While the tree dump looked fine to a C coder, passing an array
argument requires the '&' to take the address. Additionally,
it makes sense to associate the call line number to the library
calls - making the stepping through in the debugger more consistent.
I was considering putting the init/cleanup library calls for dispatch
into the '#pragma omp dispatch' line, but that gave odd jumps with
gimplificated line numbers - admittedly only for generated assignments.
As using the 'omp dispatch' line has as many pros as cons, I settled
using the function call location – which is consistent enough.
I intent to commit this later today, unless there are more comments.
* * *
The testcase is somewhat comprehensive and motivated by the checks
inlibgomp.c/interop-fr-1.c for 'omp interop'. I intent to add some more
checks using the same infrastructure; hence, the header file.
* * *
Note: The omp_get_default_device() call has no location; I was thinking
of adding one, but the real solution seems to remove it in favor or
using the omp_default_device constant instead. To do so cleanly,
requires a bit more update to the API routines; hence, I deferred
this to PR119677, https://gcc.gnu.org/119677 (see attached patch for
the modify_call_for_omp_dispatch part (+ tons of testcase changes).
'omp dispatch' itself (with interop or device clause) also adds
libgomp calls; those currently also have no line location, but this
shouldn't matter much.
Tobias
OpenMP: Fix append_args handling in modify_call_for_omp_dispatch
At tree level, the addr ref is also required for array dummy arguments,
contrary to C; the GOMP_interop calls in modify_call_for_omp_dispatch
were updated accordingly (using build_fold_addr_expr).
As the GOMP_interop calls had no location data associated with them,
the init call happened as soon as executing the previous line of code,
which was confusing; solution: use the location data of the function
call itself.
PR middle-end/119662
gcc/ChangeLog:
* gimplify.cc (modify_call_for_omp_dispatch): Fix GOMP_interop
arg passing; add location info to function calls.
libgomp/ChangeLog:
* testsuite/libgomp.c/append-args-fr-1.c: New test.
* testsuite/libgomp.c/append-args-fr.h: New test.
gcc/testsuite/ChangeLog:
* c-c++-common/gomp/append-args-interop.c: Update for fixed
GOMP_interop call.
* g++.dg/gomp/append-args-8.C: Likewise.
* gfortran.dg/gomp/append-args-interop.f90: Likewise.
gcc/gimplify.cc | 32 ++-
.../c-c++-common/gomp/append-args-interop.c | 4 +-
gcc/testsuite/g++.dg/gomp/append-args-8.C | 8 +-
.../gfortran.dg/gomp/append-args-interop.f90 | 4 +-
libgomp/testsuite/libgomp.c/append-args-fr-1.c | 232 ++++++++++++++++
libgomp/testsuite/libgomp.c/append-args-fr.h | 305 +++++++++++++++++++++
6 files changed, 563 insertions(+), 22 deletions(-)
diff --git a/gcc/gimplify.cc b/gcc/gimplify.cc
index 51595f6f51e..4f385b1b779 100644
--- a/gcc/gimplify.cc
+++ b/gcc/gimplify.cc
@@ -3887,6 +3887,7 @@ static tree
modify_call_for_omp_dispatch (tree expr, tree dispatch_clauses,
bool want_value, bool pointerize)
{
+ location_t loc = EXPR_LOCATION (expr);
tree fndecl = get_callee_fndecl (expr);
/* Skip processing if we don't get the expected call form. */
@@ -4081,14 +4082,17 @@ modify_call_for_omp_dispatch (tree expr, tree dispatch_clauses,
}
}
+ objs = build_fold_addr_expr (objs);
+ target_tgtsync = build_fold_addr_expr (target_tgtsync);
+ prefer_type = prefer_type ? build_fold_addr_expr (prefer_type)
+ : null_pointer_node;
tree fn = builtin_decl_explicit (BUILT_IN_GOMP_INTEROP);
tree create
- = build_call_expr (fn, 11, dispatch_device_num,
- nobjs, objs, target_tgtsync,
- prefer_type ? prefer_type : null_pointer_node,
- integer_zero_node, null_pointer_node,
- integer_zero_node, null_pointer_node,
- integer_zero_node, null_pointer_node);
+ = build_call_expr_loc (loc, fn, 11, dispatch_device_num,
+ nobjs, objs, target_tgtsync, prefer_type,
+ integer_zero_node, null_pointer_node,
+ integer_zero_node, null_pointer_node,
+ integer_zero_node, null_pointer_node);
if (init_code)
init_code = build2 (COMPOUND_EXPR, TREE_TYPE (create),
init_code, create);
@@ -4096,12 +4100,12 @@ modify_call_for_omp_dispatch (tree expr, tree dispatch_clauses,
init_code = create;
cleanup
- = build_call_expr (fn, 11, dispatch_device_num,
- integer_zero_node, null_pointer_node,
- null_pointer_node, null_pointer_node,
- integer_zero_node, null_pointer_node,
- nobjs, objs,
- integer_zero_node, null_pointer_node);
+ = build_call_expr_loc (loc, fn, 11, dispatch_device_num,
+ integer_zero_node, null_pointer_node,
+ null_pointer_node, null_pointer_node,
+ integer_zero_node, null_pointer_node,
+ nobjs, objs,
+ integer_zero_node, null_pointer_node);
if (clobbers)
cleanup = build2 (COMPOUND_EXPR, TREE_TYPE (clobbers),
cleanup, clobbers);
@@ -4282,8 +4286,8 @@ modify_call_for_omp_dispatch (tree expr, tree dispatch_clauses,
actual_ptr);
}
tree fn = builtin_decl_explicit (BUILT_IN_OMP_GET_MAPPED_PTR);
- tree mapped_arg = build_call_expr (fn, 2, actual_ptr,
- dispatch_device_num);
+ tree mapped_arg = build_call_expr_loc (loc, fn, 2, actual_ptr,
+ dispatch_device_num);
if (TREE_CODE (*arg_p) == ADDR_EXPR
|| (TREE_CODE (TREE_TYPE (actual_ptr)) == REFERENCE_TYPE))
diff --git a/gcc/testsuite/c-c++-common/gomp/append-args-interop.c b/gcc/testsuite/c-c++-common/gomp/append-args-interop.c
index 1211450ce63..e9b1ed43c07 100644
--- a/gcc/testsuite/c-c++-common/gomp/append-args-interop.c
+++ b/gcc/testsuite/c-c++-common/gomp/append-args-interop.c
@@ -39,6 +39,6 @@ test (int *a, int *b)
return x;
}
-/* { dg-final { scan-tree-dump "__builtin_GOMP_interop \\(D\.\[0-9\]+, 2, interopobjs\.\[0-9\]+, tgt_tgtsync\.\[0-9\]+," "gimple" } } */
-/* { dg-final { scan-tree-dump "__builtin_GOMP_interop \\(D\.\[0-9\]+, 0, 0B, 0B, 0B, 0, 0B, 2, interopobjs\.\[0-9\]+," "gimple" } } */
+/* { dg-final { scan-tree-dump "__builtin_GOMP_interop \\(D\.\[0-9\]+, 2, &interopobjs\.\[0-9\]+, &tgt_tgtsync\.\[0-9\]+," "gimple" } } */
+/* { dg-final { scan-tree-dump "__builtin_GOMP_interop \\(D\.\[0-9\]+, 0, 0B, 0B, 0B, 0, 0B, 2, &interopobjs\.\[0-9\]+," "gimple" } } */
/* { dg-final { scan-tree-dump "repl1 \\(obj1, interop\.\[0-9\]+, interop\.\[0-9\]+\\)" "gimple" } } */
diff --git a/gcc/testsuite/g++.dg/gomp/append-args-8.C b/gcc/testsuite/g++.dg/gomp/append-args-8.C
index 379c7677c95..3052c5b75b3 100644
--- a/gcc/testsuite/g++.dg/gomp/append-args-8.C
+++ b/gcc/testsuite/g++.dg/gomp/append-args-8.C
@@ -82,11 +82,11 @@ test (int *a, int *b)
/* { dg-final { scan-tree-dump-times "__builtin_omp_set_default_device \\(D\.\[0-9\]+\\);" 2 "gimple" } } */
/* { dg-final { scan-tree-dump-times "D\.\[0-9\]+ = __builtin_omp_get_mapped_ptr \\(a, D\.\[0-9\]+\\);" 1 "gimple" } } */
-/* { dg-final { scan-tree-dump-times "__builtin_GOMP_interop \\(D\.\[0-9\]+, 1, interopobjs\.\[0-9\], tgt_tgtsync\.\[0-9\]+, pref_type.2, 0, 0B, 0, 0B, 0, 0B\\);" 1 "gimple" } } */
+/* { dg-final { scan-tree-dump-times "__builtin_GOMP_interop \\(D\.\[0-9\]+, 1, &interopobjs\.\[0-9\], &tgt_tgtsync\.\[0-9\]+, &pref_type.2, 0, 0B, 0, 0B, 0, 0B\\);" 1 "gimple" } } */
/* { dg-final { scan-tree-dump-times "repl2<int\\*, int\\*, omp_interop_t> \\(b, D\.\[0-9\]+, obj1, interop\.\[0-9\]+\\);" 1 "gimple" } } */
-/* { dg-final { scan-tree-dump-times "__builtin_GOMP_interop \\(D\.\[0-9\]+, 0, 0B, 0B, 0B, 0, 0B, 1, interopobjs\.\[0-9\]+, 0, 0B\\);" 1 "gimple" } } */
+/* { dg-final { scan-tree-dump-times "__builtin_GOMP_interop \\(D\.\[0-9\]+, 0, 0B, 0B, 0B, 0, 0B, 1, &interopobjs\.\[0-9\]+, 0, 0B\\);" 1 "gimple" } } */
-/* { dg-final { scan-tree-dump-times "__builtin_GOMP_interop \\(-5, 0, 0B, 0B, 0B, 0, 0B, 3, interopobjs\.\[0-9\]+, 0, 0B\\);" 1 "gimple" } } */
+/* { dg-final { scan-tree-dump-times "__builtin_GOMP_interop \\(-5, 0, 0B, 0B, 0B, 0, 0B, 3, &interopobjs\.\[0-9\]+, 0, 0B\\);" 1 "gimple" } } */
/* { dg-final { scan-tree-dump-times "repl3<int\\*, omp_interop_t> \\(a, interop\.\[0-9\]+, interop\.\[0-9\]+, interop\.\[0-9\]+, 1, 2, \"abc\"\\);" 1 "gimple" } } */
-/* { dg-final { scan-tree-dump-times "__builtin_GOMP_interop \\(-5, 3, interopobjs\.\[0-9\]+, tgt_tgtsync\.\[0-9\]+, pref_type\.\[0-9\]+, 0, 0B, 0, 0B, 0, 0B\\);" 1 "gimple" } } */
+/* { dg-final { scan-tree-dump-times "__builtin_GOMP_interop \\(-5, 3, &interopobjs\.\[0-9\]+, &tgt_tgtsync\.\[0-9\]+, &pref_type\.\[0-9\]+, 0, 0B, 0, 0B, 0, 0B\\);" 1 "gimple" } } */
diff --git a/gcc/testsuite/gfortran.dg/gomp/append-args-interop.f90 b/gcc/testsuite/gfortran.dg/gomp/append-args-interop.f90
index f2c4d971464..540079a28ac 100644
--- a/gcc/testsuite/gfortran.dg/gomp/append-args-interop.f90
+++ b/gcc/testsuite/gfortran.dg/gomp/append-args-interop.f90
@@ -23,6 +23,6 @@ use m
call f()
end
-! { dg-final { scan-tree-dump "__builtin_GOMP_interop \\(99, 3, interopobjs\.\[0-9\]+, tgt_tgtsync\.\[0-9\]+, pref_type\.\[0-9\]+, " "gimple" } }
-! { dg-final { scan-tree-dump "__builtin_GOMP_interop \\(99, 0, 0B, 0B, 0B, 0, 0B, 3, interopobjs\.\[0-9\]+," "gimple" } }
+! { dg-final { scan-tree-dump "__builtin_GOMP_interop \\(99, 3, &interopobjs\.\[0-9\]+, &tgt_tgtsync\.\[0-9\]+, &pref_type\.\[0-9\]+, " "gimple" } }
+! { dg-final { scan-tree-dump "__builtin_GOMP_interop \\(99, 0, 0B, 0B, 0B, 0, 0B, 3, &interopobjs\.\[0-9\]+," "gimple" } }
! { dg-final { scan-tree-dump "g \\(&interop\.\[0-9\]+, interop\.\[0-9\]+, &interop\.\[0-9\]+\\)" "gimple" } }
diff --git a/libgomp/testsuite/libgomp.c/append-args-fr-1.c b/libgomp/testsuite/libgomp.c/append-args-fr-1.c
new file mode 100644
index 00000000000..2fd7edad9f6
--- /dev/null
+++ b/libgomp/testsuite/libgomp.c/append-args-fr-1.c
@@ -0,0 +1,232 @@
+/* { dg-do run } */
+
+#include "append-args-fr.h"
+
+enum { host_device, nvptx_device, gcn_device } used_device_type, used_device_type2;
+static int used_device_num, used_device_num2;
+static omp_interop_fr_t expected_fr, expected_fr2;
+static _Bool is_targetsync, is_targetsync2;
+
+void
+check_interop (omp_interop_t obj)
+{
+ if (used_device_type == host_device)
+ check_host (obj);
+ else if (used_device_type == nvptx_device)
+ check_nvptx (obj, used_device_num, expected_fr, is_targetsync);
+ else if (used_device_type == gcn_device)
+ check_gcn (obj, used_device_num, expected_fr, is_targetsync);
+ else
+ __builtin_unreachable ();
+
+ #pragma omp interop use(obj)
+}
+
+void
+check_interop2 (omp_interop_t obj, omp_interop_t obj2)
+{
+ check_interop (obj);
+
+ #pragma omp interop use(obj2)
+
+ if (used_device_type2 == host_device)
+ check_host (obj2);
+ else if (used_device_type2 == nvptx_device)
+ check_nvptx (obj2, used_device_num2, expected_fr2, is_targetsync2);
+ else if (used_device_type2 == gcn_device)
+ check_gcn (obj2, used_device_num2, expected_fr2, is_targetsync2);
+ else
+ __builtin_unreachable ();
+}
+
+
+/* Check no args + one interop arg - and no prefer_type. */
+
+int f0_1_tg_ (omp_interop_t obj) { check_interop (obj); return 4242; }
+#pragma omp declare variant(f0_1_tg_) match(construct={dispatch}) append_args(interop(target))
+int f0_1_tg () { assert (false); return 42; }
+
+void f0_1_tgsy_ (omp_interop_t obj) { check_interop (obj); }
+#pragma omp declare variant(f0_1_tgsy_) match(construct={dispatch}) append_args(interop(targetsync))
+void f0_1_tgsy () { assert (false); }
+
+int f0_1_tgtgsy_ (omp_interop_t obj) { check_interop (obj); return 3333; }
+#pragma omp declare variant(f0_1_tgtgsy_) match(construct={dispatch}) append_args(interop(targetsync,target))
+int f0_1_tgtgsy () { assert (false); return 33; }
+
+
+/* And with PREFER_TYPE. */
+
+// nv: cuda, gcn: -, -, hip
+void f0_1_tgsy_c_cd_hi_hs_ (omp_interop_t obj) { check_interop (obj); }
+#pragma omp declare variant(f0_1_tgsy_c_cd_hi_hs_) match(construct={dispatch}) \
+ append_args(interop(targetsync, prefer_type("cuda","cuda_driver", "hip", "hsa")))
+void f0_1_tgsy_c_cd_hi_hs () { assert (false); }
+
+// nv: -, cuda_driver, gcn: hsa
+void f0_1_tgsy_hs_cd_c_hi_ (omp_interop_t obj) { check_interop (obj); }
+#pragma omp declare variant(f0_1_tgsy_hs_cd_c_hi_) match(construct={dispatch}) \
+ append_args(interop(targetsync, prefer_type({attr("ompx_foo")}, {fr("hsa")}, {attr("ompx_bar"), fr("cuda_driver"), attr("ompx_foobar")},{fr("cuda")}, {fr("hip")})))
+void f0_1_tgsy_hs_cd_c_hi () { assert (false); }
+
+// nv: -, hip, gcn: hsa
+void f0_1_tgsy_hs_hi_cd_c_ (omp_interop_t obj) { check_interop (obj); }
+#pragma omp declare variant(f0_1_tgsy_hs_hi_cd_c_) match(construct={dispatch}) \
+ append_args(interop(targetsync, prefer_type("hsa", "hip", "cuda_driver", "cuda")))
+void f0_1_tgsy_hs_hi_cd_c () { assert (false); }
+
+
+void
+check_f0 ()
+{
+ if (used_device_type == nvptx_device)
+ expected_fr = omp_ifr_cuda;
+ else if (used_device_type == gcn_device)
+ expected_fr = omp_ifr_hip;
+ else /* host; variable shall not be accessed */
+ expected_fr = omp_ifr_level_zero;
+
+ int i;
+ if (used_device_num == DEFAULT_DEVICE)
+ {
+ is_targetsync = 0;
+ #pragma omp dispatch
+ i = f0_1_tg ();
+ assert (i == 4242);
+
+ is_targetsync = 1;
+ #pragma omp dispatch
+ f0_1_tgsy ();
+
+ #pragma omp dispatch
+ i = f0_1_tgtgsy ();
+ assert (i == 3333);
+
+
+ if (used_device_type == nvptx_device)
+ expected_fr = omp_ifr_cuda;
+ else if (used_device_type == gcn_device)
+ expected_fr = omp_ifr_hip;
+ #pragma omp dispatch
+ f0_1_tgsy_c_cd_hi_hs ();
+
+ if (used_device_type == nvptx_device)
+ expected_fr = omp_ifr_cuda_driver;
+ else if (used_device_type == gcn_device)
+ expected_fr = omp_ifr_hsa;
+ #pragma omp dispatch
+ f0_1_tgsy_hs_cd_c_hi ();
+
+ if (used_device_type == nvptx_device)
+ expected_fr = omp_ifr_hip;
+ else if (used_device_type == gcn_device)
+ expected_fr = omp_ifr_hsa;
+ #pragma omp dispatch
+ f0_1_tgsy_hs_hi_cd_c ();
+ }
+ else
+ {
+ is_targetsync = 0;
+ #pragma omp dispatch device(used_device_num)
+ i = f0_1_tg ();
+ assert (i == 4242);
+
+ is_targetsync = 1;
+ #pragma omp dispatch device(used_device_num)
+ f0_1_tgsy ();
+
+ #pragma omp dispatch device(used_device_num)
+ i = f0_1_tgtgsy ();
+ assert (i == 3333);
+
+
+ if (used_device_type == nvptx_device)
+ expected_fr = omp_ifr_cuda;
+ else if (used_device_type == gcn_device)
+ expected_fr = omp_ifr_hip;
+ #pragma omp dispatch device(used_device_num)
+ f0_1_tgsy_c_cd_hi_hs ();
+
+ if (used_device_type == nvptx_device)
+ expected_fr = omp_ifr_cuda_driver;
+ else if (used_device_type == gcn_device)
+ expected_fr = omp_ifr_hsa;
+ #pragma omp dispatch device(used_device_num)
+ f0_1_tgsy_hs_cd_c_hi ();
+
+ if (used_device_type == nvptx_device)
+ expected_fr = omp_ifr_hip;
+ else if (used_device_type == gcn_device)
+ expected_fr = omp_ifr_hsa;
+ #pragma omp dispatch device(used_device_num)
+ f0_1_tgsy_hs_hi_cd_c ();
+ }
+}
+
+
+
+void
+do_check (int dev)
+{
+ int num_dev = omp_get_num_devices ();
+ const char *dev_type;
+ if (dev != DEFAULT_DEVICE)
+ omp_set_default_device (dev);
+ int is_nvptx = on_device_arch_nvptx ();
+ int is_gcn = on_device_arch_gcn ();
+ int is_host;
+
+ if (dev != DEFAULT_DEVICE)
+ is_host = dev == -1 || dev == num_dev;
+ else
+ {
+ int def_dev = omp_get_default_device ();
+ is_host = def_dev == -1 || def_dev == num_dev;
+ }
+
+ assert (is_nvptx + is_gcn + is_host == 1);
+
+ if (num_dev > 0 && dev != DEFAULT_DEVICE)
+ {
+ if (is_host)
+ omp_set_default_device (0);
+ else
+ omp_set_default_device (-1);
+ }
+
+ used_device_num = dev;
+ if (is_host)
+ {
+ dev_type = "host";
+ used_device_type = host_device;
+ }
+ else if (is_nvptx)
+ {
+ dev_type = "nvptx";
+ used_device_type = nvptx_device;
+ }
+ else if (is_gcn)
+ {
+ dev_type = "gcn";
+ used_device_type = gcn_device;
+ }
+
+ printf ("Running on the %s device (%d)\n", dev_type, dev);
+ check_f0 ();
+}
+
+
+
+int
+main ()
+{
+ do_check (DEFAULT_DEVICE);
+ int ndev = omp_get_num_devices ();
+ for (int dev = -1; dev < ndev; dev++)
+ do_check (dev);
+ for (int dev = -1; dev < ndev; dev++)
+ {
+ omp_set_default_device (dev);
+ do_check (DEFAULT_DEVICE);
+ }
+}
diff --git a/libgomp/testsuite/libgomp.c/append-args-fr.h b/libgomp/testsuite/libgomp.c/append-args-fr.h
new file mode 100644
index 00000000000..9f6ca04bbe9
--- /dev/null
+++ b/libgomp/testsuite/libgomp.c/append-args-fr.h
@@ -0,0 +1,305 @@
+#include <assert.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+#include <omp.h>
+#include "../libgomp.c-c++-common/on_device_arch.h"
+
+/* Provides: */
+
+#define DEFAULT_DEVICE -99
+
+void check_host (omp_interop_t obj);
+void check_nvptx (omp_interop_t obj, int dev, omp_interop_fr_t expected_fr, _Bool is_targetsync);
+void check_gcn (omp_interop_t obj, int dev, omp_interop_fr_t expected_fr, _Bool is_targetsync);
+
+
+/* The following assumes that when a nvptx device is available,
+ cuda/cuda_driver/hip are supported.
+ And that likewise when a gcn device is available that the
+ plugin also can not only the HSA but also the HIP library
+ such that hsa/hip are supported.
+ For the host, omp_interop_none is expected.
+
+ Otherwise, it only does some basic tests without checking
+ that the returned result really makes sense. */
+
+void check_type (omp_interop_t obj)
+{
+ const char *type;
+
+ type = omp_get_interop_type_desc (obj, omp_ipr_fr_id);
+ if (obj != omp_interop_none)
+ assert (strcmp (type, "omp_interop_t") == 0);
+ else
+ assert (type == NULL);
+
+ type = omp_get_interop_type_desc (obj, omp_ipr_fr_name);
+ if (obj != omp_interop_none)
+ assert (strcmp (type, "const char *") == 0);
+ else
+ assert (type == NULL);
+
+ type = omp_get_interop_type_desc (obj, omp_ipr_vendor);
+ if (obj != omp_interop_none)
+ assert (strcmp (type, "int") == 0);
+ else
+ assert (type == NULL);
+
+ type = omp_get_interop_type_desc (obj, omp_ipr_vendor_name);
+ if (obj != omp_interop_none)
+ assert (strcmp (type, "const char *") == 0);
+ else
+ assert (type == NULL);
+
+ type = omp_get_interop_type_desc (obj, omp_ipr_device_num);
+ if (obj != omp_interop_none)
+ assert (strcmp (type, "int") == 0);
+ else
+ assert (type == NULL);
+
+ if (obj != omp_interop_none)
+ return;
+ assert (omp_get_interop_type_desc (obj, omp_ipr_platform) == NULL);
+ assert (omp_get_interop_type_desc (obj, omp_ipr_device) == NULL);
+ assert (omp_get_interop_type_desc (obj, omp_ipr_device_context) == NULL);
+ assert (omp_get_interop_type_desc (obj, omp_ipr_targetsync) == NULL);
+}
+
+
+void
+check_host (omp_interop_t obj)
+{
+ assert (obj == omp_interop_none);
+ check_type (obj);
+}
+
+
+void
+check_nvptx (omp_interop_t obj, int dev, omp_interop_fr_t expected_fr, _Bool is_targetsync)
+{
+ assert (obj != omp_interop_none && obj != (omp_interop_t) -1L);
+
+ omp_interop_rc_t ret_code = omp_irc_no_value;
+ omp_interop_fr_t fr = (omp_interop_fr_t) omp_get_interop_int (obj, omp_ipr_fr_id, &ret_code);
+
+ assert (ret_code == omp_irc_success);
+ assert (fr == expected_fr);
+
+ ret_code = omp_irc_no_value;
+ const char *fr_name = omp_get_interop_str (obj, omp_ipr_fr_name, &ret_code);
+
+ assert (ret_code == omp_irc_success);
+ if (fr == omp_ifr_cuda)
+ assert (strcmp (fr_name, "cuda") == 0);
+ else if (fr == omp_ifr_cuda_driver)
+ assert (strcmp (fr_name, "cuda_driver") == 0);
+ else if (fr == omp_ifr_hip)
+ assert (strcmp (fr_name, "hip") == 0);
+ else
+ assert (0);
+
+ ret_code = omp_irc_no_value;
+ int vendor = (int) omp_get_interop_int (obj, omp_ipr_vendor, &ret_code);
+ assert (ret_code == omp_irc_success);
+ assert (vendor == 11); /* Nvidia */
+
+ ret_code = omp_irc_no_value;
+ const char *vendor_name = omp_get_interop_str (obj, omp_ipr_vendor_name, &ret_code);
+ assert (ret_code == omp_irc_success);
+ assert (strcmp (vendor_name, "nvidia") == 0);
+
+ ret_code = omp_irc_no_value;
+ int dev_num = (int) omp_get_interop_int (obj, omp_ipr_device_num, &ret_code);
+ assert (ret_code == omp_irc_success);
+ if (dev == DEFAULT_DEVICE)
+ assert (dev_num == omp_get_default_device ());
+ else
+ assert (dev_num == dev);
+
+ /* Platform: N/A. */
+ ret_code = omp_irc_success;
+ (void) omp_get_interop_int (obj, omp_ipr_platform, &ret_code);
+ assert (ret_code == omp_irc_no_value);
+ ret_code = omp_irc_success;
+ (void) omp_get_interop_ptr (obj, omp_ipr_platform, &ret_code);
+ assert (ret_code == omp_irc_no_value);
+ ret_code = omp_irc_success;
+ (void) omp_get_interop_str (obj, omp_ipr_platform, &ret_code);
+ assert (ret_code == omp_irc_no_value);
+
+ /* Device: int / CUdevice / hipDevice_t -- all internally an 'int'. */
+ ret_code = omp_irc_no_value;
+ int fr_device = (int) omp_get_interop_int (obj, omp_ipr_device, &ret_code);
+
+ /* CUDA also starts from 0 and goes to < n with cudaGetDeviceCount(&cn). */
+ assert (ret_code == omp_irc_success);
+ assert (fr_device >= 0 && fr_device < omp_get_num_devices ());
+
+ /* Device context: N/A / CUcontext / hipCtx_t -- a pointer. */
+ ret_code = omp_irc_out_of_range;
+ void *ctx = omp_get_interop_ptr (obj, omp_ipr_device_context, &ret_code);
+
+ if (fr == omp_ifr_cuda)
+ {
+ assert (ret_code == omp_irc_no_value);
+ assert (ctx == NULL);
+ }
+ else
+ {
+ assert (ret_code == omp_irc_success);
+ assert (ctx != NULL);
+ }
+
+ /* Stream/targetsync: cudaStream_t / CUstream / hipStream_t -- a pointer. */
+ ret_code = omp_irc_out_of_range;
+ void *stream = omp_get_interop_ptr (obj, omp_ipr_targetsync, &ret_code);
+
+ if (is_targetsync) /* no targetsync */
+ {
+ assert (ret_code == omp_irc_success);
+ assert (stream != NULL);
+ }
+ else
+ {
+ assert (ret_code == omp_irc_no_value);
+ assert (stream == NULL);
+ }
+
+ check_type (obj);
+ if (fr == omp_ifr_cuda)
+ {
+ assert (strcmp (omp_get_interop_type_desc (obj, omp_ipr_platform), "N/A") == 0);
+ assert (strcmp (omp_get_interop_type_desc (obj, omp_ipr_device), "int") == 0);
+ assert (strcmp (omp_get_interop_type_desc (obj, omp_ipr_device_context), "N/A") == 0);
+ assert (strcmp (omp_get_interop_type_desc (obj, omp_ipr_targetsync), "cudaStream_t") == 0);
+ }
+ else if (fr == omp_ifr_cuda_driver)
+ {
+ assert (strcmp (omp_get_interop_type_desc (obj, omp_ipr_platform), "N/A") == 0);
+ assert (strcmp (omp_get_interop_type_desc (obj, omp_ipr_device), "CUdevice") == 0);
+ assert (strcmp (omp_get_interop_type_desc (obj, omp_ipr_device_context), "CUcontext") == 0);
+ assert (strcmp (omp_get_interop_type_desc (obj, omp_ipr_targetsync), "CUstream") == 0);
+ }
+ else
+ {
+ assert (strcmp (omp_get_interop_type_desc (obj, omp_ipr_platform), "N/A") == 0);
+ assert (strcmp (omp_get_interop_type_desc (obj, omp_ipr_device), "hipDevice_t") == 0);
+ assert (strcmp (omp_get_interop_type_desc (obj, omp_ipr_device_context), "hipCtx_t") == 0);
+ assert (strcmp (omp_get_interop_type_desc (obj, omp_ipr_targetsync), "hipStream_t") == 0);
+ }
+}
+
+
+void
+check_gcn (omp_interop_t obj, int dev, omp_interop_fr_t expected_fr, _Bool is_targetsync)
+{
+ assert (obj != omp_interop_none && obj != (omp_interop_t) -1L);
+
+ omp_interop_rc_t ret_code = omp_irc_no_value;
+ omp_interop_fr_t fr = (omp_interop_fr_t) omp_get_interop_int (obj, omp_ipr_fr_id, &ret_code);
+
+ assert (ret_code == omp_irc_success);
+ assert (fr == expected_fr);
+
+ ret_code = omp_irc_no_value;
+ const char *fr_name = omp_get_interop_str (obj, omp_ipr_fr_name, &ret_code);
+
+ assert (ret_code == omp_irc_success);
+ if (fr == omp_ifr_hip)
+ assert (strcmp (fr_name, "hip") == 0);
+ else if (fr == omp_ifr_hsa)
+ assert (strcmp (fr_name, "hsa") == 0);
+ else
+ assert (0);
+
+ ret_code = omp_irc_no_value;
+ int vendor = (int) omp_get_interop_int (obj, omp_ipr_vendor, &ret_code);
+ assert (ret_code == omp_irc_success);
+ assert (vendor == 1); /* Amd */
+
+ ret_code = omp_irc_no_value;
+ const char *vendor_name = omp_get_interop_str (obj, omp_ipr_vendor_name, &ret_code);
+ assert (ret_code == omp_irc_success);
+ assert (strcmp (vendor_name, "amd") == 0);
+
+ ret_code = omp_irc_no_value;
+ int dev_num = (int) omp_get_interop_int (obj, omp_ipr_device_num, &ret_code);
+ assert (ret_code == omp_irc_success);
+ if (dev == DEFAULT_DEVICE)
+ assert (dev_num == omp_get_default_device ());
+ else
+ assert (dev_num == dev);
+
+ /* Platform: N/A. */
+ ret_code = omp_irc_success;
+ (void) omp_get_interop_int (obj, omp_ipr_platform, &ret_code);
+ assert (ret_code == omp_irc_no_value);
+ ret_code = omp_irc_success;
+ (void) omp_get_interop_ptr (obj, omp_ipr_platform, &ret_code);
+ assert (ret_code == omp_irc_no_value);
+ ret_code = omp_irc_success;
+ (void) omp_get_interop_str (obj, omp_ipr_platform, &ret_code);
+ assert (ret_code == omp_irc_no_value);
+
+ /* Device: hipDevice_t / hsa_agent_t* -- hip is internally an 'int'. */
+ ret_code = omp_irc_no_value;
+ if (fr == omp_ifr_hip)
+ {
+ /* HIP also starts from 0 and goes to < n as with cudaGetDeviceCount(&cn). */
+ int fr_device = (int) omp_get_interop_int (obj, omp_ipr_device, &ret_code);
+ assert (ret_code == omp_irc_success);
+ assert (fr_device >= 0 && fr_device < omp_get_num_devices ());
+ }
+ else
+ {
+ void *agent = omp_get_interop_ptr (obj, omp_ipr_device, &ret_code);
+ assert (ret_code == omp_irc_success);
+ assert (agent != NULL);
+ }
+
+ /* Device context: hipCtx_t / N/A -- a pointer. */
+ ret_code = omp_irc_out_of_range;
+ void *ctx = omp_get_interop_ptr (obj, omp_ipr_device_context, &ret_code);
+ if (fr == omp_ifr_hip)
+ {
+ assert (ret_code == omp_irc_success);
+ assert (ctx != NULL);
+ }
+ else
+ {
+ assert (ret_code == omp_irc_no_value);
+ assert (ctx == NULL);
+ }
+
+ /* Stream/targetsync: cudaStream_t / CUstream / hipStream_t -- a pointer. */
+ ret_code = omp_irc_out_of_range;
+ void *stream = omp_get_interop_ptr (obj, omp_ipr_targetsync, &ret_code);
+
+ if (is_targetsync)
+ {
+ assert (ret_code == omp_irc_success);
+ assert (stream != NULL);
+ }
+ else
+ {
+ assert (ret_code == omp_irc_no_value);
+ assert (stream == NULL);
+ }
+
+ check_type (obj);
+ if (fr == omp_ifr_hip)
+ {
+ assert (strcmp (omp_get_interop_type_desc (obj, omp_ipr_platform), "N/A") == 0);
+ assert (strcmp (omp_get_interop_type_desc (obj, omp_ipr_device), "hipDevice_t") == 0);
+ assert (strcmp (omp_get_interop_type_desc (obj, omp_ipr_device_context), "hipCtx_t") == 0);
+ assert (strcmp (omp_get_interop_type_desc (obj, omp_ipr_targetsync), "hipStream_t") == 0);
+ }
+ else
+ {
+ assert (strcmp (omp_get_interop_type_desc (obj, omp_ipr_platform), "N/A") == 0);
+ assert (strcmp (omp_get_interop_type_desc (obj, omp_ipr_device), "hsa_agent_t *") == 0);
+ assert (strcmp (omp_get_interop_type_desc (obj, omp_ipr_device_context), "N/A") == 0);
+ assert (strcmp (omp_get_interop_type_desc (obj, omp_ipr_targetsync), "hsa_queue_t *") == 0);
+ }
+}
