SWDEV-235579
Move the lock before destroying the queue as there's a multithreaded race condition if the queue
is being destroy and right after we set queue_ to nullptr, another thread can call ihipWaitStreams
which will then call create on that same stream because queue is now nullptr.
Moving the lock on streamSet prevents this from happening because we would remove the stream from that
list and therefore ihipWait will not try to call asHostQueue which tries to create the queue if not created yet
since the stream won't be in the list anymore
Change-Id: I3108657ab403d39d4123e83294fcf1f0880e5563
The purpose of this KFDTest is to investigate the behaviour of an
SDMA queue when an invalid memory address is used.
v2: Don't wait for SDMA queue to finish - it won't finish because of
the gpuvm fault.
v3: Create kfd event before SDMA queue submission. This fix the issue
that gpuvm fault happens earlier than kfd event is created then KFD
exception handler can't find the kfd event (to wake up kfd test)
v4: Instead of using 0x12345678 as the invalid VA, map one page of
FB to gpu and unmap it. Use the mapped GPUVA as the
invalid address
Change-Id: I58af1511f75d869adddede302b238c2725f3fe5a
Signed-off-by: Ori Messinger <Ori.Messinger@amd.com>
Signed-off-by: Oak Zeng <Oak.Zeng@amd.com>
The purpose of this KFDTest is to investigate the behaviour of an
SDMA queue when an invalid memory address is used.
v2: Don't wait for SDMA queue to finish - it won't finish because of
the gpuvm fault.
v3: Create kfd event before SDMA queue submission. This fix the issue
that gpuvm fault happens earlier than kfd event is created then KFD
exception handler can't find the kfd event (to wake up kfd test)
v4: Instead of using 0x12345678 as the invalid VA, map one page of
FB to gpu and unmap it. Use the mapped GPUVA as the
invalid address
Change-Id: I58af1511f75d869adddede302b238c2725f3fe5a
Signed-off-by: Ori Messinger <Ori.Messinger@amd.com>
Signed-off-by: Oak Zeng <Oak.Zeng@amd.com>
[ROCm/ROCR-Runtime commit: 46f5e83066]
Peer accessibility query was not previously directed at the peer agent.
Change-Id: I259f0afac827a6e4778a56419a3acd296d00391b
[ROCm/ROCR-Runtime commit: 9c9064c2b7]
These APIs launche RCCL kernel implementation by default. If environmental
variable RCCL_ALLTOALL_KERNEL_DISABLE=1, then the APIs use wrapper around
ncclSend and ncclRecv.
[ROCm/rccl-tests commit: 2813c96826]
These APIs launche RCCL kernel implementation by default. If environmental
variable RCCL_ALLTOALL_KERNEL_DISABLE=1, then the APIs use wrapper around
ncclSend and ncclRecv.
Move libOpenCL.so from rocm-opencl-dev package to rocm-opencl.
Having the libOpenCL.so alone and that too just a symlink pointing
to a file in another package is not the right way.
Adding a symlink for libOpenCL... to rocm/lib will make package
deployment easier.
Change-Id: Icc98d2280253c8a485ef906e97fc75129ad6d064
[ROCm/clr commit: e43d427cdd]
Move libOpenCL.so from rocm-opencl-dev package to rocm-opencl.
Having the libOpenCL.so alone and that too just a symlink pointing
to a file in another package is not the right way.
Adding a symlink for libOpenCL... to rocm/lib will make package
deployment easier.
Change-Id: Icc98d2280253c8a485ef906e97fc75129ad6d064
If HIP_VISIBLE_DEVICES is active, then make sure the list of HSA
agents contains the valid agents
Change-Id: I584aad999a230ab7f88a0cfe20dcd0abe79c43a5
[ROCm/clr commit: ae4aceb55e]
This technique should never be used, and only accessed through
__builtins.
There's currently no builtin for groupstaticsize. I left ds_swizzle
since for some reason it switches to the builtin based on __HCC__ or
not.
Change-Id: If1e1394221dba83ea4add6db5e94d6b715552044
[ROCm/clr commit: 03ebfd2d49]
This technique should never be used, and only accessed through
__builtins.
There's currently no builtin for groupstaticsize. I left ds_swizzle
since for some reason it switches to the builtin based on __HCC__ or
not.
Change-Id: If1e1394221dba83ea4add6db5e94d6b715552044
This technique should never be used, and only accessed through
__builtins.
There's currently no builtin for groupstaticsize. I left ds_swizzle
since for some reason it switches to the builtin based on __HCC__ or
not.
Change-Id: If1e1394221dba83ea4add6db5e94d6b715552044
[ROCm/hip commit: d2dd307c7d]
This technique should never be used, and only accessed through
__builtins.
There's currently no builtin for groupstaticsize. I left ds_swizzle
since for some reason it switches to the builtin based on __HCC__ or
not.
Change-Id: If1e1394221dba83ea4add6db5e94d6b715552044