From mboxrd@z Thu Jan 1 00:00:00 1970 Authentication-Results: smtp.subspace.kernel.org; arc=none smtp.remote-ip=100.103.45.18 ARC-Seal: i=1; d=subspace.kernel.org; s=arc20261015; a=rsa-sha256; cv=none; t=1793211912; b=apTDTsFmarriQ1SpuveVnWrnB7qwgQDQ8gy3svgoMQjrAqy6AlBK1RPUTrt2Rt6wSOVo 2tVnHyPYl8+l+JGo7gabiJoy8DcEk/E8lcndlTbuhmsMVEyWu8W9xByTGIEt/aJaS8Blt ecKwSlxmBvVU+UtUi7jOJ5FgnLL+53gkcLW5hmz46yjXXcuNq1Re0SOYLh5yWhoKl7lDC f7DRGeWEfJheRoNqCeXmERnMpasTiiDFBuDq+d7TadpflDmQhAg8mw8/+MGLPoIeCZmBc s6PxyTIgkr3DvEUsfYf+YLogFqilfpkJ6aFqDEx/c8yujJdcSizp+Hi9AjJh6yzmITA== ARC-Message-Signature: i=1; d=subspace.kernel.org; s=arc20261015; a=rsa-sha256; c=relaxed/relaxed; t=1793211912; h=Received:DKIM-Signature:From:To:Cc:Subject:Date:Message-ID:X-Mailer: In-Reply-To:References:User-Agent:X-stable:X-Patchwork-Hint: MIME-Version:Content-Transfer-Encoding; bh=Dn4QYnsWCQDCTSYGWltXj6qL87m4ELq32rYdyIFN37o=; b=daS1F9awFS5Sa/iwhEOH1/Gt19sjZgb2fuBIIn4OvybA9NW2Syc+FcKz9dZNA6RN90bE jy4Up1jEKYrNMIFOIt86WOt1c4cyYFiOI0HOqljPcaxmuB2AIzKfuL5UM/51G3pJXFPMe yxce99JPvrxQNGFihX9WiaKddW8jZJJqMw7JTX/afUSzDyS9xaLcXyVMkjeocp/428zSF sCFA9sKWOCARrXmn/c+LbIY/Jdsd0ELkR6xlwcWphf+epDSHO3dm4FnFdahfiOpCplHFI ZDO+6xqpCFV7oilnHRO1O68MBcVyzoqT1eUXanz1EodTRruJsACA/8nTwSqFM12goIA== ARC-Authentication-Results: i=1; smtp.subspace.kernel.org; dkim=pass header.d=linuxfoundation.org header.i=@linuxfoundation.org header.a=rsa-sha256 header.s=korg header.b=X5v1bAgv; arc=none smtp.remote-ip=100.103.45.18 Received: from smtp.kernel.org (quasi.space.kernel.org [100.103.45.18]) (using TLSv1.3 with cipher TLS_AES_256_GCM_SHA384 (256/256 bits) key-exchange X25519 server-signature ECDSA (prime256v1)) (No client certificate requested) by smtp.subspace.kernel.org (Postfix) with ESMTPS id 40165606A31A; Wed, 28 Oct 2026 18:25:12 +0000 (UTC) Authentication-Results: smtp.subspace.kernel.org; dkim=pass (1024-bit key, unprotected) header.d=linuxfoundation.org header.i=@linuxfoundation.org header.a=rsa-sha256 header.s=korg header.b=X5v1bAgv Received: by smtp.kernel.org (Postfix) with ESMTPSA id 908A91F01825; Wed, 28 Oct 2026 18:25:11 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=linuxfoundation.org; s=korg; t=1793211912; bh=Dn4QYnsWCQDCTSYGWltXj6qL87m4ELq32rYdyIFN37o=; h=From:To:Cc:Subject:Date:In-Reply-To:References; b=X5v1bAgvZMwn7s0vDuUvA/XdGKuTHe2ugCmWD2ZInhHw54pG3wjyb5H4bb+eWTGqC qKFJISD8A47gsriiozeaD0wsHZwX4DAt1hpqdxyc4y/M28cKPbfnB3o7F0X7pJY8Km 1Cr2cifheInyeQ7/DLVQPNYTngR05cSF5fDC6i+I= From: Greg Kroah-Hartman To: stable@vger.kernel.org Cc: Greg Kroah-Hartman , patches@lists.linux.dev, Jann Horn , "Peter Zijlstra (Intel)" , Sasha Levin Subject: [PATCH 5.15 543/571] futex: Fix mm reuse handling for FUT_OFF_MMSHARED Date: Wed, 28 Oct 2026 18:01:30 +0100 Message-ID: <20261028170020.921920827@linuxfoundation.org> X-Mailer: git-send-email 2.56.0 In-Reply-To: <20261028170008.009143722@linuxfoundation.org> References: <20261028170008.009143722@linuxfoundation.org> User-Agent: quilt/0.69 X-stable: review X-Patchwork-Hint: ignore X-Mailing-List: patches@lists.linux.dev List-Id: List-Post: List-Help: List-Subscribe: List-Unsubscribe: List-Owner: Precedence: list MIME-Version: 1.0 Content-Transfer-Encoding: 8bit 5.15-stable review patch. If anyone has any objections, please let me know. ------------------ From: Jann Horn [ Upstream commit 3267512ad1c2cc0c7768c5b014dc2cceb398a801 ] A FUT_OFF_MMSHARED futex is a shared futex that refers to an MM. It is possible for a process to wait on a shared futex with a different MM because a FUT_OFF_INODE waiter can be requeued onto a FUT_OFF_MMSHARED futex by another process. This can cause FUT_OFF_MMSHARED waiters on a freed MM to consume wakeups intended for a newly allocated MM at the same address. Fix it by keying FUT_OFF_MMSHARED using a unique 64-bit per-MM ID. Leave private futexes as before to avoid influencing the performance of the hotpath. (Multi-threaded processes typically implicitly use FUT_OFF_MMSHARED by setting clear_child_tid such that it points into anonymous memory, which causes mm_release() in a multi-threaded mm to perform FUTEX_WAKE.) Fixes: 222993395ed3 ("futex: Remove pointless mmgrap() + mmdrop()") Closes: https://lore.kernel.org/r/CAG48ez0dLBpc3QtbAhMMVNHwHL94iHh2G+h-=BVFR4dDuzZr1g@mail.gmail.com/ Signed-off-by: Jann Horn Signed-off-by: Peter Zijlstra (Intel) Link: https://patch.msgid.link/20260929-futex-mmshared-fix-v1-1-262b1ffeb3d0@google.com Cc: stable@vger.kernel.org [ sashal: Adapted futex ID storage to a direct mm_struct field initialized in mm_init(). ] Signed-off-by: Sasha Levin --- include/linux/futex.h | 14 ++++++++---- include/linux/mm_types.h | 4 ++++ kernel/fork.c | 3 +++ kernel/futex/core.c | 49 +++++++++++++++++++++------------------- 4 files changed, 42 insertions(+), 28 deletions(-) diff --git a/include/linux/futex.h b/include/linux/futex.h index b70df27d7e85c..09f757f14d01f 100644 --- a/include/linux/futex.h +++ b/include/linux/futex.h @@ -23,26 +23,30 @@ struct task_struct; * 01 : Shared futex (PTHREAD_PROCESS_SHARED) * mapped on a file (reference on the underlying inode) * 10 : Shared futex (PTHREAD_PROCESS_SHARED) - * (but private mapping on an mm, and reference taken on it) + * (but private mapping on an mm) */ -#define FUT_OFF_INODE 1 /* We set bit 0 if key has a reference on inode */ -#define FUT_OFF_MMSHARED 2 /* We set bit 1 if key has a reference on mm */ +#define FUT_OFF_INODE 1 /* We set bit 0 if shared key identifies an inode */ +#define FUT_OFF_MMSHARED 2 /* We set bit 1 if shared key identifies an mm */ union futex_key { + /* For FUT_OFF_INODE */ struct { u64 i_seq; unsigned long pgoff; unsigned int offset; } shared; + + /* For FUT_OFF_MMSHARED or private */ struct { union { - struct mm_struct *mm; - u64 __tmp; + struct mm_struct *mm; /* for private */ + u64 mm_seq; /* for FUT_OFF_MMSHARED */ }; unsigned long address; unsigned int offset; } private; + struct { u64 ptr; unsigned long word; diff --git a/include/linux/mm_types.h b/include/linux/mm_types.h index 07285b44d831a..4a663b39585b0 100644 --- a/include/linux/mm_types.h +++ b/include/linux/mm_types.h @@ -580,6 +580,10 @@ struct mm_struct { #endif struct work_struct async_put_work; +#ifdef CONFIG_FUTEX + atomic64_t futex_unique_id; +#endif + #ifdef CONFIG_IOMMU_SUPPORT u32 pasid; #endif diff --git a/kernel/fork.c b/kernel/fork.c index 81355f325758c..25fb13362d38c 100644 --- a/kernel/fork.c +++ b/kernel/fork.c @@ -1082,6 +1082,9 @@ static struct mm_struct *mm_init(struct mm_struct *mm, struct task_struct *p, #endif mm_init_uprobes_state(mm); hugetlb_count_init(mm); +#ifdef CONFIG_FUTEX + atomic64_set(&mm->futex_unique_id, 0); +#endif if (current->mm) { mm->flags = current->mm->flags & MMF_INIT_MASK; diff --git a/kernel/futex/core.c b/kernel/futex/core.c index 8a3a27967973a..0055c446b0c60 100644 --- a/kernel/futex/core.c +++ b/kernel/futex/core.c @@ -473,39 +473,27 @@ futex_setup_timer(ktime_t *time, struct hrtimer_sleeper *timeout, } /* - * Generate a machine wide unique identifier for this inode. + * Generate a machine wide unique identifier for this object (inode or mm). * * This relies on u64 not wrapping in the life-time of the machine; which with * 1ns resolution means almost 585 years. - * - * This further relies on the fact that a well formed program will not unmap - * the file while it has a (shared) futex waiting on it. This mapping will have - * a file reference which pins the mount and inode. - * - * If for some reason an inode gets evicted and read back in again, it will get - * a new sequence number and will _NOT_ match, even though it is the exact same - * file. - * - * It is important that match_futex() will never have a false-positive, esp. - * for PI futexes that can mess up the state. The above argues that false-negatives - * are only possible for malformed programs. */ -static u64 get_inode_sequence_number(struct inode *inode) +static u64 get_object_id(atomic64_t *object_seq) { - static atomic64_t i_seq; + static atomic64_t last_assigned_seq; u64 old; - /* Does the inode already have a sequence number? */ - old = atomic64_read(&inode->i_sequence); + /* Does the object already have a sequence number? */ + old = atomic64_read(object_seq); if (likely(old)) return old; for (;;) { - u64 new = atomic64_add_return(1, &i_seq); + u64 new = atomic64_add_return(1, &last_assigned_seq); if (WARN_ON_ONCE(!new)) continue; - old = atomic64_cmpxchg_relaxed(&inode->i_sequence, 0, new); + old = atomic64_cmpxchg_relaxed(object_seq, 0, new); if (old) return old; return new; @@ -528,7 +516,7 @@ static u64 get_inode_sequence_number(struct inode *inode) * * ( inode->i_sequence, page->index, offset_within_page ) * - * [ also see get_inode_sequence_number() ] + * [ also see get_object_id() ] * * For private mappings (or when !@fshared), the key is: * @@ -680,8 +668,8 @@ static int get_futex_key(u32 __user *uaddr, bool fshared, union futex_key *key, goto out; } - key->both.offset |= FUT_OFF_MMSHARED; /* ref taken on mm */ - key->private.mm = mm; + key->both.offset |= FUT_OFF_MMSHARED; + key->private.mm_seq = get_object_id(&mm->futex_unique_id); key->private.address = address; } else { @@ -716,7 +704,22 @@ static int get_futex_key(u32 __user *uaddr, bool fshared, union futex_key *key, } key->both.offset |= FUT_OFF_INODE; /* inode-based key */ - key->shared.i_seq = get_inode_sequence_number(inode); + /* + * This relies on the fact that a well formed program will not + * unmap the file while it has a (shared) futex waiting on it. + * This mapping will have a file reference which pins the mount + * and inode. + * + * If for some reason an inode gets evicted and read back in + * again, it will get a new sequence number and will _NOT_ match, + * even though it is the exact same file. + * + * It is important that match_futex() will never have a + * false-positive, esp. for PI futexes that can mess up the state. + * The above argues that false-negatives are only possible for + * malformed programs. + */ + key->shared.i_seq = get_object_id(&inode->i_sequence); key->shared.pgoff = page_to_pgoff(tail); rcu_read_unlock(); } -- 2.53.0