byte boundaries or a multiple of four bytes in length,
> > the memory region just after may be read (but not used in the
> > computation).
> > rte_lpm6 uses 17 bytes keys consisting of IPv6 address (16 bytes) +
> > depth (1 byte).
> > 
> > This patch increases the size of the depth field up to uint32_t
> > and sets the alignment to 4 bytes.
> > 
> > Bugzilla ID: 819
> > Fixes: 86b3b21952a8 ("lpm6: store rules in hash table")
> > Cc: alex@therouter.net
> > Cc: stable@dpdk.org
> > 
> > Signed-off-by: Vladimir Medvedkin <vladimir.medvedkin@intel.com>
> 
> Acked-by: Bruce Richardson <bruce.richardson@intel.com>

Applied, thanks.




^ permalink raw reply	[flat|nested] 6+ messages in thread

end of thread, other threads:[~2021-10-25 17:11 UTC | newest]

Thread overview: 6+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2021-10-08 21:28 [dpdk-stable] [PATCH] lpm: fix buffer overflow Vladimir Medvedkin
2021-10-20 19:55 ` David Marchand
2021-10-21 17:15   ` Medvedkin, Vladimir
2021-10-21 17:15 ` [dpdk-stable] [PATCH v2] " Vladimir Medvedkin
2021-10-22  9:07   ` Bruce Richardson
2021-10-25 17:10     ` Thomas Monjalon

patches for DPDK stable branches

This inbox may be cloned and mirrored by anyone:

	git clone --mirror http://inbox.dpdk.org/stable/0 stable/git/0.git

	# If you have public-inbox 1.1+ installed, you may
	# initialize and index your mirror using the following commands:
	public-inbox-init -V2 stable stable/ http://inbox.dpdk.org/stable \
		stable@dpdk.org
	public-inbox-index stable

Example config snippet for mirrors.
Newsgroup available over NNTP:
	nntp://inbox.dpdk.org/inbox.dpdk.stable


AGPL code for this site: git clone https://public-inbox.org/public-inbox.git