dev.dpdk.org archive mirror
 help / color / mirror / Atom feed
* [dpdk-dev] [PATCH v1 1/3] eal/arm64: add 128-bit atomic compare exchange
@ 2019-06-23  2:41 Phil Yang
  2019-06-23  2:41 ` [dpdk-dev] [PATCH v1 2/3] test/atomic: add 128b compare and swap test Phil Yang
                   ` (7 more replies)
  0 siblings, 8 replies; 88+ messages in thread
From: Phil Yang @ 2019-06-23  2:41 UTC (permalink / raw)
  To: dev
  Cc: thomas, jerinj, hemant.agrawal, Honnappa.Nagarahalli, gavin.hu,
	nd, gage.eads

Add 128-bit atomic compare exchange on aarch64.

Signed-off-by: Phil Yang <phil.yang@arm.com>
Tested-by: Honnappa Nagarahalli <honnappa.nagarahalli@arm.com>
Reviewed-by: Honnappa Nagarahalli <honnappa.nagarahalli@arm.com>
---
This patch depends on
http://patchwork.dpdk.org/patch/54840/

v2:
eal/stack: fix 'pointer-sign' warning

 .../common/include/arch/arm/rte_atomic_64.h        | 184 +++++++++++++++++++++
 .../common/include/arch/x86/rte_atomic_64.h        |  12 --
 lib/librte_eal/common/include/generic/rte_atomic.h |  15 +-
 3 files changed, 198 insertions(+), 13 deletions(-)


diff --git a/lib/librte_eal/common/include/arch/arm/rte_atomic_64.h b/lib/librte_eal/common/include/arch/arm/rte_atomic_64.h
index 97060e4..4ef900c 100644
--- a/lib/librte_eal/common/include/arch/arm/rte_atomic_64.h
+++ b/lib/librte_eal/common/include/arch/arm/rte_atomic_64.h
@@ -1,5 +1,6 @@
 /* SPDX-License-Identifier: BSD-3-Clause
  * Copyright(c) 2015 Cavium, Inc
+ * Copyright(c) 2019 Arm Limited
  */
 
 #ifndef _RTE_ATOMIC_ARM64_H_
@@ -14,6 +15,9 @@ extern "C" {
 #endif
 
 #include "generic/rte_atomic.h"
+#include <rte_branch_prediction.h>
+#include <rte_compat.h>
+#include <rte_debug.h>
 
 #define dsb(opt) asm volatile("dsb " #opt : : : "memory")
 #define dmb(opt) asm volatile("dmb " #opt : : : "memory")
@@ -40,6 +44,186 @@ extern "C" {
 
 #define rte_cio_rmb() dmb(oshld)
 
+/*----------------------- 128 bit atomic operations -------------------------*/
+
+#define RTE_HAS_ACQ(mo) ((mo) != __ATOMIC_RELAXED && (mo) != __ATOMIC_RELEASE)
+#define RTE_HAS_RLS(mo) ((mo) == __ATOMIC_RELEASE || \
+			 (mo) == __ATOMIC_ACQ_REL || \
+			 (mo) == __ATOMIC_SEQ_CST)
+
+#define RTE_MO_LOAD(mo)  (RTE_HAS_ACQ((mo)) \
+		? __ATOMIC_ACQUIRE : __ATOMIC_RELAXED)
+#define RTE_MO_STORE(mo) (RTE_HAS_RLS((mo)) \
+		? __ATOMIC_RELEASE : __ATOMIC_RELAXED)
+
+#ifdef __ARM_FEATURE_ATOMICS
+static inline rte_int128_t
+__rte_casp(rte_int128_t *dst, rte_int128_t old, rte_int128_t updated, int mo)
+{
+
+	/* caspX instructions register pair must start from even-numbered
+	 * register at operand 1.
+	 * So, specify registers for local variables here.
+	 */
+	register uint64_t x0 __asm ("x0") = (uint64_t)old.val[0];
+	register uint64_t x1 __asm ("x1") = (uint64_t)old.val[1];
+	register uint64_t x2 __asm ("x2") = (uint64_t)updated.val[0];
+	register uint64_t x3 __asm ("x3") = (uint64_t)updated.val[1];
+
+	if (mo ==  __ATOMIC_RELAXED) {
+		asm volatile(
+				"casp %[old0], %[old1], %[upd0], %[upd1], [%[dst]]"
+				: [old0] "+r" (x0),
+				  [old1] "+r" (x1)
+				: [upd0] "r" (x2),
+				  [upd1] "r" (x3),
+				  [dst] "r" (dst)
+				: "memory");
+	} else if (mo == __ATOMIC_ACQUIRE) {
+		asm volatile(
+				"caspa %[old0], %[old1], %[upd0], %[upd1], [%[dst]]"
+				: [old0] "+r" (x0),
+				  [old1] "+r" (x1)
+				: [upd0] "r" (x2),
+				  [upd1] "r" (x3),
+				  [dst] "r" (dst)
+				: "memory");
+	} else if (mo == __ATOMIC_ACQ_REL) {
+		asm volatile(
+				"caspal %[old0], %[old1], %[upd0], %[upd1], [%[dst]]"
+				: [old0] "+r" (x0),
+				  [old1] "+r" (x1)
+				: [upd0] "r" (x2),
+				  [upd1] "r" (x3),
+				  [dst] "r" (dst)
+				: "memory");
+	} else if (mo == __ATOMIC_RELEASE) {
+		asm volatile(
+				"caspl %[old0], %[old1], %[upd0], %[upd1], [%[dst]]"
+				: [old0] "+r" (x0),
+				  [old1] "+r" (x1)
+				: [upd0] "r" (x2),
+				  [upd1] "r" (x3),
+				  [dst] "r" (dst)
+				: "memory");
+	} else {
+		rte_panic("Invalid memory order\n");
+	}
+
+	old.val[0] = x0;
+	old.val[1] = x1;
+
+	return old;
+}
+#else
+static inline rte_int128_t
+__rte_ldx128(const rte_int128_t *src, int mo)
+{
+	rte_int128_t ret;
+	if (mo == __ATOMIC_ACQUIRE)
+		asm volatile(
+				"ldaxp %0, %1, %2"
+				: "=&r" (ret.val[0]),
+				  "=&r" (ret.val[1])
+				: "Q" (src->val[0])
+				: "memory");
+	else if (mo == __ATOMIC_RELAXED)
+		asm volatile(
+				"ldxp %0, %1, %2"
+				: "=&r" (ret.val[0]),
+				  "=&r" (ret.val[1])
+				: "Q" (src->val[0])
+				: "memory");
+	else
+		rte_panic("Invalid memory order\n");
+
+	return ret;
+}
+
+static inline uint32_t
+__rte_stx128(rte_int128_t *dst, const rte_int128_t src, int mo)
+{
+	uint32_t ret;
+	if (mo == __ATOMIC_RELEASE)
+		asm volatile(
+				"stlxp %w0, %1, %2, %3"
+				: "=&r" (ret)
+				: "r" (src.val[0]),
+				  "r" (src.val[1]),
+				  "Q" (dst->val[0])
+				: "memory");
+	else if (mo == __ATOMIC_RELAXED)
+		asm volatile(
+				"stxp %w0, %1, %2, %3"
+				: "=&r" (ret)
+				: "r" (src.val[0]),
+				  "r" (src.val[1]),
+				  "Q" (dst->val[0])
+				: "memory");
+	else
+		rte_panic("Invalid memory order\n");
+
+	/* Return 0 on success, 1 on failure */
+	return ret;
+}
+#endif
+
+static inline int __rte_experimental
+rte_atomic128_cmp_exchange(rte_int128_t *dst,
+				rte_int128_t *exp,
+				const rte_int128_t *src,
+				unsigned int weak,
+				int success,
+				int failure)
+{
+	// Always do strong CAS
+	RTE_SET_USED(weak);
+	/* Ignore memory ordering for failure, memory order for
+	 * success must be stronger or equal
+	 */
+	RTE_SET_USED(failure);
+
+#ifdef __ARM_FEATURE_ATOMICS
+	rte_int128_t expected = *exp;
+	rte_int128_t desired = *src;
+	rte_int128_t old;
+
+	old = __rte_casp(dst, expected, desired, success);
+#else
+	int ldx_mo = RTE_MO_LOAD(success);
+	int stx_mo = RTE_MO_STORE(success);
+	uint32_t ret = 1;
+	register rte_int128_t expected = *exp;
+	register rte_int128_t desired = *src;
+	register rte_int128_t old;
+
+	/* ldx128 can not guarantee atomic,
+	 * Must write back src or old to verify atomicity of ldx128;
+	 */
+	do {
+		old = __rte_ldx128(dst, ldx_mo);
+		if (likely(old.int128 == expected.int128))
+			ret = __rte_stx128(dst, desired, stx_mo);
+		else
+			/* In the failure case (since 'weak' is ignored and only
+			 * weak == 0 is implemented), expected should contain the
+			 * atomically read value of dst. This means, 'old' needs
+			 * to be stored back to ensure it was read atomically.
+			 */
+			ret = __rte_stx128(dst, old, stx_mo);
+	} while (unlikely(ret));
+#endif
+
+	/* Unconditionally updating expected removes
+	 * an 'if' statement.
+	 * expected should already be in register if
+	 * not in the cache.
+	 */
+	*exp = old;
+
+	return (old.int128 == expected.int128);
+}
+
 #ifdef __cplusplus
 }
 #endif
diff --git a/lib/librte_eal/common/include/arch/x86/rte_atomic_64.h b/lib/librte_eal/common/include/arch/x86/rte_atomic_64.h
index 6232c57..23cf48f 100644
--- a/lib/librte_eal/common/include/arch/x86/rte_atomic_64.h
+++ b/lib/librte_eal/common/include/arch/x86/rte_atomic_64.h
@@ -212,18 +212,6 @@ static inline void rte_atomic64_clear(rte_atomic64_t *v)
 
 /*------------------------ 128 bit atomic operations -------------------------*/
 
-/**
- * 128-bit integer structure.
- */
-RTE_STD_C11
-typedef struct {
-	RTE_STD_C11
-	union {
-		uint64_t val[2];
-		__extension__ __int128 int128;
-	};
-} __rte_aligned(16) rte_int128_t;
-
 static inline int __rte_experimental
 rte_atomic128_cmp_exchange(rte_int128_t *dst,
 			   rte_int128_t *exp,
diff --git a/lib/librte_eal/common/include/generic/rte_atomic.h b/lib/librte_eal/common/include/generic/rte_atomic.h
index 9958543..7dd1aa4 100644
--- a/lib/librte_eal/common/include/generic/rte_atomic.h
+++ b/lib/librte_eal/common/include/generic/rte_atomic.h
@@ -1081,6 +1081,18 @@ static inline void rte_atomic64_clear(rte_atomic64_t *v)
 
 /*------------------------ 128 bit atomic operations -------------------------*/
 
+/**
+ * 128-bit integer structure.
+ */
+RTE_STD_C11
+typedef struct {
+	RTE_STD_C11
+	union {
+		uint64_t val[2];
+		__extension__ __int128 int128;
+	};
+} __rte_aligned(16) rte_int128_t;
+
 #ifdef __DOXYGEN__
 
 /**
@@ -1093,7 +1105,8 @@ static inline void rte_atomic64_clear(rte_atomic64_t *v)
  *     *exp = *dst
  * @endcode
  *
- * @note This function is currently only available for the x86-64 platform.
+ * @note This function is currently available for the x86-64 and aarch64
+ * platforms.
  *
  * @note The success and failure arguments must be one of the __ATOMIC_* values
  * defined in the C++11 standard. For details on their behavior, refer to the
-- 
2.7.4


^ permalink raw reply related	[flat|nested] 88+ messages in thread

end of thread, other threads:[~2019-10-21  8:27 UTC | newest]

Thread overview: 88+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2019-06-23  2:41 [dpdk-dev] [PATCH v1 1/3] eal/arm64: add 128-bit atomic compare exchange Phil Yang
2019-06-23  2:41 ` [dpdk-dev] [PATCH v1 2/3] test/atomic: add 128b compare and swap test Phil Yang
2019-06-23  2:41 ` [dpdk-dev] [PATCH v1 3/3] eal/stack: enable lock-free stack for aarch64 Phil Yang
2019-06-23  3:15 ` [dpdk-dev] [PATCH v2 1/3] eal/arm64: add 128-bit atomic compare exchange Phil Yang
2019-06-23  3:15   ` [dpdk-dev] [PATCH v2 2/3] test/atomic: add 128b compare and swap test Phil Yang
2019-06-24 15:09     ` Eads, Gage
2019-06-24 15:29       ` Phil Yang (Arm Technology China)
2019-06-23  3:15   ` [dpdk-dev] [PATCH v2 3/3] eal/stack: enable lock-free stack for aarch64 Phil Yang
2019-06-24 15:15     ` Eads, Gage
2019-06-24 15:22       ` Phil Yang (Arm Technology China)
2019-06-24 14:46   ` [dpdk-dev] [PATCH v2 1/3] eal/arm64: add 128-bit atomic compare exchange Eads, Gage
2019-06-24 15:35     ` Phil Yang (Arm Technology China)
2019-06-28  8:11 ` [dpdk-dev] [PATCH v3 " Phil Yang
2019-06-28  8:11   ` [dpdk-dev] [PATCH v3 2/3] test/atomic: add 128b compare and swap test Phil Yang
2019-06-29  0:17     ` Eads, Gage
2019-07-19  4:03     ` [dpdk-dev] [EXT] " Jerin Jacob Kollanukkaran
2019-06-28  8:11   ` [dpdk-dev] [PATCH v3 3/3] eal/stack: enable lock-free stack for aarch64 Phil Yang
2019-06-29  0:18     ` Eads, Gage
2019-07-19  4:18     ` [dpdk-dev] [EXT] " Jerin Jacob Kollanukkaran
2019-07-19  4:42       ` Eads, Gage
2019-07-19  5:02         ` Jerin Jacob Kollanukkaran
2019-07-19  5:15           ` Phil Yang (Arm Technology China)
2019-07-03 12:25   ` [dpdk-dev] [EXT] [PATCH v3 1/3] eal/arm64: add 128-bit atomic compare exchange Jerin Jacob Kollanukkaran
2019-07-03 13:07     ` Jerin Jacob Kollanukkaran
2019-07-05  4:20       ` Honnappa Nagarahalli
2019-07-05  4:37         ` Pavan Nikhilesh Bhagavatula
2019-07-09  9:27           ` Phil Yang (Arm Technology China)
2019-07-09 11:14             ` Jerin Jacob Kollanukkaran
2019-07-19  6:24   ` Jerin Jacob Kollanukkaran
2019-07-19 11:01     ` Phil Yang (Arm Technology China)
2019-07-19 12:35       ` Jerin Jacob Kollanukkaran
2019-07-19 13:56         ` Phil Yang (Arm Technology China)
2019-07-19 14:50           ` Eads, Gage
2019-07-22  8:44 ` [dpdk-dev] [PATCH v4 " Phil Yang
2019-07-22  8:44   ` [dpdk-dev] [PATCH v4 2/3] test/atomic: add 128b compare and swap test Phil Yang
2019-07-22  8:44   ` [dpdk-dev] [PATCH v4 3/3] eal/stack: enable lock-free stack for aarch64 Phil Yang
2019-07-22 10:22     ` [dpdk-dev] [EXT] " Jerin Jacob Kollanukkaran
2019-07-22 11:51       ` Phil Yang (Arm Technology China)
2019-07-22 10:20   ` [dpdk-dev] [EXT] [PATCH v4 1/3] eal/arm64: add 128-bit atomic compare exchange Jerin Jacob Kollanukkaran
2019-07-22 11:50     ` Phil Yang (Arm Technology China)
2019-07-22 13:06 ` [dpdk-dev] [PATCH v5 " Phil Yang
2019-07-22 13:06   ` [dpdk-dev] [PATCH v5 2/3] test/atomic: add 128b compare and swap test Phil Yang
2019-07-22 13:06   ` [dpdk-dev] [PATCH v5 3/3] eal/stack: enable lock-free stack for aarch64 Phil Yang
2019-07-22 14:14     ` [dpdk-dev] [EXT] " Jerin Jacob Kollanukkaran
2019-07-22 15:19       ` Phil Yang (Arm Technology China)
2019-07-22 14:34     ` [dpdk-dev] " Eads, Gage
2019-07-22 14:43       ` Phil Yang (Arm Technology China)
2019-07-22 14:19   ` [dpdk-dev] [EXT] [PATCH v5 1/3] eal/arm64: add 128-bit atomic compare exchange Jerin Jacob Kollanukkaran
2019-07-22 16:23     ` Phil Yang (Arm Technology China)
2019-07-22 16:22 ` [dpdk-dev] [PATCH v6 " Phil Yang
2019-07-22 16:22   ` [dpdk-dev] [PATCH v6 2/3] test/atomic: add 128b compare and swap test Phil Yang
2019-07-22 16:22   ` [dpdk-dev] [PATCH v6 3/3] eal/stack: enable lock-free stack for aarch64 Phil Yang
2019-07-22 16:59     ` [dpdk-dev] [EXT] " Jerin Jacob Kollanukkaran
2019-07-22 16:57   ` [dpdk-dev] [EXT] [PATCH v6 1/3] eal/arm64: add 128-bit atomic compare exchange Jerin Jacob Kollanukkaran
2019-07-23  3:28     ` Phil Yang (Arm Technology China)
2019-07-23  7:09       ` Jerin Jacob Kollanukkaran
2019-07-23  7:53         ` Phil Yang (Arm Technology China)
2019-07-23  5:57 ` [dpdk-dev] [PATCH v7 " Phil Yang
2019-07-23  5:57   ` [dpdk-dev] [PATCH v7 2/3] test/atomic: add 128b compare and swap test Phil Yang
2019-07-23  5:57   ` [dpdk-dev] [PATCH v7 3/3] eal/stack: enable lock-free stack for aarch64 Phil Yang
2019-07-23  7:05   ` [dpdk-dev] [PATCH v8 1/3] eal/arm64: add 128-bit atomic compare exchange jerinj
2019-07-23  7:05     ` [dpdk-dev] [PATCH v8 2/3] test/atomic: add 128b compare and swap test jerinj
2019-07-23  7:05     ` [dpdk-dev] [PATCH v8 3/3] eal/stack: enable lock-free stack for aarch64 jerinj
2019-08-14  8:27     ` [dpdk-dev] [PATCH v9 1/3] eal/arm64: add 128-bit atomic compare exchange Phil Yang
2019-08-14  8:27       ` [dpdk-dev] [PATCH v9 2/3] test/atomic: add 128b compare and swap test Phil Yang
2019-10-14 15:45         ` David Marchand
2019-10-15 11:32           ` Phil Yang (Arm Technology China)
2019-08-14  8:27       ` [dpdk-dev] [PATCH v9 3/3] eal/stack: enable lock-free stack for aarch64 Phil Yang
2019-10-14 15:45         ` David Marchand
2019-10-15 11:32           ` Phil Yang (Arm Technology China)
2019-10-14 15:43       ` [dpdk-dev] [PATCH v9 1/3] eal/arm64: add 128-bit atomic compare exchange David Marchand
2019-10-15 11:32         ` Phil Yang (Arm Technology China)
2019-10-15 12:16           ` David Marchand
2019-10-16  9:04             ` Phil Yang (Arm Technology China)
2019-10-17 12:45               ` David Marchand
2019-10-15 11:38       ` [dpdk-dev] [PATCH v10 " Phil Yang
2019-10-15 11:38         ` [dpdk-dev] [PATCH v10 2/3] test/atomic: add 128b compare and swap test Phil Yang
2019-10-15 11:38         ` [dpdk-dev] [PATCH v10 3/3] eal/stack: enable lock-free stack for aarch64 Phil Yang
2019-10-18 11:21         ` [dpdk-dev] [PATCH v11 1/3] eal/arm64: add 128-bit atomic compare exchange Phil Yang
2019-10-18 11:21           ` [dpdk-dev] [PATCH v11 2/3] test/atomic: add 128b compare and swap test Phil Yang
2019-10-21  8:25             ` David Marchand
2019-10-18 11:21           ` [dpdk-dev] [PATCH v11 3/3] eal/stack: enable lock-free stack for aarch64 Phil Yang
2019-10-21  8:26             ` David Marchand
2019-10-18 14:16           ` [dpdk-dev] [PATCH v11 1/3] eal/arm64: add 128-bit atomic compare exchange David Marchand
2019-10-18 14:24             ` Jerin Jacob
2019-10-18 14:33               ` David Marchand
2019-10-18 14:36                 ` Jerin Jacob
2019-10-21  8:24                   ` David Marchand

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).