Media/compute-runtime
1
0
Fork 0
mirror of https://github.com/intel/compute-runtime.git synced 2025-09-15 13:01:45 +08:00
Code Issues Packages Projects Releases Wiki Activity

Detect GPU hangs in appendMemoryRangesBarrier()

Browse Source 
This change introduces checking the error code returned by
executeCommandListImmediate() inside appendMemoryRangesBarrier().
When GPU hang is encountered, the error code is propagated.
This change contains also ULTs to cover the new code and improve
coverage of that function.

Related-To: NEO-6681
Signed-off-by: Patryk Wrobel <patryk.wrobel@intel.com>
This commit is contained in:
Patryk Wrobel
2022-04-27 13:33:21 +00:00
committed by Compute-Runtime-Automation
parent 35a1a3fc2a
commit 96dd1de20c
2 changed files with 87 additions and 2 deletions
Download Patch File Download Diff File Expand all files Collapse all files

5
level_zero/core/source/cmdlist/cmdlist_hw.inl
View File

@ -371,7 +371,10 @@ ze_result_t CommandListCoreFamily<gfxCoreFamily>::appendMemoryRangesBarrier(uint
appendSignalEventPostWalker(hSignalEvent, workloadPartition);
if (this->cmdListType == CommandListType::TYPE_IMMEDIATE) {
executeCommandListImmediate(true);
const auto executionResult = executeCommandListImmediate(true);
if (executionResult == ZE_RESULT_ERROR_DEVICE_LOST) {
return ZE_RESULT_ERROR_DEVICE_LOST;
}
}
return ZE_RESULT_SUCCESS;

84
level_zero/core/test/unit_tests/gen12lp/test_cmdlist_gen12lp.cpp
View File

@ -1,5 +1,5 @@
/*
* Copyright (C) 2021 Intel Corporation
* Copyright (C) 2021-2022 Intel Corporation
*
* SPDX-License-Identifier: MIT
*
@ -29,6 +29,21 @@ struct CommandListAdjustStateComputeMode : public WhiteBox<::L0::CommandListProd
using ::L0::CommandListProductFamily<productFamily>::applyMemoryRangesBarrier;
};
template <PRODUCT_FAMILY productFamily>
class MockCommandListHw : public WhiteBox<::L0::CommandListProductFamily<productFamily>> {
public:
MockCommandListHw() : WhiteBox<::L0::CommandListProductFamily<productFamily>>(1) {}
using ::L0::CommandListProductFamily<productFamily>::applyMemoryRangesBarrier;
ze_result_t executeCommandListImmediate(bool performMigration) override {
++executeCommandListImmediateCalledCount;
return executeCommandListImmediateReturnValue;
}
ze_result_t executeCommandListImmediateReturnValue{};
int executeCommandListImmediateCalledCount{};
};
HWTEST2_F(CommandListCreate, givenAllocationsWhenApplyRangesBarrierThenCheckWhetherL3ControlIsProgrammed, IsGen12LP) {
using GfxFamily = typename NEO::GfxFamilyMapper<gfxCoreFamily>::GfxFamily;
using L3_CONTROL = typename GfxFamily::L3_CONTROL;
@ -58,6 +73,73 @@ HWTEST2_F(CommandListCreate, givenAllocationsWhenApplyRangesBarrierThenCheckWhet
}
}
HWTEST2_F(CommandListCreate, GivenNullptrWaitEventsArrayAndCountGreaterThanZeroWhenAppendingMemoryBarrierThenInvalidArgumentErrorIsReturned,
IsDG1) {
ze_result_t result;
uint32_t numRanges = 1;
const size_t pRangeSizes = 1;
const char *_pRanges[pRangeSizes];
const void **pRanges = reinterpret_cast<const void **>(&_pRanges[0]);
auto commandList = new CommandListAdjustStateComputeMode<productFamily>();
bool ret = commandList->initialize(device, NEO::EngineGroupType::RenderCompute, 0u);
ASSERT_FALSE(ret);
uint32_t numWaitEvents{1};
ze_event_handle_t *phWaitEvents{nullptr};
result = commandList->appendMemoryRangesBarrier(numRanges, &pRangeSizes,
pRanges, nullptr, numWaitEvents,
phWaitEvents);
EXPECT_EQ(ZE_RESULT_ERROR_INVALID_ARGUMENT, result);
commandList->destroy();
}
HWTEST2_F(CommandListCreate, GivenImmediateListAndExecutionSuccessWhenAppendingMemoryBarrierThenExecuteCommandListImmediateCalledAndSuccessReturned,
IsDG1) {
ze_result_t result;
uint32_t numRanges = 1;
const size_t pRangeSizes = 1;
const char *_pRanges[pRangeSizes];
const void **pRanges = reinterpret_cast<const void **>(&_pRanges[0]);
auto cmdList = new MockCommandListHw<productFamily>;
cmdList->cmdListType = CommandList::CommandListType::TYPE_IMMEDIATE;
cmdList->initialize(device, NEO::EngineGroupType::RenderCompute, 0u);
cmdList->executeCommandListImmediateReturnValue = ZE_RESULT_SUCCESS;
result = cmdList->appendMemoryRangesBarrier(numRanges, &pRangeSizes,
pRanges, nullptr, 0,
nullptr);
EXPECT_EQ(1, cmdList->executeCommandListImmediateCalledCount);
EXPECT_EQ(ZE_RESULT_SUCCESS, result);
cmdList->destroy();
}
HWTEST2_F(CommandListCreate, GivenImmediateListAndGpuFailureWhenAppendingMemoryBarrierThenExecuteCommandListImmediateCalledAndDeviceLostReturned,
IsDG1) {
ze_result_t result;
uint32_t numRanges = 1;
const size_t pRangeSizes = 1;
const char *_pRanges[pRangeSizes];
const void **pRanges = reinterpret_cast<const void **>(&_pRanges[0]);
auto cmdList = new MockCommandListHw<productFamily>;
cmdList->cmdListType = CommandList::CommandListType::TYPE_IMMEDIATE;
cmdList->initialize(device, NEO::EngineGroupType::RenderCompute, 0u);
cmdList->executeCommandListImmediateReturnValue = ZE_RESULT_ERROR_DEVICE_LOST;
result = cmdList->appendMemoryRangesBarrier(numRanges, &pRangeSizes,
pRanges, nullptr, 0,
nullptr);
EXPECT_EQ(1, cmdList->executeCommandListImmediateCalledCount);
EXPECT_EQ(ZE_RESULT_ERROR_DEVICE_LOST, result);
cmdList->destroy();
}
HWTEST2_F(CommandListCreate, GivenHostMemoryNotInSvmManagerWhenAppendingMemoryBarrierThenAdditionalCommandsNotAdded,
IsDG1) {
ze_result_t result;