https://jira.whamcloud.com This file is an XML representation of an issue en-us 9.4.14 940014 05-12-2023 https://jira.whamcloud.com/browse/LU-2073 Lustre <p>It looks like we need to somehow free lustre subdevice symlinks. After running with memleak detector I see a bunch of entries like this:</p> <div class="code panel" style="border-width: 1px;"><div class="codeContent panelContent"> <pre class="code-java">unreferenced object 0xffff880131381458 (size 64): comm <span class="code-quote">"llog_process_th"</span>, pid 12528, jiffies 4296618724 hex dump (first 32 bytes): 2e 2e 2f 6f 73 70 2f 6c 75 73 74 72 65 2d 4f 53 ../osp/lustre-OS 54 30 30 30 31 2d 6f 73 63 2d 4d 44 54 30 30 30 T0001-osc-MDT000 backtrace: [&lt;ffffffff814e4aee&gt;] kmemleak_alloc+0x5e/0xd0 [&lt;ffffffff81164533&gt;] __kmalloc+0x1c3/0x2f0 [&lt;ffffffff811eb433&gt;] proc_symlink+0x53/0xb0 [&lt;ffffffffa0370c5a&gt;] 0xffffffffa0370c5a ... unreferenced object 0xffff8800dc8cc2a0 (size 64): comm <span class="code-quote">"llog_process_th"</span>, pid 12528, jiffies 4296618722 hex dump (first 32 bytes): 2e 2e 2f 6f 73 70 2f 6c 75 73 74 72 65 2d 4f 53 ../osp/lustre-OS 54 30 30 30 30 2d 6f 73 63 2d 4d 44 54 30 30 30 T0000-osc-MDT000 backtrace: [&lt;ffffffff814e4aee&gt;] kmemleak_alloc+0x5e/0xd0 [&lt;ffffffff81164533&gt;] __kmalloc+0x1c3/0x2f0 [&lt;ffffffff811eb433&gt;] proc_symlink+0x53/0xb0 [&lt;ffffffffa0370c5a&gt;] 0xffffffffa0370c5a ... unreferenced object 0xffff880250630400 (size 64): comm <span class="code-quote">"mount.lustre"</span>, pid 12525, jiffies 4296617672 hex dump (first 32 bytes): 2e 2e 2f 2e 2e 2f 2e 2e 2f 6d 64 63 2f 6c 75 73 ../../../mdc/lus 74 72 65 2d 4d 44 54 30 30 30 30 2d 6d 64 63 2d tre-MDT0000-mdc- backtrace: [&lt;ffffffff814e4aee&gt;] kmemleak_alloc+0x5e/0xd0 [&lt;ffffffff81164533&gt;] __kmalloc+0x1c3/0x2f0 [&lt;ffffffff811eb433&gt;] proc_symlink+0x53/0xb0 [&lt;ffffffffa0370c5a&gt;] 0xffffffffa0370c5a ... unreferenced object 0xffff88014c3e7ea0 (size 32): comm <span class="code-quote">"llog_process_th"</span>, pid 12225, jiffies 4296616700 hex dump (first 32 bytes): 2e 2e 2f 6c 6f 64 2f 6c 75 73 74 72 65 2d 4d 44 ../lod/lustre-MD 54 30 30 30 30 2d 6d 64 74 6c 6f 76 00 5a 5a a5 T0000-mdtlov.ZZ. backtrace: [&lt;ffffffff814e4aee&gt;] kmemleak_alloc+0x5e/0xd0 [&lt;ffffffff81164533&gt;] __kmalloc+0x1c3/0x2f0 [&lt;ffffffff811eb433&gt;] proc_symlink+0x53/0xb0 [&lt;ffffffffa0370c5a&gt;] 0xffffffffa0370c5a ... unreferenced object 0xffff8800d85ab820 (size 64): comm <span class="code-quote">"mount.lustre"</span>, pid 12178, jiffies 4296616645 hex dump (first 32 bytes): 2e 2e 2f 2e 2e 2f 6f 73 64 2d 6c 64 69 73 6b 66 ../../osd-ldiskf 73 2f 6c 75 73 74 72 65 2d 4d 44 54 30 30 30 30 s/lustre-MDT0000 backtrace: [&lt;ffffffff814e4aee&gt;] kmemleak_alloc+0x5e/0xd0 [&lt;ffffffff81164533&gt;] __kmalloc+0x1c3/0x2f0 [&lt;ffffffff811eb433&gt;] proc_symlink+0x53/0xb0 [&lt;ffffffffa0370c5a&gt;] 0xffffffffa0370c5a ... </pre> </div></div> <p>and so on.</p> LU-2073

procfs symlinks are apparently never freed

Bug Minor Resolved Fixed Yang Sheng Oleg Drokin Tue, 2 Oct 2012 00:43:04 +0000 Wed, 7 Nov 2012 13:45:07 +0000 Wed, 7 Nov 2012 13:45:07 +0000 Lustre 2.4.0 0 6 <p>there is the following lines in osp:</p> <p>if (m-&gt;opd_symlink)<br/> lprocfs_remove(&amp;m-&gt;opd_symlink);</p> <p>Looking at lprocfs_remove_nolock we see this:</p> <div class="code panel" style="border-width: 1px;"><div class="codeContent panelContent"> <pre class="code-java">#ifdef HAVE_PROCFS_USERS /* <span class="code-keyword">if</span> procfs uses user count to synchronize deletion of * proc entry, there is no protection <span class="code-keyword">for</span> rm_entry-&gt;data, * then lprocfs_fops_read and lprocfs_fops_write maybe * call proc_dir_entry-&gt;read_proc (or write_proc) with * proc_dir_entry-&gt;data == NULL, then cause kernel Oops. * see bug19706 <span class="code-keyword">for</span> detailed information */ /* procfs won<span class="code-quote">'t free rm_entry-&gt;data <span class="code-keyword">if</span> it isn'</span>t a LINK, * and Lustre won't use rm_entry-&gt;data <span class="code-keyword">if</span> it is a LINK */ <span class="code-keyword">if</span> (S_ISLNK(rm_entry-&gt;mode)) rm_entry-&gt;data = NULL; #<span class="code-keyword">else</span> </pre> </div></div> <p>Now, if we look into proc_symlink, we see that entry-&gt;data is kmalloced so that's how the leak appears.</p> <p>In free_proc_entry we can see:</p> <div class="code panel" style="border-width: 1px;"><div class="codeContent panelContent"> <pre class="code-java"> <span class="code-keyword">if</span> (S_ISLNK(de-&gt;mode)) kfree(de-&gt;data); </pre> </div></div> <p>As such I think the assignment of NULL is in error.</p> <p>Last person touching this code (and probably even adding the NULL assignment) was YangSheng, so perhaps he can chime in here?</p> <p>Yangsheng</p> <p>Could you please look into this one?</p> <p>Thanks</p> <p>Peter</p> <p>Patch commit to :<a href="http://review.whamcloud.com/4434" class="external-link" target="_blank" rel="nofollow noopener">http://review.whamcloud.com/4434</a></p> <p>Landed for 2.4</p> Development Rank 1|hzv4wf: Rank (Obsolete) 4327 Severity