CINXE.COM
LKML: Thomas Gleixner: [patch V3 00/18] posix-timers: Rework the global hash table and provide a sane mechanism for CRIU
<?xml version="1.0" encoding="UTF-8" standalone="yes"?> <!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd"><html xmlns="http://www.w3.org/1999/xhtml"><head><meta http-equiv="Content-Type" content="text/html; charset=UTF-8" /><title>LKML: Thomas Gleixner: [patch V3 00/18] posix-timers: Rework the global hash table and provide a sane mechanism for CRIU</title><link href="/css/message.css" rel="stylesheet" type="text/css" /><link href="/css/wrap.css" rel="alternate stylesheet" type="text/css" title="wrap" /><link href="/css/nowrap.css" rel="stylesheet" type="text/css" title="nowrap" /><link href="/favicon.ico" rel="shortcut icon" /><script src="/js/simple-calendar.js" type="text/javascript"></script><script src="/js/styleswitcher.js" type="text/javascript"></script><link rel="alternate" type="application/rss+xml" title="lkml.org : last 100 messages" href="/rss.php" /><link rel="alternate" type="application/rss+xml" title="lkml.org : last messages by Thomas Gleixner" href="/groupie.php?aid=" /><!--Matomo--><script> var _paq = window._paq = window._paq || []; /* tracker methods like "setCustomDimension" should be called before "trackPageView" */ _paq.push(["setDoNotTrack", true]); _paq.push(["disableCookies"]); _paq.push(['trackPageView']); _paq.push(['enableLinkTracking']); (function() { var u="//m.lkml.org/"; _paq.push(['setTrackerUrl', u+'matomo.php']); _paq.push(['setSiteId', '1']); var d=document, g=d.createElement('script'), s=d.getElementsByTagName('script')[0]; g.async=true; g.src=u+'matomo.js'; s.parentNode.insertBefore(g,s); })(); </script><!--End Matomo Code--></head><body onload="es.jasper.simpleCalendar.init();" itemscope="itemscope" itemtype="http://schema.org/BlogPosting"><table border="0" cellpadding="0" cellspacing="0"><tr><td width="180" align="center"><a href="/"><img style="border:0;width:135px;height:32px" src="/images/toprowlk.gif" alt="lkml.org" /></a></td><td width="32">聽</td><td class="nb"><div><a class="nb" href="/lkml"> [lkml]</a> 聽 <a class="nb" href="/lkml/2025"> [2025]</a> 聽 <a class="nb" href="/lkml/2025/3"> [Mar]</a> 聽 <a class="nb" href="/lkml/2025/3/8"> [8]</a> 聽 <a class="nb" href="/lkml/last100"> [last100]</a> 聽 <a href="/rss.php"><img src="/images/rss-or.gif" border="0" alt="RSS Feed" /></a></div><div>Views: <a href="#" class="nowrap" onclick="setActiveStyleSheet('wrap');return false;">[wrap]</a><a href="#" class="wrap" onclick="setActiveStyleSheet('nowrap');return false;">[no wrap]</a> 聽 <a class="nb" href="/lkml/mheaders/2025/3/8/426" onclick="this.href='/lkml/headers'+'/2025/3/8/426';">[headers]</a>聽 <a href="/lkml/bounce/2025/3/8/426">[forward]</a>聽 </div></td><td width="32">聽</td></tr><tr><td valign="top"><div class="es-jasper-simpleCalendar" baseurl="/lkml/"></div><div class="threadlist">Messages in this thread</div><ul class="threadlist"><li class="root"><a href="/lkml/2025/3/8/426">First message in thread</a></li><li class="origin"><a href="/lkml/2025/3/8/427">Thomas Gleixner</a><ul><li><a href="/lkml/2025/3/8/427">Thomas Gleixner</a><ul><li><a href="/lkml/2025/3/8/616">Frederic Weisbecker</a></li><li><a href="/lkml/2025/3/13/634">"tip-bot2 for Thomas Gleixner"</a></li></ul></li><li><a href="/lkml/2025/3/8/428">Thomas Gleixner</a><ul><li><a href="/lkml/2025/3/11/914">Frederic Weisbecker</a><ul><li><a href="/lkml/2025/3/11/977">Thomas Gleixner</a></li></ul></li><li><a href="/lkml/2025/3/13/633">"tip-bot2 for Eric Dumazet"</a></li></ul></li><li><a href="/lkml/2025/3/8/430">Thomas Gleixner</a><ul><li><a href="/lkml/2025/3/13/631">"tip-bot2 for Eric Dumazet"</a></li></ul></li><li><a href="/lkml/2025/3/8/431">Thomas Gleixner</a><ul><li><a href="/lkml/2025/3/13/630">"tip-bot2 for Thomas Gleixner"</a></li></ul></li><li><a href="/lkml/2025/3/8/432">Thomas Gleixner</a><ul><li><a href="/lkml/2025/3/13/632">"tip-bot2 for Thomas Gleixner"</a></li></ul></li><li><a href="/lkml/2025/3/8/433">Thomas Gleixner</a><ul><li><a href="/lkml/2025/3/13/629">"tip-bot2 for Thomas Gleixner"</a></li></ul></li><li><a href="/lkml/2025/3/8/434">Thomas Gleixner</a><ul><li><a href="/lkml/2025/3/13/627">"tip-bot2 for Thomas Gleixner"</a></li></ul></li><li><a href="/lkml/2025/3/8/435">Thomas Gleixner</a><ul><li><a href="/lkml/2025/3/13/626">"tip-bot2 for Thomas Gleixner"</a></li></ul></li><li><a href="/lkml/2025/3/8/436">Thomas Gleixner</a><ul><li><a href="/lkml/2025/3/10/13">Frederic Weisbecker</a><ul><li><a href="/lkml/2025/3/10/148">Thomas Gleixner</a></li></ul></li><li><a href="/lkml/2025/3/10/256">Thomas Gleixner</a><ul><li><a href="/lkml/2025/3/13/628">"tip-bot2 for Thomas Gleixner"</a></li></ul></li></ul></li><li><a href="/lkml/2025/3/8/437">Thomas Gleixner</a><ul><li><a href="/lkml/2025/3/10/634">Frederic Weisbecker</a><ul><li><a href="/lkml/2025/3/10/1329">Thomas Gleixner</a><ul><li><a href="/lkml/2025/3/10/1664">Frederic Weisbecker</a></li></ul></li></ul></li><li><a href="/lkml/2025/3/13/625">"tip-bot2 for Peter Zijlstra"</a></li></ul></li><li><a href="/lkml/2025/3/8/438">Thomas Gleixner</a><ul><li><a href="/lkml/2025/3/10/1716">Frederic Weisbecker</a></li><li><a href="/lkml/2025/3/11/927">Frederic Weisbecker</a></li><li><a href="/lkml/2025/3/13/623">"tip-bot2 for Eric Dumazet"</a></li></ul></li><li><a href="/lkml/2025/3/8/439">Thomas Gleixner</a><ul><li><a href="/lkml/2025/3/11/932">Frederic Weisbecker</a></li><li><a href="/lkml/2025/3/13/624">"tip-bot2 for Thomas Gleixner"</a></li></ul></li><li><a href="/lkml/2025/3/8/440">Thomas Gleixner</a><ul><li><a href="/lkml/2025/3/13/622">"tip-bot2 for Thomas Gleixner"</a></li></ul></li><li><a href="/lkml/2025/3/8/441">Thomas Gleixner</a><ul><li><a href="/lkml/2025/3/11/940">Frederic Weisbecker</a></li><li><a href="/lkml/2025/3/13/621">"tip-bot2 for Thomas Gleixner"</a></li><li><a href="/lkml/2025/3/13/1624">David Laight</a></li><li><a href="/lkml/2025/3/17/208">"Nysal Jan K.A."</a></li></ul></li><li><a href="/lkml/2025/3/8/442">Thomas Gleixner</a><ul><li><a href="/lkml/2025/3/11/1110">Frederic Weisbecker</a></li><li><a href="/lkml/2025/3/13/620">"tip-bot2 for Thomas Gleixner"</a></li></ul></li><li><a href="/lkml/2025/3/8/443">Thomas Gleixner</a><ul><li><a href="/lkml/2025/3/9/14">Cyrill Gorcunov</a></li><li><a href="/lkml/2025/3/11/1103">Frederic Weisbecker</a></li><li><a href="/lkml/2025/3/13/619">"tip-bot2 for Thomas Gleixner"</a></li></ul></li><li><a href="/lkml/2025/3/8/444">Thomas Gleixner</a><ul><li><a href="/lkml/2025/3/9/13">Cyrill Gorcunov</a></li><li><a href="/lkml/2025/3/11/1588">Frederic Weisbecker</a><ul><li><a href="/lkml/2025/3/11/1615">Thomas Gleixner</a><ul><li><a href="/lkml/2025/3/11/1618">Thomas Gleixner</a></li></ul></li><li><a href="/lkml/2025/3/12/649">Cyrill Gorcunov</a></li></ul></li></ul></li><li><a href="/lkml/2025/3/8/445">Thomas Gleixner</a><ul><li><a href="/lkml/2025/3/10/254">Thomas Gleixner</a><ul><li><a href="/lkml/2025/3/11/1599">Frederic Weisbecker</a></li><li><a href="/lkml/2025/3/13/617">"tip-bot2 for Thomas Gleixner"</a></li></ul></li></ul></li></ul></li></ul></td><td width="32" rowspan="2" class="c" valign="top"><img src="/images/icornerl.gif" width="32" height="32" alt="/" /></td><td class="c" rowspan="2" valign="top" style="padding-top: 1em"><table><tr><td><table><tr><td class="lp">From</td><td class="rp" itemprop="author">Thomas Gleixner <></td></tr><tr><td class="lp">Subject</td><td class="rp" itemprop="name">[patch V3 00/18] posix-timers: Rework the global hash table and provide a sane mechanism for CRIU</td></tr><tr><td class="lp">Date</td><td class="rp" itemprop="datePublished">Sat, 8 Mar 2025 17:48:08 +0100 (CET)</td></tr></table></td><td></td></tr></table><pre itemprop="articleBody">This is a follow up on V2 of this work, which can be found here:<br /><br /> <a href="https://lore.kernel.org/all/20250224095736.145530367@linutronix.de">https://lore.kernel.org/all/20250224095736.145530367@linutronix.de</a><br /><br />It addresses the scalability problem of the posix timer hash and provides a<br />performant mechanism to restore Posix timers with a given ID along with a<br />couple of preperatory cleanups and enhancements. More details about<br />implementation choices are in the change logs and the cover letter of V1:<br /><br /> <a href="https://lore.kernel.org/all/20250302185753.311903554@linutronix.de">https://lore.kernel.org/all/20250302185753.311903554@linutronix.de</a><br /><br />Changes vs. V2:<br /><br /> - Ensure consistency on timer_create() (new patch) - Frederic<br /> - Pick up the lock_timer() conditional unlock fix (was V2a)<br /> - Use proper defines in selftests<br /> - Pick up review/ack tags<br /><br />The series survives all posix timer tests and did not show any regressions<br />so far.<br /><br />The series is based on:<br /><br /> git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip timers/core<br /><br />and is also available from git:<br /><br /> git://git.kernel.org/pub/scm/linux/kernel/git/tglx/devel.git timers/posix<br /><br />Thanks,<br /><br /> tglx<br />---<br />Eric Dumazet (3):<br /> posix-timers: Initialise timer before adding it to the hash table<br /> posix-timers: Add cond_resched() to posix_timer_add() search loop<br /> posix-timers: Make signal_struct:: Next_posix_timer_id an atomic_t<br /><br />Peter Zijlstra (1):<br /> posix-timers: Make lock_timer() use guard()<br /><br />Thomas Gleixner (14):<br /> posix-timers: Ensure that timer initialization is fully visible<br /> posix-timers: Cleanup includes<br /> posix-timers: Remove a few paranoid warnings<br /> posix-timers: Remove SLAB_PANIC from kmem cache<br /> posix-timers: Use guards in a few places<br /> posix-timers: Simplify lock/unlock_timer()<br /> posix-timers: Rework timer removal<br /> posix-timers: Improve hash table performance<br /> posix-timers: Switch to jhash32()<br /> posix-timers: Avoid false cacheline sharing<br /> posix-timers: Make per process list RCU safe<br /> posix-timers: Dont iterate /proc/$PID/timers with sighand:: Siglock held<br /> posix-timers: Provide a mechanism to allocate a given timer ID<br /> selftests/timers/posix-timers: Add a test for exact allocation mode<br /><br /><br /> fs/proc/base.c | 48 --<br /> include/linux/cleanup.h | 22 -<br /> include/linux/posix-timers.h | 30 +<br /> include/linux/sched/signal.h | 3 <br /> include/uapi/linux/prctl.h | 10 <br /> kernel/signal.c | 2 <br /> kernel/sys.c | 5 <br /> kernel/time/posix-timers.c | 540 +++++++++++++-------------<br /> tools/testing/selftests/timers/posix_timers.c | 66 +++<br /> 9 files changed, 418 insertions(+), 308 deletions(-)<br /><br /></pre></td><td width="32" rowspan="2" class="c" valign="top"><img src="/images/icornerr.gif" width="32" height="32" alt="\" /></td></tr><tr><td align="right" valign="bottom"> 聽 </td></tr><tr><td align="right" valign="bottom">聽</td><td class="c" valign="bottom" style="padding-bottom: 0px"><img src="/images/bcornerl.gif" width="32" height="32" alt="\" /></td><td class="c">聽</td><td class="c" valign="bottom" style="padding-bottom: 0px"><img src="/images/bcornerr.gif" width="32" height="32" alt="/" /></td></tr><tr><td align="right" valign="top" colspan="2"> 聽 </td><td class="lm">Last update: 2025-03-08 17:49 聽聽 [W:1.231 / U:2.238 seconds]<br />漏2003-2020 <a href="http://blog.jasper.es/"><span itemprop="editor">Jasper Spaans</span></a>|hosted at <a href="https://www.digitalocean.com/?refcode=9a8e99d24cf9">Digital Ocean</a> and my Meterkast|<a href="http://blog.jasper.es/categories.html#lkml-ref">Read the blog</a></td><td>聽</td></tr></table><script language="javascript" src="/js/styleswitcher.js" type="text/javascript"></script></body></html>