[PATCH] Free cookie before we memdup a new one

[PATCH] Free cookie before we memdup a new one

[Neil Horman]

Based on comments from Xin, even after fixes for our recent syzbot
report of cookie memory leaks, its possible to get a resend of an INIT
chunk which would lead to us leaking cookie memory.

To ensure that we don't leak cookie memory, free any previously
allocated cookie first.

Signed-off-by: Neil Horman <nhorman@tuxdriver.com>
CC: Marcelo Ricardo Leitner <marcelo.leitner@gmail.com>
CC: Xin Long <lucien.xin@gmail.com>
CC: "David S. Miller" <davem@davemloft.net>
CC: netdev@vger.kernel.org
---
 net/sctp/sm_make_chunk.c | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/net/sctp/sm_make_chunk.c b/net/sctp/sm_make_chunk.c
index f17908f5c4f3..21f7faf032e5 100644
--- a/net/sctp/sm_make_chunk.c
+++ b/net/sctp/sm_make_chunk.c
@@ -2583,6 +2583,8 @@ static int sctp_process_param(struct sctp_association *asoc,
 	case SCTP_PARAM_STATE_COOKIE:
 		asoc->peer.cookie_len =
 			ntohs(param.p->length) - sizeof(struct sctp_paramhdr);
+		if (asoc->peer.cookie)
+			kfree(asoc->peer.cookie);
 		asoc->peer.cookie = kmemdup(param.cookie->body, asoc->peer.cookie_len, gfp);
 		if (!asoc->peer.cookie)
 			retval = 0;
-- 
2.20.1

Re: [PATCH] Free cookie before we memdup a new one

[David Miller]

From: Neil Horman <nhorman@tuxdriver.com>
Date: Mon, 10 Jun 2019 12:34:56 -0400

> Based on comments from Xin, even after fixes for our recent syzbot
> report of cookie memory leaks, its possible to get a resend of an INIT
> chunk which would lead to us leaking cookie memory.
> 
> To ensure that we don't leak cookie memory, free any previously
> allocated cookie first.
> 
> Signed-off-by: Neil Horman <nhorman@tuxdriver.com>

Please post with an appropriate subsystem prefix in the Subject.

Thank you.

Re: [PATCH] Free cookie before we memdup a new one

[Xin Long]

On Tue, Jun 11, 2019 at 12:35 AM Neil Horman <nhorman@tuxdriver.com> wrote:
>
> Based on comments from Xin, even after fixes for our recent syzbot
> report of cookie memory leaks, its possible to get a resend of an INIT
> chunk which would lead to us leaking cookie memory.
>
> To ensure that we don't leak cookie memory, free any previously
> allocated cookie first.
Hi, Neil,

I think we should also do the same thing to the other 2 kmemdups in
sctp_process_param():
asoc->peer.peer_random
asoc->peer.peer_hmacs


>
> Signed-off-by: Neil Horman <nhorman@tuxdriver.com>
> CC: Marcelo Ricardo Leitner <marcelo.leitner@gmail.com>
> CC: Xin Long <lucien.xin@gmail.com>
> CC: "David S. Miller" <davem@davemloft.net>
> CC: netdev@vger.kernel.org
> ---
>  net/sctp/sm_make_chunk.c | 2 ++
>  1 file changed, 2 insertions(+)
>
> diff --git a/net/sctp/sm_make_chunk.c b/net/sctp/sm_make_chunk.c
> index f17908f5c4f3..21f7faf032e5 100644
> --- a/net/sctp/sm_make_chunk.c
> +++ b/net/sctp/sm_make_chunk.c
> @@ -2583,6 +2583,8 @@ static int sctp_process_param(struct sctp_association *asoc,
>         case SCTP_PARAM_STATE_COOKIE:
>                 asoc->peer.cookie_len =
>                         ntohs(param.p->length) - sizeof(struct sctp_paramhdr);
> +               if (asoc->peer.cookie)
> +                       kfree(asoc->peer.cookie);
>                 asoc->peer.cookie = kmemdup(param.cookie->body, asoc->peer.cookie_len, gfp);
>                 if (!asoc->peer.cookie)
>                         retval = 0;
> --
> 2.20.1
>

[PATCH v2] [net] Free cookie before we memdup a new one

[Neil Horman]

Based on comments from Xin, even after fixes for our recent syzbot
report of cookie memory leaks, its possible to get a resend of an INIT
chunk which would lead to us leaking cookie memory.

To ensure that we don't leak cookie memory, free any previously
allocated cookie first.

---
Change notes
v1->v2
update subsystem tag in subject (davem)
repeat kfree check for peer_random and peer_hmacs (xin)

Signed-off-by: Neil Horman <nhorman@tuxdriver.com>
CC: Marcelo Ricardo Leitner <marcelo.leitner@gmail.com>
CC: Xin Long <lucien.xin@gmail.com>
CC: "David S. Miller" <davem@davemloft.net>
CC: netdev@vger.kernel.org
---
 net/sctp/sm_make_chunk.c | 6 ++++++
 1 file changed, 6 insertions(+)

diff --git a/net/sctp/sm_make_chunk.c b/net/sctp/sm_make_chunk.c
index f17908f5c4f3..0992ec0395f8 100644
--- a/net/sctp/sm_make_chunk.c
+++ b/net/sctp/sm_make_chunk.c
@@ -2583,6 +2583,8 @@ static int sctp_process_param(struct sctp_association *asoc,
 	case SCTP_PARAM_STATE_COOKIE:
 		asoc->peer.cookie_len =
 			ntohs(param.p->length) - sizeof(struct sctp_paramhdr);
+		if (asoc->peer.cookie)
+			kfree(asoc->peer.cookie);
 		asoc->peer.cookie = kmemdup(param.cookie->body, asoc->peer.cookie_len, gfp);
 		if (!asoc->peer.cookie)
 			retval = 0;
@@ -2647,6 +2649,8 @@ static int sctp_process_param(struct sctp_association *asoc,
 			goto fall_through;
 
 		/* Save peer's random parameter */
+		if (asoc->peer.peer_random)
+			kfree(asoc->peer.peer_random);
 		asoc->peer.peer_random = kmemdup(param.p,
 					    ntohs(param.p->length), gfp);
 		if (!asoc->peer.peer_random) {
@@ -2660,6 +2664,8 @@ static int sctp_process_param(struct sctp_association *asoc,
 			goto fall_through;
 
 		/* Save peer's HMAC list */
+		if (asoc->peer.peer_hmacs)
+			kfree(asoc->peer.peer_hmacs);
 		asoc->peer.peer_hmacs = kmemdup(param.p,
 					    ntohs(param.p->length), gfp);
 		if (!asoc->peer.peer_hmacs) {
-- 
2.20.1

Re: [PATCH v2] [net] Free cookie before we memdup a new one

[Marcelo Ricardo Leitner]

On Tue, Jun 11, 2019 at 07:21:28AM -0400, Neil Horman wrote:
> Based on comments from Xin, even after fixes for our recent syzbot
> report of cookie memory leaks, its possible to get a resend of an INIT
> chunk which would lead to us leaking cookie memory.
> 
> To ensure that we don't leak cookie memory, free any previously
> allocated cookie first.
> 
> ---
> Change notes
> v1->v2
> update subsystem tag in subject (davem)
> repeat kfree check for peer_random and peer_hmacs (xin)

They are actually 4 vars. The 4th one being peer_chunks.
And a syzkaller Fixes tag would be welcomed as well, so that if
someone backports the fix for it will have a hint to backport this
patch also.

> 
> Signed-off-by: Neil Horman <nhorman@tuxdriver.com>
> CC: Marcelo Ricardo Leitner <marcelo.leitner@gmail.com>
> CC: Xin Long <lucien.xin@gmail.com>
> CC: "David S. Miller" <davem@davemloft.net>
> CC: netdev@vger.kernel.org
> ---
>  net/sctp/sm_make_chunk.c | 6 ++++++
>  1 file changed, 6 insertions(+)
> 
> diff --git a/net/sctp/sm_make_chunk.c b/net/sctp/sm_make_chunk.c
> index f17908f5c4f3..0992ec0395f8 100644
> --- a/net/sctp/sm_make_chunk.c
> +++ b/net/sctp/sm_make_chunk.c
> @@ -2583,6 +2583,8 @@ static int sctp_process_param(struct sctp_association *asoc,
>  	case SCTP_PARAM_STATE_COOKIE:
>  		asoc->peer.cookie_len =
>  			ntohs(param.p->length) - sizeof(struct sctp_paramhdr);
> +		if (asoc->peer.cookie)
> +			kfree(asoc->peer.cookie);
>  		asoc->peer.cookie = kmemdup(param.cookie->body, asoc->peer.cookie_len, gfp);
>  		if (!asoc->peer.cookie)
>  			retval = 0;
> @@ -2647,6 +2649,8 @@ static int sctp_process_param(struct sctp_association *asoc,
>  			goto fall_through;
>  
>  		/* Save peer's random parameter */
> +		if (asoc->peer.peer_random)
> +			kfree(asoc->peer.peer_random);
>  		asoc->peer.peer_random = kmemdup(param.p,
>  					    ntohs(param.p->length), gfp);
>  		if (!asoc->peer.peer_random) {
> @@ -2660,6 +2664,8 @@ static int sctp_process_param(struct sctp_association *asoc,
>  			goto fall_through;
>  
>  		/* Save peer's HMAC list */
> +		if (asoc->peer.peer_hmacs)
> +			kfree(asoc->peer.peer_hmacs);
>  		asoc->peer.peer_hmacs = kmemdup(param.p,
>  					    ntohs(param.p->length), gfp);
>  		if (!asoc->peer.peer_hmacs) {
> -- 
> 2.20.1
> 

Re: [PATCH v2] [net] Free cookie before we memdup a new one

[Marcelo Ricardo Leitner]

On Tue, Jun 11, 2019 at 07:21:28AM -0400, Neil Horman wrote:

Btw, I guess DaveM had meant to add "sctp: " in the subject.

  Marcelo

Re: [PATCH v2] [net] Free cookie before we memdup a new one

[David Miller]

From: Neil Horman <nhorman@tuxdriver.com>
Date: Tue, 11 Jun 2019 07:21:28 -0400

> Based on comments from Xin, even after fixes for our recent syzbot
> report of cookie memory leaks, its possible to get a resend of an INIT
> chunk which would lead to us leaking cookie memory.
> 
> To ensure that we don't leak cookie memory, free any previously
> allocated cookie first.
> 
> ---
> Change notes
> v1->v2
> update subsystem tag in subject (davem)

Subsystem tag is "sctp: " which still isn't there :-)

[PATCH v3] [sctp] Free cookie before we memdup a new one

[Neil Horman]

Based on comments from Xin, even after fixes for our recent syzbot
report of cookie memory leaks, its possible to get a resend of an INIT
chunk which would lead to us leaking cookie memory.

To ensure that we don't leak cookie memory, free any previously
allocated cookie first.

---
Change notes
v1->v2
update subsystem tag in subject (davem)
repeat kfree check for peer_random and peer_hmacs (xin)

v2->v3
net->sctp
also free peer_chunks

Signed-off-by: Neil Horman <nhorman@tuxdriver.com>
Reported-by: syzbot+f7e9153b037eac9b1df8@syzkaller.appspotmail.com
CC: Marcelo Ricardo Leitner <marcelo.leitner@gmail.com>
CC: Xin Long <lucien.xin@gmail.com>
CC: "David S. Miller" <davem@davemloft.net>
CC: netdev@vger.kernel.org
---
 net/sctp/sm_make_chunk.c | 8 ++++++++
 1 file changed, 8 insertions(+)

diff --git a/net/sctp/sm_make_chunk.c b/net/sctp/sm_make_chunk.c
index f17908f5c4f3..9b0e5b0d701a 100644
--- a/net/sctp/sm_make_chunk.c
+++ b/net/sctp/sm_make_chunk.c
@@ -2583,6 +2583,8 @@ static int sctp_process_param(struct sctp_association *asoc,
 	case SCTP_PARAM_STATE_COOKIE:
 		asoc->peer.cookie_len =
 			ntohs(param.p->length) - sizeof(struct sctp_paramhdr);
+		if (asoc->peer.cookie)
+			kfree(asoc->peer.cookie);
 		asoc->peer.cookie = kmemdup(param.cookie->body, asoc->peer.cookie_len, gfp);
 		if (!asoc->peer.cookie)
 			retval = 0;
@@ -2647,6 +2649,8 @@ static int sctp_process_param(struct sctp_association *asoc,
 			goto fall_through;
 
 		/* Save peer's random parameter */
+		if (asoc->peer.peer_random)
+			kfree(asoc->peer.peer_random);
 		asoc->peer.peer_random = kmemdup(param.p,
 					    ntohs(param.p->length), gfp);
 		if (!asoc->peer.peer_random) {
@@ -2660,6 +2664,8 @@ static int sctp_process_param(struct sctp_association *asoc,
 			goto fall_through;
 
 		/* Save peer's HMAC list */
+		if (asoc->peer.peer_hmacs)
+			kfree(asoc->peer.peer_hmacs);
 		asoc->peer.peer_hmacs = kmemdup(param.p,
 					    ntohs(param.p->length), gfp);
 		if (!asoc->peer.peer_hmacs) {
@@ -2675,6 +2681,8 @@ static int sctp_process_param(struct sctp_association *asoc,
 		if (!ep->auth_enable)
 			goto fall_through;
 
+		if (asoc->peer.peer_chunks)
+			kfree(asoc->peer.peer_chunks);
 		asoc->peer.peer_chunks = kmemdup(param.p,
 					    ntohs(param.p->length), gfp);
 		if (!asoc->peer.peer_chunks)
-- 
2.20.1

Re: [PATCH v3] [sctp] Free cookie before we memdup a new one

[David Miller]

From: Neil Horman <nhorman@tuxdriver.com>
Date: Tue, 11 Jun 2019 15:22:45 -0400

> v2->v3
> net->sctp
> also free peer_chunks

Neil this isn't the first time you're submitting sctp patches right? ;-)

Subject: "[PATCH v3 net] sctp: Free cookie before we memdup a new one"

It's "subsystem_prefix: " and I even stated this explicitly yesterday.

Re: [PATCH v3] [sctp] Free cookie before we memdup a new one

[Neil Horman]

On Tue, Jun 11, 2019 at 01:08:56PM -0700, David Miller wrote:
> From: Neil Horman <nhorman@tuxdriver.com>
> Date: Tue, 11 Jun 2019 15:22:45 -0400
> 
> > v2->v3
> > net->sctp
> > also free peer_chunks
> 
> Neil this isn't the first time you're submitting sctp patches right? ;-)
> 
> Subject: "[PATCH v3 net] sctp: Free cookie before we memdup a new one"
> 
> It's "subsystem_prefix: " and I even stated this explicitly yesterday.
> 
No, sorry, I'm trying to do this while I get a CI mess sorted out at the same
time, and I'm not close attention here.

If you want to take over a few dozen user space packages for me in RHEL, I'll
gladly get this right on the first try :)

Neil

[PATCH v4 net] sctp: Free cookie before we memdup a new one

[Neil Horman]

Based on comments from Xin, even after fixes for our recent syzbot
report of cookie memory leaks, its possible to get a resend of an INIT
chunk which would lead to us leaking cookie memory.

To ensure that we don't leak cookie memory, free any previously
allocated cookie first.

---
Change notes
v1->v2
update subsystem tag in subject (davem)
repeat kfree check for peer_random and peer_hmacs (xin)

v2->v3
net->sctp
also free peer_chunks

v3->v4
fix subject tags

Signed-off-by: Neil Horman <nhorman@tuxdriver.com>
Reported-by: syzbot+f7e9153b037eac9b1df8@syzkaller.appspotmail.com
CC: Marcelo Ricardo Leitner <marcelo.leitner@gmail.com>
CC: Xin Long <lucien.xin@gmail.com>
CC: "David S. Miller" <davem@davemloft.net>
CC: netdev@vger.kernel.org
---
 net/sctp/sm_make_chunk.c | 8 ++++++++
 1 file changed, 8 insertions(+)

diff --git a/net/sctp/sm_make_chunk.c b/net/sctp/sm_make_chunk.c
index f17908f5c4f3..9b0e5b0d701a 100644
--- a/net/sctp/sm_make_chunk.c
+++ b/net/sctp/sm_make_chunk.c
@@ -2583,6 +2583,8 @@ static int sctp_process_param(struct sctp_association *asoc,
 	case SCTP_PARAM_STATE_COOKIE:
 		asoc->peer.cookie_len =
 			ntohs(param.p->length) - sizeof(struct sctp_paramhdr);
+		if (asoc->peer.cookie)
+			kfree(asoc->peer.cookie);
 		asoc->peer.cookie = kmemdup(param.cookie->body, asoc->peer.cookie_len, gfp);
 		if (!asoc->peer.cookie)
 			retval = 0;
@@ -2647,6 +2649,8 @@ static int sctp_process_param(struct sctp_association *asoc,
 			goto fall_through;
 
 		/* Save peer's random parameter */
+		if (asoc->peer.peer_random)
+			kfree(asoc->peer.peer_random);
 		asoc->peer.peer_random = kmemdup(param.p,
 					    ntohs(param.p->length), gfp);
 		if (!asoc->peer.peer_random) {
@@ -2660,6 +2664,8 @@ static int sctp_process_param(struct sctp_association *asoc,
 			goto fall_through;
 
 		/* Save peer's HMAC list */
+		if (asoc->peer.peer_hmacs)
+			kfree(asoc->peer.peer_hmacs);
 		asoc->peer.peer_hmacs = kmemdup(param.p,
 					    ntohs(param.p->length), gfp);
 		if (!asoc->peer.peer_hmacs) {
@@ -2675,6 +2681,8 @@ static int sctp_process_param(struct sctp_association *asoc,
 		if (!ep->auth_enable)
 			goto fall_through;
 
+		if (asoc->peer.peer_chunks)
+			kfree(asoc->peer.peer_chunks);
 		asoc->peer.peer_chunks = kmemdup(param.p,
 					    ntohs(param.p->length), gfp);
 		if (!asoc->peer.peer_chunks)
-- 
2.20.1

Re: [PATCH v4 net] sctp: Free cookie before we memdup a new one

[Xin Long]

On Wed, Jun 12, 2019 at 8:38 AM Neil Horman <nhorman@tuxdriver.com> wrote:
>
> Based on comments from Xin, even after fixes for our recent syzbot
> report of cookie memory leaks, its possible to get a resend of an INIT
> chunk which would lead to us leaking cookie memory.
>
> To ensure that we don't leak cookie memory, free any previously
> allocated cookie first.
>
> ---
> Change notes
> v1->v2
> update subsystem tag in subject (davem)
> repeat kfree check for peer_random and peer_hmacs (xin)
>
> v2->v3
> net->sctp
> also free peer_chunks
>
> v3->v4
> fix subject tags
>
> Signed-off-by: Neil Horman <nhorman@tuxdriver.com>
> Reported-by: syzbot+f7e9153b037eac9b1df8@syzkaller.appspotmail.com
> CC: Marcelo Ricardo Leitner <marcelo.leitner@gmail.com>
> CC: Xin Long <lucien.xin@gmail.com>
> CC: "David S. Miller" <davem@davemloft.net>
> CC: netdev@vger.kernel.org
> ---
>  net/sctp/sm_make_chunk.c | 8 ++++++++
>  1 file changed, 8 insertions(+)
>
> diff --git a/net/sctp/sm_make_chunk.c b/net/sctp/sm_make_chunk.c
> index f17908f5c4f3..9b0e5b0d701a 100644
> --- a/net/sctp/sm_make_chunk.c
> +++ b/net/sctp/sm_make_chunk.c
> @@ -2583,6 +2583,8 @@ static int sctp_process_param(struct sctp_association *asoc,
>         case SCTP_PARAM_STATE_COOKIE:
>                 asoc->peer.cookie_len =
>                         ntohs(param.p->length) - sizeof(struct sctp_paramhdr);
> +               if (asoc->peer.cookie)
> +                       kfree(asoc->peer.cookie);
>                 asoc->peer.cookie = kmemdup(param.cookie->body, asoc->peer.cookie_len, gfp);
>                 if (!asoc->peer.cookie)
>                         retval = 0;
> @@ -2647,6 +2649,8 @@ static int sctp_process_param(struct sctp_association *asoc,
>                         goto fall_through;
>
>                 /* Save peer's random parameter */
> +               if (asoc->peer.peer_random)
> +                       kfree(asoc->peer.peer_random);
>                 asoc->peer.peer_random = kmemdup(param.p,
>                                             ntohs(param.p->length), gfp);
>                 if (!asoc->peer.peer_random) {
> @@ -2660,6 +2664,8 @@ static int sctp_process_param(struct sctp_association *asoc,
>                         goto fall_through;
>
>                 /* Save peer's HMAC list */
> +               if (asoc->peer.peer_hmacs)
> +                       kfree(asoc->peer.peer_hmacs);
>                 asoc->peer.peer_hmacs = kmemdup(param.p,
>                                             ntohs(param.p->length), gfp);
>                 if (!asoc->peer.peer_hmacs) {
> @@ -2675,6 +2681,8 @@ static int sctp_process_param(struct sctp_association *asoc,
>                 if (!ep->auth_enable)
>                         goto fall_through;
>
> +               if (asoc->peer.peer_chunks)
> +                       kfree(asoc->peer.peer_chunks);
>                 asoc->peer.peer_chunks = kmemdup(param.p,
>                                             ntohs(param.p->length), gfp);
>                 if (!asoc->peer.peer_chunks)
> --
> 2.20.1
>
Reviewed-by: Xin Long <lucien.xin@gmail.com>

Re: [PATCH v4 net] sctp: Free cookie before we memdup a new one

[Marcelo Ricardo Leitner]

On Tue, Jun 11, 2019 at 08:38:14PM -0400, Neil Horman wrote:
> Based on comments from Xin, even after fixes for our recent syzbot
> report of cookie memory leaks, its possible to get a resend of an INIT
> chunk which would lead to us leaking cookie memory.
> 
> To ensure that we don't leak cookie memory, free any previously
> allocated cookie first.
> 
> ---

This marker can't be here, as it causes git to loose everything below.

> Change notes
> v1->v2
> update subsystem tag in subject (davem)
> repeat kfree check for peer_random and peer_hmacs (xin)
> 
> v2->v3
> net->sctp
> also free peer_chunks
> 
> v3->v4
> fix subject tags
> 
> Signed-off-by: Neil Horman <nhorman@tuxdriver.com>
> Reported-by: syzbot+f7e9153b037eac9b1df8@syzkaller.appspotmail.com
> CC: Marcelo Ricardo Leitner <marcelo.leitner@gmail.com>
> CC: Xin Long <lucien.xin@gmail.com>
> CC: "David S. Miller" <davem@davemloft.net>
> CC: netdev@vger.kernel.org

Anyhow, LGTM and reproducer didn't give any hits in 2 runs of 50mins.
Acked-by: Marcelo Ricardo Leitner <marcelo.leitner@gmail.com>

Re: [PATCH v4 net] sctp: Free cookie before we memdup a new one

[Neil Horman]

On Wed, Jun 12, 2019 at 03:07:15PM -0300, Marcelo Ricardo Leitner wrote:
> On Tue, Jun 11, 2019 at 08:38:14PM -0400, Neil Horman wrote:
> > Based on comments from Xin, even after fixes for our recent syzbot
> > report of cookie memory leaks, its possible to get a resend of an INIT
> > chunk which would lead to us leaking cookie memory.
> > 
> > To ensure that we don't leak cookie memory, free any previously
> > allocated cookie first.
> > 
> > ---
> 
> This marker can't be here, as it causes git to loose everything below.
> 
thats intentional so that, when Dave commits it, the change notes arent carried
into the changelog (I.e. the change notes are useful for email review, but not
especially useful in the permanent commit history).

Neil

> > Change notes
> > v1->v2
> > update subsystem tag in subject (davem)
> > repeat kfree check for peer_random and peer_hmacs (xin)
> > 
> > v2->v3
> > net->sctp
> > also free peer_chunks
> > 
> > v3->v4
> > fix subject tags
> > 
> > Signed-off-by: Neil Horman <nhorman@tuxdriver.com>
> > Reported-by: syzbot+f7e9153b037eac9b1df8@syzkaller.appspotmail.com
> > CC: Marcelo Ricardo Leitner <marcelo.leitner@gmail.com>
> > CC: Xin Long <lucien.xin@gmail.com>
> > CC: "David S. Miller" <davem@davemloft.net>
> > CC: netdev@vger.kernel.org
> 
> Anyhow, LGTM and reproducer didn't give any hits in 2 runs of 50mins.
> Acked-by: Marcelo Ricardo Leitner <marcelo.leitner@gmail.com>
> 

Re: [PATCH v4 net] sctp: Free cookie before we memdup a new one

[David Miller]

From: Neil Horman <nhorman@tuxdriver.com>
Date: Wed, 12 Jun 2019 16:32:30 -0400

> On Wed, Jun 12, 2019 at 03:07:15PM -0300, Marcelo Ricardo Leitner wrote:
>> On Tue, Jun 11, 2019 at 08:38:14PM -0400, Neil Horman wrote:
>> > Based on comments from Xin, even after fixes for our recent syzbot
>> > report of cookie memory leaks, its possible to get a resend of an INIT
>> > chunk which would lead to us leaking cookie memory.
>> > 
>> > To ensure that we don't leak cookie memory, free any previously
>> > allocated cookie first.
>> > 
>> > ---
>> 
>> This marker can't be here, as it causes git to loose everything below.
>> 
> thats intentional so that, when Dave commits it, the change notes arent carried
> into the changelog (I.e. the change notes are useful for email review, but not
> especially useful in the permanent commit history).

1) I like the changelog notes to be included

2) Your signoff etc. was in that area too so would have been lost as well

[PATCH v5 net] sctp: Free cookie before we memdup a new one

[Neil Horman]

Based on comments from Xin, even after fixes for our recent syzbot
report of cookie memory leaks, its possible to get a resend of an INIT
chunk which would lead to us leaking cookie memory.

To ensure that we don't leak cookie memory, free any previously
allocated cookie first.

Change notes
v1->v2
update subsystem tag in subject (davem)
repeat kfree check for peer_random and peer_hmacs (xin)

v2->v3
net->sctp
also free peer_chunks

v3->v4
fix subject tags

v4->v5
remove cut line

Signed-off-by: Neil Horman <nhorman@tuxdriver.com>
Reported-by: syzbot+f7e9153b037eac9b1df8@syzkaller.appspotmail.com
CC: Marcelo Ricardo Leitner <marcelo.leitner@gmail.com>
CC: Xin Long <lucien.xin@gmail.com>
CC: "David S. Miller" <davem@davemloft.net>
CC: netdev@vger.kernel.org
---
 net/sctp/sm_make_chunk.c | 8 ++++++++
 1 file changed, 8 insertions(+)

diff --git a/net/sctp/sm_make_chunk.c b/net/sctp/sm_make_chunk.c
index f17908f5c4f3..9b0e5b0d701a 100644
--- a/net/sctp/sm_make_chunk.c
+++ b/net/sctp/sm_make_chunk.c
@@ -2583,6 +2583,8 @@ static int sctp_process_param(struct sctp_association *asoc,
 	case SCTP_PARAM_STATE_COOKIE:
 		asoc->peer.cookie_len =
 			ntohs(param.p->length) - sizeof(struct sctp_paramhdr);
+		if (asoc->peer.cookie)
+			kfree(asoc->peer.cookie);
 		asoc->peer.cookie = kmemdup(param.cookie->body, asoc->peer.cookie_len, gfp);
 		if (!asoc->peer.cookie)
 			retval = 0;
@@ -2647,6 +2649,8 @@ static int sctp_process_param(struct sctp_association *asoc,
 			goto fall_through;
 
 		/* Save peer's random parameter */
+		if (asoc->peer.peer_random)
+			kfree(asoc->peer.peer_random);
 		asoc->peer.peer_random = kmemdup(param.p,
 					    ntohs(param.p->length), gfp);
 		if (!asoc->peer.peer_random) {
@@ -2660,6 +2664,8 @@ static int sctp_process_param(struct sctp_association *asoc,
 			goto fall_through;
 
 		/* Save peer's HMAC list */
+		if (asoc->peer.peer_hmacs)
+			kfree(asoc->peer.peer_hmacs);
 		asoc->peer.peer_hmacs = kmemdup(param.p,
 					    ntohs(param.p->length), gfp);
 		if (!asoc->peer.peer_hmacs) {
@@ -2675,6 +2681,8 @@ static int sctp_process_param(struct sctp_association *asoc,
 		if (!ep->auth_enable)
 			goto fall_through;
 
+		if (asoc->peer.peer_chunks)
+			kfree(asoc->peer.peer_chunks);
 		asoc->peer.peer_chunks = kmemdup(param.p,
 					    ntohs(param.p->length), gfp);
 		if (!asoc->peer.peer_chunks)
-- 
2.20.1

Re: [PATCH v5 net] sctp: Free cookie before we memdup a new one

[Marcelo Ricardo Leitner]

On Thu, Jun 13, 2019 at 06:35:59AM -0400, Neil Horman wrote:
> Based on comments from Xin, even after fixes for our recent syzbot
> report of cookie memory leaks, its possible to get a resend of an INIT
> chunk which would lead to us leaking cookie memory.
> 
> To ensure that we don't leak cookie memory, free any previously
> allocated cookie first.
> 
> Change notes
> v1->v2
> update subsystem tag in subject (davem)
> repeat kfree check for peer_random and peer_hmacs (xin)
> 
> v2->v3
> net->sctp
> also free peer_chunks
> 
> v3->v4
> fix subject tags
> 
> v4->v5
> remove cut line
> 
> Signed-off-by: Neil Horman <nhorman@tuxdriver.com>
> Reported-by: syzbot+f7e9153b037eac9b1df8@syzkaller.appspotmail.com
> CC: Marcelo Ricardo Leitner <marcelo.leitner@gmail.com>
> CC: Xin Long <lucien.xin@gmail.com>
> CC: "David S. Miller" <davem@davemloft.net>
> CC: netdev@vger.kernel.org

Acked-by: Marcelo Ricardo Leitner <marcelo.leitner@gmail.com>

> ---
>  net/sctp/sm_make_chunk.c | 8 ++++++++
>  1 file changed, 8 insertions(+)
> 
> diff --git a/net/sctp/sm_make_chunk.c b/net/sctp/sm_make_chunk.c
> index f17908f5c4f3..9b0e5b0d701a 100644
> --- a/net/sctp/sm_make_chunk.c
> +++ b/net/sctp/sm_make_chunk.c
> @@ -2583,6 +2583,8 @@ static int sctp_process_param(struct sctp_association *asoc,
>  	case SCTP_PARAM_STATE_COOKIE:
>  		asoc->peer.cookie_len =
>  			ntohs(param.p->length) - sizeof(struct sctp_paramhdr);
> +		if (asoc->peer.cookie)
> +			kfree(asoc->peer.cookie);
>  		asoc->peer.cookie = kmemdup(param.cookie->body, asoc->peer.cookie_len, gfp);
>  		if (!asoc->peer.cookie)
>  			retval = 0;
> @@ -2647,6 +2649,8 @@ static int sctp_process_param(struct sctp_association *asoc,
>  			goto fall_through;
>  
>  		/* Save peer's random parameter */
> +		if (asoc->peer.peer_random)
> +			kfree(asoc->peer.peer_random);
>  		asoc->peer.peer_random = kmemdup(param.p,
>  					    ntohs(param.p->length), gfp);
>  		if (!asoc->peer.peer_random) {
> @@ -2660,6 +2664,8 @@ static int sctp_process_param(struct sctp_association *asoc,
>  			goto fall_through;
>  
>  		/* Save peer's HMAC list */
> +		if (asoc->peer.peer_hmacs)
> +			kfree(asoc->peer.peer_hmacs);
>  		asoc->peer.peer_hmacs = kmemdup(param.p,
>  					    ntohs(param.p->length), gfp);
>  		if (!asoc->peer.peer_hmacs) {
> @@ -2675,6 +2681,8 @@ static int sctp_process_param(struct sctp_association *asoc,
>  		if (!ep->auth_enable)
>  			goto fall_through;
>  
> +		if (asoc->peer.peer_chunks)
> +			kfree(asoc->peer.peer_chunks);
>  		asoc->peer.peer_chunks = kmemdup(param.p,
>  					    ntohs(param.p->length), gfp);
>  		if (!asoc->peer.peer_chunks)
> -- 
> 2.20.1
> 

Re: [PATCH v5 net] sctp: Free cookie before we memdup a new one

[David Miller]

From: Neil Horman <nhorman@tuxdriver.com>
Date: Thu, 13 Jun 2019 06:35:59 -0400

> Based on comments from Xin, even after fixes for our recent syzbot
> report of cookie memory leaks, its possible to get a resend of an INIT
> chunk which would lead to us leaking cookie memory.
> 
> To ensure that we don't leak cookie memory, free any previously
> allocated cookie first.
> 
> Change notes
> v1->v2
> update subsystem tag in subject (davem)
> repeat kfree check for peer_random and peer_hmacs (xin)
> 
> v2->v3
> net->sctp
> also free peer_chunks
> 
> v3->v4
> fix subject tags
> 
> v4->v5
> remove cut line
> 
> Signed-off-by: Neil Horman <nhorman@tuxdriver.com>
> Reported-by: syzbot+f7e9153b037eac9b1df8@syzkaller.appspotmail.com

Applied, thanks.