<feed xmlns='http://www.w3.org/2005/Atom'>
<title>linux-toradex.git/net/rds/ib_cm.c, branch v4.17</title>
<subtitle>Linux kernel for Apalis and Colibri modules</subtitle>
<link rel='alternate' type='text/html' href='https://git.toradex.cn/cgit/linux-toradex.git/'/>
<entry>
<title>rds: ib: Fix missing call to rds_ib_dev_put in rds_ib_setup_qp</title>
<updated>2018-04-25T18:34:08+00:00</updated>
<author>
<name>Dag Moxnes</name>
<email>dag.moxnes@oracle.com</email>
</author>
<published>2018-04-25T11:22:01+00:00</published>
<link rel='alternate' type='text/html' href='https://git.toradex.cn/cgit/linux-toradex.git/commit/?id=91a825290ca4eae88603bc811bf74a45f94a3f46'/>
<id>91a825290ca4eae88603bc811bf74a45f94a3f46</id>
<content type='text'>
The function rds_ib_setup_qp is calling rds_ib_get_client_data and
should correspondingly call rds_ib_dev_put. This call was lost in
the non-error path with the introduction of error handling done in
commit 3b12f73a5c29 ("rds: ib: add error handle")

Signed-off-by: Dag Moxnes &lt;dag.moxnes@oracle.com&gt;
Reviewed-by: Håkon Bugge &lt;haakon.bugge@oracle.com&gt;
Acked-by: Santosh Shilimkar &lt;santosh.shilimkar@oracle.com&gt;
Signed-off-by: David S. Miller &lt;davem@davemloft.net&gt;
</content>
<content type='xhtml'>
<div xmlns='http://www.w3.org/1999/xhtml'>
<pre>
The function rds_ib_setup_qp is calling rds_ib_get_client_data and
should correspondingly call rds_ib_dev_put. This call was lost in
the non-error path with the introduction of error handling done in
commit 3b12f73a5c29 ("rds: ib: add error handle")

Signed-off-by: Dag Moxnes &lt;dag.moxnes@oracle.com&gt;
Reviewed-by: Håkon Bugge &lt;haakon.bugge@oracle.com&gt;
Acked-by: Santosh Shilimkar &lt;santosh.shilimkar@oracle.com&gt;
Signed-off-by: David S. Miller &lt;davem@davemloft.net&gt;
</pre>
</div>
</content>
</entry>
<entry>
<title>rds: tcp: use rds_destroy_pending() to synchronize netns/module teardown and rds connection/workq management</title>
<updated>2018-02-08T20:23:52+00:00</updated>
<author>
<name>Sowmini Varadhan</name>
<email>sowmini.varadhan@oracle.com</email>
</author>
<published>2018-02-03T12:26:51+00:00</published>
<link rel='alternate' type='text/html' href='https://git.toradex.cn/cgit/linux-toradex.git/commit/?id=ebeeb1ad9b8adcc37c2ec21a96f39e9d35199b46'/>
<id>ebeeb1ad9b8adcc37c2ec21a96f39e9d35199b46</id>
<content type='text'>
An rds_connection can get added during netns deletion between lines 528
and 529 of

  506 static void rds_tcp_kill_sock(struct net *net)
  :
  /* code to pull out all the rds_connections that should be destroyed */
  :
  528         spin_unlock_irq(&amp;rds_tcp_conn_lock);
  529         list_for_each_entry_safe(tc, _tc, &amp;tmp_list, t_tcp_node)
  530                 rds_conn_destroy(tc-&gt;t_cpath-&gt;cp_conn);

Such an rds_connection would miss out the rds_conn_destroy()
loop (that cancels all pending work) and (if it was scheduled
after netns deletion) could trigger the use-after-free.

A similar race-window exists for the module unload path
in rds_tcp_exit -&gt; rds_tcp_destroy_conns

Concurrency with netns deletion (rds_tcp_kill_sock()) must be handled
by checking check_net() before enqueuing new work or adding new
connections.

Concurrency with module-unload is handled by maintaining a module
specific flag that is set at the start of the module exit function,
and must be checked before enqueuing new work or adding new connections.

This commit refactors existing RDS_DESTROY_PENDING checks added by
commit 3db6e0d172c9 ("rds: use RCU to synchronize work-enqueue with
connection teardown") and consolidates all the concurrency checks
listed above into the function rds_destroy_pending().

Signed-off-by: Sowmini Varadhan &lt;sowmini.varadhan@oracle.com&gt;
Acked-by: Santosh Shilimkar &lt;santosh.shilimkar@oracle.com&gt;
Signed-off-by: David S. Miller &lt;davem@davemloft.net&gt;
</content>
<content type='xhtml'>
<div xmlns='http://www.w3.org/1999/xhtml'>
<pre>
An rds_connection can get added during netns deletion between lines 528
and 529 of

  506 static void rds_tcp_kill_sock(struct net *net)
  :
  /* code to pull out all the rds_connections that should be destroyed */
  :
  528         spin_unlock_irq(&amp;rds_tcp_conn_lock);
  529         list_for_each_entry_safe(tc, _tc, &amp;tmp_list, t_tcp_node)
  530                 rds_conn_destroy(tc-&gt;t_cpath-&gt;cp_conn);

Such an rds_connection would miss out the rds_conn_destroy()
loop (that cancels all pending work) and (if it was scheduled
after netns deletion) could trigger the use-after-free.

A similar race-window exists for the module unload path
in rds_tcp_exit -&gt; rds_tcp_destroy_conns

Concurrency with netns deletion (rds_tcp_kill_sock()) must be handled
by checking check_net() before enqueuing new work or adding new
connections.

Concurrency with module-unload is handled by maintaining a module
specific flag that is set at the start of the module exit function,
and must be checked before enqueuing new work or adding new connections.

This commit refactors existing RDS_DESTROY_PENDING checks added by
commit 3db6e0d172c9 ("rds: use RCU to synchronize work-enqueue with
connection teardown") and consolidates all the concurrency checks
listed above into the function rds_destroy_pending().

Signed-off-by: Sowmini Varadhan &lt;sowmini.varadhan@oracle.com&gt;
Acked-by: Santosh Shilimkar &lt;santosh.shilimkar@oracle.com&gt;
Signed-off-by: David S. Miller &lt;davem@davemloft.net&gt;
</pre>
</div>
</content>
</entry>
<entry>
<title>Merge git://git.kernel.org/pub/scm/linux/kernel/git/davem/net</title>
<updated>2017-03-15T18:59:10+00:00</updated>
<author>
<name>David S. Miller</name>
<email>davem@davemloft.net</email>
</author>
<published>2017-03-15T18:59:10+00:00</published>
<link rel='alternate' type='text/html' href='https://git.toradex.cn/cgit/linux-toradex.git/commit/?id=101c431492d297dd0d111b461d8d324895676bee'/>
<id>101c431492d297dd0d111b461d8d324895676bee</id>
<content type='text'>
Conflicts:
	drivers/net/ethernet/broadcom/genet/bcmgenet.c
	net/core/sock.c

Conflicts were overlapping changes in bcmgenet and the
lockdep handling of sockets.

Signed-off-by: David S. Miller &lt;davem@davemloft.net&gt;
</content>
<content type='xhtml'>
<div xmlns='http://www.w3.org/1999/xhtml'>
<pre>
Conflicts:
	drivers/net/ethernet/broadcom/genet/bcmgenet.c
	net/core/sock.c

Conflicts were overlapping changes in bcmgenet and the
lockdep handling of sockets.

Signed-off-by: David S. Miller &lt;davem@davemloft.net&gt;
</pre>
</div>
</content>
</entry>
<entry>
<title>rds: ib: drop unnecessary rdma_reject</title>
<updated>2017-03-14T06:20:05+00:00</updated>
<author>
<name>Zhu Yanjun</name>
<email>yanjun.zhu@oracle.com</email>
</author>
<published>2017-03-13T05:43:45+00:00</published>
<link rel='alternate' type='text/html' href='https://git.toradex.cn/cgit/linux-toradex.git/commit/?id=b418c5276a41c821c25410542df83d283bbc0b2a'/>
<id>b418c5276a41c821c25410542df83d283bbc0b2a</id>
<content type='text'>
When rdma_accept fails, rdma_reject is called in it. As such, it is
not necessary to execute rdma_reject again.

Cc: Joe Jin &lt;joe.jin@oracle.com&gt;
Cc: Junxiao Bi &lt;junxiao.bi@oracle.com&gt;
Acked-by: Santosh Shilimkar &lt;santosh.shilimkar@oracle.com&gt;
Signed-off-by: Zhu Yanjun &lt;yanjun.zhu@oracle.com&gt;
Signed-off-by: David S. Miller &lt;davem@davemloft.net&gt;
</content>
<content type='xhtml'>
<div xmlns='http://www.w3.org/1999/xhtml'>
<pre>
When rdma_accept fails, rdma_reject is called in it. As such, it is
not necessary to execute rdma_reject again.

Cc: Joe Jin &lt;joe.jin@oracle.com&gt;
Cc: Junxiao Bi &lt;junxiao.bi@oracle.com&gt;
Acked-by: Santosh Shilimkar &lt;santosh.shilimkar@oracle.com&gt;
Signed-off-by: Zhu Yanjun &lt;yanjun.zhu@oracle.com&gt;
Signed-off-by: David S. Miller &lt;davem@davemloft.net&gt;
</pre>
</div>
</content>
</entry>
<entry>
<title>rds: ib: add error handle</title>
<updated>2017-03-09T21:09:18+00:00</updated>
<author>
<name>Zhu Yanjun</name>
<email>yanjun.zhu@oracle.com</email>
</author>
<published>2017-03-07T07:48:36+00:00</published>
<link rel='alternate' type='text/html' href='https://git.toradex.cn/cgit/linux-toradex.git/commit/?id=3b12f73a5c2977153f28a224392fd4729b50d1dc'/>
<id>3b12f73a5c2977153f28a224392fd4729b50d1dc</id>
<content type='text'>
In the function rds_ib_setup_qp, the error handle is missing. When some
error occurs, it is possible that memory leak occurs. As such, error
handle is added.

Cc: Joe Jin &lt;joe.jin@oracle.com&gt;
Reviewed-by: Junxiao Bi &lt;junxiao.bi@oracle.com&gt;
Reviewed-by: Guanglei Li &lt;guanglei.li@oracle.com&gt;
Signed-off-by: Zhu Yanjun &lt;yanjun.zhu@oracle.com&gt;
Acked-by: Santosh Shilimkar &lt;santosh.shilimkar@oracle.com&gt;
Signed-off-by: David S. Miller &lt;davem@davemloft.net&gt;
</content>
<content type='xhtml'>
<div xmlns='http://www.w3.org/1999/xhtml'>
<pre>
In the function rds_ib_setup_qp, the error handle is missing. When some
error occurs, it is possible that memory leak occurs. As such, error
handle is added.

Cc: Joe Jin &lt;joe.jin@oracle.com&gt;
Reviewed-by: Junxiao Bi &lt;junxiao.bi@oracle.com&gt;
Reviewed-by: Guanglei Li &lt;guanglei.li@oracle.com&gt;
Signed-off-by: Zhu Yanjun &lt;yanjun.zhu@oracle.com&gt;
Acked-by: Santosh Shilimkar &lt;santosh.shilimkar@oracle.com&gt;
Signed-off-by: David S. Miller &lt;davem@davemloft.net&gt;
</pre>
</div>
</content>
</entry>
<entry>
<title>RDS: IB: fix panic due to handlers running post teardown</title>
<updated>2017-01-02T22:02:55+00:00</updated>
<author>
<name>Santosh Shilimkar</name>
<email>santosh.shilimkar@oracle.com</email>
</author>
<published>2016-09-29T18:07:11+00:00</published>
<link rel='alternate' type='text/html' href='https://git.toradex.cn/cgit/linux-toradex.git/commit/?id=cf657269d311d575eb196c7045579b3443631b8b'/>
<id>cf657269d311d575eb196c7045579b3443631b8b</id>
<content type='text'>
Shutdown code reaping loop takes care of emptying the
CQ's before they being destroyed. And once tasklets are
killed, the hanlders are not expected to run.

But because of core tasklet code issues, tasklet handler could
still run even after tasklet_kill,
RDS IB shutdown code already reaps the CQs before freeing
cq/qp resources so as such the handlers have nothing left
to do post shutdown.

On other hand any handler running after teardown and trying
to access already freed qp/cq resources causes issues
Patch fixes this race by  makes sure that handlers returns
without any action post teardown.

Reviewed-by: Wengang &lt;wen.gang.wang@oracle.com&gt;
Signed-off-by: Santosh Shilimkar &lt;santosh.shilimkar@oracle.com&gt;
</content>
<content type='xhtml'>
<div xmlns='http://www.w3.org/1999/xhtml'>
<pre>
Shutdown code reaping loop takes care of emptying the
CQ's before they being destroyed. And once tasklets are
killed, the hanlders are not expected to run.

But because of core tasklet code issues, tasklet handler could
still run even after tasklet_kill,
RDS IB shutdown code already reaps the CQs before freeing
cq/qp resources so as such the handlers have nothing left
to do post shutdown.

On other hand any handler running after teardown and trying
to access already freed qp/cq resources causes issues
Patch fixes this race by  makes sure that handlers returns
without any action post teardown.

Reviewed-by: Wengang &lt;wen.gang.wang@oracle.com&gt;
Signed-off-by: Santosh Shilimkar &lt;santosh.shilimkar@oracle.com&gt;
</pre>
</div>
</content>
</entry>
<entry>
<title>RDS: IB: Add vector spreading for cqs</title>
<updated>2017-01-02T22:02:52+00:00</updated>
<author>
<name>Santosh Shilimkar</name>
<email>santosh.shilimkar@oracle.com</email>
</author>
<published>2016-07-04T23:16:36+00:00</published>
<link rel='alternate' type='text/html' href='https://git.toradex.cn/cgit/linux-toradex.git/commit/?id=be2f76eacc278c272f26d46e4168efe5a55f5383'/>
<id>be2f76eacc278c272f26d46e4168efe5a55f5383</id>
<content type='text'>
Based on available device vectors, allocate cqs accordingly to
get better spread of completion vectors which helps performace
great deal..

Signed-off-by: Santosh Shilimkar &lt;santosh.shilimkar@oracle.com&gt;
</content>
<content type='xhtml'>
<div xmlns='http://www.w3.org/1999/xhtml'>
<pre>
Based on available device vectors, allocate cqs accordingly to
get better spread of completion vectors which helps performace
great deal..

Signed-off-by: Santosh Shilimkar &lt;santosh.shilimkar@oracle.com&gt;
</pre>
</div>
</content>
</entry>
<entry>
<title>RDS: IB: track and log active side endpoint in connection</title>
<updated>2017-01-02T22:02:50+00:00</updated>
<author>
<name>Santosh Shilimkar</name>
<email>santosh.shilimkar@oracle.com</email>
</author>
<published>2016-07-10T01:31:38+00:00</published>
<link rel='alternate' type='text/html' href='https://git.toradex.cn/cgit/linux-toradex.git/commit/?id=581d53c91cbf7b31415a9ed5e9a8b89d6af609b3'/>
<id>581d53c91cbf7b31415a9ed5e9a8b89d6af609b3</id>
<content type='text'>
Useful to know the active and passive end points in a
RDS IB connection.

Signed-off-by: Santosh Shilimkar &lt;santosh.shilimkar@oracle.com&gt;
</content>
<content type='xhtml'>
<div xmlns='http://www.w3.org/1999/xhtml'>
<pre>
Useful to know the active and passive end points in a
RDS IB connection.

Signed-off-by: Santosh Shilimkar &lt;santosh.shilimkar@oracle.com&gt;
</pre>
</div>
</content>
</entry>
<entry>
<title>RDS: IB: split the mr registration and invalidation path</title>
<updated>2017-01-02T22:02:47+00:00</updated>
<author>
<name>Santosh Shilimkar</name>
<email>santosh.shilimkar@oracle.com</email>
</author>
<published>2016-03-08T17:19:01+00:00</published>
<link rel='alternate' type='text/html' href='https://git.toradex.cn/cgit/linux-toradex.git/commit/?id=56012459310a1dbcc55c2dbf5500a9f7571402cb'/>
<id>56012459310a1dbcc55c2dbf5500a9f7571402cb</id>
<content type='text'>
MR invalidation in RDS is done in background thread and not in
data path like registration. So break the dependency between them
which helps to remove the performance bottleneck.

Signed-off-by: Santosh Shilimkar &lt;santosh.shilimkar@oracle.com&gt;
</content>
<content type='xhtml'>
<div xmlns='http://www.w3.org/1999/xhtml'>
<pre>
MR invalidation in RDS is done in background thread and not in
data path like registration. So break the dependency between them
which helps to remove the performance bottleneck.

Signed-off-by: Santosh Shilimkar &lt;santosh.shilimkar@oracle.com&gt;
</pre>
</div>
</content>
</entry>
<entry>
<title>RDS: IB: include faddr in connection log</title>
<updated>2017-01-02T22:02:41+00:00</updated>
<author>
<name>Santosh Shilimkar</name>
<email>santosh.shilimkar@oracle.com</email>
</author>
<published>2016-03-14T14:43:55+00:00</published>
<link rel='alternate' type='text/html' href='https://git.toradex.cn/cgit/linux-toradex.git/commit/?id=ff3f19a2f608ee406331e8c7b60d7376e75c2157'/>
<id>ff3f19a2f608ee406331e8c7b60d7376e75c2157</id>
<content type='text'>
Also use pr_* for it.

Signed-off-by: Santosh Shilimkar &lt;santosh.shilimkar@oracle.com&gt;
</content>
<content type='xhtml'>
<div xmlns='http://www.w3.org/1999/xhtml'>
<pre>
Also use pr_* for it.

Signed-off-by: Santosh Shilimkar &lt;santosh.shilimkar@oracle.com&gt;
</pre>
</div>
</content>
</entry>
</feed>
