Shared Library Constructor (rocprofv3 deadlock fix) (#599)
* Moved tests/apps to tests/bin * Renamed cmake project in tests/bin * Update samples - Use ROCPROFILER_DEFAULT_FAIL_REGEX - tweaks to stdout messages * Update tests - Use ROCPROFILER_DEFAULT_FAIL_REGEX * Add tests/lib - libraries with HIP code * Update PTL submodule - remove atexit delete of thread_id_map * Update cmake/rocprofiler_options.cmake - Set ROCPROFILER_DEFAULT_FAIL_REGEX * Update common lib: env + logging - improved customization of logging settings - default to disabling logging to files - install failure handler for rocprofv3 - set_env support in environment.* * Add lib/rocprofiler-sdk/shared_library.cpp - shared library constructor * Update lib/rocprofiler-sdk-tool/tool.cpp - destructor thread safety - convert callback_name_info and buffered_name_info to pointers - install failure handler for logging * Add tests/bin/hip-in-libraries - hip-in-libraries is an exe which uses two shared libraries where each shared library contains HIP kernels - used for testing deadlocking within __hipRegisterFatBinary * Update bin/rocprofv3 - reorganized the env variables - use exec to launch command - set ROCPROFILER_LIBRARY_CTOR=1 * Add tests/rocprofv3/tracing-hip-in-libraries - uses hip-in-libraries exe for exe which uses shared libraries to launch HIP kernels * Update bin/rocprofv3 - fix counter collection (no exec) * Update lib/rocprofiler-sdk-tool/tool.cpp - replace "Kernel-Name" with "Kernel_Name" * Update lib/rocprofiler-sdk/registration.cpp Use RTLD_LOCAL instead of RTLD_GLOBAL for env libraries * Update tests/rocprofv3 - replace "Kernel-Name" with "Kernel_Name" * Update tests - vector-ops (bin) stream syncs + runs with 4 queues per device - improve counter-collection/input1 validation - rocprofv3/tracing-hip-in-libraries does not do sys-trace - improved validation script for tracing-hip-in-libraries - updated dispatch_callback in json-tool.cpp following reworking of prototypes for counter collection * Update samples/counter_collection - updated dispatch_callback(s) and record_callback(s) following reworking of prototypes * Update bin/rocprofv3 - reorganized help menu - added options for sub-HSA tables - added --hip-runtime-trace - changed --hip-trace to include --hip-compiler-trace * Update lib/rocprofiler-sdk-tool - improved kernel filtering - removed arch_vgpr, accum_vgpr, sgpr code (in rocprofiler-sdk) - fixed issue with counter-collection w/o tracing - added support for fine grained HSA API tracing - removed directly linking to HSA-runtime * Update lib/rocprofiler-sdk/agent.cpp - rocp_agents != hsa_agents is non-fatal when ROCPROFILER_BUILD_CI=OFF (CMake option) * GPR (vector and scalar) info in kernel symbol data - rocprofiler_callback_tracing_code_object_kernel_symbol_register_data_t contains general purpose register info * Header include order fix - Include repo headers first - Third party library headers next - standard library headers last * Update dispatch profiling public API - introduce rocprofiler_profile_counting_dispatch_data_t - change signature of rocprofiler_profile_counting_dispatch_callback_t and rocprofiler_profile_counting_record_callback_t - provide rocprofiler_user_data_t pointer in dispatch callback - provide rocprofiler_user_data_t value (from dispatch cb) in record callback * Update tests/bin/CMakeLists.txt - fix add_subdirectory(hip-in-libraries) order * Update VERSION - bump to 0.2.0 in prep for AFAR
이 커밋은 다음에 포함됨:
Jonathan R. Madsen
GitHub
@@ -24,6 +24,7 @@
|
||||
|
||||
#include "lib/common/container/small_vector.hpp"
|
||||
#include "lib/common/synchronized.hpp"
|
||||
#include "lib/common/utility.hpp"
|
||||
#include "lib/rocprofiler-sdk/agent.hpp"
|
||||
#include "lib/rocprofiler-sdk/aql/helpers.hpp"
|
||||
#include "lib/rocprofiler-sdk/aql/packet_construct.hpp"
|
||||
@@ -32,6 +33,7 @@
|
||||
#include "lib/rocprofiler-sdk/hsa/queue_controller.hpp"
|
||||
#include "lib/rocprofiler-sdk/registration.hpp"
|
||||
|
||||
#include <rocprofiler-sdk/fwd.h>
|
||||
#include <rocprofiler-sdk/rocprofiler.h>
|
||||
|
||||
namespace rocprofiler
|
||||
@@ -255,6 +257,7 @@ queue_cb(const std::shared_ptr<counter_callback_info>& info,
|
||||
const hsa::Queue& queue,
|
||||
const hsa::rocprofiler_packet& pkt,
|
||||
uint64_t kernel_id,
|
||||
rocprofiler_user_data_t* user_data,
|
||||
const hsa::Queue::queue_info_session_t::external_corr_id_map_t& extern_corr_ids,
|
||||
const context::correlation_id* correlation_id)
|
||||
{
|
||||
@@ -272,14 +275,25 @@ queue_cb(const std::shared_ptr<counter_callback_info>& info,
|
||||
}
|
||||
}
|
||||
|
||||
rocprofiler_profile_config_id_t req_profile = {.handle = 0};
|
||||
info->user_cb(queue.get_id(),
|
||||
queue.get_agent().get_rocp_agent(),
|
||||
_corr_id_v,
|
||||
&pkt.kernel_dispatch,
|
||||
kernel_id,
|
||||
info->callback_args,
|
||||
&req_profile);
|
||||
auto req_profile = rocprofiler_profile_config_id_t{.handle = 0};
|
||||
auto dispatch_data =
|
||||
common::init_public_api_struct(rocprofiler_profile_counting_dispatch_data_t{});
|
||||
|
||||
dispatch_data.kernel_id = kernel_id;
|
||||
dispatch_data.agent_id = CHECK_NOTNULL(queue.get_agent().get_rocp_agent())->id;
|
||||
dispatch_data.queue_id = queue.get_id();
|
||||
dispatch_data.correlation_id = _corr_id_v;
|
||||
dispatch_data.private_segment_size = pkt.kernel_dispatch.private_segment_size;
|
||||
dispatch_data.group_segment_size = pkt.kernel_dispatch.group_segment_size;
|
||||
dispatch_data.workgroup_size = {pkt.kernel_dispatch.workgroup_size_x,
|
||||
pkt.kernel_dispatch.workgroup_size_y,
|
||||
pkt.kernel_dispatch.workgroup_size_z};
|
||||
dispatch_data.grid_size = {pkt.kernel_dispatch.grid_size_x,
|
||||
pkt.kernel_dispatch.grid_size_y,
|
||||
pkt.kernel_dispatch.grid_size_z};
|
||||
|
||||
info->user_cb(dispatch_data, &req_profile, user_data, info->callback_args);
|
||||
|
||||
if(req_profile.handle == 0) return nullptr;
|
||||
|
||||
auto prof_config = get_controller().get_profile_cfg(req_profile);
|
||||
@@ -407,13 +421,27 @@ completed_cb(const std::shared_ptr<counter_callback_info>& info,
|
||||
if(!out.empty())
|
||||
{
|
||||
CHECK(info->record_callback);
|
||||
info->record_callback(queue.get_id(),
|
||||
queue.get_agent().get_rocp_agent()->id,
|
||||
_corr_id_v,
|
||||
session.kernel_id,
|
||||
info->record_callback_args,
|
||||
out.size(),
|
||||
out.data());
|
||||
|
||||
auto dispatch_data =
|
||||
common::init_public_api_struct(rocprofiler_profile_counting_dispatch_data_t{});
|
||||
|
||||
const auto& kernel_dispatch_pkt = session.kernel_pkt.kernel_dispatch;
|
||||
|
||||
dispatch_data.kernel_id = session.kernel_id;
|
||||
dispatch_data.agent_id = CHECK_NOTNULL(queue.get_agent().get_rocp_agent())->id;
|
||||
dispatch_data.queue_id = queue.get_id();
|
||||
dispatch_data.correlation_id = _corr_id_v;
|
||||
dispatch_data.private_segment_size = kernel_dispatch_pkt.private_segment_size;
|
||||
dispatch_data.group_segment_size = kernel_dispatch_pkt.group_segment_size;
|
||||
dispatch_data.workgroup_size = {kernel_dispatch_pkt.workgroup_size_x,
|
||||
kernel_dispatch_pkt.workgroup_size_y,
|
||||
kernel_dispatch_pkt.workgroup_size_z};
|
||||
dispatch_data.grid_size = {kernel_dispatch_pkt.grid_size_x,
|
||||
kernel_dispatch_pkt.grid_size_y,
|
||||
kernel_dispatch_pkt.grid_size_z};
|
||||
|
||||
info->record_callback(
|
||||
dispatch_data, out.data(), out.size(), session.user_data, info->record_callback_args);
|
||||
}
|
||||
}
|
||||
|
||||
@@ -436,9 +464,11 @@ start_context(const context::context* ctx)
|
||||
[=](const hsa::Queue& q,
|
||||
const hsa::rocprofiler_packet& kern_pkt,
|
||||
uint64_t kernel_id,
|
||||
rocprofiler_user_data_t* user_data,
|
||||
const hsa::Queue::queue_info_session_t::external_corr_id_map_t& extern_corr_ids,
|
||||
const context::correlation_id* correlation_id) {
|
||||
return queue_cb(cb, q, kern_pkt, kernel_id, extern_corr_ids, correlation_id);
|
||||
return queue_cb(
|
||||
cb, q, kern_pkt, kernel_id, user_data, extern_corr_ids, correlation_id);
|
||||
},
|
||||
// Completion CB
|
||||
[=](const hsa::Queue& q,
|
||||
|
||||
새 이슈에서 참조
사용자 차단