There are a couple of fixes required to enable gfx11 debugging.
First, ADD_QUEUE.trap_en is an inappropriate place to toggle
a per-process register so move it to SET_SHADER_DEBUGGER.trap_en.
When ADD_QUEUE.skip_process_ctx_clear is set, MES will prioritize
the SET_SHADER_DEBUGGER.trap_en setting.
Second, to preserve correct save/restore priviledged wave states
in coordination with the trap enablement setting, resume suspended
waves early in the disable call.
Signed-off-by: Jonathan Kim <jonathan.kim@amd.com>
Reviewed-by: Felix Kuehling <felix.kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
This patch fixes a regression caused by a bad merge where
the handling of reserved SDMA queues was accidentally removed.
With the fix, the reserved SDMA queues are again correctly
marked as unavailable for allocation.
Fixes: a805889a15 ("drm/amdkfd: Update SDMA queue management for GFX9.4.3")
Signed-off-by: Mukul Joshi <mukul.joshi@amd.com>
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Allow the debugger to get a snapshot of a specified number of queues
containing various queue property information that is copied to the
debugger.
Since the debugger doesn't know how many queues exist at any given time,
allow the debugger to pass the requested number of snapshots as 0 to get
the actual number of potential snapshots to use for a subsequent snapshot
request for actual information.
To prevent future ABI breakage, pass in the requested entry_size.
The KFD will return it's own entry_size in case the debugger still wants
log the information in a core dump on sizing failure.
Also allow the debugger to clear exceptions when doing a snapshot.
Signed-off-by: Jonathan Kim <jonathan.kim@amd.com>
Reviewed-by: Felix Kuehling <felix.kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
In order to inspect waves from the saved context at any point during a
debug session, the debugger must be able to preempt queues to trigger
context save by suspending them.
On queue suspend, the KFD will copy the context save header information
so that the debugger can correctly crawl the appropriate size of the saved
context. The debugger must then also be allowed to resume suspended queues.
A queue that is newly created cannot be suspended because queue ids are
recycled after destruction so the debugger needs to know that this has
occurred. Query functions will be later added that will clear a given
queue of its new queue status.
A queue cannot be destroyed while it is suspended to preserve its saved
context during debugger inspection. Have queue destruction block while
a queue is suspended and unblocked when it is resumed. Likewise, if a
queue is about to be destroyed, it cannot be suspended.
Return the number of queues successfully suspended or resumed along with
a per queue status array where the upper bits per queue status show that
the request was invalid (new/destroyed queue suspend request, missing
queue) or an error occurred (HWS in a fatal state so it can't suspend or
resume queues).
Signed-off-by: Jonathan Kim <jonathan.kim@amd.com>
Reviewed-by: Felix Kuehling <felix.kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Due to a HW bug, waves in only half the shader arrays can enter trap.
When starting a debug session, relocate all waves to the first shader
array of each shader engine and mask off the 2nd shader array as
unavailable.
When ending a debug session, re-enable the 2nd shader array per
shader engine.
User CU masking per queue cannot be guaranteed to remain functional
if requested during debugging (e.g. user cu mask requests only 2nd shader
array as an available resource leading to zero HW resources available)
nor can runtime be alerted of any of these changes during execution.
Make user CU masking and debugging mutual exclusive with respect to
availability.
If the debugger tries to attach to a process with a user cu masked
queue, return the runtime status as enabled but busy.
If the debugger tries to attach and fails to reallocate queue waves to
the first shader array of each shader engine, return the runtime status
as enabled but with an error.
In addition, like any other mutli-process debug supported devices,
disable trap temporary setup per-process to avoid performance impact from
setup overhead.
Signed-off-by: Jonathan Kim <jonathan.kim@amd.com>
Reviewed-by: Felix Kuehling <felix.kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Unlike single process debug devices, multi-process debug devices allow
debug mode setting per-VMID (non-device-global).
Because the HWS manages PASID-VMID mapping, the new MAP_PROCESS API allows
the KFD to forward the required SPI debug register write requests.
To request a new debug mode setting change, the KFD must be able to
preempt all queues then remap all queues with these new setting
requests for MAP_PROCESS to take effect.
Note that by default, trap enablement in non-debug mode must be disabled
for performance reasons for multi-process debug devices due to setup
overhead in FW.
Signed-off-by: Jonathan Kim <jonathan.kim@amd.com>
Reviewed-by: Felix Kuehling <felix.kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Older HW only supports debugging on a single process because the
SPI debug mode setting registers are device global.
The HWS has supplied a single pinned VMID (0xf) for MAP_PROCESS
for debug purposes. To pin the VMID, the KFD will remove the VMID from
the HWS dynamic VMID allocation via SET_RESOUCES so that a debugged
process will never migrate away from its pinned VMID.
The KFD is responsible for reserving and releasing this pinned VMID
accordingly whenever the debugger attaches and detaches respectively.
Signed-off-by: Jonathan Kim <jonathan.kim@amd.com>
Reviewed-by: Felix Kuehling <felix.kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
The HWS schedule allows a grace period for wave completion prior to
preemption for better performance by avoiding CWSR on waves that can
potentially complete quickly. The debugger, on the other hand, will
want to inspect wave status immediately after it actively triggers
preemption (a suspend function to be provided).
To minimize latency between preemption and debugger wave inspection, allow
immediate preemption by setting the grace period to 0.
Note that setting the preepmtion grace period to 0 will result in an
infinite grace period being set due to a CP FW bug so set it to 1 for now.
Signed-off-by: Jonathan Kim <jonathan.kim@amd.com>
Reviewed-by: Felix Kuehling <felix.kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
The ROCm debugger will attach to a process to debug by PTRACE and will
expect the KFD to prepare a process for the target PID, whether the
target PID has opened the KFD device or not.
This patch is to explicity handle this requirement. Further HW mode
setting and runtime coordination requirements will be handled in
following patches.
In the case where the target process has not opened the KFD device,
a new KFD process must be created for the target PID.
The debugger as well as the target process for this case will have not
acquired any VMs so handle process restoration to correctly account for
this.
To coordinate with HSA runtime, the debugger must be aware of the target
process' runtime enablement status and will copy the runtime status
information into the debugged KFD process for later query.
On enablement, the debugger will subscribe to a set of exceptions where
each exception events will notify the debugger through a pollable FIFO
file descriptor that the debugger provides to the KFD to manage.
Finally on process termination of either the debugger or the target,
debugging must be disabled if it has not been done so.
Signed-off-by: Jonathan Kim <jonathan.kim@amd.com>
Reviewed-by: Felix Kuehling <felix.kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
for_each_inst modifies xcc_mask and therefore the loop doesn't
initialize properly interrupts on all pipes. Keep looping through xcc as
the outer loop to fix this issue.
Fixes: c4050ff1a4 ("drm/amdkfd: Use xcc mask for identifying xcc")
Signed-off-by: Lijo Lazar <lijo.lazar@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Instead of start xcc id and number of xcc per node, use the xcc mask
which is the mask of logical ids of xccs belonging to a parition.
Signed-off-by: Lijo Lazar <lijo.lazar@amd.com>
Reviewed-by: Le Ma <le.ma@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
When creating a user-mode SDMA queue, CP FW expects
driver to use/set virtual SDMA engine id in MAP_QUEUES
packet instead of using the physical SDMA engine id.
Each partition node's virtual SDMA number should start
from 0. However, when allocating doorbell for the queue,
KFD needs to allocate the doorbell from doorbell space
corresponding to the physical SDMA engine id, otherwise
the hwardware will not see the doorbell press.
Signed-off-by: Mukul Joshi <mukul.joshi@amd.com>
Reviewed-by: Amber Lin <Amber.Lin@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
This patch updates SDMA queue management for multi XCC in GFX9.4.3.
- Allocate/deallocate SDMA queues from the correct SDMA engines
based on the partition mode.
- Updates the kgd2kfd interface to fetch the correct SDMA register
addresses.
- It also fixes dumping correct SDMA queue info in debugfs.
v2: squash in fix "drm/amdkfd: Fix XGMI SDMA user-mode queue allocation"
Signed-off-by: Mukul Joshi <mukul.joshi@amd.com>
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
During DQM tear down, call DQM stop to unitialize HIQ and
associated memory allocated during packet manager init.
Signed-off-by: Mukul Joshi <mukul.joshi@amd.com>
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Context save handling needs to be updated for a multi XCC
setup:
- On a multi XCC setup, KFD needs to report context save base
address and size for each XCC in MQD.
- Thunk will allocate a large context save area covering all
XCCs which will be equal to: num_of_xccs in a partition * size
of context save area for 1 XCC. However, it will report only the
size of context save area for 1 XCC only in the ioctl call.
- Driver then setups the MQD correctly using the size passed from
Thunk and information about number of XCCs in a partition.
- Update get_wave_state function to return context save area
for all XCCs in the partition.
v2: update the get_wave_state function for mqd manager v11 (Morris)
Signed-off-by: Mukul Joshi <mukul.joshi@amd.com>
Tested-by: Amber Lin <Amber.Lin@amd.com>
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Morris Zhang <Shiwu.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Gfx 9 starts to have multiple XCC instances in one device. Add instance
parameter to kgd2kfd functions where XCC instance was hard coded as 0.
Also, update code to pass the correct instance number when running
on a multi-XCC setup.
v2: introduce the XCC instance to gfx v11 (Morris)
v3: rebase (Alex)
Signed-off-by: Amber Lin <Amber.Lin@amd.com>
Signed-off-by: Mukul Joshi <mukul.joshi@amd.com>
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com>
Tested-by: Amber Lin <Amber.Lin@amd.com>
Signed-off-by: Morris Zhang <Shiwu.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Update MQD management for both HIQ and user-mode compute
queues on a multi XCC setup. MQDs needs to be allocated,
initialized, loaded and destroyed for each XCC in the KFD
node.
v2: squash in fix "drm/amdkfd: Fix SDMA+HIQ HQD allocation on GFX9.4.3"
Signed-off-by: Mukul Joshi <mukul.joshi@amd.com>
Signed-off-by: Amber Lin <Amber.Lin@amd.com>
Tested-by: Amber Lin <Amber.Lin@amd.com>
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
This patch introduces multi-partition support in KFD.
This patch includes:
- Support for maximum 8 spatial partitions in KFD.
- Initialize one HIQ per partition.
- Management of VMID range depending on partition mode.
- Management of doorbell aperture space between all
partitions.
- Each partition does its own queue management, interrupt
handling, SMI event reporting.
- IOMMU, if enabled with multiple partitions, will only work
on first partition.
- SPM is only supported on the first partition.
- Currently, there is no support for resetting individual
partitions. All partitions will reset together.
Signed-off-by: Mukul Joshi <mukul.joshi@amd.com>
Tested-by: Amber Lin <Amber.Lin@amd.com>
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Introduce a new structure, kfd_node, which will now represent
a compute node. kfd_node is carved out of kfd_dev structure.
kfd_dev struct now will become the parent of kfd_node, and will
store common resources such as doorbells, GTT sub-alloctor etc.
kfd_node struct will store all resources specific to a compute
node, such as device queue manager, interrupt handling etc.
This is the first step in adding compute partition support in KFD.
v2: introduce kfd_node struct to gc v11 (Hawking)
v3: make reference to kfd_dev struct through kfd_node (Morris)
v4: use kfd_node instead for kfd isr/mqd functions (Morris)
v5: rebase (Alex)
Signed-off-by: Mukul Joshi <mukul.joshi@amd.com>
Tested-by: Amber Lin <Amber.Lin@amd.com>
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Morris Zhang <Shiwu.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
For the MQD memory, KMD would always allocate 4K memory,
and mes scheduler would write to the end of MQD for unmap flag.
Signed-off-by: Ruili Ji <ruiliji2@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
The point bo->kfd_bo is NULL for queue's write pointer BO
when creating queue on mGPU. To avoid using the pointer
fixes the error.
Signed-off-by: Eric Huang <jinhuieric.huang@amd.com>
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
This was fixed in initialize_cpsch before, but not in initialize_nocpsch.
Factor sdma bitmap initialization into a helper function to apply the
correct implementation in both cases without duplicating it.
v2: Added a range check
Reported-by: Ellis Michael <ellis@ellismichael.com>
Signed-off-by: Felix Kuehling <Felix.Kuehling@amd.com>
Reviewed-by: Graham Sider <Graham.Sider@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Update mes_v11_api_def.h add_queue API with is_aql_queue parameter. Also
re-use gds_size for the queue size (unused for KFD). MES requires the
queue size in order to compute the actual wptr offset within the queue
RB since it increases monotonically for AQL queues.
v2: Make is_aql_queue assign clearer
Signed-off-by: Graham Sider <Graham.Sider@amd.com>
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Queue would be freed when create_queue_cpsch fails
So lets do queue cleanup otherwise various list and memory issues
happen.
Signed-off-by: xinhui pan <xinhui.pan@amd.com>
Reviewed-by: Philip Yang <Philip.Yang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Update MES API to support oversubscription without aggregated doorbell
for usermode queues.
v2: Change oversubscription_no_aggregated_en to is_kfd_process (align
with MES)
Signed-off-by: Graham Sider <Graham.Sider@amd.com>
Acked-by: Felix Kuehling <Felix.Kuehling@amd.com>
Reviewed-by: Jack Xiao <Jack.Xiao@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Starting with GFX11, MES requires wptr BOs to be GTT allocated/mapped to
GART for usermode queues in order to support oversubscription. In the
case that work is submitted to an unmapped queue, MES must have a GART
wptr address to determine whether the queue should be mapped.
This change is accompanied with changes in MES and is applicable for
MES_API_VERSION >= 2.
v3:
- Use amdgpu_vm_bo_lookup_mapping for wptr_bo mapping lookup
- Move wptr_bo refcount increment to amdgpu_amdkfd_map_gtt_bo_to_gart
- Remove list_del_init from amdgpu_amdkfd_map_gtt_bo_to_gart
- Cleanup/fix create_queue wptr_bo error handling
v4:
- Add MES version shift/mask defines to amdgpu_mes.h
- Change version check from MES_VERSION to MES_API_VERSION
- Add check in kfd_ioctl_create_queue before wptr bo pin/GART map to
ensure bo is a single page.
Signed-off-by: Graham Sider <Graham.Sider@amd.com>
Acked-by: Alex Deucher <alexander.deucher@amd.com>
Acked-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Philip Yang <Philip.Yang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
After queue unmap or remove from MES successfully, free queue sysfs
entries, doorbell and remove from queue list. Otherwise, application may
destroy queue again, cause below kernel warning or crash backtrace.
For outstanding queues, either application forget to destroy or failed
to destroy, kfd_process_notifier_release will remove queue sysfs
entries, kfd_process_wq_release will free queue doorbell.
v2: decrement_queue_count for MES queue
refcount_t: underflow; use-after-free.
WARNING: CPU: 7 PID: 3053 at lib/refcount.c:28
Call Trace:
kobject_put+0xd6/0x1a0
kfd_procfs_del_queue+0x27/0x30 [amdgpu]
pqm_destroy_queue+0xeb/0x240 [amdgpu]
kfd_ioctl_destroy_queue+0x32/0x70 [amdgpu]
kfd_ioctl+0x27d/0x500 [amdgpu]
do_syscall_64+0x35/0x80
WARNING: CPU: 2 PID: 3053 at drivers/gpu/drm/amd/amdgpu/../amdkfd/kfd_device_queue_manager.c:400
Call Trace:
deallocate_doorbell.isra.0+0x39/0x40 [amdgpu]
destroy_queue_cpsch+0xb3/0x270 [amdgpu]
pqm_destroy_queue+0x108/0x240 [amdgpu]
kfd_ioctl_destroy_queue+0x32/0x70 [amdgpu]
kfd_ioctl+0x27d/0x500 [amdgpu]
general protection fault, probably for non-canonical address
0xdead000000000108:
Call Trace:
pqm_destroy_queue+0xf0/0x200 [amdgpu]
kfd_ioctl_destroy_queue+0x2f/0x60 [amdgpu]
kfd_ioctl+0x19b/0x600 [amdgpu]
Signed-off-by: Philip Yang <Philip.Yang@amd.com>
Reviewed-by: Graham Sider <Graham.Sider@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
We remove the user queue from MES scheduler to update queue properties.
If the queue becomes active after updating, add the user queue to MES
scheduler, to be able to handle command packet submission.
v2: don't break pqm_set_gws
Signed-off-by: Philip Yang <Philip.Yang@amd.com>
Reviewed-by: Graham Sider <Graham.Sider@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Add initial support for soc21 in KFD compute
driver (Mukul)
- Add new definition for soc21 device.
- Add new file for amdgpu-kfd interface for GFX11 family.
- Add new file for queue management, interrupt handling,
mqd management for GFX11 family in KFD driver.
- Related changes/updates for soc21 device in
KFD driver.
- Repurpose last 2 entries of SDMA MQD for driver use.
v2: Add an optional argument into update queue operation (Mukul)
v3: Switch to ip version check, replace kgd_dev with
amdgpu_device (Hawking)
Signed-off-by: Mukul Joshi <mukul.joshi@amd.com>
Signed-off-by: Hawking Zhang <Hawking.Zhang@amd.com>
Reviewed-by: Oak Zeng <Oak.Zeng@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
dqm->gws_queue_count and pdd->qpd.mapped_gws_queue need to be updated
each time the queue gets evicted.
Fixes: b8020b0304 ("drm/amdkfd: Enable over-subscription with >1 GWS queue")
Signed-off-by: David Yat Sin <david.yatsin@amd.com>
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
The KFD only unmaps all queues, all dynamics queues or all process queues
since RUN_LIST is mapped with all KFD queues.
There's no need to provide a single type unmap so remove this option.
Signed-off-by: Jonathan Kim <jonathan.kim@amd.com>
Reviewed-by: Felix Kuehling <felix.kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
A bunch of errors and warnings are leftover KFD over the years, attempt
to fix the errors and most warnings reported by checkpatch tool. Still a
few warnings remain which may be false positives so ignore them for now.
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Rajneesh Bhardwaj <rajneesh.bhardwaj@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Cleanup the kfd code by removing the unused old debugger
implementation.
The address watch was only ever implemented in the upstream
driver for GFXv7 (Kaveri). The user mode tools runtime using
this API was never open-sourced. Work on the old debugger
prototype that used this API has been discontinued years ago.
Only a small piece of resetting wavefronts is kept and
is moved to kfd_device_queue_manager.c.
Signed-off-by: Mukul Joshi <mukul.joshi@amd.com>
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
The new interface unmaps queues with reset mode for the process consumes
RAS poison, it's only for compute queue.
v2: rename the function to reset_queues.
Signed-off-by: Tao Zhou <tao.zhou1@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Acked-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Initializes kfd->device_info given either asic_type (enum) if GFX
version is less than GFX9, or GC IP version if greater. Also takes in vf
and the target compiler gfx version. Uses SDMA version to determine
num_sdma_queues_per_engine.
Convert device_info to a non-pointer member of kfd, change references
accordingly.
Change unsupported asic condition to only probe f2g, move device_info
initialization post-switch.
Signed-off-by: Graham Sider <Graham.Sider@amd.com>
Reviewed-by: Felix Kuehling <Felix.Kuehling@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
