AI/rocm-systems
2
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

SWDEV-257787 - Reset active signal if ROCR call failed

Browse Source 
- ROCR fails the call for some reason, then the signal will
become invalid and can hang on a wait. The logic will reset the
active signal in such cases

Change-Id: Ia131420200f1bbd7c9a162b8f1b06db8cecf41c6
This commit is contained in:
German Andryeyev
2021-01-21 17:29:34 -05:00
parent de3e8b092c
commit ce2e5eba6b
3 changed files with 18 additions and 0 deletions
Download Patch File Download Diff File Expand all files Collapse all files
rocclr/device/rocm/rocblit.cpp
+4
View File
@@ -450,6 +450,7 @@ bool DmaBlitManager::copyBufferRect(device::Memory& srcMemory, device::Memory& d
hsa_status_t status = hsa_amd_memory_async_copy_rect(&dstMem, &offset,
&srcMem, &offset, &dim, agent, direction, num_wait_events, wait_event, active);
if (status != HSA_STATUS_SUCCESS) {
gpu().Barriers().ResetCurrentSignal();
LogPrintfError("DMA buffer failed with code %d", status);
return false;
}
@@ -473,6 +474,7 @@ bool DmaBlitManager::copyBufferRect(device::Memory& srcMemory, device::Memory& d
size[0], num_wait_events, wait_event, active);
gpu().setLastCommandSDMA(true) ;
if (status != HSA_STATUS_SUCCESS) {
gpu().Barriers().ResetCurrentSignal();
LogPrintfError("DMA buffer failed with code %d", status);
return false;
}
@@ -656,6 +658,7 @@ bool DmaBlitManager::hsaCopy(const Memory& srcMemory, const Memory& dstMemory,
if (status == HSA_STATUS_SUCCESS) {
gpu().addSystemScope();
} else {
gpu().Barriers().ResetCurrentSignal();
LogPrintfError("Hsa copy from host to device failed with code %d", status);
}
@@ -702,6 +705,7 @@ bool DmaBlitManager::hsaCopyStaged(const_address hostSrc, address hostDst, size_
srcAgent, size, 0, nullptr, active);
gpu().setLastCommandSDMA(true);
if (status != HSA_STATUS_SUCCESS) {
gpu().Barriers().ResetCurrentSignal();
LogPrintfError("Hsa copy from host to device failed with code %d", status);
return false;
}