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-4231 Lustre <p>[ 1007.910111] LustreError: 9425:0:(llite_internal.h:1091:ll_inode2fid()) ASSERTION( inode != (<br/> (void *)0) ) failed:<br/> [ 1007.920862] LustreError: 9425:0:(llite_internal.h:1091:ll_inode2fid()) LBUG<br/> [ 1007.929703] Kernel panic - not syncing: LBUG<br/> [ 1007.933988] CPU: 0 PID: 9425 Comm: rpc.mountd Tainted: P C O 3.12.0-rc7-ql-generic-10 #1<br/> [ 1007.942781] Hardware name: Supermicro X8DTN/X8DTN, BIOS 080015 05/04/2009<br/> [ 1007.949668] 0000000000000000 ffff8801b20b3a38 ffffffff814fd482 0000000000000007<br/> [ 1007.957158] ffffffffa04e66bd ffff8801b20b3ab8 ffffffff814faa1f ffffffff815eb50a<br/> [ 1007.964647] ffffffff00000008 ffff8801b20b3ac8 ffff8801b20b3a68 ffffffffa0b6ba0e<br/> [ 1007.972144] Call Trace:<br/> [ 1007.974607] <span class="error">&#91;&lt;ffffffff814fd482&gt;&#93;</span> dump_stack+0x46/0x58<br/> [ 1007.979759] <span class="error">&#91;&lt;ffffffff814faa1f&gt;&#93;</span> panic+0xb6/0x1c6<br/> [ 1007.984632] <span class="error">&#91;&lt;ffffffffa04cab50&gt;&#93;</span> lbug_with_loc+0xb0/0xb0 <span class="error">&#91;libcfs&#93;</span><br/> [ 1007.990892] <span class="error">&#91;&lt;ffffffffa0b40635&gt;&#93;</span> ll_encode_fh+0x165/0x170 <span class="error">&#91;lustre&#93;</span><br/> [ 1007.997244] <span class="error">&#91;&lt;ffffffffa03b601b&gt;&#93;</span> exportfs_encode_inode_fh+0x1b/0xa0 <span class="error">&#91;exportfs&#93;</span><br/> [ 1008.004707] <span class="error">&#91;&lt;ffffffffa03b60f0&gt;&#93;</span> exportfs_encode_fh+0x50/0x70 <span class="error">&#91;exportfs&#93;</span><br/> [ 1008.011581] <span class="error">&#91;&lt;ffffffffa03bd4bb&gt;&#93;</span> _fh_update.isra.9.part.10+0x3b/0x50 <span class="error">&#91;nfsd&#93;</span><br/> [ 1008.018716] <span class="error">&#91;&lt;ffffffffa03be011&gt;&#93;</span> fh_compose+0x241/0x520 <span class="error">&#91;nfsd&#93;</span><br/> [ 1008.024719] <span class="error">&#91;&lt;ffffffffa03c4244&gt;&#93;</span> exp_rootfh+0x104/0x2a0 <span class="error">&#91;nfsd&#93;</span><br/> [ 1008.030725] <span class="error">&#91;&lt;ffffffffa03bb7a9&gt;&#93;</span> write_filehandle+0x179/0x200 <span class="error">&#91;nfsd&#93;</span><br/> [ 1008.037250] <span class="error">&#91;&lt;ffffffff811156a5&gt;&#93;</span> ? alloc_pages_current+0xb5/0x180<br/> [ 1008.043505] <span class="error">&#91;&lt;ffffffff810d8839&gt;&#93;</span> ? __get_free_pages+0x9/0x40<br/> [ 1008.049329] <span class="error">&#91;&lt;ffffffff810d8881&gt;&#93;</span> ? get_zeroed_page+0x11/0x20<br/> [ 1008.055156] <span class="error">&#91;&lt;ffffffffa03bb630&gt;&#93;</span> ? write_unlock_ip+0xd0/0xd0 <span class="error">&#91;nfsd&#93;</span><br/> [ 1008.061596] <span class="error">&#91;&lt;ffffffffa03baff8&gt;&#93;</span> nfsctl_transaction_write+0x58/0x80 <span class="error">&#91;nfsd&#93;</span><br/> [ 1008.068642] <span class="error">&#91;&lt;ffffffff81137995&gt;&#93;</span> vfs_write+0xc5/0x1d0<br/> [ 1008.073854] <span class="error">&#91;&lt;ffffffff81137e3d&gt;&#93;</span> SyS_write+0x4d/0xa0<br/> [ 1008.078985] <span class="error">&#91;&lt;ffffffff81507479&gt;&#93;</span> ? do_page_fault+0x9/0x10<br/> [ 1008.084549] <span class="error">&#91;&lt;ffffffff8150a066&gt;&#93;</span> system_call_fastpath+0x1a/0x1f<br/> [ 1008.091486] -----------<del>[ cut here ]</del>-----------</p> - Client mainline kernel 3.12 / lustre-utils 2.4.0<br/> - Servers lustre 2.4.0/ZFS OSDs<br/> - ko2iblnd<br/> - reexport works with 2.6.32/2.4.0 client LU-4231

NFS reexport leads to LBUG in mainline 3.12 client

Bug Major Resolved Fixed Dmitry Eremin Roland Fehrenbacher Fri, 8 Nov 2013 17:12:44 +0000 Tue, 4 Feb 2014 08:10:17 +0000 Tue, 4 Feb 2014 08:10:17 +0000 Lustre 2.4.1 Lustre 2.5.0 Lustre 2.6.0 Lustre 2.5.1 0 9 <p>I hope the patch <a href="http://review.whamcloud.com/8347" class="external-link" target="_blank" rel="nofollow noopener">http://review.whamcloud.com/8347</a> should resolve this.</p> <p>The backtrace shows it assert on exp_rootfh(), I suspect this is not a common issue for all fh encode, could you make sure of this?</p> <p>This is general case. From the code of function <b>exportfs_encode_fh()</b> we can see the path when the <b>parent == NULL</b>.</p> <div class="code panel" style="border-width: 1px;"><div class="codeContent panelContent"> <pre class="code-java"><span class="code-object">int</span> exportfs_encode_fh(struct dentry *dentry, struct fid *fid, <span class="code-object">int</span> *max_len, <span class="code-object">int</span> connectable) { <span class="code-object">int</span> error; struct dentry *p = NULL; struct inode *inode = dentry-&gt;d_inode, *parent = NULL; <span class="code-keyword">if</span> (connectable &amp;&amp; !S_ISDIR(inode-&gt;i_mode)) { p = dget_parent(dentry); /* * note that <span class="code-keyword">while</span> p might've ceased to be our parent already, * it's still pinned by and still positive. */ parent = p-&gt;d_inode; } error = exportfs_encode_inode_fh(inode, fid, max_len, parent); dput(p); <span class="code-keyword">return</span> error; } </pre> </div></div> <p>The flag <b>connectable</b> depends from export flag <b>NFSEXP_NOSUBTREECHECK</b> only.</p> <div class="code panel" style="border-width: 1px;"><div class="codeContent panelContent"> <pre class="code-java">{ struct fid *fid = (struct fid *) (fhp-&gt;fh_handle.fh_auth + fhp-&gt;fh_handle.fh_size/4 - 1); <span class="code-object">int</span> maxsize = (fhp-&gt;fh_maxsize - fhp-&gt;fh_handle.fh_size)/4; <span class="code-object">int</span> subtreecheck = !(exp-&gt;ex_flags &amp; NFSEXP_NOSUBTREECHECK); fhp-&gt;fh_handle.fh_fileid_type = exportfs_encode_fh(dentry, fid, &amp;maxsize, subtreecheck); fhp-&gt;fh_handle.fh_size += maxsize * 4; } </pre> </div></div> <p>In other place we have explicit call:</p> <div class="code panel" style="border-width: 1px;"><div class="codeContent panelContent"> <pre class="code-java"><span class="code-comment">/* we ask <span class="code-keyword">for</span> a non connected handle */</span> retval = exportfs_encode_fh(path-&gt;dentry, (struct fid *)handle-&gt;f_handle, &amp;handle_dwords, 0); </pre> </div></div> <p>Did you disable subtree_check in this test?</p> <p>According my playing with kernel 3.12.0+ with staging lustre client my patch works. NFSD always asked for encode inode without parent independent from "no_subtree_check" option.</p> <p>The patch <a href="http://review.whamcloud.com/8347" class="external-link" target="_blank" rel="nofollow noopener">http://review.whamcloud.com/8347</a> with little adoption enable NFS re-export for Lustre staging client (Lustre: Build Version: v2_3_64_0-g6e62c21-CHANGED-3.9.0).</p> <p>It seems a duplication of <a href="https://jira.whamcloud.com/browse/LU-3717" title="Kernel panic in ll_encode_fh() while testing file handle syscalls on FC18 client" class="issue-link" data-issue-key="LU-3717"><del>LU-3717</del></a>. Please check the patch <a href="http://review.whamcloud.com/#/c/8072/" class="external-link" target="_blank" rel="nofollow noopener">http://review.whamcloud.com/#/c/8072/</a>. You can work more based on such patch if more issues found on 3.12 (or newer) client.</p> <p>Patch landed to master. Do we need to back port this to b2_5 or b2_4?</p> <p>I think we need it in b2_5. Not sure about b2_4. I will back port it to b2_5 only.</p> <p>This is a patch that fixes the issue for the 3.12 mainline kernel client.</p> <p>Landed b2_5</p> Duplicate LU-3717 Related LU-4416 Development Rank 1|hzw8gn: Rank (Obsolete) 11523 Severity