CINXE.COM

<?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 16/18] posix-timers: Dont iterate /proc/$PID/timers with sighand:: Siglock held</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=" /><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></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/2025/3/8/443" onclick="this.href='/lkml/headers'+'/2025/3/8/443';">[headers]</a>聽 <a href="/lkml/bounce/2025/3/8/443">[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><a href="/lkml/2025/3/8/426">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 class="origin"><a href="/lkml/2025/3/9/14">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><div class="threadlist">Patch in this message</div><ul class="threadlist"><li><a href="/lkml/diff/2025/3/8/443/1">Get diff 1</a></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 16/18] posix-timers: Dont iterate /proc/$PID/timers with sighand:: Siglock held</td></tr><tr><td class="lp">Date</td><td class="rp" itemprop="datePublished">Sat, 8 Mar 2025 17:48:45 +0100 (CET)</td></tr></table></td><td></td></tr></table><pre itemprop="articleBody">The readout of /proc/$PID/timers holds sighand::siglock with interrupts disabled. That is required to protect against concurrent modifications of the task::signal::posix_timers list because the list is not RCU safe. With the conversion of the timer storage to a RCU protected hlist, this is not longer required. The only requirement is to protect the returned entry against a concurrent free, which is trivial as the timers are RCU protected. Removing the trylock of sighand::siglock is benign because the life time of task_struct::signal is bound to the life time of the task_struct itself. There are two scenarios where this matters: 1) The process is life and not about to be checkpointed 2) The process is stopped via ptrace for checkpointing #1 is a racy snapshot of the armed timers and nothing can rely on it. It's not more than debug information and it has been that way before because sighand lock is dropped when the buffer is full and the restart of the iteration might find a completely different set of timers. The task and therefore task::signal cannot be freed as timers_start() acquired a reference count via get_pid_task(). #2 the process is stopped for checkpointing so nothing can delete or create timers at this point. Neither can the process exit during the traversal. If CRIU fails to observe an exit in progress prior to the dissimination of the timers, then there are more severe problems to solve in the CRIU mechanics as they can't rely on posix timers being enabled in the first place. Therefore replace the lock acquisition with rcu_read_lock() and switch the timer storage traversal over to seq_hlist_*_rcu(). Signed-off-by: Thomas Gleixner <tglx@linutronix.de> --- fs/proc/base.c | 48 ++++++++++++++++++++---------------------------- 1 file changed, 20 insertions(+), 28 deletions(-) --- a/fs/proc/base.c +++ b/fs/proc/base.c @@ -2497,11 +2497,9 @@ static const struct file_operations proc #if defined(CONFIG_CHECKPOINT_RESTORE) && defined(CONFIG_POSIX_TIMERS) struct timers_private { - struct pid *pid; - struct task_struct *task; - struct sighand_struct *sighand; - struct pid_namespace *ns; - unsigned long flags; + struct pid *pid; + struct task_struct *task; + struct pid_namespace *ns; }; static void *timers_start(struct seq_file *m, loff_t *pos) @@ -2512,54 +2510,48 @@ static void *timers_start(struct seq_fil if (!tp->task) return ERR_PTR(-ESRCH); - tp->sighand = lock_task_sighand(tp->task, &tp->flags); - if (!tp->sighand) - return ERR_PTR(-ESRCH); - - return seq_hlist_start(&tp->task->signal->posix_timers, *pos); + rcu_read_lock(); + return seq_hlist_start_rcu(&tp->task->signal->posix_timers, *pos); } static void *timers_next(struct seq_file *m, void *v, loff_t *pos) { struct timers_private *tp = m->private; - return seq_hlist_next(v, &tp->task->signal->posix_timers, pos); + + return seq_hlist_next_rcu(v, &tp->task->signal->posix_timers, pos); } static void timers_stop(struct seq_file *m, void *v) { struct timers_private *tp = m->private; - if (tp->sighand) { - unlock_task_sighand(tp->task, &tp->flags); - tp->sighand = NULL; - } - if (tp->task) { put_task_struct(tp->task); tp->task = NULL; + rcu_read_unlock(); } } static int show_timer(struct seq_file *m, void *v) { - struct k_itimer *timer; - struct timers_private *tp = m->private; - int notify; static const char * const nstr[] = { - [SIGEV_SIGNAL] = "signal", - [SIGEV_NONE] = "none", - [SIGEV_THREAD] = "thread", + [SIGEV_SIGNAL] = "signal", + [SIGEV_NONE] = "none", + [SIGEV_THREAD] = "thread", }; - timer = hlist_entry((struct hlist_node *)v, struct k_itimer, list); - notify = timer->it_sigev_notify; + struct k_itimer *timer = hlist_entry((struct hlist_node *)v, struct k_itimer, list); + struct timers_private *tp = m->private; + int notify = timer->it_sigev_notify; + + guard(spinlock_irq)(&timer->it_lock); + if (!posixtimer_valid(timer)) + return 0; seq_printf(m, "ID: %d\n", timer->it_id); - seq_printf(m, "signal: %d/%px\n", - timer->sigq.info.si_signo, + seq_printf(m, "signal: %d/%px\n", timer->sigq.info.si_signo, timer->sigq.info.si_value.sival_ptr); - seq_printf(m, "notify: %s/%s.%d\n", - nstr[notify & ~SIGEV_THREAD_ID], + seq_printf(m, "notify: %s/%s.%d\n", nstr[notify & ~SIGEV_THREAD_ID], (notify & SIGEV_THREAD_ID) ? "tid" : "pid", pid_nr_ns(timer->it_pid, tp->ns)); seq_printf(m, "ClockID: %d\n", timer->it_clock); </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 17:52 聽聽 [W:1.188 / U:1.121 seconds] 漏2003-2020 <a href="http://blog.jasper.es/">Jasper Spaans</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>