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-16160 Lustre <p>osc_queue_sync_pages() add osc_extent to osc_object's IO extent list without taking ldlm locks, and then it calls osc_io_unplug_async() to queue the IO work for the client.</p> <p>I think the IO extent should take ldlm locks while waiting in the IO work queue.</p> LU-16160

take ldlm lock when queue sync pages

Bug Major Resolved Fixed Zhenyu Xu Zhenyu Xu Thu, 15 Sep 2022 06:45:34 +0000 Fri, 19 Jan 2024 03:14:49 +0000 Tue, 14 Mar 2023 13:11:29 +0000 Lustre 2.16.0 Lustre 2.15.3 0 13 <p>"Bobi Jam &lt;bobijam@hotmail.com&gt;" uploaded a new patch: <a href="https://review.whamcloud.com/48557" class="external-link" target="_blank" rel="nofollow noopener">https://review.whamcloud.com/48557</a><br/> Subject: <a href="https://jira.whamcloud.com/browse/LU-16160" title="take ldlm lock when queue sync pages" class="issue-link" data-issue-key="LU-16160"><del>LU-16160</del></a> osc: take ldlm lock when queue sync pages<br/> Project: fs/lustre-release<br/> Branch: master<br/> Current Patch Set: 1<br/> Commit: 5fddc18876c2cdc7321bfeb9a1e8e30f733a4aa7</p> <p>"Zhenyu Xu &lt;bobijam@hotmail.com&gt;" uploaded a new patch: <a href="https://review.whamcloud.com/48607" class="external-link" target="_blank" rel="nofollow noopener">https://review.whamcloud.com/48607</a><br/> Subject: <a href="https://jira.whamcloud.com/browse/LU-16160" title="take ldlm lock when queue sync pages" class="issue-link" data-issue-key="LU-16160"><del>LU-16160</del></a> llite: clear stale page's uptodate bit<br/> Project: fs/lustre-release<br/> Branch: master<br/> Current Patch Set: 1<br/> Commit: c04d9374b0bf05cccc84f353200ff29cc65b2af7</p> <p>"Oleg Drokin &lt;green@whamcloud.com&gt;" merged in patch <a href="https://review.whamcloud.com/48557/" class="external-link" target="_blank" rel="nofollow noopener">https://review.whamcloud.com/48557/</a><br/> Subject: <a href="https://jira.whamcloud.com/browse/LU-16160" title="take ldlm lock when queue sync pages" class="issue-link" data-issue-key="LU-16160"><del>LU-16160</del></a> osc: take ldlm lock when queue sync pages<br/> Project: fs/lustre-release<br/> Branch: master<br/> Current Patch Set: <br/> Commit: 67aca1fcc6bed20794832decdba590a758d67d8f</p> <p>Just for record.<br/> Cray hits an stale page read now. but it's not related to the race described in the "Subject: <a href="https://jira.whamcloud.com/browse/LU-16160" title="take ldlm lock when queue sync pages" class="issue-link" data-issue-key="LU-16160"><del>LU-16160</del></a> llite: clear stale page's uptodate bit" patch.</p> <p>Based on current logs - large time between lock cancel and next read. So high likely old race (from <a href="https://jira.whamcloud.com/browse/LU-14541" title="Memory reclaim caused a stale data read" class="issue-link" data-issue-key="LU-14541"><del>LU-14541</del></a>) returned. Not a race between swapd and read. I work on grab additional details now.</p> <p>"Zhenyu Xu &lt;bobijam@hotmail.com&gt;" uploaded a new patch: <a href="https://review.whamcloud.com/48673" class="external-link" target="_blank" rel="nofollow noopener">https://review.whamcloud.com/48673</a><br/> Subject: <a href="https://jira.whamcloud.com/browse/LU-16160" title="take ldlm lock when queue sync pages" class="issue-link" data-issue-key="LU-16160"><del>LU-16160</del></a> llite: clear page uptodate bit on cache drop<br/> Project: fs/lustre-release<br/> Branch: master<br/> Current Patch Set: 1<br/> Commit: 265f4a8a08f299a5de10bdc07167d875d8fb2531</p> <p>"Zhenyu Xu &lt;bobijam@hotmail.com&gt;" uploaded a new patch: <a href="https://review.whamcloud.com/c/fs/lustre-release/+/48804" class="external-link" target="_blank" rel="nofollow noopener">https://review.whamcloud.com/c/fs/lustre-release/+/48804</a><br/> Subject: <a href="https://jira.whamcloud.com/browse/LU-16160" title="take ldlm lock when queue sync pages" class="issue-link" data-issue-key="LU-16160"><del>LU-16160</del></a> llite: clear stale page's uptodate bit<br/> Project: fs/lustre-release<br/> Branch: b2_15<br/> Current Patch Set: 1<br/> Commit: d2dc3487cdb845dd79124856101d62ae6b9f8f10</p> <p>"Zhenyu Xu &lt;bobijam@hotmail.com&gt;" uploaded a new patch: <a href="https://review.whamcloud.com/c/fs/lustre-release/+/48805" class="external-link" target="_blank" rel="nofollow noopener">https://review.whamcloud.com/c/fs/lustre-release/+/48805</a><br/> Subject: <a href="https://jira.whamcloud.com/browse/LU-16160" title="take ldlm lock when queue sync pages" class="issue-link" data-issue-key="LU-16160"><del>LU-16160</del></a> llite: clear page uptodate bit on cache drop<br/> Project: fs/lustre-release<br/> Branch: b2_15<br/> Current Patch Set: 1<br/> Commit: 47eb0bee9bd74e6f2472b1b4a372aadc2c591ad7</p> <p>I replicated a situation when page live in cache with uptodate flag and without cl_page.<br/> Coredump + modules + kernel available at <a href="http://shadowland.me:8080/webshare/WhamCloud/16150/log1/" class="external-link" target="_blank" rel="nofollow noopener">http://shadowland.me:8080/webshare/WhamCloud/16150/log1/</a>.</p> <p>"Oleg Drokin &lt;green@whamcloud.com&gt;" merged in patch <a href="https://review.whamcloud.com/c/fs/lustre-release/+/48607/" class="external-link" target="_blank" rel="nofollow noopener">https://review.whamcloud.com/c/fs/lustre-release/+/48607/</a><br/> Subject: <a href="https://jira.whamcloud.com/browse/LU-16160" title="take ldlm lock when queue sync pages" class="issue-link" data-issue-key="LU-16160"><del>LU-16160</del></a> llite: clear stale page's uptodate bit<br/> Project: fs/lustre-release<br/> Branch: master<br/> Current Patch Set: <br/> Commit: 5b911e03261c3de6b0c2934c86dd191f01af4f2f</p> <p>All existing patches have landed. Please reopen if more work is to be tracked under this ticket.</p> <p>"Zhenyu Xu &lt;bobijam@hotmail.com&gt;" uploaded a new patch: <a href="https://review.whamcloud.com/c/fs/lustre-release/+/49372" class="external-link" target="_blank" rel="nofollow noopener">https://review.whamcloud.com/c/fs/lustre-release/+/49372</a><br/> Subject: <a href="https://jira.whamcloud.com/browse/LU-16160" title="take ldlm lock when queue sync pages" class="issue-link" data-issue-key="LU-16160"><del>LU-16160</del></a> llite: make page state consistent<br/> Project: fs/lustre-release<br/> Branch: master<br/> Current Patch Set: 1<br/> Commit: 7307e91ee06955c3fafad60e0a6c6f329736ccfd</p> <p>"Zhenyu Xu &lt;bobijam@hotmail.com&gt;" uploaded a new patch: <a href="https://review.whamcloud.com/c/fs/lustre-release/+/49534" class="external-link" target="_blank" rel="nofollow noopener">https://review.whamcloud.com/c/fs/lustre-release/+/49534</a><br/> Subject: <a href="https://jira.whamcloud.com/browse/LU-16160" title="take ldlm lock when queue sync pages" class="issue-link" data-issue-key="LU-16160"><del>LU-16160</del></a> llite: handle filemap_fault() returns SIGBUS error<br/> Project: fs/lustre-release<br/> Branch: master<br/> Current Patch Set: 1<br/> Commit: 437ee8064530b3db7a7c8a8a26e0694273bf0919</p> <p>"Zhenyu Xu &lt;bobijam@hotmail.com&gt;" uploaded a new patch: <a href="https://review.whamcloud.com/c/fs/lustre-release/+/49541" class="external-link" target="_blank" rel="nofollow noopener">https://review.whamcloud.com/c/fs/lustre-release/+/49541</a><br/> Subject: <a href="https://jira.whamcloud.com/browse/LU-16160" title="take ldlm lock when queue sync pages" class="issue-link" data-issue-key="LU-16160"><del>LU-16160</del></a> llite: revert commit 5b911e03261c3<br/> Project: fs/lustre-release<br/> Branch: master<br/> Current Patch Set: 1<br/> Commit: 4e63d5c7b94b42c8b8936a591a967b8138ad04fc</p> <p>&gt; Please reopen if more work is to be tracked under this ticket<br/> Reopening since <a href="https://jira.whamcloud.com/secure/ViewProfile.jspa?name=bobijam" class="user-hover" rel="bobijam">bobijam</a> has posted a revert and more work against this ticket.</p> <p>"Zhenyu Xu &lt;bobijam@hotmail.com&gt;" uploaded a new patch: <a href="https://review.whamcloud.com/c/fs/lustre-release/+/49549" class="external-link" target="_blank" rel="nofollow noopener">https://review.whamcloud.com/c/fs/lustre-release/+/49549</a><br/> Subject: <a href="https://jira.whamcloud.com/browse/LU-16160" title="take ldlm lock when queue sync pages" class="issue-link" data-issue-key="LU-16160"><del>LU-16160</del></a> llite: handle filemap_fault() returns SIGBUS error<br/> Project: fs/lustre-release<br/> Branch: b2_12<br/> Current Patch Set: 1<br/> Commit: 2e20813f4f355d6c8d0cf0ac861fefbf35d7f6f3</p> <p>"Zhenyu Xu &lt;bobijam@hotmail.com&gt;" uploaded a new patch: <a href="https://review.whamcloud.com/c/fs/lustre-release/+/49553" class="external-link" target="_blank" rel="nofollow noopener">https://review.whamcloud.com/c/fs/lustre-release/+/49553</a><br/> Subject: <a href="https://jira.whamcloud.com/browse/LU-16160" title="take ldlm lock when queue sync pages" class="issue-link" data-issue-key="LU-16160"><del>LU-16160</del></a> llite: handle filemap_fault() returns SIGBUS error<br/> Project: fs/lustre-release<br/> Branch: b2_15<br/> Current Patch Set: 1<br/> Commit: 7b9342563fc08f800fa4379b189f3b5df6ebd570</p> <p>I've uploaded the workaround for stale data+SIGBUS issue that we are currently using, as an attachment. We're going to replace it with a better (design-wise/performance-wise) solution if/when we have it. Not sure if it's useful but we'd like to share.</p> <p>Thanks very much Andrew. Is it possible to push the patch into gerrit so it is easier for us to provide testing/review feedback?</p> <p>Andrew</p> <p>Bobijam feels that your attached patch takes a similar approach to <a href="https://review.whamcloud.com/c/fs/lustre-release/+/49534" class="external-link" target="_blank" rel="nofollow noopener">his latest LU-16160 patch</a>. Could you please review the latter in gerrit to flag any issues that should deter us from landing this to master?</p> <p>Thanks</p> <p>Peter</p> <p>Peter, sure, I'll do that. Thank you</p> <p>"Patrick Farrell &lt;farr0186@gmail.com&gt;" uploaded a new patch: <a href="https://review.whamcloud.com/c/fs/lustre-release/+/49647" class="external-link" target="_blank" rel="nofollow noopener">https://review.whamcloud.com/c/fs/lustre-release/+/49647</a><br/> Subject: <a href="https://jira.whamcloud.com/browse/LU-16160" title="take ldlm lock when queue sync pages" class="issue-link" data-issue-key="LU-16160"><del>LU-16160</del></a> llite: SIGBUS is possible on a race with page reclaim<br/> Project: fs/lustre-release<br/> Branch: master<br/> Current Patch Set: 1<br/> Commit: 371d94fe71b103c1712bf8f2b1bb1c026cf31de4</p> <p><a href="https://jira.whamcloud.com/secure/ViewProfile.jspa?name=panda" class="user-hover" rel="panda">panda</a> Patrick ported your attached patch to master and pushed it into gerrit, so we can compare and contrast both similar approaches. To that end, please can you confirm that nothing was altered during the porting? Thanks!</p> <p>"Oleg Drokin &lt;green@whamcloud.com&gt;" merged in patch <a href="https://review.whamcloud.com/c/fs/lustre-release/+/49541/" class="external-link" target="_blank" rel="nofollow noopener">https://review.whamcloud.com/c/fs/lustre-release/+/49541/</a><br/> Subject: <a href="https://jira.whamcloud.com/browse/LU-16160" title="take ldlm lock when queue sync pages" class="issue-link" data-issue-key="LU-16160"><del>LU-16160</del></a> revert: "llite: clear stale page's uptodate bit"<br/> Project: fs/lustre-release<br/> Branch: master<br/> Current Patch Set: <br/> Commit: 84c9618190f9e3a526ce51dc4995fcfa3a9ed265</p> <p>"Oleg Drokin &lt;green@whamcloud.com&gt;" merged in patch <a href="https://review.whamcloud.com/c/fs/lustre-release/+/49647/" class="external-link" target="_blank" rel="nofollow noopener">https://review.whamcloud.com/c/fs/lustre-release/+/49647/</a><br/> Subject: <a href="https://jira.whamcloud.com/browse/LU-16160" title="take ldlm lock when queue sync pages" class="issue-link" data-issue-key="LU-16160"><del>LU-16160</del></a> llite: SIGBUS is possible on a race with page reclaim<br/> Project: fs/lustre-release<br/> Branch: master<br/> Current Patch Set: <br/> Commit: b4da788a819f82d35b685d6ee7f02809c05ca005</p> <p>"Patrick Farrell &lt;pfarrell@whamcloud.com&gt;" uploaded a new patch: <a href="https://review.whamcloud.com/c/fs/lustre-release/+/50202" class="external-link" target="_blank" rel="nofollow noopener">https://review.whamcloud.com/c/fs/lustre-release/+/50202</a><br/> Subject: <a href="https://jira.whamcloud.com/browse/LU-16160" title="take ldlm lock when queue sync pages" class="issue-link" data-issue-key="LU-16160"><del>LU-16160</del></a> llite: SIGBUS is possible on a race with page reclaim<br/> Project: fs/lustre-release<br/> Branch: b2_15<br/> Current Patch Set: 1<br/> Commit: f817fcd2c553c223fc89087f62c9d59517bb8e59</p> <p>Looks like everything tracked under this ticket has landed for 2.16</p> <p>"Oleg Drokin &lt;green@whamcloud.com&gt;" merged in patch <a href="https://review.whamcloud.com/c/fs/lustre-release/+/50202/" class="external-link" target="_blank" rel="nofollow noopener">https://review.whamcloud.com/c/fs/lustre-release/+/50202/</a><br/> Subject: <a href="https://jira.whamcloud.com/browse/LU-16160" title="take ldlm lock when queue sync pages" class="issue-link" data-issue-key="LU-16160"><del>LU-16160</del></a> llite: SIGBUS is possible on a race with page reclaim<br/> Project: fs/lustre-release<br/> Branch: b2_15<br/> Current Patch Set: <br/> Commit: b0a6d4d08e19d06661deabdb7278f07662d8b6e8</p> Related LU-16156 LU-14541 LU-15815 LU-16401 LU-16224 Development Rank 1|i0307r: Rank (Obsolete) 9223372036854775807 Severity