On 2018/12/15 1:52 AM, Thomas Schwinge wrote:
As for the following changes, will you please make sure that there is one
common order for these, used in "libgomp/libgomp-plugin.h" function
prototypes, "libgomp/libgomp.h:acc_dispatch_t",
"libgomp/target.c:gomp_load_plugin_for_device", "libgomp/oacc-host.c"
function definitions as well as in "host_dispatch", and the
libgomp-plugin(s) themselves (that's all, I think?).
--- a/libgomp/libgomp-plugin.h
+++ b/libgomp/libgomp-plugin.h
@@ -93,22 +107,31 @@ extern bool GOMP_OFFLOAD_dev2dev (int, void *, const void
*, size_t);
extern bool GOMP_OFFLOAD_can_run (void *);
extern void GOMP_OFFLOAD_run (int, void *, void *, void **);
extern void GOMP_OFFLOAD_async_run (int, void *, void *, void **, void *);
+
extern void GOMP_OFFLOAD_openacc_exec (void (*) (void *), size_t, void **,
- void **, int, unsigned *, void *);
-extern void GOMP_OFFLOAD_openacc_register_async_cleanup (void *, int);
-extern int GOMP_OFFLOAD_openacc_async_test (int);
-extern int GOMP_OFFLOAD_openacc_async_test_all (void);
-extern void GOMP_OFFLOAD_openacc_async_wait (int);
-extern void GOMP_OFFLOAD_openacc_async_wait_async (int, int);
-extern void GOMP_OFFLOAD_openacc_async_wait_all (void);
-extern void GOMP_OFFLOAD_openacc_async_wait_all_async (int);
-extern void GOMP_OFFLOAD_openacc_async_set_async (int);
+ void **, unsigned *, void *);
+extern void GOMP_OFFLOAD_openacc_async_exec (void (*) (void *), size_t, void
**,
+ void **, unsigned *, void *,
+ struct goacc_asyncqueue *);
+extern struct goacc_asyncqueue *GOMP_OFFLOAD_openacc_async_construct (void);
+extern bool GOMP_OFFLOAD_openacc_async_destruct (struct goacc_asyncqueue *);
+extern int GOMP_OFFLOAD_openacc_async_test (struct goacc_asyncqueue *);
+extern void GOMP_OFFLOAD_openacc_async_synchronize (struct goacc_asyncqueue *);
+extern void GOMP_OFFLOAD_openacc_async_serialize (struct goacc_asyncqueue *,
+ struct goacc_asyncqueue *);
+extern void GOMP_OFFLOAD_openacc_async_queue_callback (struct goacc_asyncqueue
*,
+ void (*)(void *), void
*);
+extern bool GOMP_OFFLOAD_openacc_async_host2dev (int, void *, const void *,
size_t,
+ struct goacc_asyncqueue *);
+extern bool GOMP_OFFLOAD_openacc_async_dev2host (int, void *, const void *,
size_t,
+ struct goacc_asyncqueue *);
This patch revises the ordering of the above functions/hooks to be consistent
across libgomp, and un-deletes goacc_async_unmap_vars in libgomp.map.
Chung-Lin
Index: libgomp/libgomp.h
===================================================================
--- libgomp/libgomp.h (revision 267226)
+++ libgomp/libgomp.h (working copy)
@@ -949,25 +949,29 @@ typedef struct acc_dispatch_t
/* Execute. */
__typeof (GOMP_OFFLOAD_openacc_exec) *exec_func;
- /* Async cleanup callback registration. */
- __typeof (GOMP_OFFLOAD_openacc_register_async_cleanup)
- *register_async_cleanup_func;
-
- /* Asynchronous routines. */
- __typeof (GOMP_OFFLOAD_openacc_async_test) *async_test_func;
- __typeof (GOMP_OFFLOAD_openacc_async_test_all) *async_test_all_func;
- __typeof (GOMP_OFFLOAD_openacc_async_wait) *async_wait_func;
- __typeof (GOMP_OFFLOAD_openacc_async_wait_async) *async_wait_async_func;
- __typeof (GOMP_OFFLOAD_openacc_async_wait_all) *async_wait_all_func;
- __typeof (GOMP_OFFLOAD_openacc_async_wait_all_async)
- *async_wait_all_async_func;
- __typeof (GOMP_OFFLOAD_openacc_async_set_async) *async_set_async_func;
-
/* Create/destroy TLS data. */
__typeof (GOMP_OFFLOAD_openacc_create_thread_data) *create_thread_data_func;
__typeof (GOMP_OFFLOAD_openacc_destroy_thread_data)
*destroy_thread_data_func;
+
+ struct {
+ gomp_mutex_t lock;
+ int nasyncqueue;
+ struct goacc_asyncqueue **asyncqueue;
+ struct goacc_asyncqueue_list *active;
+ __typeof (GOMP_OFFLOAD_openacc_async_construct) *construct_func;
+ __typeof (GOMP_OFFLOAD_openacc_async_destruct) *destruct_func;
+ __typeof (GOMP_OFFLOAD_openacc_async_test) *test_func;
+ __typeof (GOMP_OFFLOAD_openacc_async_synchronize) *synchronize_func;
+ __typeof (GOMP_OFFLOAD_openacc_async_serialize) *serialize_func;
+ __typeof (GOMP_OFFLOAD_openacc_async_queue_callback) *queue_callback_func;
+
+ __typeof (GOMP_OFFLOAD_openacc_async_exec) *exec_func;
+ __typeof (GOMP_OFFLOAD_openacc_async_dev2host) *dev2host_func;
+ __typeof (GOMP_OFFLOAD_openacc_async_host2dev) *host2dev_func;
+ } async;
+
/* NVIDIA target specific routines. */
struct {
__typeof (GOMP_OFFLOAD_openacc_cuda_get_current_device)
@@ -1053,17 +1057,33 @@ enum gomp_map_vars_kind
GOMP_MAP_VARS_ENTER_DATA
};
-extern void gomp_acc_insert_pointer (size_t, void **, size_t *, void *);
+extern void gomp_acc_insert_pointer (size_t, void **, size_t *, void *, int);
extern void gomp_acc_remove_pointer (void *, size_t, bool, int, int, int);
extern void gomp_acc_declare_allocate (bool, size_t, void **, size_t *,
unsigned short *);
+struct gomp_coalesce_buf;
+extern void gomp_copy_host2dev (struct gomp_device_descr *,
+ struct goacc_asyncqueue *, void *, const void *,
+ size_t, struct gomp_coalesce_buf *);
+extern void gomp_copy_dev2host (struct gomp_device_descr *,
+ struct goacc_asyncqueue *, void *, const void *,
+ size_t);
extern struct target_mem_desc *gomp_map_vars (struct gomp_device_descr *,
size_t, void **, void **,
size_t *, void *, bool,
enum gomp_map_vars_kind);
+extern struct target_mem_desc *gomp_map_vars_async (struct gomp_device_descr *,
+ struct goacc_asyncqueue *,
+ size_t, void **, void **,
+ size_t *, void *, bool,
+ enum gomp_map_vars_kind);
+extern void gomp_unmap_tgt (struct target_mem_desc *);
extern void gomp_unmap_vars (struct target_mem_desc *, bool);
+extern void gomp_unmap_vars_async (struct target_mem_desc *, bool,
+ struct goacc_asyncqueue *);
extern void gomp_init_device (struct gomp_device_descr *);
+extern bool gomp_fini_device (struct gomp_device_descr *);
extern void gomp_free_memmap (struct splay_tree_s *);
extern void gomp_unload_device (struct gomp_device_descr *);
extern bool gomp_remove_var (struct gomp_device_descr *, splay_tree_key);
Index: libgomp/libgomp.map
===================================================================
--- libgomp/libgomp.map (revision 267226)
+++ libgomp/libgomp.map (working copy)
@@ -464,8 +464,12 @@ OACC_2.5 {
acc_delete_finalize_async_32_h_;
acc_delete_finalize_async_64_h_;
acc_delete_finalize_async_array_h_;
+ acc_get_default_async;
+ acc_get_default_async_h_;
acc_memcpy_from_device_async;
acc_memcpy_to_device_async;
+ acc_set_default_async;
+ acc_set_default_async_h_;
acc_update_device_async;
acc_update_device_async_32_h_;
acc_update_device_async_64_h_;
Index: libgomp/libgomp-plugin.h
===================================================================
--- libgomp/libgomp-plugin.h (revision 267226)
+++ libgomp/libgomp-plugin.h (working copy)
@@ -53,6 +53,20 @@ enum offload_target_type
OFFLOAD_TARGET_TYPE_HSA = 7
};
+/* Opaque type to represent plugin-dependent implementation of an
+ OpenACC asynchronous queue. */
+struct goacc_asyncqueue;
+
+/* Used to keep a list of active asynchronous queues. */
+struct goacc_asyncqueue_list
+{
+ struct goacc_asyncqueue *aq;
+ struct goacc_asyncqueue_list *next;
+};
+
+typedef struct goacc_asyncqueue *goacc_aq;
+typedef struct goacc_asyncqueue_list *goacc_aq_list;
+
/* Auxiliary struct, used for transferring pairs of addresses from plugin
to libgomp. */
struct addr_pair
@@ -93,22 +107,31 @@ extern bool GOMP_OFFLOAD_dev2dev (int, void *, con
extern bool GOMP_OFFLOAD_can_run (void *);
extern void GOMP_OFFLOAD_run (int, void *, void *, void **);
extern void GOMP_OFFLOAD_async_run (int, void *, void *, void **, void *);
+
extern void GOMP_OFFLOAD_openacc_exec (void (*) (void *), size_t, void **,
- void **, int, unsigned *, void *);
-extern void GOMP_OFFLOAD_openacc_register_async_cleanup (void *, int);
-extern int GOMP_OFFLOAD_openacc_async_test (int);
-extern int GOMP_OFFLOAD_openacc_async_test_all (void);
-extern void GOMP_OFFLOAD_openacc_async_wait (int);
-extern void GOMP_OFFLOAD_openacc_async_wait_async (int, int);
-extern void GOMP_OFFLOAD_openacc_async_wait_all (void);
-extern void GOMP_OFFLOAD_openacc_async_wait_all_async (int);
-extern void GOMP_OFFLOAD_openacc_async_set_async (int);
+ void **, unsigned *, void *);
extern void *GOMP_OFFLOAD_openacc_create_thread_data (int);
extern void GOMP_OFFLOAD_openacc_destroy_thread_data (void *);
+extern struct goacc_asyncqueue *GOMP_OFFLOAD_openacc_async_construct (void);
+extern bool GOMP_OFFLOAD_openacc_async_destruct (struct goacc_asyncqueue *);
+extern int GOMP_OFFLOAD_openacc_async_test (struct goacc_asyncqueue *);
+extern void GOMP_OFFLOAD_openacc_async_synchronize (struct goacc_asyncqueue *);
+extern void GOMP_OFFLOAD_openacc_async_serialize (struct goacc_asyncqueue *,
+ struct goacc_asyncqueue *);
+extern void GOMP_OFFLOAD_openacc_async_queue_callback (struct goacc_asyncqueue
*,
+ void (*)(void *), void
*);
+extern void GOMP_OFFLOAD_openacc_async_exec (void (*) (void *), size_t, void
**,
+ void **, unsigned *, void *,
+ struct goacc_asyncqueue *);
+extern bool GOMP_OFFLOAD_openacc_async_dev2host (int, void *, const void *,
size_t,
+ struct goacc_asyncqueue *);
+extern bool GOMP_OFFLOAD_openacc_async_host2dev (int, void *, const void *,
size_t,
+ struct goacc_asyncqueue *);
extern void *GOMP_OFFLOAD_openacc_cuda_get_current_device (void);
extern void *GOMP_OFFLOAD_openacc_cuda_get_current_context (void);
-extern void *GOMP_OFFLOAD_openacc_cuda_get_stream (int);
-extern int GOMP_OFFLOAD_openacc_cuda_set_stream (int, void *);
+extern void *GOMP_OFFLOAD_openacc_cuda_get_stream (struct goacc_asyncqueue *);
+extern int GOMP_OFFLOAD_openacc_cuda_set_stream (struct goacc_asyncqueue *,
+ void *);
#ifdef __cplusplus
}
diff -ru trunk-orig/libgomp/libgomp.h trunk-work/libgomp/libgomp.h
--- trunk-orig/libgomp/libgomp.h 2018-12-14 18:31:07.487203770 +0800
+++ trunk-work/libgomp/libgomp.h 2018-12-18 22:24:06.351190428 +0800
@@ -949,6 +949,11 @@
/* Execute. */
__typeof (GOMP_OFFLOAD_openacc_exec) *exec_func;
+ /* Create/destroy TLS data. */
+ __typeof (GOMP_OFFLOAD_openacc_create_thread_data) *create_thread_data_func;
+ __typeof (GOMP_OFFLOAD_openacc_destroy_thread_data)
+ *destroy_thread_data_func;
+
struct {
gomp_mutex_t lock;
int nasyncqueue;
@@ -963,15 +968,10 @@
__typeof (GOMP_OFFLOAD_openacc_async_queue_callback) *queue_callback_func;
__typeof (GOMP_OFFLOAD_openacc_async_exec) *exec_func;
- __typeof (GOMP_OFFLOAD_openacc_async_host2dev) *host2dev_func;
__typeof (GOMP_OFFLOAD_openacc_async_dev2host) *dev2host_func;
+ __typeof (GOMP_OFFLOAD_openacc_async_host2dev) *host2dev_func;
} async;
- /* Create/destroy TLS data. */
- __typeof (GOMP_OFFLOAD_openacc_create_thread_data) *create_thread_data_func;
- __typeof (GOMP_OFFLOAD_openacc_destroy_thread_data)
- *destroy_thread_data_func;
-
/* NVIDIA target specific routines. */
struct {
__typeof (GOMP_OFFLOAD_openacc_cuda_get_current_device)
diff -ru trunk-orig/libgomp/libgomp.map trunk-work/libgomp/libgomp.map
--- trunk-orig/libgomp/libgomp.map 2018-12-14 18:31:07.487203770 +0800
+++ trunk-work/libgomp/libgomp.map 2018-12-18 22:28:18.103210976 +0800
@@ -506,6 +506,7 @@
GOMP_PLUGIN_debug;
GOMP_PLUGIN_error;
GOMP_PLUGIN_fatal;
+ GOMP_PLUGIN_async_unmap_vars;
GOMP_PLUGIN_acc_thread;
};
diff -ru trunk-orig/libgomp/libgomp-plugin.h trunk-work/libgomp/libgomp-plugin.h
--- trunk-orig/libgomp/libgomp-plugin.h 2018-12-14 18:31:07.487203770 +0800
+++ trunk-work/libgomp/libgomp-plugin.h 2018-12-18 22:56:12.338389017 +0800
@@ -110,9 +110,8 @@
extern void GOMP_OFFLOAD_openacc_exec (void (*) (void *), size_t, void **,
void **, unsigned *, void *);
-extern void GOMP_OFFLOAD_openacc_async_exec (void (*) (void *), size_t, void
**,
- void **, unsigned *, void *,
- struct goacc_asyncqueue *);
+extern void *GOMP_OFFLOAD_openacc_create_thread_data (int);
+extern void GOMP_OFFLOAD_openacc_destroy_thread_data (void *);
extern struct goacc_asyncqueue *GOMP_OFFLOAD_openacc_async_construct (void);
extern bool GOMP_OFFLOAD_openacc_async_destruct (struct goacc_asyncqueue *);
extern int GOMP_OFFLOAD_openacc_async_test (struct goacc_asyncqueue *);
@@ -121,12 +120,13 @@
struct goacc_asyncqueue *);
extern void GOMP_OFFLOAD_openacc_async_queue_callback (struct goacc_asyncqueue
*,
void (*)(void *), void
*);
-extern bool GOMP_OFFLOAD_openacc_async_host2dev (int, void *, const void *,
size_t,
- struct goacc_asyncqueue *);
+extern void GOMP_OFFLOAD_openacc_async_exec (void (*) (void *), size_t, void
**,
+ void **, unsigned *, void *,
+ struct goacc_asyncqueue *);
extern bool GOMP_OFFLOAD_openacc_async_dev2host (int, void *, const void *,
size_t,
struct goacc_asyncqueue *);
-extern void *GOMP_OFFLOAD_openacc_create_thread_data (int);
-extern void GOMP_OFFLOAD_openacc_destroy_thread_data (void *);
+extern bool GOMP_OFFLOAD_openacc_async_host2dev (int, void *, const void *,
size_t,
+ struct goacc_asyncqueue *);
extern void *GOMP_OFFLOAD_openacc_cuda_get_current_device (void);
extern void *GOMP_OFFLOAD_openacc_cuda_get_current_context (void);
extern void *GOMP_OFFLOAD_openacc_cuda_get_stream (struct goacc_asyncqueue *);
