Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1758044Ab3CDEEw (ORCPT ); Sun, 3 Mar 2013 23:04:52 -0500 Received: from shadbolt.e.decadent.org.uk ([88.96.1.126]:32833 "EHLO shadbolt.e.decadent.org.uk" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1756258Ab3CDDnc (ORCPT ); Sun, 3 Mar 2013 22:43:32 -0500 Message-Id: <20130304033717.200389731@decadent.org.uk> User-Agent: quilt/0.60-1 Date: Mon, 04 Mar 2013 03:38:44 +0000 From: Ben Hutchings To: linux-kernel@vger.kernel.org, stable@vger.kernel.org Cc: akpm@linux-foundation.org, Sagi Grimberg , Andrea Arcangeli , Peter Zijlstra , Haggai Eran , "Paul E. McKenney" , Linus Torvalds Subject: [ 097/153] mm: mmu_notifier: have mmu_notifiers use a global SRCU so they may safely schedule In-Reply-To: <20130304033707.648729212@decadent.org.uk> X-SA-Exim-Connect-IP: 2001:470:1f08:1539:a11:96ff:fec6:70c4 X-SA-Exim-Mail-From: ben@decadent.org.uk X-SA-Exim-Scanned: No (on shadbolt.decadent.org.uk); SAEximRunCond expanded to false Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Content-Length: 9250 Lines: 293 3.2-stable review patch. If anyone has any objections, please let me know. ------------------ From: Sagi Grimberg commit 21a92735f660eaecf69a6f2e777f18463760ec32 upstream. With an RCU based mmu_notifier implementation, any callout to mmu_notifier_invalidate_range_{start,end}() or mmu_notifier_invalidate_page() would not be allowed to call schedule() as that could potentially allow a modification to the mmu_notifier structure while it is currently being used. Since srcu allocs 4 machine words per instance per cpu, we may end up with memory exhaustion if we use srcu per mm. So all mms share a global srcu. Note that during large mmu_notifier activity exit & unregister paths might hang for longer periods, but it is tolerable for current mmu_notifier clients. Signed-off-by: Sagi Grimberg Signed-off-by: Andrea Arcangeli Cc: Peter Zijlstra Cc: Haggai Eran Cc: "Paul E. McKenney" Signed-off-by: Andrew Morton Signed-off-by: Linus Torvalds Signed-off-by: Ben Hutchings --- include/linux/mmu_notifier.h | 1 + mm/mmu_notifier.c | 73 +++++++++++++++++++++++++++--------------- 2 files changed, 49 insertions(+), 25 deletions(-) --- a/include/linux/mmu_notifier.h +++ b/include/linux/mmu_notifier.h @@ -4,6 +4,7 @@ #include #include #include +#include struct mmu_notifier; struct mmu_notifier_ops; --- a/mm/mmu_notifier.c +++ b/mm/mmu_notifier.c @@ -14,10 +14,14 @@ #include #include #include +#include #include #include #include +/* global SRCU for all MMs */ +struct srcu_struct srcu; + /* * This function can't run concurrently against mmu_notifier_register * because mm->mm_users > 0 during mmu_notifier_register and exit_mmap @@ -25,8 +29,8 @@ * in parallel despite there being no task using this mm any more, * through the vmas outside of the exit_mmap context, such as with * vmtruncate. This serializes against mmu_notifier_unregister with - * the mmu_notifier_mm->lock in addition to RCU and it serializes - * against the other mmu notifiers with RCU. struct mmu_notifier_mm + * the mmu_notifier_mm->lock in addition to SRCU and it serializes + * against the other mmu notifiers with SRCU. struct mmu_notifier_mm * can't go away from under us as exit_mmap holds an mm_count pin * itself. */ @@ -34,12 +38,13 @@ void __mmu_notifier_release(struct mm_st { struct mmu_notifier *mn; struct hlist_node *n; + int id; /* * RCU here will block mmu_notifier_unregister until * ->release returns. */ - rcu_read_lock(); + id = srcu_read_lock(&srcu); hlist_for_each_entry_rcu(mn, n, &mm->mmu_notifier_mm->list, hlist) /* * if ->release runs before mmu_notifier_unregister it @@ -50,7 +55,7 @@ void __mmu_notifier_release(struct mm_st */ if (mn->ops->release) mn->ops->release(mn, mm); - rcu_read_unlock(); + srcu_read_unlock(&srcu, id); spin_lock(&mm->mmu_notifier_mm->lock); while (unlikely(!hlist_empty(&mm->mmu_notifier_mm->list))) { @@ -68,7 +73,7 @@ void __mmu_notifier_release(struct mm_st spin_unlock(&mm->mmu_notifier_mm->lock); /* - * synchronize_rcu here prevents mmu_notifier_release to + * synchronize_srcu here prevents mmu_notifier_release to * return to exit_mmap (which would proceed freeing all pages * in the mm) until the ->release method returns, if it was * invoked by mmu_notifier_unregister. @@ -76,7 +81,7 @@ void __mmu_notifier_release(struct mm_st * The mmu_notifier_mm can't go away from under us because one * mm_count is hold by exit_mmap. */ - synchronize_rcu(); + synchronize_srcu(&srcu); } /* @@ -89,14 +94,14 @@ int __mmu_notifier_clear_flush_young(str { struct mmu_notifier *mn; struct hlist_node *n; - int young = 0; + int young = 0, id; - rcu_read_lock(); + id = srcu_read_lock(&srcu); hlist_for_each_entry_rcu(mn, n, &mm->mmu_notifier_mm->list, hlist) { if (mn->ops->clear_flush_young) young |= mn->ops->clear_flush_young(mn, mm, address); } - rcu_read_unlock(); + srcu_read_unlock(&srcu, id); return young; } @@ -106,9 +111,9 @@ int __mmu_notifier_test_young(struct mm_ { struct mmu_notifier *mn; struct hlist_node *n; - int young = 0; + int young = 0, id; - rcu_read_lock(); + id = srcu_read_lock(&srcu); hlist_for_each_entry_rcu(mn, n, &mm->mmu_notifier_mm->list, hlist) { if (mn->ops->test_young) { young = mn->ops->test_young(mn, mm, address); @@ -116,7 +121,7 @@ int __mmu_notifier_test_young(struct mm_ break; } } - rcu_read_unlock(); + srcu_read_unlock(&srcu, id); return young; } @@ -126,8 +131,9 @@ void __mmu_notifier_change_pte(struct mm { struct mmu_notifier *mn; struct hlist_node *n; + int id; - rcu_read_lock(); + id = srcu_read_lock(&srcu); hlist_for_each_entry_rcu(mn, n, &mm->mmu_notifier_mm->list, hlist) { if (mn->ops->change_pte) mn->ops->change_pte(mn, mm, address, pte); @@ -138,7 +144,7 @@ void __mmu_notifier_change_pte(struct mm else if (mn->ops->invalidate_page) mn->ops->invalidate_page(mn, mm, address); } - rcu_read_unlock(); + srcu_read_unlock(&srcu, id); } void __mmu_notifier_invalidate_page(struct mm_struct *mm, @@ -146,13 +152,14 @@ void __mmu_notifier_invalidate_page(stru { struct mmu_notifier *mn; struct hlist_node *n; + int id; - rcu_read_lock(); + id = srcu_read_lock(&srcu); hlist_for_each_entry_rcu(mn, n, &mm->mmu_notifier_mm->list, hlist) { if (mn->ops->invalidate_page) mn->ops->invalidate_page(mn, mm, address); } - rcu_read_unlock(); + srcu_read_unlock(&srcu, id); } void __mmu_notifier_invalidate_range_start(struct mm_struct *mm, @@ -160,13 +167,14 @@ void __mmu_notifier_invalidate_range_sta { struct mmu_notifier *mn; struct hlist_node *n; + int id; - rcu_read_lock(); + id = srcu_read_lock(&srcu); hlist_for_each_entry_rcu(mn, n, &mm->mmu_notifier_mm->list, hlist) { if (mn->ops->invalidate_range_start) mn->ops->invalidate_range_start(mn, mm, start, end); } - rcu_read_unlock(); + srcu_read_unlock(&srcu, id); } void __mmu_notifier_invalidate_range_end(struct mm_struct *mm, @@ -174,13 +182,14 @@ void __mmu_notifier_invalidate_range_end { struct mmu_notifier *mn; struct hlist_node *n; + int id; - rcu_read_lock(); + id = srcu_read_lock(&srcu); hlist_for_each_entry_rcu(mn, n, &mm->mmu_notifier_mm->list, hlist) { if (mn->ops->invalidate_range_end) mn->ops->invalidate_range_end(mn, mm, start, end); } - rcu_read_unlock(); + srcu_read_unlock(&srcu, id); } static int do_mmu_notifier_register(struct mmu_notifier *mn, @@ -192,6 +201,12 @@ static int do_mmu_notifier_register(stru BUG_ON(atomic_read(&mm->mm_users) <= 0); + /* + * Verify that mmu_notifier_init() already run and the global srcu is + * initialized. + */ + BUG_ON(!srcu.per_cpu_ref); + ret = -ENOMEM; mmu_notifier_mm = kmalloc(sizeof(struct mmu_notifier_mm), GFP_KERNEL); if (unlikely(!mmu_notifier_mm)) @@ -274,8 +289,8 @@ void __mmu_notifier_mm_destroy(struct mm /* * This releases the mm_count pin automatically and frees the mm * structure if it was the last user of it. It serializes against - * running mmu notifiers with RCU and against mmu_notifier_unregister - * with the unregister lock + RCU. All sptes must be dropped before + * running mmu notifiers with SRCU and against mmu_notifier_unregister + * with the unregister lock + SRCU. All sptes must be dropped before * calling mmu_notifier_unregister. ->release or any other notifier * method may be invoked concurrently with mmu_notifier_unregister, * and only after mmu_notifier_unregister returned we're guaranteed @@ -290,8 +305,9 @@ void mmu_notifier_unregister(struct mmu_ * RCU here will force exit_mmap to wait ->release to finish * before freeing the pages. */ - rcu_read_lock(); + int id; + id = srcu_read_lock(&srcu); /* * exit_mmap will block in mmu_notifier_release to * guarantee ->release is called before freeing the @@ -299,7 +315,7 @@ void mmu_notifier_unregister(struct mmu_ */ if (mn->ops->release) mn->ops->release(mn, mm); - rcu_read_unlock(); + srcu_read_unlock(&srcu, id); spin_lock(&mm->mmu_notifier_mm->lock); hlist_del_rcu(&mn->hlist); @@ -310,10 +326,17 @@ void mmu_notifier_unregister(struct mmu_ * Wait any running method to finish, of course including * ->release if it was run by mmu_notifier_relase instead of us. */ - synchronize_rcu(); + synchronize_srcu(&srcu); BUG_ON(atomic_read(&mm->mm_count) <= 0); mmdrop(mm); } EXPORT_SYMBOL_GPL(mmu_notifier_unregister); + +static int __init mmu_notifier_init(void) +{ + return init_srcu_struct(&srcu); +} + +module_init(mmu_notifier_init); -- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.kernel.org More majordomo info at http://vger.kernel.org/majordomo-info.html Please read the FAQ at http://www.tux.org/lkml/