Adds a bool to the GPU agent and a public member method to
check if the GPU supports large BAR. This is needed so we can
check if large BAR is supported when a user tries to allocate
an AQL queue in device memory on a given GPU agent.
Also adds an exception to the AQL queue if device-side AQL queues
are requested and the GPU owner of the AQL doesn't support large
BAR. Otherwise, ROCr will currently allow device-side queues
that can cause faults when the user tries to touch their ring
buffers and the user will not know why the faults are occuring.
This relies on the fact that the KFD does not exposed any links
from the CPU to the GPU if large BAR is not enabled (though
links from the GPU to the CPU may still be exposed by the KFD).
This builds on a prior change that allowed for allocating
a user-mode queue's packet buffer in device memory to also
allocate the queue struct in device memory. This provides
additional latency benefits particularly for cases where
dispatches are performed from the GPU itself. Flags are
added to support the various use cases.
This patch will adds recommended sdma supports with
limited XGMI SDMA engine. It will use one PCIe SDMA
to do gpu <-> gpu copies which will help improve all
to all copy performance.
Signed-off-by: Shane Xiao <shane.xiao@amd.com>
Adding a general stage for agents to release their resources on
shutdown. This avoids a circular dependency during shutdown because
we have to delete allocated resources before deleting memory pools, but
we also have to delete memory pools before destroying agents.
The over arching goal it so provide an API that pre-silicon models can latch into for software bring up.# Please enter the commit message for your changes. Lines starting
Use the core Driver object in the CPU agent to make it OS/driver
agnostic.
Implement the GetMemoryProperties() and GetCacheProperties methods
for the KFD driver.
Updating ROCr code to match new handshake protocol with CP FW for
asynchronous scratch reclaim.
Increase previous limits when scratch reclaim feature is available.
- When waiting on non-interrupt signals, do not uSleep. This causes
regressions compared to interrupt signal usage.
- Cleanup code.
Change-Id: I706bda0b13e64ffec0b607c1915d8380a2ce0dea
Replaces WaitAny with WaitMultiple to more closely align with the
underlying driver API for waiting on multiple events.
WaitMultiple adds a single parameter, wait_on_all, to the WaitAny
interface providing a single function for waiting on multiple
events when we only need AND and OR semantics for the signal
checking logic.
Change-Id: I68a4a45d48151d9d69aef02fd8f7263b9e6c0e75
Generalize the driver discovery and move driver-specific
functionality to the concrete driver implementations.
Currently, this process is tightly coupled to the hsakmt
which is GPU and OS specific.
Change-Id: Ie1c53fef407a71b5ec4c6eaf3a3ed00871184409
Minor modifications to multiple source and header
files based on Coverity report
Change-Id: I4a73d0f56640983c4d5124e13c8c280245cca672
Signed-off-by: Apurv Mishra <apurv.mishra@amd.com>
Add return checks, initialization and clean
redundant memory operations
fix 1: check return value of 'setsockopt' for error
fix 2: check return value of 'PtrInfo' for error
fix 3: move 'tool_names' instead of copying
fix 4: call 'munmap' for 'va' only once
fix 5: use 'ssize' for possible return values of -1 (err)
fix 6: add missing initialization in constructors
fix 7: add initialization for some scalars and pointers
Change-Id: I07d90e36d4e1fe48c4de4f44e18083e5ed4c5fbc
Signed-off-by: Apurv Mishra <apurv.mishra@amd.com>
To allow non-POD global variables to last until the last thread
has exited, use "new" to allocate the memory instead of static
allocation.
Change-Id: Ica571b61ff8068a52e472c49cb1c44917e60c8c8
An ASAN run of the release build revealed some elements of
the supported_isas static map were still using stack data. This
change makes it use heap data so it will persist.
Change-Id: Ie51887e88b9e2dec27acfc97ea45a6219fea971c
- Add the new path to avoid WaitAny() calls in AsyncEventsLoopp() with
HSA_WAIT_ANY_DEBUG key. The new path is selected by default.
The optimizaiton combines all logic of WaitAny() in a single processing loop
and avoids extra memory allocations or ref counting. Also it won't spin
on the CPU if all events are busy.
Change-Id: I197ce60d0d023fbb672f700d6e87702686f1f55a
On GPUs where EOP is handled in asic, the read_dispatch_id is not always
updated after each packet. Look for the first dispatch packet that needs
scratch memory before allocating scratch.
Change-Id: Ibf4b4b485f99bf2fabfe48e9609ca99111fdafbe
The supported_isas static unordered_map was adding stack
allocated Isa objects. Instead, make the objects statically
allocated, as supported_isas itself is.
Change-Id: I23405e218290d48deea6f984f76c57e7b43e314e
When ROCr is built as a static library, global variables
were often not initialized to valid values at their first
use. This change addresses that problem.
Change-Id: I550fa41feb3bc04b9cc686bcfb4acf2a7b651a88
Socket server accept calls do not guarantee synchronous actions
post-accept. This can result in a race condition.
To resolve this, first limit the socket server's listen backlog to a
single connection. This will force competing clients to busy-retry
until timeout.
Second, make the DMABUF IPC file descriptor send-receive and import
calls into an atomic routine per connection.
By doing these fixes, not only to we resolve potential races but
we guarantee that any exporter process will create at most one
file descriptor that will only last for the duration of the import
transaction. This alleviates any concern on running into system
limits for the number of open file descriptors per process.
Change-Id: I6d8b14795a680d89a2707e082fa027d525792e05
Discarding blocks for reallocation on IPC export for better memory
performance trigger memory violations with DMA BUF exports so bypass
this for now as application performance drops haven't been observed
with the bypass.
The raw fragment should be passed to the DMA Buf export call as well
since offsets will be implicitly applied in the Thunk/KFD for
export/import calls.
Also, use the agent information directly from the pointer
information so that the export call doesn't have to scan memory to find
this. Pass the node ID in the handle so that the import call doesn't
have to make two thunk imports to fetch the node ID for GPU memory
imports.
Finally, allow the user to use DMA Buf IPC via
HSA_ENABLE_IPC_MODE_LEGACY=0 for developer testing as legacy mode will
be applied by default.
Change-Id: Ie8fe267f8768fa5df37126078406f7065f69ff4e
Adds support for AllocateMemoryOnly inside XDNA driver.
Move the IsLocalMemory() check inside the KFD driver
since the XDNA driver can, and needs to, create handles
on system memory buffer objects.
Changed handle variable name from thunk_handle to user_mode_driver_handle,
which is more representative if we support non-GPU drivers.
Change-Id: I95db9d575afd1ab0ff2de74cea5175d9a12a721b
Adds support for initialzing the XDNA driver so that
a hardware context can be created for an AIE queue.
Right now this initializes the device heap in the driver,
gets the relevant tile parameters for the AIE agent,
and creates a hardware context that backs the AIE queue.
Change-Id: Ib90e1bc67a8637f6db3ff2bebe34677843796417
- Use HSA_ALLOCATE_QUEUE_DEV_MEM=1 to create AQL queue in device
memory.
- Before writing AQL packet header to the queue use an SFENCE to ensure
that there is no reodering of the writes over PCIE
Change-Id: I5eacdc35108c4a1e245c75ae349b7495451aa60d
Remove KFD-specific Allocate/Free calls from the AMD::MemoryRegion.
The KFD-driver-specific Allocate/Free calls are now implemented in
the KfdDriver. Future changes will migrate the remaining KFD-specific
calls out of AMD::MemoryRegion.
This allows the MemoryRegion to be used across AMD drivers like the
XDNA driver.
Change-Id: Ib6a2a9e5e1a15e61644d2592beb3a8e6578c3010
Tony Gutierrez
Saleel Kudchadker
Shane Xiao
Yiannis Papadopoulos
David Yat Sin
jordans
Khatri, Shweta
Ben Vanik
Apurv Mishra
Eddie Richter
Chris Freehill
German Andryeyev
Jonathan Kim