From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: from mails.dpdk.org (mails.dpdk.org [217.70.189.124]) by inbox.dpdk.org (Postfix) with ESMTP id 2AF8C48AF1 for ; Wed, 12 Nov 2025 20:57:39 +0100 (CET) Received: from mails.dpdk.org (localhost [127.0.0.1]) by mails.dpdk.org (Postfix) with ESMTP id 2311B40151; Wed, 12 Nov 2025 20:57:39 +0100 (CET) Received: from foss.arm.com (foss.arm.com [217.140.110.172]) by mails.dpdk.org (Postfix) with ESMTP id 856A140151 for ; Wed, 12 Nov 2025 20:57:37 +0100 (CET) Received: from usa-sjc-imap-foss1.foss.arm.com (unknown [10.121.207.14]) by usa-sjc-mx-foss1.foss.arm.com (Postfix) with ESMTP id DC35C1515; Wed, 12 Nov 2025 11:57:28 -0800 (PST) Received: from ampere-altra-2-1.usa.arm.com (ampere-altra-2-1.usa.arm.com [10.118.91.158]) by usa-sjc-imap-foss1.foss.arm.com (Postfix) with ESMTPA id 5786B3F63F; Wed, 12 Nov 2025 11:57:36 -0800 (PST) From: Wathsala Vithanage To: Honnappa Nagarahalli , Konstantin Ananyev Cc: stable@dpdk.org, Wathsala Vithanage , ola.liljedahl@arm.com, dhruv.tripathi@arm.com, konstantin.ananyev@huawei.com Subject: [PATCH 22.11 v2 1/1] ring: establish safe partial order in default mode Date: Wed, 12 Nov 2025 19:53:56 +0000 Message-ID: <20251112195715.1169873-1-wathsala.vithanage@arm.com> X-Mailer: git-send-email 2.43.0 MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit X-BeenThere: stable@dpdk.org X-Mailman-Version: 2.1.29 Precedence: list List-Id: patches for DPDK stable branches List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: stable-bounces@dpdk.org [ upstream commit a4ad0eba9def1d1d071da8afe5e96eb2a2e0d71f] The function __rte_ring_headtail_move_head() assumes that the barrier (fence) between the load of the head and the load-acquire of the opposing tail guarantees the following: if a first thread reads tail and then writes head and a second thread reads the new value of head and then reads tail, then it should observe the same (or a later) value of tail. This assumption is incorrect under the C11 memory model. If the barrier (fence) is intended to establish a total ordering of ring operations, it fails to do so. Instead, the current implementation only enforces a partial ordering, which can lead to unsafe interleavings. In particular, some partial orders can cause underflows in free slot or available element computations, potentially resulting in data corruption. The issue manifests when a CPU first acts as a producer and later as a consumer. In this scenario, the barrier assumption may fail when another core takes the consumer role. A Herd7 litmus test in C11 can demonstrate this violation. The problem has not been widely observed so far because: (a) on strong memory models (e.g., x86-64) the assumption holds, and (b) on relaxed models with RCsc semantics the ordering is still strong enough to prevent hazards. The problem becomes visible only on weaker models, when load-acquire is implemented with RCpc semantics (e.g. some AArch64 CPUs which support the LDAPR and LDAPUR instructions). Three possible solutions exist: 1. Strengthen ordering by upgrading release/acquire semantics to sequential consistency. This requires using seq-cst for stores, loads, and CAS operations. However, this approach introduces a significant performance penalty on relaxed-memory architectures. 2. Establish a safe partial order by enforcing a pair-wise happens-before relationship between thread of same role by changing the CAS and the preceding load of the head by converting them to release and acquire respectively. This approach makes the original barrier assumption unnecessary and allows its removal. 3. Retain partial ordering but ensure only safe partial orders are committed. This can be done by detecting underflow conditions (producer < consumer) and quashing the update in such cases. This approach makes the original barrier assumption unnecessary and allows its removal. This patch implements solution (2) to preserve the “enqueue always succeeds” contract expected by dependent libraries (e.g., mempool). While solution (3) offers higher performance, adopting it now would break that assumption. Signed-off-by: Wathsala Vithanage Signed-off-by: Ola Liljedahl Reviewed-by: Honnappa Nagarahalli Reviewed-by: Dhruv Tripathi Acked-by: Konstantin Ananyev Tested-by: Konstantin Ananyev --- lib/ring/rte_ring_c11_pvt.h | 58 +++++++++++++++++++++++-------------- 1 file changed, 36 insertions(+), 22 deletions(-) diff --git a/lib/ring/rte_ring_c11_pvt.h b/lib/ring/rte_ring_c11_pvt.h index 5c04a001e1..82f0ce4946 100644 --- a/lib/ring/rte_ring_c11_pvt.h +++ b/lib/ring/rte_ring_c11_pvt.h @@ -25,7 +25,8 @@ __rte_ring_update_tail(struct rte_ring_headtail *ht, uint32_t old_val, rte_wait_until_equal_32(&ht->tail, old_val, __ATOMIC_RELAXED); /* - * R0: Establishes a synchronizing edge with load-acquire of tail at A1. + * R0: Establishes a synchronizing edge with load-acquire of + * cons_tail at A1 or prod_tail at A4. * Ensures that memory effects by this thread on ring elements array * is observed by a different thread of the other type. */ @@ -69,8 +70,8 @@ __rte_ring_move_prod_head(struct rte_ring *r, unsigned int is_sp, /* * A0: Establishes a synchronizing edge with R1. * Ensure that this thread observes same values - * to stail observed by the thread that updated - * d->head. + * to cons_tail observed by the thread that + * updated r->prod.head. * If not, an unsafe partial order may ensue. */ *old_head = __atomic_load_n(&r->prod.head, __ATOMIC_ACQUIRE); @@ -84,8 +85,7 @@ __rte_ring_move_prod_head(struct rte_ring *r, unsigned int is_sp, * ring elements array is observed by the time * this thread observes its tail update. */ - cons_tail = __atomic_load_n(&r->cons.tail, - __ATOMIC_ACQUIRE); + cons_tail = __atomic_load_n(&r->cons.tail, __ATOMIC_ACQUIRE); /* The subtraction is done between two unsigned 32bits value * (the result is always modulo 32 bits even if we have @@ -107,10 +107,19 @@ __rte_ring_move_prod_head(struct rte_ring *r, unsigned int is_sp, r->prod.head = *new_head, success = 1; else /* on failure, *old_head is updated */ + /* + * R1/A2. + * R1: Establishes a synchronizing edge with A0 of a + * different thread. + * A2: Establishes a synchronizing edge with R1 of a + * different thread to observe same value for + * cons_tail observed by that thread on CAS failure + * (to retry with an updated *old_head). + */ success = __atomic_compare_exchange_n(&r->prod.head, old_head, *new_head, - 0, __ATOMIC_RELAXED, - __ATOMIC_RELAXED); + 0, __ATOMIC_RELEASE, + __ATOMIC_ACQUIRE); } while (unlikely(success == 0)); return n; } @@ -148,20 +157,25 @@ __rte_ring_move_cons_head(struct rte_ring *r, int is_sc, uint32_t prod_tail; int success; - /* move cons.head atomically */ - *old_head = __atomic_load_n(&r->cons.head, __ATOMIC_RELAXED); + /* + * A3: Establishes a synchronizing edge with R2. + * Ensure that this thread observes same values + * to prod_tail observed by the thread that + * updated r->cons.head. + * If not, an unsafe partial order may ensue. + */ + *old_head = __atomic_load_n(&r->cons.head, __ATOMIC_ACQUIRE); do { /* Restore n as it may change every loop */ n = max; - /* Ensure the head is read before tail */ - __atomic_thread_fence(__ATOMIC_ACQUIRE); - - /* this load-acquire synchronize with store-release of ht->tail - * in update_tail. + /* + * A4: Establishes a synchronizing edge with R0. + * Ensures that other thread's memory effects on + * ring elements array is observed by the time + * this thread observes its tail update. */ - prod_tail = __atomic_load_n(&r->prod.tail, - __ATOMIC_ACQUIRE); + prod_tail = __atomic_load_n(&r->prod.tail, __ATOMIC_ACQUIRE); /* The subtraction is done between two unsigned 32bits value * (the result is always modulo 32 bits even if we have @@ -183,13 +197,13 @@ __rte_ring_move_cons_head(struct rte_ring *r, int is_sc, else /* on failure, *old_head will be updated */ /* - * R1/A2. - * R1: Establishes a synchronizing edge with A0 of a + * R2/A5. + * R2: Establishes a synchronizing edge with A3 of a * different thread. - * A2: Establishes a synchronizing edge with R1 of a - * different thread to observe same value for stail - * observed by that thread on CAS failure (to retry - * with an updated *old_head). + * A5: Establishes a synchronizing edge with R2 of a + * different thread to observe same value for + * prod_tail observed by that thread on CAS failure + * (to retry with an updated *old_head). */ success = __atomic_compare_exchange_n(&r->cons.head, old_head, *new_head, -- 2.43.0