<html>

  <head>

    <meta content="text/html; charset=ISO-8859-1"

      http-equiv="Content-Type">

  </head>

  <body bgcolor="#FFFFFF" text="#000000">

    <div class="moz-cite-prefix">On 02/25/2014 07:30 AM, Srinivas Eeda

      wrote:<br>

    </div>

    <blockquote cite="mid:530BD609.8080104@oracle.com" type="cite">

      <meta content="text/html; charset=ISO-8859-1"

        http-equiv="Content-Type">

      Junxiao, thanks for looking into this issue. Please see my comment

      below<br>

      <br>

      On 02/24/2014 01:07 AM, Junxiao Bi wrote:<br>

      <blockquote cite="mid:530B0BB5.90600@oracle.com" type="cite">

        <meta content="text/html; charset=ISO-8859-1"

          http-equiv="Content-Type">

        <div class="moz-cite-prefix">Hi,<br>

          <br>

          On 07/19/2012 09:59 AM, Sunil Mushran wrote:<br>

        </div>

        <blockquote

cite="mid:CAEeiSHV+TVsnwqnsi0u4r=ucBoddo8wD8DcqbsCn1UoA3xjtdg@mail.gmail.com"

          type="cite">Different issues.<br>

          <br>

          <div class="gmail_quote">On Wed, Jul 18, 2012 at 6:34 PM,

            Junxiao Bi <span dir="ltr">&lt;<a moz-do-not-send="true"

                href="mailto:junxiao.bi@oracle.com" target="_blank">junxiao.bi@oracle.com</a>&gt;</span>

            wrote:<br>

            <blockquote class="gmail_quote" style="margin:0 0 0

              .8ex;border-left:1px #ccc solid;padding-left:1ex">

              <div bgcolor="#FFFFFF" text="#000000">

                <div>On 07/19/2012 12:36 AM, Sunil Mushran wrote:<br>

                </div>

                <blockquote type="cite">

                  <div>This bug was detected during code audit. Never

                    seen a crash. If it does hit,</div>

                  <div>then we have bigger problems. So no point posting

                    to stable.</div>

                </blockquote>

              </div>

            </blockquote>

          </div>

        </blockquote>

        I read a lot of dlm recovery code recently, I found this bug

        could happen at the following scenario.<br>

        <br>

        node 1:&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; migrate

        target node x:<br>

        dlm_unregister_domain()<br>

        &nbsp;dlm_migrate_all_locks()<br>

        &nbsp; dlm_empty_lockres()<br>

        &nbsp;&nbsp; select node x as migrate target node<br>

        &nbsp;&nbsp; since there is a node x lock on the granted list.<br>

        &nbsp;&nbsp; dlm_migrate_lockres()<br>

        &nbsp;&nbsp;&nbsp; dlm_mark_lockres_migrating() {<br>

        &nbsp;&nbsp;&nbsp;&nbsp; wait_event(dlm-&gt;ast_wq, !dlm_lockres_is_dirty(dlm,

        res));<br>

        &lt;&lt;&lt; node x unlock may happen here, res-&gt;granted list

        can be empty.<br>

      </blockquote>

      If the unlock request got sent at this point, and if the request

      was *processed*, lock must have been removed from the

      granted_list. If the request was *not yet processed*, then the

      DLM_LOCK_RES_MIGRATING set in dlm_lockres_release_ast would make

      dlm_unlock handler to return DLM_MIGRATING to the caller (in this

      case node x). So I don't see how granted_list could have stale

      lock. Am I missing something ?<br>

      <br>

      I do think there is such race that you pointed below exist, but I

      am not sure if it was due to the above race described.<br>

    </blockquote>

    Outside the windows from set RES_BLOCK_DIRTY flag and wait_event()

    to dlm_lockres_release_ast(), granted_list can not be empty, since

    wait_event will wait until dlm_thread clear the dirty flag where

    shuffle list will pick another lock to the granted list. After the

    window, DLM_MIGRATING flag will stop other node unlock to the

    granted list. So I think this cause the empty granted list and cause

    the migrate target panic. I didn't see any other harm of this since

    the migrate target node will shuffle the list and send the ast

    message later.<br>

    <br>

    Thanks,<br>

    Junxiao.<br>

    <blockquote cite="mid:530BD609.8080104@oracle.com" type="cite"> <br>

      <blockquote cite="mid:530B0BB5.90600@oracle.com" type="cite"> &nbsp;&nbsp;&nbsp;&nbsp;

        dlm_lockres_release_ast(dlm, res);<br>

        &nbsp;&nbsp;&nbsp; }&nbsp;&nbsp; <br>

        &nbsp;&nbsp;&nbsp; dlm_send_one_lockres()<br>

        &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;

        dlm_process_recovery_data() {<br>

        &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; tmpq is

        res-&gt;granted list and is empty.<br>

        &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;

        list_for_each_entry(lock, tmpq, list) {<br>

        &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; if

        (lock-&gt;ml.cookie != ml-&gt;cookie)<br>

        &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; lock =

        NULL;<br>

        &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; else <br>

        &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; break;<br>

        &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; }&nbsp;&nbsp; <br>

        &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; lock will be

        invalid here.<br>

        &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; if

        (lock-&gt;ml.node != ml-&gt;node)<br>

        &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; BUG()

        --&gt; crash here.<br>

        &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; }<br>

        <br>

        Thanks,<br>

        Junxiao.<br>

        <blockquote

cite="mid:CAEeiSHV+TVsnwqnsi0u4r=ucBoddo8wD8DcqbsCn1UoA3xjtdg@mail.gmail.com"

          type="cite">

          <div class="gmail_quote">

            <blockquote class="gmail_quote" style="margin:0 0 0

              .8ex;border-left:1px #ccc solid;padding-left:1ex">

              <div bgcolor="#FFFFFF" text="#000000">

                <blockquote type="cite"> </blockquote>

                Our customer can reproduce it. Also I saw you were

                assigned a similar bug before, see <a

                  moz-do-not-send="true"

                  href="https://oss.oracle.com/bugzilla/show_bug.cgi?id=1220"

                  target="_blank">https://oss.oracle.com/bugzilla/show_bug.cgi?id=1220</a>,

                is it the same BUG?<br>

                <blockquote type="cite"><br>

                  <div class="gmail_quote">On Tue, Jul 17, 2012 at 6:36

                    PM, Junxiao Bi <span dir="ltr">&lt;<a

                        moz-do-not-send="true"

                        href="mailto:junxiao.bi@oracle.com"

                        target="_blank">junxiao.bi@oracle.com</a>&gt;</span>

                    wrote:<br>

                    <blockquote class="gmail_quote" style="margin:0 0 0

                      .8ex;border-left:1px #ccc solid;padding-left:1ex">

                      <div bgcolor="#FFFFFF" text="#000000">

                        <div>Hi Sunil,<br>

                          <br>

                          On 07/18/2012 03:49 AM, Sunil Mushran wrote:<br>

                        </div>

                        <blockquote type="cite">

                          <div>

                            <div class="gmail_quote">On Tue, Jul 17,

                              2012 at 12:10 AM, Junxiao Bi <span

                                dir="ltr">&lt;<a moz-do-not-send="true"

                                  href="mailto:junxiao.bi@oracle.com"

                                  target="_blank">junxiao.bi@oracle.com</a>&gt;</span>

                              wrote:<br>

                              <blockquote class="gmail_quote"

                                style="margin:0 0 0 .8ex;border-left:1px

                                #ccc solid;padding-left:1ex"> In the

                                target node of the dlm lock migration,

                                the logic to find<br>

                                the local dlm lock is wrong, it

                                shouldn't change the loop variable<br>

                                "lock" in the list_for_each_entry loop.

                                This will cause a NULL-pointer<br>

                                accessing crash.<br>

                                <br>

                                Signed-off-by: Junxiao Bi &lt;<a

                                  moz-do-not-send="true"

                                  href="mailto:junxiao.bi@oracle.com"

                                  target="_blank">junxiao.bi@oracle.com</a>&gt;<br>

                                Cc: <a moz-do-not-send="true"

                                  href="mailto:stable@vger.kernel.org"

                                  target="_blank">stable@vger.kernel.org</a><br>

                                ---<br>

                                &nbsp;fs/ocfs2/dlm/dlmrecovery.c | &nbsp; 12

                                +++++++-----<br>

                                &nbsp;1 file changed, 7 insertions(+), 5

                                deletions(-)<br>

                                <br>

                                diff --git a/fs/ocfs2/dlm/dlmrecovery.c

                                b/fs/ocfs2/dlm/dlmrecovery.c<br>

                                index 01ebfd0..0b9cc88 100644<br>

                                --- a/fs/ocfs2/dlm/dlmrecovery.c<br>

                                +++ b/fs/ocfs2/dlm/dlmrecovery.c<br>

                                @@ -1762,6 +1762,7 @@ static int

                                dlm_process_recovery_data(struct

                                dlm_ctxt *dlm,<br>

                                &nbsp; &nbsp; &nbsp; &nbsp; u8 from = O2NM_MAX_NODES;<br>

                                &nbsp; &nbsp; &nbsp; &nbsp; unsigned int added = 0;<br>

                                &nbsp; &nbsp; &nbsp; &nbsp; __be64 c;<br>

                                + &nbsp; &nbsp; &nbsp; int found;<br>

                                <br>

                                &nbsp; &nbsp; &nbsp; &nbsp; mlog(0, "running %d locks for

                                this lockres\n", mres-&gt;num_locks);<br>

                                &nbsp; &nbsp; &nbsp; &nbsp; for (i=0;

                                i&lt;mres-&gt;num_locks; i++) {<br>

                                @@ -1793,22 +1794,23 @@ static int

                                dlm_process_recovery_data(struct

                                dlm_ctxt *dlm,<br>

                                &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; /* MIGRATION

                                ONLY! */<br>

                                &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp;

                                BUG_ON(!(mres-&gt;flags &amp;

                                DLM_MRES_MIGRATION));<br>

                                <br>

                                + &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; found = 0;<br>

                                &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp;

                                spin_lock(&amp;res-&gt;spinlock);<br>

                                &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; for (j =

                                DLM_GRANTED_LIST; j &lt;=

                                DLM_BLOCKED_LIST; j++) {<br>

                                &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; tmpq =

                                dlm_list_idx_to_ptr(res, j);<br>

                                &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp;

                                list_for_each_entry(lock, tmpq, list) {<br>

                                - &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp;

                                if (lock-&gt;ml.cookie != ml-&gt;cookie)<br>

                                - &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp;

                                &nbsp; &nbsp; &nbsp; &nbsp; lock = NULL;<br>

                                - &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp;

                                else<br>

                                + &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp;

                                if (lock-&gt;ml.cookie == ml-&gt;cookie)

                                {<br>

                                + &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp;

                                &nbsp; &nbsp; &nbsp; &nbsp; found = 1;<br>

                                &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp;

                                &nbsp; &nbsp; &nbsp; &nbsp; break;<br>

                                + &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp;

                                }<br>

                                &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; }<br>

                                - &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; if

                                (lock)<br>

                                + &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; if

                                (found)<br>

                                &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp;

                                break;<br>

                                &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; }<br>

                                <br>

                                &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; /* lock is

                                always created locally first, and<br>

                                &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp;* destroyed

                                locally last. &nbsp;it must be on the list */<br>

                                - &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; if (!lock) {<br>

                                + &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; if (!found) {<br>

                                &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; c =

                                ml-&gt;cookie;<br>

                                &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp;

                                mlog(ML_ERROR, "Could not find local

                                lock "<br>

                                &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp; &nbsp;

                                &nbsp; &nbsp; &nbsp; &nbsp;"with cookie %u:%llu, node %u, "<br>

                              </blockquote>

                              <div><br>

                              </div>

                              <div><br>

                              </div>

                              <a moz-do-not-send="true"

href="https://oss.oracle.com/git/?p=smushran/linux-2.6.git;a=blobdiff;f=fs/ocfs2/dlm/dlmrecovery.c;h=c881be6043a8c27c26ee44d217fb8ecf1eb37e02;hp=01ebfd0bdad72264b99345378f0c6febe246503d;hb=13279667cc8bbaf901591dee96f762d4aab8b307;hpb=a5ae0116eb56ec7c128e84fe15646a5cb9a8cb47"

                                target="_blank">https://oss.oracle.com/git/?p=smushran/linux-2.6.git;a=blobdiff;f=fs/ocfs2/dlm/dlmrecovery.c;h=c881be6043a8c27c26ee44d217fb8ecf1eb37e02;hp=01ebfd0bdad72264b99345378f0c6febe246503d;hb=13279667cc8bbaf901591dee96f762d4aab8b307;hpb=a5ae0116eb56ec7c128e84fe15646a5cb9a8cb47</a>

                              <div> <br>

                              </div>

                              <div>We had decided to go back to

                                list_for_each().</div>

                            </div>

                          </div>

                        </blockquote>

                        <br>

                        OK, thank you. It's OK to revert it back for a

                        introduced bug. But I think you'd better cc

                        stable branch.<br>

                      </div>

                    </blockquote>

                  </div>

                  <br>

                </blockquote>

                <br>

                <br>

              </div>

            </blockquote>

          </div>

          <br>

        </blockquote>

        <br>

        <br>

        <fieldset class="mimeAttachmentHeader"></fieldset>

        <br>

        <pre wrap="">_______________________________________________

Ocfs2-devel mailing list

<a moz-do-not-send="true" class="moz-txt-link-abbreviated" href="mailto:Ocfs2-devel@oss.oracle.com">Ocfs2-devel@oss.oracle.com</a>

<a moz-do-not-send="true" class="moz-txt-link-freetext" href="https://oss.oracle.com/mailman/listinfo/ocfs2-devel">https://oss.oracle.com/mailman/listinfo/ocfs2-devel</a></pre>

      </blockquote>

      <br>

      <br>

      <fieldset class="mimeAttachmentHeader"></fieldset>

      <br>

      <pre wrap="">_______________________________________________

Ocfs2-devel mailing list

<a class="moz-txt-link-abbreviated" href="mailto:Ocfs2-devel@oss.oracle.com">Ocfs2-devel@oss.oracle.com</a>

<a class="moz-txt-link-freetext" href="https://oss.oracle.com/mailman/listinfo/ocfs2-devel">https://oss.oracle.com/mailman/listinfo/ocfs2-devel</a></pre>

    </blockquote>

    <br>

  </body>

</html>