Added if conditions to enable useChunking flag by checking
with ExecutionEnvironment::isDebuggingEnabled.
Related-To: NEO-8164
Signed-off-by: Young Jin Yoon <young.jin.yoon@intel.com>
The previous name "pageSize2Mb" defined in
shared/source/helpers/constant.h is inconsistent to other variable,
i.e. pageSize64k.
Furthermore, it's a bit misleading because the page size is defined in
Megabytes (MB), not in Megabits (Mb).
Related-to: NEO-7695
Signed-off-by: Young Jin Yoon <young.jin.yoon@intel.com>
Related-to: NEO-7695
New debug keys added:
EnableBOChunking is now a mask
0 = no chunking (default).
1 = shared allocations only
2 = device allocations only
3 = shared and device allocations
MinimalAllocationSizeForChunking sets the minimum allocation
size to apply chunking. Default is 2MB.
Signed-off-by: Jaime Arteaga <jaime.a.arteaga.molina@intel.com>
When upstream ioctl helper is created it will try to create small
allocation, adding I915_GEM_CREATE_EXT_SET_PAT extension. If it
succeeds, for all resources with valid pat index value it will then
explicitly program pat index value with gem create ext call.
PrintBOCreateDestroyResult value can be used to:
- print whether the set pat extension is supported by the kernel, when
ioctl helper is created
- print whether set pat extension was added for a given gem create ext
call and what pat index value was programmed
Resolves: NEO-7896
Signed-off-by: Filip Hazubski <filip.hazubski@intel.com>
- Abstracts product helpers logic for uuid
- Add UUID retrieval for Linux for Sysman via zesInit path
Related-To: LOCI-4137
Signed-off-by: Aravind Gopalakrishnan <aravind.gopalakrishnan@intel.com>
The Memory Info object is used in the getState function for memory.
Some of the ULTS in the memory modules has been modified.
A function to return the sysfs nodes for the Memory address range has
been added in the IoctlHelper class corresponding to the XE and i915
driver.
Related-To: LOCI-4397
Signed-off-by: Bari, Pratik <pratik.bari@intel.com>
Set no preference for KMD migrated shared allocation and rely on ACC to migrate.
Related-To: NEO-6839
Signed-off-by: Milczarek, Slawomir <slawomir.milczarek@intel.com>
- by default ZE_ENABLE_PCI_ID_DEVICE_ORDER is disabled
- by default devices are sorted by type (discrete first), then by pci order
- when ZE_ENABLE_PCI_ID_DEVICE_ORDER is enabled, devices are sorted by pci id
Related-To: LOCI-4520
Signed-off-by: Mateusz Jablonski <mateusz.jablonski@intel.com>
Use flushStamp=taskCount when passed flushStamp==0.
This will cause driver to busy wait for a short while before falling
back to use kmd notify.
Related-To: GSD-3612
Signed-off-by: Dominik Dabek <dominik.dabek@intel.com>
remove method to removing duplicates from StackVec as the method
implicitly sorted the vector
Related-To: GSD-4692
Signed-off-by: Mateusz Jablonski <mateusz.jablonski@intel.com>
When receiving EAGAIN, UMD should retry the submission again.
On scenarios where KMD runs out of ring-space, an error
should be propagated up to the user. Those scenarios are
more rare, but still we will need KMD support in the future
to differentiate those two cases: retry and propagate error to
user, since right now, KMD returns the same error code.
Related-To: NEO-8040
Signed-off-by: Jaime Arteaga <jaime.a.arteaga.molina@intel.com>
Have makeResident return error to the caller, instead of always
SUCCESS. This will allow interfaces like zeContextMakeMemoryResident
to fail properly.
Additionally, change the parsing of MemoryOperationsStatus from
ZE_RESULT_ERROR_OUT_OF_HOST_MEMORY to
ZE_RESULT_ERROR_OUT_OF_DEVICE_MEMORY, since when making resources
resident, it is the device running out of memory, instead of the
host.
Related-To: LOCI-4443
Signed-off-by: Jaime Arteaga <jaime.a.arteaga.molina@intel.com>
Related-To: LOCI-4176
- Given a Base Pointer passed into Get Peer Allocation, then the base
pointer is used in the map of the new allocation to the virtual memory.
- Enables users to use the same pointer for all devices in Peer To Peer.
- Currently unsupported on reserved memory due to mapped and exec
resiedency of Virtual addresses.
Signed-off-by: Neil R Spruit <neil.r.spruit@intel.com>
- vm_bind with user fence updates fence value independently for every
VM hence with per-context VMs, every context needs its unique fence
address. This prevents 2 contexts from updating value possibly
writing lower value than the one that was already stored
Resolves: NEO-8004
Signed-off-by: Mateusz Hoppe <mateusz.hoppe@intel.com>
Get ipVersion from productHelper function on xe and upstream.
On prelim first try to query ipVersion from kmd,
if it fails, get ipVersion from productHelper function.
Related-To: NEO-7786
Signed-off-by: Kamil Kopryk <kamil.kopryk@intel.com>
This fixes illegal memory accesses by the job submitted to the GuC.
Also some unit tests are added to harness the vmBind operation.
Related-To: NEO-7996
Signed-off-by: Francois Dugast <francois.dugast@intel.com>
Signed-off-by: Mateusz Jablonski <mateusz.jablonski@intel.com>
populate memory info based on mem usage and gts info
propagate error from xeWaitUserFence function
Related-To: NEO-7931
Co-authored-by: Francois Dugast <francois.dugast@intel.com>
Signed-off-by: Mateusz Jablonski <mateusz.jablonski@intel.com>
Allows the user to use alignments > 64KB in `createUnifiedMemoryAllocation`
So that the restriction in `piextUSMDeviceAlloc` of the DPC++ runtime
could be lifted
Related-To: LOCI-4168
Signed-off-by: Lu, Wenbin <wenbin.lu@intel.com>
