Add experimental extension to set global work offest in L0.
Current L0 specification does not have interface to export
experimental function symbols, so for now, applications need
to find the symbol like with dlsym on Linux.
A blackbox test showing functionality is also added.
Signed-off-by: Jaime Arteaga <jaime.a.arteaga.molina@intel.com>
Add support for device and shared allocations that use the
ZE_DEVICE_MEM_ALLOC_FLAG_BIAS_UNCACHED flag, whether the
kernel using the memory is stateless or statefull.
Signed-off-by: Jaime Arteaga <jaime.a.arteaga.molina@intel.com>
Update usage of SUPPORTED_IMAGES flag and do not use images when disabled.
Use SUPPORTED_2_0 only on fully OCL 2.1 conformant platforms.
Signed-off-by: Filip Hazubski <filip.hazubski@intel.com>
Add ZE_experimental_module_program extension to L0 driver to support
multiple input modules being passed to zeModuleCreate().
Signed-off-by: Jaime Arteaga <jaime.a.arteaga.molina@intel.com>
Add support for passing ze_external_memory_import_fd_t extended
structure to zeMemAllocDevice() to allocate a device memory
out of an fd handle.
Signed-off-by: Jaime Arteaga <jaime.a.arteaga.molina@intel.com>
Add support for passing ze_external_memory_export_fd_t extended
structure to zeMemGetAllocProperties() to retrieve an fd handle
for a device allocation.
Signed-off-by: Jaime Arteaga <jaime.a.arteaga.molina@intel.com>
Add support for reading ze_external_memory_export_desc_t passed
to calls to zeMemAllocDevice. Current driver implementation
only supports handles as dma-buf, so add validation that only
that flag is being used.
Signed-off-by: Jaime Arteaga <jaime.a.arteaga.molina@intel.com>
appendMemoryCopy has wait events since v1.0, so add the corresponding
support. Other copy operations were missing passing this
in the entry points, but already had support internally.
Also, rename some variables for consistency.
Also, remove signal event from appendMemoryCopyBlit since it is not
used.
Signed-off-by: Jaime Arteaga <jaime.a.arteaga.molina@intel.com>
- Fixed Memory Fill to use aligned allocations in Destination and Source
buffer kernel arguments to ensure proper alignment of memory in the
surface state.
- Fixed patternSize == 1 case where simd size is > the size being
written resulting in the inital kernel append with group count of 0 and
the remainder kernel being executed incorrectly.
- Fixed offset calculation in the remainder kernel to be based off of
the aligned allocation offset.
- Fixed memory overrun case where size < patternsize with the minimum
size being used in the kernel.
Change-Id: I39bd83b3a83ceb6df05d374238f85f7fdf0bd09a
Signed-off-by: Spruit, Neil R <neil.r.spruit@intel.com>
Move location of assigning sysman handles, such that sysman handle
will be assigned to device and then assign that sysman handle
to subdevices of that device.
Change-Id: Ia606954a866c491e45803b903c8419ee09b1a939
Signed-off-by: Jitendra Sharma <jitendra.sharma@intel.com>
