static int
hwloc_solaris_get_sth_cpubind(hwloc_topology_t topology, idtype_t idtype, id_t id, hwloc_bitmap_t hwloc_set, int flags __hwloc_attribute_unused)
{
processorid_t binding;
int depth = hwloc_get_type_depth(topology, HWLOC_OBJ_NUMANODE);
int n;
int i;
if (depth < 0) {
errno = ENOSYS;
return -1;
}
/* first check if processor_bind() was used to bind to a single processor rather than to an lgroup */
if ( processor_bind(idtype, id, PBIND_QUERY, &binding) == 0 && binding != PBIND_NONE ) {
hwloc_bitmap_only(hwloc_set, binding);
return 0;
}
/* if not, check lgroups */
hwloc_bitmap_zero(hwloc_set);
n = hwloc_get_nbobjs_by_depth(topology, depth);
for (i = 0; i < n; i++) {
hwloc_obj_t obj = hwloc_get_obj_by_depth(topology, depth, i);
lgrp_affinity_t aff = lgrp_affinity_get(idtype, id, obj->os_index);
if (aff == LGRP_AFF_STRONG)
hwloc_bitmap_or(hwloc_set, hwloc_set, obj->cpuset);
}
if (hwloc_bitmap_iszero(hwloc_set))
hwloc_bitmap_copy(hwloc_set, hwloc_topology_get_complete_cpuset(topology));
return 0;
}
static HYD_status handle_rr_binding(void)
{
int i;
HYD_status status = HYD_SUCCESS;
HYDU_FUNC_ENTER();
HYDU_ASSERT(hwloc_initialized, status);
/* initialize bitmaps */
HYDT_topo_hwloc_info.num_bitmaps = hwloc_get_nbobjs_by_type(topology, HWLOC_OBJ_PU);
HYDU_MALLOC_OR_JUMP(HYDT_topo_hwloc_info.bitmap, hwloc_bitmap_t *,
HYDT_topo_hwloc_info.num_bitmaps * sizeof(hwloc_bitmap_t), status);
for (i = 0; i < HYDT_topo_hwloc_info.num_bitmaps; i++) {
HYDT_topo_hwloc_info.bitmap[i] = hwloc_bitmap_alloc();
hwloc_bitmap_only(HYDT_topo_hwloc_info.bitmap[i], i);
}
fn_exit:
HYDU_FUNC_EXIT();
return status;
fn_fail:
goto fn_exit;
}
static int
hwloc_win_get_area_membind(hwloc_topology_t topology __hwloc_attribute_unused, const void *addr, size_t len, hwloc_nodeset_t nodeset, hwloc_membind_policy_t * policy, int flags)
{
SYSTEM_INFO SystemInfo;
DWORD page_size;
uintptr_t start;
unsigned nb;
GetSystemInfo(&SystemInfo);
page_size = SystemInfo.dwPageSize;
start = (((uintptr_t) addr) / page_size) * page_size;
nb = (unsigned)((((uintptr_t) addr + len - start) + page_size - 1) / page_size);
if (!nb)
nb = 1;
{
PSAPI_WORKING_SET_EX_INFORMATION *pv;
unsigned i;
pv = calloc(nb, sizeof(*pv));
for (i = 0; i < nb; i++)
pv[i].VirtualAddress = (void*) (start + i * page_size);
if (!QueryWorkingSetExProc(GetCurrentProcess(), pv, nb * sizeof(*pv))) {
free(pv);
return -1;
}
*policy = HWLOC_MEMBIND_BIND;
if (flags & HWLOC_MEMBIND_STRICT) {
unsigned node = pv[0].VirtualAttributes.Node;
for (i = 1; i < nb; i++) {
if (pv[i].VirtualAttributes.Node != node) {
errno = EXDEV;
free(pv);
return -1;
}
}
hwloc_bitmap_only(nodeset, node);
free(pv);
return 0;
}
hwloc_bitmap_zero(nodeset);
for (i = 0; i < nb; i++)
hwloc_bitmap_set(nodeset, pv[i].VirtualAttributes.Node);
free(pv);
return 0;
}
}
static int
hwloc_aix_get_thisthread_last_cpu_location(hwloc_topology_t topology, hwloc_bitmap_t hwloc_set, int flags __hwloc_attribute_unused)
{
cpu_t cpu;
if (topology->pid) {
errno = ENOSYS;
return -1;
}
cpu = mycpu();
if (cpu < 0)
return -1;
hwloc_bitmap_only(hwloc_set, cpu);
return 0;
}
std::size_t thread_affinity_worker(std::size_t desired)
{
// Returns the OS-thread number of the worker that is running this
// PX-thread.
std::size_t current = hpx::get_worker_thread_num();
bool numa_sensitive = hpx::is_scheduler_numa_sensitive();
if (current == desired)
{
// extract the desired affinity mask
hpx::threads::topology const& t = hpx::get_runtime().get_topology();
hpx::threads::mask_type desired_mask = t.get_thread_affinity_mask(current, numa_sensitive);
#if defined(HPX_HAVE_HWLOC)
std::size_t idx = hpx::threads::find_first(desired_mask);
hwloc_topology_t topo;
hwloc_topology_init(&topo);
hwloc_topology_load(topo);
// retrieve the current affinity mask
hwloc_cpuset_t cpuset = hwloc_bitmap_alloc();
hwloc_bitmap_zero(cpuset);
if (0 == hwloc_get_cpubind(topo, cpuset, HWLOC_CPUBIND_THREAD)) {
// sadly get_cpubind is not implemented for Windows based systems
hwloc_cpuset_t cpuset_cmp = hwloc_bitmap_alloc();
hwloc_bitmap_zero(cpuset_cmp);
hwloc_bitmap_only(cpuset_cmp, unsigned(idx));
HPX_TEST(hwloc_bitmap_compare(cpuset, cpuset_cmp) == 0);
hwloc_bitmap_free(cpuset_cmp);
}
else
{
HPX_TEST(false && "hwloc_get_cpubind(topo, cpuset, HWLOC_CPUBIND_THREAD) failed!");
}
hwloc_bitmap_free(cpuset);
hwloc_topology_destroy(topo);
#endif
return desired;
}
// This PX-thread has been run by the wrong OS-thread, make the foreman
// try again by rescheduling it.
return std::size_t(-1);
}
void UseOnly(ui32 id) {
hwloc_bitmap_only(Set, id);
}
static int
hwloc_look_windows(struct hwloc_backend *backend)
{
struct hwloc_topology *topology = backend->topology;
hwloc_bitmap_t groups_pu_set = NULL;
SYSTEM_INFO SystemInfo;
DWORD length;
if (topology->levels[0][0]->cpuset)
/* somebody discovered things */
return -1;
hwloc_alloc_obj_cpusets(topology->levels[0][0]);
GetSystemInfo(&SystemInfo);
if (!GetLogicalProcessorInformationExProc && GetLogicalProcessorInformationProc) {
PSYSTEM_LOGICAL_PROCESSOR_INFORMATION procInfo, tmpprocInfo;
unsigned id;
unsigned i;
struct hwloc_obj *obj;
hwloc_obj_type_t type;
length = 0;
procInfo = NULL;
while (1) {
if (GetLogicalProcessorInformationProc(procInfo, &length))
break;
if (GetLastError() != ERROR_INSUFFICIENT_BUFFER)
return -1;
tmpprocInfo = realloc(procInfo, length);
if (!tmpprocInfo) {
free(procInfo);
goto out;
}
procInfo = tmpprocInfo;
}
assert(!length || procInfo);
for (i = 0; i < length / sizeof(*procInfo); i++) {
/* Ignore unknown caches */
if (procInfo->Relationship == RelationCache
&& procInfo->Cache.Type != CacheUnified
&& procInfo->Cache.Type != CacheData
&& procInfo->Cache.Type != CacheInstruction)
continue;
id = -1;
switch (procInfo[i].Relationship) {
case RelationNumaNode:
type = HWLOC_OBJ_NUMANODE;
id = procInfo[i].NumaNode.NodeNumber;
break;
case RelationProcessorPackage:
type = HWLOC_OBJ_PACKAGE;
break;
case RelationCache:
type = (procInfo[i].Cache.Type == CacheInstruction ? HWLOC_OBJ_L1ICACHE : HWLOC_OBJ_L1CACHE) + procInfo[i].Cache.Level - 1;
break;
case RelationProcessorCore:
type = HWLOC_OBJ_CORE;
break;
case RelationGroup:
default:
type = HWLOC_OBJ_GROUP;
break;
}
if (!hwloc_filter_check_keep_object_type(topology, type))
continue;
obj = hwloc_alloc_setup_object(topology, type, id);
obj->cpuset = hwloc_bitmap_alloc();
hwloc_debug("%s#%u mask %lx\n", hwloc_type_name(type), id, procInfo[i].ProcessorMask);
/* ProcessorMask is a ULONG_PTR */
hwloc_bitmap_set_ith_ULONG_PTR(obj->cpuset, 0, procInfo[i].ProcessorMask);
hwloc_debug_2args_bitmap("%s#%u bitmap %s\n", hwloc_type_name(type), id, obj->cpuset);
switch (type) {
case HWLOC_OBJ_NUMANODE:
{
ULONGLONG avail;
obj->nodeset = hwloc_bitmap_alloc();
hwloc_bitmap_set(obj->nodeset, id);
if ((GetNumaAvailableMemoryNodeExProc && GetNumaAvailableMemoryNodeExProc(id, &avail))
|| (GetNumaAvailableMemoryNodeProc && GetNumaAvailableMemoryNodeProc(id, &avail)))
obj->memory.local_memory = avail;
obj->memory.page_types_len = 2;
obj->memory.page_types = malloc(2 * sizeof(*obj->memory.page_types));
memset(obj->memory.page_types, 0, 2 * sizeof(*obj->memory.page_types));
obj->memory.page_types_len = 1;
obj->memory.page_types[0].size = SystemInfo.dwPageSize;
#if HAVE_DECL__SC_LARGE_PAGESIZE
obj->memory.page_types_len++;
obj->memory.page_types[1].size = sysconf(_SC_LARGE_PAGESIZE);
#endif
break;
}
case HWLOC_OBJ_L1CACHE:
case HWLOC_OBJ_L2CACHE:
case HWLOC_OBJ_L3CACHE:
case HWLOC_OBJ_L4CACHE:
case HWLOC_OBJ_L5CACHE:
case HWLOC_OBJ_L1ICACHE:
case HWLOC_OBJ_L2ICACHE:
case HWLOC_OBJ_L3ICACHE:
obj->attr->cache.size = procInfo[i].Cache.Size;
obj->attr->cache.associativity = procInfo[i].Cache.Associativity == CACHE_FULLY_ASSOCIATIVE ? -1 : procInfo[i].Cache.Associativity ;
obj->attr->cache.linesize = procInfo[i].Cache.LineSize;
obj->attr->cache.depth = procInfo[i].Cache.Level;
switch (procInfo->Cache.Type) {
case CacheUnified:
obj->attr->cache.type = HWLOC_OBJ_CACHE_UNIFIED;
break;
case CacheData:
obj->attr->cache.type = HWLOC_OBJ_CACHE_DATA;
break;
case CacheInstruction:
obj->attr->cache.type = HWLOC_OBJ_CACHE_INSTRUCTION;
break;
default:
hwloc_free_unlinked_object(obj);
continue;
}
break;
case HWLOC_OBJ_GROUP:
obj->attr->group.kind = procInfo[i].Relationship == RelationGroup ? HWLOC_GROUP_KIND_WINDOWS_PROCESSOR_GROUP : HWLOC_GROUP_KIND_WINDOWS_RELATIONSHIP_UNKNOWN;
break;
default:
break;
}
hwloc_insert_object_by_cpuset(topology, obj);
}
free(procInfo);
}
if (GetLogicalProcessorInformationExProc) {
PSYSTEM_LOGICAL_PROCESSOR_INFORMATION_EX procInfoTotal, tmpprocInfoTotal, procInfo;
unsigned id;
struct hwloc_obj *obj;
hwloc_obj_type_t type;
length = 0;
procInfoTotal = NULL;
while (1) {
if (GetLogicalProcessorInformationExProc(RelationAll, procInfoTotal, &length))
break;
if (GetLastError() != ERROR_INSUFFICIENT_BUFFER)
return -1;
tmpprocInfoTotal = realloc(procInfoTotal, length);
if (!tmpprocInfoTotal) {
free(procInfoTotal);
goto out;
}
procInfoTotal = tmpprocInfoTotal;
}
for (procInfo = procInfoTotal;
(void*) procInfo < (void*) ((uintptr_t) procInfoTotal + length);
procInfo = (void*) ((uintptr_t) procInfo + procInfo->Size)) {
unsigned num, i;
GROUP_AFFINITY *GroupMask;
/* Ignore unknown caches */
if (procInfo->Relationship == RelationCache
&& procInfo->Cache.Type != CacheUnified
&& procInfo->Cache.Type != CacheData
&& procInfo->Cache.Type != CacheInstruction)
continue;
id = -1;
switch (procInfo->Relationship) {
case RelationNumaNode:
type = HWLOC_OBJ_NUMANODE;
num = 1;
GroupMask = &procInfo->NumaNode.GroupMask;
id = procInfo->NumaNode.NodeNumber;
break;
case RelationProcessorPackage:
type = HWLOC_OBJ_PACKAGE;
num = procInfo->Processor.GroupCount;
GroupMask = procInfo->Processor.GroupMask;
break;
case RelationCache:
type = (procInfo->Cache.Type == CacheInstruction ? HWLOC_OBJ_L1ICACHE : HWLOC_OBJ_L1CACHE) + procInfo->Cache.Level - 1;
num = 1;
GroupMask = &procInfo->Cache.GroupMask;
break;
case RelationProcessorCore:
type = HWLOC_OBJ_CORE;
num = procInfo->Processor.GroupCount;
GroupMask = procInfo->Processor.GroupMask;
break;
case RelationGroup:
/* So strange an interface... */
for (id = 0; id < procInfo->Group.ActiveGroupCount; id++) {
KAFFINITY mask;
hwloc_bitmap_t set;
set = hwloc_bitmap_alloc();
mask = procInfo->Group.GroupInfo[id].ActiveProcessorMask;
hwloc_debug("group %u %d cpus mask %lx\n", id,
procInfo->Group.GroupInfo[id].ActiveProcessorCount, mask);
/* KAFFINITY is ULONG_PTR */
hwloc_bitmap_set_ith_ULONG_PTR(set, id, mask);
/* FIXME: what if running 32bits on a 64bits windows with 64-processor groups?
* ULONG_PTR is 32bits, so half the group is invisible?
* maybe scale id to id*8/sizeof(ULONG_PTR) so that groups are 64-PU aligned?
*/
hwloc_debug_2args_bitmap("group %u %d bitmap %s\n", id, procInfo->Group.GroupInfo[id].ActiveProcessorCount, set);
/* save the set of PUs so that we can create them at the end */
if (!groups_pu_set)
groups_pu_set = hwloc_bitmap_alloc();
hwloc_bitmap_or(groups_pu_set, groups_pu_set, set);
if (hwloc_filter_check_keep_object_type(topology, HWLOC_OBJ_GROUP)) {
obj = hwloc_alloc_setup_object(topology, HWLOC_OBJ_GROUP, id);
obj->cpuset = set;
obj->attr->group.kind = HWLOC_GROUP_KIND_WINDOWS_PROCESSOR_GROUP;
hwloc_insert_object_by_cpuset(topology, obj);
} else
hwloc_bitmap_free(set);
}
continue;
default:
/* Don't know how to get the mask. */
hwloc_debug("unknown relation %d\n", procInfo->Relationship);
continue;
}
if (!hwloc_filter_check_keep_object_type(topology, type))
continue;
obj = hwloc_alloc_setup_object(topology, type, id);
obj->cpuset = hwloc_bitmap_alloc();
for (i = 0; i < num; i++) {
hwloc_debug("%s#%u %d: mask %d:%lx\n", hwloc_type_name(type), id, i, GroupMask[i].Group, GroupMask[i].Mask);
/* GROUP_AFFINITY.Mask is KAFFINITY, which is ULONG_PTR */
hwloc_bitmap_set_ith_ULONG_PTR(obj->cpuset, GroupMask[i].Group, GroupMask[i].Mask);
/* FIXME: scale id to id*8/sizeof(ULONG_PTR) as above? */
}
hwloc_debug_2args_bitmap("%s#%u bitmap %s\n", hwloc_type_name(type), id, obj->cpuset);
switch (type) {
case HWLOC_OBJ_NUMANODE:
{
ULONGLONG avail;
obj->nodeset = hwloc_bitmap_alloc();
hwloc_bitmap_set(obj->nodeset, id);
if ((GetNumaAvailableMemoryNodeExProc && GetNumaAvailableMemoryNodeExProc(id, &avail))
|| (GetNumaAvailableMemoryNodeProc && GetNumaAvailableMemoryNodeProc(id, &avail)))
obj->memory.local_memory = avail;
obj->memory.page_types = malloc(2 * sizeof(*obj->memory.page_types));
memset(obj->memory.page_types, 0, 2 * sizeof(*obj->memory.page_types));
obj->memory.page_types_len = 1;
obj->memory.page_types[0].size = SystemInfo.dwPageSize;
#if HAVE_DECL__SC_LARGE_PAGESIZE
obj->memory.page_types_len++;
obj->memory.page_types[1].size = sysconf(_SC_LARGE_PAGESIZE);
#endif
break;
}
case HWLOC_OBJ_L1CACHE:
case HWLOC_OBJ_L2CACHE:
case HWLOC_OBJ_L3CACHE:
case HWLOC_OBJ_L4CACHE:
case HWLOC_OBJ_L5CACHE:
case HWLOC_OBJ_L1ICACHE:
case HWLOC_OBJ_L2ICACHE:
case HWLOC_OBJ_L3ICACHE:
obj->attr->cache.size = procInfo->Cache.CacheSize;
obj->attr->cache.associativity = procInfo->Cache.Associativity == CACHE_FULLY_ASSOCIATIVE ? -1 : procInfo->Cache.Associativity ;
obj->attr->cache.linesize = procInfo->Cache.LineSize;
obj->attr->cache.depth = procInfo->Cache.Level;
switch (procInfo->Cache.Type) {
case CacheUnified:
obj->attr->cache.type = HWLOC_OBJ_CACHE_UNIFIED;
break;
case CacheData:
obj->attr->cache.type = HWLOC_OBJ_CACHE_DATA;
break;
case CacheInstruction:
obj->attr->cache.type = HWLOC_OBJ_CACHE_INSTRUCTION;
break;
default:
hwloc_free_unlinked_object(obj);
continue;
}
break;
default:
break;
}
hwloc_insert_object_by_cpuset(topology, obj);
}
free(procInfoTotal);
}
if (groups_pu_set) {
/* the system supports multiple Groups.
* PU indexes may be discontiguous, especially if Groups contain less than 64 procs.
*/
hwloc_obj_t obj;
unsigned idx;
hwloc_bitmap_foreach_begin(idx, groups_pu_set) {
obj = hwloc_alloc_setup_object(topology, HWLOC_OBJ_PU, idx);
obj->cpuset = hwloc_bitmap_alloc();
hwloc_bitmap_only(obj->cpuset, idx);
hwloc_debug_1arg_bitmap("cpu %u has cpuset %s\n",
idx, obj->cpuset);
hwloc_insert_object_by_cpuset(topology, obj);
} hwloc_bitmap_foreach_end();
hwloc_bitmap_free(groups_pu_set);
} else {
