* [Ocfs2-devel] [PATCH 1/1] O2net: Disallow o2net accept connection request from itself.
@ 2010-08-02 2:00 Tristan Ye
2010-08-02 3:00 ` Sunil Mushran
2010-08-07 18:40 ` Joel Becker
0 siblings, 2 replies; 8+ messages in thread
From: Tristan Ye @ 2010-08-02 2:00 UTC (permalink / raw)
To: ocfs2-devel
Currently, o2net_accept_one() is allowed to accept a connection from
listening node itself, such a fake connection will not be successfully
established due to no handshake detected afterwards, and later end up
with triggering connecting worker in a loop.
We're going to fix this by treating such connection request as 'invalid',
since we've got no chance of requesting connection from a node to itself
in a OCFS2 cluster.
The fix doesn't hurt user's scan for o2net-listener, it always gets a
successful connection from userpace.
Signed-off-by: Tristan Ye <tristan.ye@oracle.com>
Acked-by: Sunil Mushran <sunil.mushran@oracle.com>
---
fs/ocfs2/cluster/tcp.c | 17 ++++++++++++-----
1 files changed, 12 insertions(+), 5 deletions(-)
diff --git a/fs/ocfs2/cluster/tcp.c b/fs/ocfs2/cluster/tcp.c
index aa75ca3..1361997 100644
--- a/fs/ocfs2/cluster/tcp.c
+++ b/fs/ocfs2/cluster/tcp.c
@@ -1759,6 +1759,7 @@ static int o2net_accept_one(struct socket *sock)
struct sockaddr_in sin;
struct socket *new_sock = NULL;
struct o2nm_node *node = NULL;
+ struct o2nm_node *local_node = NULL;
struct o2net_sock_container *sc = NULL;
struct o2net_node *nn;
@@ -1796,11 +1797,15 @@ static int o2net_accept_one(struct socket *sock)
goto out;
}
- if (o2nm_this_node() > node->nd_num) {
- mlog(ML_NOTICE, "unexpected connect attempted from a lower "
- "numbered node '%s' at " "%pI4:%d with num %u\n",
- node->nd_name, &sin.sin_addr.s_addr,
- ntohs(sin.sin_port), node->nd_num);
+ if (o2nm_this_node() >= node->nd_num) {
+ local_node = o2nm_get_node_by_num(o2nm_this_node());
+ mlog(ML_NOTICE, "unexpected connect attempt seen at node '%s' ("
+ "%u, %pI4:%d) from node '%s' (%u, %pI4:%d)\n",
+ local_node->nd_name, local_node->nd_num,
+ &(local_node->nd_ipv4_address),
+ ntohs(local_node->nd_ipv4_port),
+ node->nd_name, node->nd_num, &sin.sin_addr.s_addr,
+ ntohs(sin.sin_port));
ret = -EINVAL;
goto out;
}
@@ -1857,6 +1862,8 @@ out:
sock_release(new_sock);
if (node)
o2nm_node_put(node);
+ if (local_node)
+ o2nm_node_put(local_node);
if (sc)
sc_put(sc);
return ret;
--
1.5.5
^ permalink raw reply related [flat|nested] 8+ messages in thread
* [Ocfs2-devel] [PATCH 1/1] O2net: Disallow o2net accept connection request from itself.
2010-08-02 2:00 [Ocfs2-devel] [PATCH 1/1] O2net: Disallow o2net accept connection request from itself Tristan Ye
@ 2010-08-02 3:00 ` Sunil Mushran
2010-08-07 18:40 ` Joel Becker
1 sibling, 0 replies; 8+ messages in thread
From: Sunil Mushran @ 2010-08-02 3:00 UTC (permalink / raw)
To: ocfs2-devel
Acked
On Aug 1, 2010, at 7:00 PM, Tristan Ye <tristan.ye@oracle.com> wrote:
> Currently, o2net_accept_one() is allowed to accept a connection from
> listening node itself, such a fake connection will not be successfully
> established due to no handshake detected afterwards, and later end up
> with triggering connecting worker in a loop.
>
> We're going to fix this by treating such connection request as 'invalid',
> since we've got no chance of requesting connection from a node to itself
> in a OCFS2 cluster.
>
> The fix doesn't hurt user's scan for o2net-listener, it always gets a
> successful connection from userpace.
>
> Signed-off-by: Tristan Ye <tristan.ye@oracle.com>
> Acked-by: Sunil Mushran <sunil.mushran@oracle.com>
> ---
> fs/ocfs2/cluster/tcp.c | 17 ++++++++++++-----
> 1 files changed, 12 insertions(+), 5 deletions(-)
>
> diff --git a/fs/ocfs2/cluster/tcp.c b/fs/ocfs2/cluster/tcp.c
> index aa75ca3..1361997 100644
> --- a/fs/ocfs2/cluster/tcp.c
> +++ b/fs/ocfs2/cluster/tcp.c
> @@ -1759,6 +1759,7 @@ static int o2net_accept_one(struct socket *sock)
> struct sockaddr_in sin;
> struct socket *new_sock = NULL;
> struct o2nm_node *node = NULL;
> + struct o2nm_node *local_node = NULL;
> struct o2net_sock_container *sc = NULL;
> struct o2net_node *nn;
>
> @@ -1796,11 +1797,15 @@ static int o2net_accept_one(struct socket *sock)
> goto out;
> }
>
> - if (o2nm_this_node() > node->nd_num) {
> - mlog(ML_NOTICE, "unexpected connect attempted from a lower "
> - "numbered node '%s' at " "%pI4:%d with num %u\n",
> - node->nd_name, &sin.sin_addr.s_addr,
> - ntohs(sin.sin_port), node->nd_num);
> + if (o2nm_this_node() >= node->nd_num) {
> + local_node = o2nm_get_node_by_num(o2nm_this_node());
> + mlog(ML_NOTICE, "unexpected connect attempt seen at node '%s' ("
> + "%u, %pI4:%d) from node '%s' (%u, %pI4:%d)\n",
> + local_node->nd_name, local_node->nd_num,
> + &(local_node->nd_ipv4_address),
> + ntohs(local_node->nd_ipv4_port),
> + node->nd_name, node->nd_num, &sin.sin_addr.s_addr,
> + ntohs(sin.sin_port));
> ret = -EINVAL;
> goto out;
> }
> @@ -1857,6 +1862,8 @@ out:
> sock_release(new_sock);
> if (node)
> o2nm_node_put(node);
> + if (local_node)
> + o2nm_node_put(local_node);
> if (sc)
> sc_put(sc);
> return ret;
> --
> 1.5.5
>
>
> _______________________________________________
> Ocfs2-devel mailing list
> Ocfs2-devel at oss.oracle.com
> http://oss.oracle.com/mailman/listinfo/ocfs2-devel
^ permalink raw reply [flat|nested] 8+ messages in thread
* [Ocfs2-devel] [PATCH 1/1] O2net: Disallow o2net accept connection request from itself.
2010-08-02 2:00 [Ocfs2-devel] [PATCH 1/1] O2net: Disallow o2net accept connection request from itself Tristan Ye
2010-08-02 3:00 ` Sunil Mushran
@ 2010-08-07 18:40 ` Joel Becker
1 sibling, 0 replies; 8+ messages in thread
From: Joel Becker @ 2010-08-07 18:40 UTC (permalink / raw)
To: ocfs2-devel
On Mon, Aug 02, 2010 at 10:00:26AM +0800, Tristan Ye wrote:
> Currently, o2net_accept_one() is allowed to accept a connection from
> listening node itself, such a fake connection will not be successfully
> established due to no handshake detected afterwards, and later end up
> with triggering connecting worker in a loop.
>
> We're going to fix this by treating such connection request as 'invalid',
> since we've got no chance of requesting connection from a node to itself
> in a OCFS2 cluster.
>
> The fix doesn't hurt user's scan for o2net-listener, it always gets a
> successful connection from userpace.
>
> Signed-off-by: Tristan Ye <tristan.ye@oracle.com>
> Acked-by: Sunil Mushran <sunil.mushran@oracle.com>
This patch is now in the fixes branch of ocfs2.git.
Joel
--
"There are only two ways to live your life. One is as though nothing
is a miracle. The other is as though everything is a miracle."
- Albert Einstein
Joel Becker
Consulting Software Developer
Oracle
E-mail: joel.becker at oracle.com
Phone: (650) 506-8127
^ permalink raw reply [flat|nested] 8+ messages in thread
* [Ocfs2-devel] [PATCH 1/1] O2net: Disallow o2net accept connection request from itself.
2010-07-30 8:25 Tristan Ye
@ 2010-07-30 17:27 ` Sunil Mushran
0 siblings, 0 replies; 8+ messages in thread
From: Sunil Mushran @ 2010-07-30 17:27 UTC (permalink / raw)
To: ocfs2-devel
You can tighten it a little more. You can add my Acked-by
to the next patch.
On 07/30/2010 01:25 AM, Tristan Ye wrote:
> Currently, o2net_accept_one() is allowed to accept a connection from
> listening node itself, such a fake connection will not be successfully
> established due to no handshake detected afterwards, and later end up
> with triggering connecting worker in a loop.
>
> We're going to fix this by treating such connection request as 'invalid',
> since we've got no chance of requesting connection from a node to itself
> in a OCFS2 cluster.
>
> The fix doesn't hurt user's scan for o2net-listener, it always gets a
> successful connection from userpace.
>
> Signed-off-by: Tristan Ye<tristan.ye@oracle.com>
> ---
> fs/ocfs2/cluster/tcp.c | 16 +++++++++++-----
> 1 files changed, 11 insertions(+), 5 deletions(-)
>
> diff --git a/fs/ocfs2/cluster/tcp.c b/fs/ocfs2/cluster/tcp.c
> index aa75ca3..7a17e91 100644
> --- a/fs/ocfs2/cluster/tcp.c
> +++ b/fs/ocfs2/cluster/tcp.c
> @@ -1759,6 +1759,7 @@ static int o2net_accept_one(struct socket *sock)
> struct sockaddr_in sin;
> struct socket *new_sock = NULL;
> struct o2nm_node *node = NULL;
> + struct o2nm_node *local_node = o2nm_get_node_by_num(o2nm_this_node());
>
+ struct o2nm_node *local_node = NULL;
> struct o2net_sock_container *sc = NULL;
> struct o2net_node *nn;
>
> @@ -1796,11 +1797,14 @@ static int o2net_accept_one(struct socket *sock)
> goto out;
> }
>
> - if (o2nm_this_node()> node->nd_num) {
> - mlog(ML_NOTICE, "unexpected connect attempted from a lower "
> - "numbered node '%s' at " "%pI4:%d with num %u\n",
> - node->nd_name,&sin.sin_addr.s_addr,
> - ntohs(sin.sin_port), node->nd_num);
> + if (o2nm_this_node()>= node->nd_num) {
>
+ local_node = o2nm_get_node_by_num(o2nm_this_node());
> + mlog(ML_NOTICE, "unexpected connect attempt seen at node '%s' ("
> + "%u, %pI4:%d) from node '%s' (%u, %pI4:%d)\n",
> + local_node->nd_name, local_node->nd_num,
> + &(local_node->nd_ipv4_address),
> + ntohs(local_node->nd_ipv4_port),
> + node->nd_name, node->nd_num,&sin.sin_addr.s_addr,
> + ntohs(sin.sin_port));
> ret = -EINVAL;
> goto out;
> }
> @@ -1857,6 +1861,8 @@ out:
> sock_release(new_sock);
> if (node)
> o2nm_node_put(node);
> + if (local_node)
> + o2nm_node_put(local_node);
> if (sc)
> sc_put(sc);
> return ret;
>
^ permalink raw reply [flat|nested] 8+ messages in thread
* [Ocfs2-devel] [PATCH 1/1] O2net: Disallow o2net accept connection request from itself.
@ 2010-07-30 8:25 Tristan Ye
2010-07-30 17:27 ` Sunil Mushran
0 siblings, 1 reply; 8+ messages in thread
From: Tristan Ye @ 2010-07-30 8:25 UTC (permalink / raw)
To: ocfs2-devel
Currently, o2net_accept_one() is allowed to accept a connection from
listening node itself, such a fake connection will not be successfully
established due to no handshake detected afterwards, and later end up
with triggering connecting worker in a loop.
We're going to fix this by treating such connection request as 'invalid',
since we've got no chance of requesting connection from a node to itself
in a OCFS2 cluster.
The fix doesn't hurt user's scan for o2net-listener, it always gets a
successful connection from userpace.
Signed-off-by: Tristan Ye <tristan.ye@oracle.com>
---
fs/ocfs2/cluster/tcp.c | 16 +++++++++++-----
1 files changed, 11 insertions(+), 5 deletions(-)
diff --git a/fs/ocfs2/cluster/tcp.c b/fs/ocfs2/cluster/tcp.c
index aa75ca3..7a17e91 100644
--- a/fs/ocfs2/cluster/tcp.c
+++ b/fs/ocfs2/cluster/tcp.c
@@ -1759,6 +1759,7 @@ static int o2net_accept_one(struct socket *sock)
struct sockaddr_in sin;
struct socket *new_sock = NULL;
struct o2nm_node *node = NULL;
+ struct o2nm_node *local_node = o2nm_get_node_by_num(o2nm_this_node());
struct o2net_sock_container *sc = NULL;
struct o2net_node *nn;
@@ -1796,11 +1797,14 @@ static int o2net_accept_one(struct socket *sock)
goto out;
}
- if (o2nm_this_node() > node->nd_num) {
- mlog(ML_NOTICE, "unexpected connect attempted from a lower "
- "numbered node '%s' at " "%pI4:%d with num %u\n",
- node->nd_name, &sin.sin_addr.s_addr,
- ntohs(sin.sin_port), node->nd_num);
+ if (o2nm_this_node() >= node->nd_num) {
+ mlog(ML_NOTICE, "unexpected connect attempt seen at node '%s' ("
+ "%u, %pI4:%d) from node '%s' (%u, %pI4:%d)\n",
+ local_node->nd_name, local_node->nd_num,
+ &(local_node->nd_ipv4_address),
+ ntohs(local_node->nd_ipv4_port),
+ node->nd_name, node->nd_num, &sin.sin_addr.s_addr,
+ ntohs(sin.sin_port));
ret = -EINVAL;
goto out;
}
@@ -1857,6 +1861,8 @@ out:
sock_release(new_sock);
if (node)
o2nm_node_put(node);
+ if (local_node)
+ o2nm_node_put(local_node);
if (sc)
sc_put(sc);
return ret;
--
1.5.5
^ permalink raw reply related [flat|nested] 8+ messages in thread
* [Ocfs2-devel] [PATCH 1/1] O2net: Disallow o2net accept connection request from itself.
2010-07-29 12:07 ` Sunil Mushran
@ 2010-07-30 2:12 ` tristan
0 siblings, 0 replies; 8+ messages in thread
From: tristan @ 2010-07-30 2:12 UTC (permalink / raw)
To: ocfs2-devel
Sunil Mushran wrote:
> unexpected connect attempt seen at node x (num, ip:port) from node y (num, ip:port).
>
> The above reads better, don't you think. Lower or same are software bugs. Not a user error.
That makes sense.
>
> Btw, one test you could do is race nc with mount. Run nc in a loop while mounting the first device. nc should ping a node that has it mounted. Test with higher to lower and reverse.
Yeah, to stress and race nc with mount is a good way to test o2net's
listening function.
>
> On Jul 29, 2010, at 4:04 AM, Tristan Ye <tristan.ye@oracle.com> wrote:
>
>> Currently, o2net_accept_one() is allowed to accept a connection from
>> listening node itself, such a fake connection will not be successfully
>> established due to no handshake detected afterwards, and later end up
>> with triggering connecting worker in a loop.
>>
>> We're going to fix this by treating such connection request as 'invalid',
>> since we've got no chance of requesting connection from a node to itself
>> in a OCFS2 cluster.
>>
>> The fix doesn't hurt user's scan for o2net-listener, it always gets a
>> successful connection from userpace.
>>
>> Signed-off-by: Tristan Ye <tristan.ye@oracle.com>
>> ---
>> fs/ocfs2/cluster/tcp.c | 4 ++--
>> 1 files changed, 2 insertions(+), 2 deletions(-)
>>
>> diff --git a/fs/ocfs2/cluster/tcp.c b/fs/ocfs2/cluster/tcp.c
>> index aa75ca3..7e4e451 100644
>> --- a/fs/ocfs2/cluster/tcp.c
>> +++ b/fs/ocfs2/cluster/tcp.c
>> @@ -1796,9 +1796,9 @@ static int o2net_accept_one(struct socket *sock)
>> goto out;
>> }
>>
>> - if (o2nm_this_node() > node->nd_num) {
>> + if (o2nm_this_node() >= node->nd_num) {
>> mlog(ML_NOTICE, "unexpected connect attempted from a lower "
>> - "numbered node '%s' at " "%pI4:%d with num %u\n",
>> + "or same numbered node '%s' at " "%pI4:%d with num %u\n",
>> node->nd_name, &sin.sin_addr.s_addr,
>> ntohs(sin.sin_port), node->nd_num);
>> ret = -EINVAL;
>> --
>> 1.5.5
>>
>>
>> _______________________________________________
>> Ocfs2-devel mailing list
>> Ocfs2-devel at oss.oracle.com
>> http://oss.oracle.com/mailman/listinfo/ocfs2-devel
^ permalink raw reply [flat|nested] 8+ messages in thread
* [Ocfs2-devel] [PATCH 1/1] O2net: Disallow o2net accept connection request from itself.
2010-07-29 11:04 Tristan Ye
@ 2010-07-29 12:07 ` Sunil Mushran
2010-07-30 2:12 ` tristan
0 siblings, 1 reply; 8+ messages in thread
From: Sunil Mushran @ 2010-07-29 12:07 UTC (permalink / raw)
To: ocfs2-devel
unexpected connect attempt seen at node x (num, ip:port) from node y (num, ip:port).
The above reads better, don't you think. Lower or same are software bugs. Not a user error.
Btw, one test you could do is race nc with mount. Run nc in a loop while mounting the first device. nc should ping a node that has it mounted. Test with higher to lower and reverse.
On Jul 29, 2010, at 4:04 AM, Tristan Ye <tristan.ye@oracle.com> wrote:
> Currently, o2net_accept_one() is allowed to accept a connection from
> listening node itself, such a fake connection will not be successfully
> established due to no handshake detected afterwards, and later end up
> with triggering connecting worker in a loop.
>
> We're going to fix this by treating such connection request as 'invalid',
> since we've got no chance of requesting connection from a node to itself
> in a OCFS2 cluster.
>
> The fix doesn't hurt user's scan for o2net-listener, it always gets a
> successful connection from userpace.
>
> Signed-off-by: Tristan Ye <tristan.ye@oracle.com>
> ---
> fs/ocfs2/cluster/tcp.c | 4 ++--
> 1 files changed, 2 insertions(+), 2 deletions(-)
>
> diff --git a/fs/ocfs2/cluster/tcp.c b/fs/ocfs2/cluster/tcp.c
> index aa75ca3..7e4e451 100644
> --- a/fs/ocfs2/cluster/tcp.c
> +++ b/fs/ocfs2/cluster/tcp.c
> @@ -1796,9 +1796,9 @@ static int o2net_accept_one(struct socket *sock)
> goto out;
> }
>
> - if (o2nm_this_node() > node->nd_num) {
> + if (o2nm_this_node() >= node->nd_num) {
> mlog(ML_NOTICE, "unexpected connect attempted from a lower "
> - "numbered node '%s' at " "%pI4:%d with num %u\n",
> + "or same numbered node '%s' at " "%pI4:%d with num %u\n",
> node->nd_name, &sin.sin_addr.s_addr,
> ntohs(sin.sin_port), node->nd_num);
> ret = -EINVAL;
> --
> 1.5.5
>
>
> _______________________________________________
> Ocfs2-devel mailing list
> Ocfs2-devel at oss.oracle.com
> http://oss.oracle.com/mailman/listinfo/ocfs2-devel
^ permalink raw reply [flat|nested] 8+ messages in thread
* [Ocfs2-devel] [PATCH 1/1] O2net: Disallow o2net accept connection request from itself.
@ 2010-07-29 11:04 Tristan Ye
2010-07-29 12:07 ` Sunil Mushran
0 siblings, 1 reply; 8+ messages in thread
From: Tristan Ye @ 2010-07-29 11:04 UTC (permalink / raw)
To: ocfs2-devel
Currently, o2net_accept_one() is allowed to accept a connection from
listening node itself, such a fake connection will not be successfully
established due to no handshake detected afterwards, and later end up
with triggering connecting worker in a loop.
We're going to fix this by treating such connection request as 'invalid',
since we've got no chance of requesting connection from a node to itself
in a OCFS2 cluster.
The fix doesn't hurt user's scan for o2net-listener, it always gets a
successful connection from userpace.
Signed-off-by: Tristan Ye <tristan.ye@oracle.com>
---
fs/ocfs2/cluster/tcp.c | 4 ++--
1 files changed, 2 insertions(+), 2 deletions(-)
diff --git a/fs/ocfs2/cluster/tcp.c b/fs/ocfs2/cluster/tcp.c
index aa75ca3..7e4e451 100644
--- a/fs/ocfs2/cluster/tcp.c
+++ b/fs/ocfs2/cluster/tcp.c
@@ -1796,9 +1796,9 @@ static int o2net_accept_one(struct socket *sock)
goto out;
}
- if (o2nm_this_node() > node->nd_num) {
+ if (o2nm_this_node() >= node->nd_num) {
mlog(ML_NOTICE, "unexpected connect attempted from a lower "
- "numbered node '%s' at " "%pI4:%d with num %u\n",
+ "or same numbered node '%s' at " "%pI4:%d with num %u\n",
node->nd_name, &sin.sin_addr.s_addr,
ntohs(sin.sin_port), node->nd_num);
ret = -EINVAL;
--
1.5.5
^ permalink raw reply related [flat|nested] 8+ messages in thread
end of thread, other threads:[~2010-08-07 18:40 UTC | newest]
Thread overview: 8+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2010-08-02 2:00 [Ocfs2-devel] [PATCH 1/1] O2net: Disallow o2net accept connection request from itself Tristan Ye
2010-08-02 3:00 ` Sunil Mushran
2010-08-07 18:40 ` Joel Becker
-- strict thread matches above, loose matches on Subject: below --
2010-07-30 8:25 Tristan Ye
2010-07-30 17:27 ` Sunil Mushran
2010-07-29 11:04 Tristan Ye
2010-07-29 12:07 ` Sunil Mushran
2010-07-30 2:12 ` tristan
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.