[PATCH v7 10/39] eventdev: use C11 alignas

[Tyler Retzlaff <roretzla@linux.microsoft.com>]

The current location used for __rte_aligned(a) for alignment of types
and variables is not compatible with MSVC. There is only a single
location accepted by both toolchains.

For variables standard C11 offers alignas(a) supported by conformant
compilers i.e. both MSVC and GCC.

For types the standard offers no alignment facility that compatibly
interoperates with C and C++ but may be achieved by relocating the
placement of __rte_aligned(a) to the aforementioned location accepted
by all currently supported toolchains.

To allow alignment for both compilers do the following:

* Move __rte_aligned from the end of {struct,union} definitions to
  be between {struct,union} and tag.

  The placement between {struct,union} and the tag allows the desired
  alignment to be imparted on the type regardless of the toolchain being
  used for all of GCC, LLVM, MSVC compilers building both C and C++.

* Replace use of __rte_aligned(a) on variables/fields with alignas(a).

Signed-off-by: Tyler Retzlaff <roretzla@linux.microsoft.com>
Acked-by: Morten Brørup <mb@smartsharesystems.com>
---
 lib/eventdev/event_timer_adapter_pmd.h  |  4 ++--
 lib/eventdev/eventdev_pmd.h             |  8 ++++----
 lib/eventdev/rte_event_crypto_adapter.c | 16 ++++++++--------
 lib/eventdev/rte_event_dma_adapter.c    | 16 ++++++++--------
 lib/eventdev/rte_event_eth_rx_adapter.c |  8 ++++----
 lib/eventdev/rte_event_eth_tx_adapter.c |  4 ++--
 lib/eventdev/rte_event_timer_adapter.c  |  9 +++++----
 lib/eventdev/rte_event_timer_adapter.h  |  8 ++++----
 lib/eventdev/rte_eventdev.h             | 10 +++-------
 lib/eventdev/rte_eventdev_core.h        |  4 ++--
 10 files changed, 42 insertions(+), 45 deletions(-)

diff --git a/lib/eventdev/event_timer_adapter_pmd.h b/lib/eventdev/event_timer_adapter_pmd.h
index 65b421b..cd5127f 100644
--- a/lib/eventdev/event_timer_adapter_pmd.h
+++ b/lib/eventdev/event_timer_adapter_pmd.h
@@ -86,7 +86,7 @@ struct event_timer_adapter_ops {
  * @internal Adapter data; structure to be placed in shared memory to be
  * accessible by various processes in a multi-process configuration.
  */
-struct rte_event_timer_adapter_data {
+struct __rte_cache_aligned rte_event_timer_adapter_data {
 	uint8_t id;
 	/**< Event timer adapter ID */
 	uint8_t event_dev_id;
@@ -110,7 +110,7 @@ struct rte_event_timer_adapter_data {
 
 	uint8_t started : 1;
 	/**< Flag to indicate adapter started. */
-} __rte_cache_aligned;
+};
 
 #ifdef __cplusplus
 }
diff --git a/lib/eventdev/eventdev_pmd.h b/lib/eventdev/eventdev_pmd.h
index 2ad8802..7a5699f 100644
--- a/lib/eventdev/eventdev_pmd.h
+++ b/lib/eventdev/eventdev_pmd.h
@@ -105,7 +105,7 @@ struct rte_eventdev_global {
  * This structure is safe to place in shared memory to be common among
  * different processes in a multi-process configuration.
  */
-struct rte_eventdev_data {
+struct __rte_cache_aligned rte_eventdev_data {
 	int socket_id;
 	/**< Socket ID where memory is allocated */
 	uint8_t dev_id;
@@ -144,10 +144,10 @@ struct rte_eventdev_data {
 
 	uint64_t reserved_64s[4]; /**< Reserved for future fields */
 	void *reserved_ptrs[4];	  /**< Reserved for future fields */
-} __rte_cache_aligned;
+};
 
 /** @internal The data structure associated with each event device. */
-struct rte_eventdev {
+struct __rte_cache_aligned rte_eventdev {
 	struct rte_eventdev_data *data;
 	/**< Pointer to device data */
 	struct eventdev_ops *dev_ops;
@@ -187,7 +187,7 @@ struct rte_eventdev {
 
 	uint64_t reserved_64s[3]; /**< Reserved for future fields */
 	void *reserved_ptrs[3];	  /**< Reserved for future fields */
-} __rte_cache_aligned;
+};
 
 extern struct rte_eventdev *rte_eventdevs;
 /** @internal The pool of rte_eventdev structures. */
diff --git a/lib/eventdev/rte_event_crypto_adapter.c b/lib/eventdev/rte_event_crypto_adapter.c
index 9903f96..db1c7f3 100644
--- a/lib/eventdev/rte_event_crypto_adapter.c
+++ b/lib/eventdev/rte_event_crypto_adapter.c
@@ -42,7 +42,7 @@
 
 #define ECA_ADAPTER_ARRAY "crypto_adapter_array"
 
-struct crypto_ops_circular_buffer {
+struct __rte_cache_aligned crypto_ops_circular_buffer {
 	/* index of head element in circular buffer */
 	uint16_t head;
 	/* index of tail element in circular buffer */
@@ -53,9 +53,9 @@ struct crypto_ops_circular_buffer {
 	uint16_t size;
 	/* Pointer to hold rte_crypto_ops for batching */
 	struct rte_crypto_op **op_buffer;
-} __rte_cache_aligned;
+};
 
-struct event_crypto_adapter {
+struct __rte_cache_aligned event_crypto_adapter {
 	/* Event device identifier */
 	uint8_t eventdev_id;
 	/* Event port identifier */
@@ -98,10 +98,10 @@ struct event_crypto_adapter {
 	uint16_t nb_qps;
 	/* Adapter mode */
 	enum rte_event_crypto_adapter_mode mode;
-} __rte_cache_aligned;
+};
 
 /* Per crypto device information */
-struct crypto_device_info {
+struct __rte_cache_aligned crypto_device_info {
 	/* Pointer to cryptodev */
 	struct rte_cryptodev *dev;
 	/* Pointer to queue pair info */
@@ -118,15 +118,15 @@ struct crypto_device_info {
 	 * be invoked if not already invoked
 	 */
 	uint16_t num_qpairs;
-} __rte_cache_aligned;
+};
 
 /* Per queue pair information */
-struct crypto_queue_pair_info {
+struct __rte_cache_aligned crypto_queue_pair_info {
 	/* Set to indicate queue pair is enabled */
 	bool qp_enabled;
 	/* Circular buffer for batching crypto ops to cdev */
 	struct crypto_ops_circular_buffer cbuf;
-} __rte_cache_aligned;
+};
 
 static struct event_crypto_adapter **event_crypto_adapter;
 
diff --git a/lib/eventdev/rte_event_dma_adapter.c b/lib/eventdev/rte_event_dma_adapter.c
index 4196164..24dff55 100644
--- a/lib/eventdev/rte_event_dma_adapter.c
+++ b/lib/eventdev/rte_event_dma_adapter.c
@@ -26,7 +26,7 @@
 	} while (0)
 
 /* DMA ops circular buffer */
-struct dma_ops_circular_buffer {
+struct __rte_cache_aligned dma_ops_circular_buffer {
 	/* Index of head element */
 	uint16_t head;
 
@@ -41,19 +41,19 @@ struct dma_ops_circular_buffer {
 
 	/* Pointer to hold rte_event_dma_adapter_op for processing */
 	struct rte_event_dma_adapter_op **op_buffer;
-} __rte_cache_aligned;
+};
 
 /* Vchan information */
-struct dma_vchan_info {
+struct __rte_cache_aligned dma_vchan_info {
 	/* Set to indicate vchan queue is enabled */
 	bool vq_enabled;
 
 	/* Circular buffer for batching DMA ops to dma_dev */
 	struct dma_ops_circular_buffer dma_buf;
-} __rte_cache_aligned;
+};
 
 /* DMA device information */
-struct dma_device_info {
+struct __rte_cache_aligned dma_device_info {
 	/* Pointer to vchan queue info */
 	struct dma_vchan_info *vchanq;
 
@@ -81,9 +81,9 @@ struct dma_device_info {
 	 * transfer uses a hardware mechanism
 	 */
 	uint8_t internal_event_port;
-} __rte_cache_aligned;
+};
 
-struct event_dma_adapter {
+struct __rte_cache_aligned event_dma_adapter {
 	/* Event device identifier */
 	uint8_t eventdev_id;
 
@@ -145,7 +145,7 @@ struct event_dma_adapter {
 
 	/* Per instance stats structure */
 	struct rte_event_dma_adapter_stats dma_stats;
-} __rte_cache_aligned;
+};
 
 static struct event_dma_adapter **event_dma_adapter;
 
diff --git a/lib/eventdev/rte_event_eth_rx_adapter.c b/lib/eventdev/rte_event_eth_rx_adapter.c
index 1b83a55..3ee20d9 100644
--- a/lib/eventdev/rte_event_eth_rx_adapter.c
+++ b/lib/eventdev/rte_event_eth_rx_adapter.c
@@ -72,7 +72,7 @@ struct eth_rx_poll_entry {
 	uint16_t eth_rx_qid;
 };
 
-struct eth_rx_vector_data {
+struct __rte_cache_aligned eth_rx_vector_data {
 	TAILQ_ENTRY(eth_rx_vector_data) next;
 	uint16_t port;
 	uint16_t queue;
@@ -82,7 +82,7 @@ struct eth_rx_vector_data {
 	uint64_t vector_timeout_ticks;
 	struct rte_mempool *vector_pool;
 	struct rte_event_vector *vector_ev;
-} __rte_cache_aligned;
+};
 
 TAILQ_HEAD(eth_rx_vector_data_list, eth_rx_vector_data);
 
@@ -103,7 +103,7 @@ struct eth_event_enqueue_buffer {
 	uint16_t last_mask;
 };
 
-struct event_eth_rx_adapter {
+struct __rte_cache_aligned event_eth_rx_adapter {
 	/* RSS key */
 	uint8_t rss_key_be[RSS_KEY_SIZE];
 	/* Event device identifier */
@@ -188,7 +188,7 @@ struct event_eth_rx_adapter {
 	uint8_t rxa_started;
 	/* Adapter ID */
 	uint8_t id;
-} __rte_cache_aligned;
+};
 
 /* Per eth device */
 struct eth_device_info {
diff --git a/lib/eventdev/rte_event_eth_tx_adapter.c b/lib/eventdev/rte_event_eth_tx_adapter.c
index 56435be..67fff8b 100644
--- a/lib/eventdev/rte_event_eth_tx_adapter.c
+++ b/lib/eventdev/rte_event_eth_tx_adapter.c
@@ -109,7 +109,7 @@ struct txa_service_queue_info {
 };
 
 /* PMD private structure */
-struct txa_service_data {
+struct __rte_cache_aligned txa_service_data {
 	/* Max mbufs processed in any service function invocation */
 	uint32_t max_nb_tx;
 	/* Number of Tx queues in adapter */
@@ -144,7 +144,7 @@ struct txa_service_data {
 	int64_t service_id;
 	/* Memory allocation name */
 	char mem_name[TXA_MEM_NAME_LEN];
-} __rte_cache_aligned;
+};
 
 /* Per eth device structure */
 struct txa_service_ethdev {
diff --git a/lib/eventdev/rte_event_timer_adapter.c b/lib/eventdev/rte_event_timer_adapter.c
index a3c6fb7..fdeeece 100644
--- a/lib/eventdev/rte_event_timer_adapter.c
+++ b/lib/eventdev/rte_event_timer_adapter.c
@@ -4,6 +4,7 @@
  */
 
 #include <ctype.h>
+#include <stdalign.h>
 #include <string.h>
 #include <inttypes.h>
 #include <stdbool.h>
@@ -511,11 +512,11 @@ struct rte_event_timer_adapter *
 
 #define EXP_TIM_BUF_SZ 128
 
-struct event_buffer {
+struct __rte_cache_aligned event_buffer {
 	size_t head;
 	size_t tail;
 	struct rte_event events[EVENT_BUFFER_SZ];
-} __rte_cache_aligned;
+};
 
 static inline bool
 event_buffer_full(struct event_buffer *bufp)
@@ -631,9 +632,9 @@ struct swtim {
 	/* Identifier of timer data instance */
 	uint32_t timer_data_id;
 	/* Track which cores have actually armed a timer */
-	struct {
+	alignas(RTE_CACHE_LINE_SIZE) struct {
 		RTE_ATOMIC(uint16_t) v;
-	} __rte_cache_aligned in_use[RTE_MAX_LCORE];
+	} in_use[RTE_MAX_LCORE];
 	/* Track which cores' timer lists should be polled */
 	RTE_ATOMIC(unsigned int) poll_lcores[RTE_MAX_LCORE];
 	/* The number of lists that should be polled */
diff --git a/lib/eventdev/rte_event_timer_adapter.h b/lib/eventdev/rte_event_timer_adapter.h
index c133dec..0bd1b30 100644
--- a/lib/eventdev/rte_event_timer_adapter.h
+++ b/lib/eventdev/rte_event_timer_adapter.h
@@ -473,7 +473,7 @@ enum rte_event_timer_state {
  * The generic *rte_event_timer* structure to hold the event timer attributes
  * for arm and cancel operations.
  */
-struct rte_event_timer {
+struct __rte_cache_aligned rte_event_timer {
 	struct rte_event ev;
 	/**<
 	 * Expiry event attributes.  On successful event timer timeout,
@@ -504,7 +504,7 @@ struct rte_event_timer {
 	/**< Memory to store user specific metadata.
 	 * The event timer adapter implementation should not modify this area.
 	 */
-} __rte_cache_aligned;
+};
 
 typedef uint16_t (*rte_event_timer_arm_burst_t)(
 		const struct rte_event_timer_adapter *adapter,
@@ -526,7 +526,7 @@ typedef uint16_t (*rte_event_timer_cancel_burst_t)(
 /**
  * @internal Data structure associated with each event timer adapter.
  */
-struct rte_event_timer_adapter {
+struct __rte_cache_aligned rte_event_timer_adapter {
 	rte_event_timer_arm_burst_t arm_burst;
 	/**< Pointer to driver arm_burst function. */
 	rte_event_timer_arm_tmo_tick_burst_t arm_tmo_tick_burst;
@@ -540,7 +540,7 @@ struct rte_event_timer_adapter {
 
 	uint8_t allocated : 1;
 	/**< Flag to indicate that this adapter has been allocated */
-} __rte_cache_aligned;
+};
 
 #define ADAPTER_VALID_OR_ERR_RET(adapter, retval) do {		\
 	if (adapter == NULL || !adapter->allocated)		\
diff --git a/lib/eventdev/rte_eventdev.h b/lib/eventdev/rte_eventdev.h
index 3af4686..08e5f93 100644
--- a/lib/eventdev/rte_eventdev.h
+++ b/lib/eventdev/rte_eventdev.h
@@ -1338,7 +1338,7 @@ int rte_event_dev_stop_flush_callback_register(uint8_t dev_id,
 /**
  * Event vector structure.
  */
-struct rte_event_vector {
+struct __rte_aligned(16) rte_event_vector {
 	uint16_t nb_elem;
 	/**< Number of elements valid in this event vector. */
 	uint16_t elem_offset : 12;
@@ -1376,23 +1376,19 @@ struct rte_event_vector {
 	 * value to share between dequeue and enqueue operation.
 	 * The application should not modify this field.
 	 */
-	union {
+	union __rte_aligned(16) {
 #endif
 		struct rte_mbuf *mbufs[0];
 		void *ptrs[0];
 		uint64_t u64s[0];
 #ifndef __cplusplus
-	} __rte_aligned(16);
+	};
 #endif
 	/**< Start of the vector array union. Depending upon the event type the
 	 * vector array can be an array of mbufs or pointers or opaque u64
 	 * values.
 	 */
-#ifndef __DOXYGEN__
-} __rte_aligned(16);
-#else
 };
-#endif
 
 /* Scheduler type definitions */
 #define RTE_SCHED_TYPE_ORDERED          0
diff --git a/lib/eventdev/rte_eventdev_core.h b/lib/eventdev/rte_eventdev_core.h
index 5b40551..fc8e155 100644
--- a/lib/eventdev/rte_eventdev_core.h
+++ b/lib/eventdev/rte_eventdev_core.h
@@ -49,7 +49,7 @@ typedef uint16_t (*event_dma_adapter_enqueue_t)(void *port, struct rte_event ev[
 typedef int (*event_profile_switch_t)(void *port, uint8_t profile);
 /**< @internal Switch active link profile on the event port. */
 
-struct rte_event_fp_ops {
+struct __rte_cache_aligned rte_event_fp_ops {
 	void **data;
 	/**< points to array of internal port data pointers */
 	event_enqueue_t enqueue;
@@ -77,7 +77,7 @@ struct rte_event_fp_ops {
 	event_profile_switch_t profile_switch;
 	/**< PMD Event switch profile function. */
 	uintptr_t reserved[4];
-} __rte_cache_aligned;
+};
 
 extern struct rte_event_fp_ops rte_event_fp_ops[RTE_EVENT_MAX_DEVS];
 
-- 
1.8.3.1