In Xe, the perf layer allows capture of HW counter streams. These HW
counters are generally performance related but don't have to be necessarily
so. Also, the name "perf" is a carryover from i915 and is not preferred.
Here we propose the name "observation" for this common layer which allows
capture of different types of these counter streams.
v2: Rename observability layer to observation layer (Lucas/Rodrigo)
v3: Rename sysctl file to "observation_paranoid" (Jose)
Fixes: 52c2e956dc ("drm/xe/perf/uapi: "Perf" layer to support multiple perf counter stream types")
Fixes: fe8929bdf8 ("drm/xe/perf/uapi: Add perf_stream_paranoid sysctl")
Acked-by: Lucas De Marchi <lucas.demarchi@intel.com>
Acked-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Signed-off-by: Ashutosh Dixit <ashutosh.dixit@intel.com>
Reviewed-by: Umesh Nerlige Ramappa <umesh.nerlige.ramappa@intel.com>
Acked-by: José Roberto de Souza <jose.souza@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20240703164801.2561423-1-ashutosh.dixit@intel.com
(cherry picked from commit 8169b2097d)
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Implement query for properties of OA units present on a device.
v2: Clean up reserved/pad fields (Umesh)
Follow the same scheme as other query structs
v3: Skip reporting reserved engines attached to OA units
v4: Expose oa_buf_size via DRM_XE_PERF_IOCTL_INFO (Umesh)
v5: Don't expose capabilities as OR of properties (Umesh)
v6: Add extensions to query output structs: drm_xe_oa_unit,
drm_xe_query_oa_units and drm_xe_oa_stream_info
v7: Change oa_units[] array to __u64 type
Acked-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Reviewed-by: Umesh Nerlige Ramappa <umesh.nerlige.ramappa@intel.com>
Signed-off-by: Ashutosh Dixit <ashutosh.dixit@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20240618014609.3233427-13-ashutosh.dixit@intel.com
Implement the OA stream read file_operation. Both blocking and non-blocking
reads are supported. As part of read system call, the read copies OA perf
data from the OA buffer to the user buffer, after appending packet headers
for status and data packets.
v2: Drop OA report headers, implement DRM_XE_PERF_IOCTL_STATUS (Umesh)
v3: Introduce 'struct drm_xe_oa_stream_status'
v4: Define oa_status register bitfields (Umesh)
v5: Add extensions to 'struct drm_xe_oa_stream_status'
v6: Minor cleanup, eliminate report32 variable
v7: Use -EIO to signal to userspace to read OASTATUS using
DRM_XE_PERF_IOCTL_STATUS, change previous sites returning -EIO to
return -EINVAL
Make drm_xe_oa_stream_status bits contiguous (Jose, Umesh)
rmw oa_status bits (Umesh)
Acked-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Acked-by: José Roberto de Souza <jose.souza@intel.com>
Reviewed-by: Umesh Nerlige Ramappa <umesh.nerlige.ramappa@intel.com>
Signed-off-by: Ashutosh Dixit <ashutosh.dixit@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20240618014609.3233427-10-ashutosh.dixit@intel.com
The OA stream open perf op returns an fd with its own file_operations for
the newly initialized OA stream. These file_operations allow userspace to
enable or disable the stream, as well as apply a different metric
configuration for the OA stream. Userspace can also poll for data
availability. OA stream initialization is completed in this commit by
enabling the OA stream. When sampling is enabled this starts a hrtimer
which periodically checks for data availablility.
v2: Use stream properties for stream reconfiguration with
DRM_XE_PERF_IOCTL_CONFIG
v3: Hold runtime_pm reference across oa buffer alloc/free
v4: Fix 32 bit build
Acked-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Acked-by: José Roberto de Souza <jose.souza@intel.com>
Reviewed-by: Umesh Nerlige Ramappa <umesh.nerlige.ramappa@intel.com>
Signed-off-by: Ashutosh Dixit <ashutosh.dixit@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20240618014609.3233427-9-ashutosh.dixit@intel.com
Properties for OA streams are specified by user space, when the stream is
opened, as a chain of drm_xe_ext_set_property struct's. Parse and validate
these stream properties.
v2: Remove struct drm_xe_oa_open_param (Harish Chegondi)
Drop DRM_XE_OA_PROPERTY_POLL_OA_PERIOD_US (Umesh)
Eliminate comparison with xe_oa_max_sample_rate (Umesh)
Drop 'struct drm_xe_oa_record_header' (Umesh)
v3: s/DRM_XE_OA_PROPERTY_OA_EXPONENT/ \
DRM_XE_OA_PROPERTY_OA_PERIOD_EXPONENT/ (Jose)
v4: Fix 32 bit build
v5: Add non-static function kernel doc (Michal)
Acked-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Acked-by: José Roberto de Souza <jose.souza@intel.com>
Reviewed-by: Umesh Nerlige Ramappa <umesh.nerlige.ramappa@intel.com>
Signed-off-by: Ashutosh Dixit <ashutosh.dixit@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20240618014609.3233427-7-ashutosh.dixit@intel.com
Introduce add/remove config perf ops for OA. OA configurations consist of a
set of event/counter select register address/value pairs. The add_config
perf op validates and stores such configurations and also exposes them in
the metrics sysfs. These configurations will be programmed to OA unit HW
when an OA stream using a configuration is opened. The OA stream can also
switch to other stored configurations.
v2: Start config id's from 1 and other minor review comments (Umesh)
v3: Add 32 bit build
v4: Add kernel doc for non-static functions (Michal)
Acked-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Acked-by: José Roberto de Souza <jose.souza@intel.com>
Reviewed-by: Umesh Nerlige Ramappa <umesh.nerlige.ramappa@intel.com>
Signed-off-by: Ashutosh Dixit <ashutosh.dixit@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20240618014609.3233427-6-ashutosh.dixit@intel.com
In Xe, the plan is to support multiple types of perf counter streams (OA is
only one type of these streams). Rather than introduce NxM ioctls for
these (N perf streams with M ioctl's per perf stream), we decide to
multiplex these (N different stream types and the M ops for each of these
stream types) through a single PERF ioctl. This multiplexing is the purpose
of the PERF layer.
In addition to PERF DRM ioctl's, another set of ioctl's on the PERF fd are
defined. These are expected to be common to different PERF stream types and
therefore defined at the PERF layer itself.
v2: Add param_size to 'struct drm_xe_perf_param' (Umesh)
v3: Rename 'enum drm_xe_perf_ops' to
'enum drm_xe_perf_ioctls' (Guy Zadicario)
Add DRM_ prefix to ioctl names to indicate uapi names
v4: Add 'enum drm_xe_perf_op' previously missed out (Guy Zadicario)
v5: Squash the ops and PERF layer patches into a single patch (Umesh)
Remove param_size from struct 'drm_xe_perf_param' (Umesh)
v6: Add DRM_XE_PERF_IOCTL_STATUS
v7: Add DRM_XE_PERF_IOCTL_INFO
v8: Fix Copyright years, fix DRM_XE_PERF_TYPE_MAX, move '#include
"xe_perf.h"' to xe_perf.c, add kernel doc (Michal)
Acked-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Acked-by: Guy Zadicario <gzadicario@habana.ai>
Acked-by: José Roberto de Souza <jose.souza@intel.com>
Reviewed-by: Umesh Nerlige Ramappa <umesh.nerlige.ramappa@intel.com>
Signed-off-by: Ashutosh Dixit <ashutosh.dixit@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20240618014609.3233427-2-ashutosh.dixit@intel.com
For modern platforms (MTL and later), both kernel and userspace drivers
are expected to apply GT programming and workarounds based on the IP
version and stepping self-reported by the GT hardware via the GMD_ID
registers. Since userspace drivers can't access these registers
directly, pass along the version and stepping information via the GT
list query. Note that the new query fields will remain 0's when running
on pre-GMD_ID platforms. Userspace is expected to continue using PCI
devid / revid on those older platforms.
Although the hardware also has a GMD_ID register for display
version/stepping, that value is intentionally *not* included anywhere in
the Xe uapi. Display userspace should be using platform-agnostic APIs
and auto-detecting platform capabilities rather than matching specific
IP versions.
v2:
- s/revid/rev/ (Lucas)
- Fix kerneldoc copy/paste mistakes
Signed-off-by: Matt Roper <matthew.d.roper@intel.com>
Reviewed-by: Lucas De Marchi <lucas.demarchi@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20240312211229.2871288-4-matthew.d.roper@intel.com
Those cases missed in previous uAPI cleanups were mostly accidentally
brought in from i915 or created to exercise the possibilities of gpuvm
but they are not used by userspace yet, so let's remove them. They can
still be brought back later if needed.
v2:
- Fix XE_VM_FLAG_FAULT_MODE support in xe_lrc.c (Brian Welty)
- Leave DRM_XE_VM_BIND_OP_UNMAP_ALL (José Roberto de Souza)
- Ensure invalid flag values are rejected (Rodrigo Vivi)
v3: Rebase after removal of persistent exec_queues (Francois Dugast)
v4: Rodrigo: Rebase after the new dumpable flag.
Fixes: dd08ebf6c3 ("drm/xe: Introduce a new DRM driver for Intel GPUs")
Cc: Thomas Hellström <thomas.hellstrom@linux.intel.com>
Cc: Rodrigo Vivi <rodrigo.vivi@intel.com>
Signed-off-by: Francois Dugast <francois.dugast@intel.com>
Reviewed-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20240222232356.175431-1-rodrigo.vivi@intel.com
Complete the documentation of some structs by adding functional
examples of user space code. Those examples are intentionally kept
very simple. Put together, they provide a foundation for a minimal
application that executes a job using the Xe driver.
v2: Remove use of DRM_XE_VM_BIND_FLAG_ASYNC (Francois Dugast)
Reviewed-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Acked-by: José Roberto de Souza <jose.souza@intel.com>
Acked-by: Mateusz Naklicki <mateusz.naklicki@intel.com>
Signed-off-by: Francois Dugast <francois.dugast@intel.com>
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
In order to make proper use the uAPI, a prerequisite is to understand
some key concepts about the discrete GPU devices which are supported
by the Xe driver. For example, some structs defined in the uAPI are an
abstraction of a hardware component with a specific role.
This diagram helps to build a mental representation of a device how it
is seen by the Xe driver. As written in the documentation, it does not
intend to be a literal representation of an existing device. A lot
more information could be added but the intention for the overview is
to keep it simple, and go into detail as needed in other sections.
v2: Add GT1 inside Tile0 (José Roberto de Souza)
Reviewed-by: José Roberto de Souza <jose.souza@intel.com>
Acked-by: José Roberto de Souza <jose.souza@intel.com>
Acked-by: Mateusz Naklicki <mateusz.naklicki@intel.com>
Signed-off-by: Francois Dugast <francois.dugast@intel.com>
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
The uAPI should stay generic in regarding to the bitmask. It is
the userspace responsibility to check for the type/class of the
memory, without any assumption.
Also add comments inside the code to explain how it is actually
constructed so we don't accidentally change the assignment of
the instance and the masks.
No functional change in this patch. It only explains and document
the memory_region masks. A further follow-up work with the
organization of all memory regions around struct xe_mem_regions
is desired, but not part of this patch.
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Reviewed-by: Lucas De Marchi <lucas.demarchi@intel.com>
Acked-by: José Roberto de Souza <jose.souza@intel.com>
Acked-by: Mateusz Naklicki <mateusz.naklicki@intel.com>
Signed-off-by: Francois Dugast <francois.dugast@intel.com>
No functional change in this patch.
Let's ensure all of our structs are documented and with a certain
standard. Also, let's have an overview and list of IOCTLs as the
very beginning of the generated HTML doc.
v2: Nits (Lucas De Marchi)
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Reviewed-by: Lucas De Marchi <lucas.demarchi@intel.com>
Acked-by: José Roberto de Souza <jose.souza@intel.com>
Acked-by: Mateusz Naklicki <mateusz.naklicki@intel.com>
Signed-off-by: Francois Dugast <francois.dugast@intel.com>
This patch doesn't modify any text or uapi entries themselves.
It only move things up and down aiming a better organization of the uAPI.
While fixing the documentation I noticed that query_engine_cs_cycles
was in the middle of the memory_region info. Then I noticed more
mismatches on the order when compared to the order of the IOCTL
and QUERY entries declaration. So this patch aims to bring some
order to the uAPI so it gets easier to read and the documentation
generated in the end is able to tell a consistent story.
Overall order:
1. IOCTL definition
2. Extension definition and helper structs
3. IOCTL's Query structs in the order of the Query's entries.
4. The rest of IOCTL structs in the order of IOCTL declaration.
5. uEvents
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Reviewed-by: Lucas De Marchi <lucas.demarchi@intel.com>
Acked-by: José Roberto de Souza <jose.souza@intel.com>
Acked-by: Mateusz Naklicki <mateusz.naklicki@intel.com>
Signed-off-by: Francois Dugast <francois.dugast@intel.com>
As there is no direct way to make comments of constants directly
visible in the kernel doc, move them to the description of the
structure where they can be used. By doing so they appear in the
"Description" section of the struct documentation.
v2: Remove DRM_XE_UFENCE_WAIT_MASK_* (Francois Dugast)
Reviewed-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Acked-by: José Roberto de Souza <jose.souza@intel.com>
Acked-by: Mateusz Naklicki <mateusz.naklicki@intel.com>
Signed-off-by: Francois Dugast <francois.dugast@intel.com>
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
This removes the documentation build warnings below:
include/uapi/drm/xe_drm.h:828: warning: Function parameter or \
member 'pad2' not described in 'drm_xe_vm_bind_op'
include/uapi/drm/xe_drm.h:875: warning: Function parameter or \
member 'pad2' not described in 'drm_xe_vm_bind'
include/uapi/drm/xe_drm.h:1006: warning: Function parameter or \
member 'handle' not described in 'drm_xe_sync'
include/uapi/drm/xe_drm.h:1006: warning: Function parameter or \
member 'timeline_value' not described in 'drm_xe_sync'
Reviewed-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Acked-by: José Roberto de Souza <jose.souza@intel.com>
Acked-by: Mateusz Naklicki <mateusz.naklicki@intel.com>
Signed-off-by: Francois Dugast <francois.dugast@intel.com>
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
remove the num_engines/instances members from drm_xe_wait_user_fence
structure and add a exec_queue_id member
Right now this is only checking if the engine list is sane and nothing
else. In the end every operation with this IOCTL is a soft check.
So, let's formalize that and only use this IOCTL to wait on the fence.
exec_queue_id member will help to user space to get proper error code
from kernel while in exec_queue reset
Signed-off-by: Bommu Krishnaiah <krishnaiah.bommu@intel.com>
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Acked-by: Matthew Brost <matthew.brost@intel.com>
Reviewed-by: Francois Dugast <francois.dugast@intel.com>
Acked-by: José Roberto de Souza <jose.souza@intel.com>
Acked-by: Mateusz Naklicki <mateusz.naklicki@intel.com>
Signed-off-by: Francois Dugast <francois.dugast@intel.com>
Currently we're using "compute mode" for long running VMs using
preempt-fences for memory management, and "fault mode" for long
running VMs using page faults.
Change this to use the terminology "long-running" abbreviated as LR for
long-running VMs. These VMs can then either be in preempt-fence mode or
fault mode. The user can force fault mode at creation time, but otherwise
the driver can choose to use fault- or preempt-fence mode for long-running
vms depending on the device capabilities. Initially unless fault-mode is
specified, the driver uses preempt-fence mode.
v2:
- Fix commit message wording and the documentation around
CREATE_FLAG_LR_MODE and CREATE_FLAG_FAULT_MODE
Cc: Matthew Brost <matthew.brost@intel.com>
Cc: Rodrigo Vivi <rodrigo.vivi@intel.com>
Cc: Francois Dugast <francois.dugast@intel.com>
Signed-off-by: Thomas Hellström <thomas.hellstrom@linux.intel.com>
Signed-off-by: Francois Dugast <francois.dugast@intel.com>
Reviewed-by: José Roberto de Souza <jose.souza@intel.com>
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Although the exec ioctl is a very important one, it makes no sense
to explain xe_exec before explaining the exec_queue. So, let's
move this down to help bring a better flow on the documentation
and code readability.
It is important to highlight that this patch is changing all
the ioctl numbers in a non-backward compatible way. However, we
are doing this final uapi clean-up before we submit our first
pull-request to be part of the upstream Kernel. Once we get
there, no other change like this will ever happen and all the
backward compatibility will be respected.
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Signed-off-by: Francois Dugast <francois.dugast@intel.com>
Reviewed-by: José Roberto de Souza <jose.souza@intel.com>
Let's respect Documentation/process/botching-up-ioctls.rst
and add the proper padding for a 64b alignment with all as
well as all the required checks and settings for the pads
and the reserved entries.
v2: Fix remaining holes and double check with pahole (Jose)
Ensure with pahole that both 32b and 64b have exact same
layout (Thomas)
Do not set query's pad and reserved bits to zero since it
is redundant and already done by kzalloc (Matt)
v3: Fix alignment after rebase (José Roberto de Souza)
v4: Fix pad check (Francois Dugast)
Cc: Thomas Hellström <thomas.hellstrom@linux.intel.com>
Cc: Francois Dugast <francois.dugast@intel.com>
Cc: José Roberto de Souza <jose.souza@intel.com>
Cc: Matt Roper <matthew.d.roper@intel.com>
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Signed-off-by: Francois Dugast <francois.dugast@intel.com>
Reviewed-by: José Roberto de Souza <jose.souza@intel.com>
As an information only. So Userspace can use this information
and be able to correlate different GTs.
Make API symmetric between Engine and GT info.
There's no need right now to include a tile_query entry
since there's no other information that we need from tile
that is not already exposed through different queries.
However, this could be added later if we have different Tile
information that could matter to userspace. But let's keep
the API ready for a direct reference to Tile ID based on
the GT entry.
Signed-off-by: Francois Dugast <francois.dugast@intel.com>
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Reviewed-by: José Roberto de Souza <jose.souza@intel.com>
First of all, let's remove the duplication.
But also, let's rename it to remove the word 'frequency'
out of it. In general, the first thing people think of frequency
is the frequency in which the GTs are operating to execute the
GPU instructions.
While this frequency here is a crystal reference clock frequency
which is the base of everything else, and in this case of this
uAPI it is used to calculate a better and precise timestamp.
v2: (Suggested by Jose) Remove the engine_cs and keep the GT info one
since it might be useful for other SRIOV cases where the engine_cs
will be zeroed. So, grabbing from the GT_LIST should be cleaner.
v3: Keep comment on put_user() call (José Roberto de Souza)
Cc: Matt Roper <matthew.d.roper@intel.com>
Umesh Nerlige Ramappa <umesh.nerlige.ramappa@intel.com>
Cc: Jose Souza <jose.souza@intel.com>
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Signed-off-by: Francois Dugast <francois.dugast@intel.com>
Reviewed-by: José Roberto de Souza <jose.souza@intel.com>
It is currently unused, so by the rules it cannot go upstream.
Also there was the desire to convert that to align with the
engine_class_instance selection, but the consensus on that one
is to remain with the global gt_id. So we are keeping the gt_id
there, not converting to a generic sched_group and also killing
this tile_mask and only using the default behavior of 0 that is
to create a mapping / page_table entry on every tile, similar
to what i915.
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Signed-off-by: Francois Dugast <francois.dugast@intel.com>
Reviewed-by: Matthew Brost <matthew.brost@intel.com>
Reviewed-by: José Roberto de Souza <jose.souza@intel.com>
