drm/i915: rename I915_PMU_MAX_GTS to I915_PMU_MAX_GT

_GTS as an abbreviation here leads to some confusion, match other
definitions and drop the s.

Cc: Matt Roper <matthew.d.roper@intel.com>
Cc: Ashutosh Dixit <ashutosh.dixit@linux.intel.com>
Cc: Andi Shyti <andy.shyti@linux.intel.com>
Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Cc: Umesh Nerlige Ramappa <umesh.nerlige.ramappa@intel.com>

Signed-off-by: Matt Atwood <matthew.s.atwood@intel.com>
Reviewed-by: Matt Roper <matthew.d.roper@intel.com>
Signed-off-by: Matt Roper <matthew.d.roper@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20230602231754.1596433-3-matthew.s.atwood@intel.com

drm/i915/pmu: Make PMU sample array two-dimensional

No functional changes but we can remove some unsightly index computation
and read/write functions if we convert the PMU sample array from a
one-dimensional to a two-dimensional array.

v2: Retain read/store helpers (Tvrtko)

Suggested-by: Tvrtko Ursulin <tvrtko.ursulin@linux.intel.com>
Reviewed-by: Andrzej Hajda <andrzej.hajda@intel.com>
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Signed-off-by: Ashutosh Dixit <ashutosh.dixit@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20230524215629.97920-3-ashutosh.dixit@intel.com

drm/i915/pmu: Prepare for multi-tile non-engine counters

Reserve some bits in the counter config namespace which will carry the
tile id and prepare the code to handle this.

No per tile counters have been added yet.

v2:
- Fix checkpatch issues
- Use 4 bits for gt id in non-engine counters. Drop FIXME.
- Set MAX GTs to 4. Drop FIXME.

v3: (Ashutosh, Tvrtko)
- Drop BUG_ON that would never fire
- Make enable u64
- Pull in some code from next patch

v4: Set I915_PMU_MAX_GTS to 2 (Tvrtko)

v5: s/u64/u32 where needed (Ashutosh)

Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Signed-off-by: Umesh Nerlige Ramappa <umesh.nerlige.ramappa@intel.com>
Reviewed-by: Ashutosh Dixit <ashutosh.dixit@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20230519154946.3751971-7-umesh.nerlige.ramappa@intel.com

drm/i915/pmu: Add reference counting to the sampling timer

We do not want to have timers per tile and waste CPU cycles and energy via
multiple wake-up sources, for a relatively un-important task of PMU
sampling, so keeping a single timer works well. But we also do not want
the first GT which goes idle to turn off the timer.

Add some reference counting, via a mask of unparked GTs, to solve this.

v2: Drop the check for unparked in i915_sample (Ashutosh)
v3: Revert v2 (Tvrtko)

Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Reviewed-by: Umesh Nerlige Ramappa <umesh.nerlige.ramappa@intel.com>
Signed-off-by: Umesh Nerlige Ramappa <umesh.nerlige.ramappa@intel.com>
Reviewed-by: Ashutosh Dixit <ashutosh.dixit@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20230519154946.3751971-6-umesh.nerlige.ramappa@intel.com

drm/i915/pmu: Transform PMU parking code to be GT based

Trivial prep work for full multi-tile enablement later.

Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Signed-off-by: Vinay Belgaumkar <vinay.belgaumkar@intel.com>
Reviewed-by: Umesh Nerlige Ramappa <umesh.nerlige.ramappa@intel.com>
Signed-off-by: Umesh Nerlige Ramappa <umesh.nerlige.ramappa@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20230519154946.3751971-5-umesh.nerlige.ramappa@intel.com

drm/i915/pmu: drop kernel-doc

The comments are closer to not being kernel-doc.

drivers/gpu/drm/i915/i915_pmu.h:21: warning: cannot understand function prototype: 'enum i915_pmu_tracked_events '
drivers/gpu/drm/i915/i915_pmu.h:32: warning: cannot understand function prototype: 'enum '
drivers/gpu/drm/i915/i915_pmu.h:41: warning: This comment starts with '/**', but isn't a kernel-doc comment. Refer Documentation/doc-guide/kernel-doc.rst
* How many different events we track in the global PMU mask.

Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Reviewed-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/7656b8f58b088c108a2a32f7089329740efabba1.1683041799.git.jani.nikula@intel.com

drm/i915: Use a table for i915_init/exit (v2)

If the driver was not fully loaded, we may still have globals lying
around. If we don't tear those down in i915_exit(), we'll leak a bunch
of memory slabs. This can happen two ways: use_kms = false and if we've
run mock selftests. In either case, we have an early exit from
i915_init which happens after i915_globals_init() and we need to clean
up those globals.

The mock selftests case is especially sticky. The load isn't entirely
a no-op. We actually do quite a bit inside those selftests including
allocating a bunch of mock objects and running tests on them. Once all
those tests are complete, we exit early from i915_init(). Perviously,
i915_init() would return a non-zero error code on failure and a zero
error code on success. In the success case, we would get to i915_exit()
and check i915_pci_driver.driver.owner to detect if i915_init exited early
and do nothing. In the failure case, we would fail i915_init() but
there would be no opportunity to clean up globals.

The most annoying part is that you don't actually notice the failure as
part of the self-tests since leaking a bit of memory, while bad, doesn't
result in anything observable from userspace. Instead, the next time we
load the driver (usually for next IGT test), i915_globals_init() gets
invoked again, we go to allocate a bunch of new memory slabs, those
implicitly create debugfs entries, and debugfs warns that we're trying
to create directories and files that already exist. Since this all
happens as part of the next driver load, it shows up in the dmesg-warn
of whatever IGT test ran after the mock selftests.

While the obvious thing to do here might be to call i915_globals_exit()
after selftests, that's not actually safe. The dma-buf selftests call
i915_gem_prime_export which creates a file. We call dma_buf_put() on
the resulting dmabuf which calls fput() on the file. However, fput()
isn't immediate and gets flushed right before syscall returns. This
means that all the fput()s from the selftests don't happen until right
before the module load syscall used to fire off the selftests returns
which is after i915_init(). If we call i915_globals_exit() in
i915_init() after selftests, we end up freeing slabs out from under
objects which won't get released until fput() is flushed at the end of
the module load syscall.

The solution here is to let i915_init() return success early and detect
the early success in i915_exit() and only tear down globals and nothing
else. This way the module loads successfully, regardless of the success
or failure of the tests. Because we've not enumerated any PCI devices,
no device nodes are created and it's entirely useless from userspace.
The only thing the module does at that point is hold on to a bit of
memory until we unload it and i915_exit() is called. Importantly, this
means that everything from our selftests has the ability to properly
flush out between i915_init() and i915_exit() because there is at least
one syscall boundary in between.

In order to handle all the delicate init/exit cases, we convert the
whole thing to a table of init/exit pairs and track the init status in
the new init_progress global. This allows us to ensure that i915_exit()
always tears down exactly the things that i915_init() successfully
initialized. We also allow early-exit of i915_init() without failure by
an init function returning > 0. This is useful for nomodeset, and
selftests. For the mock selftests, we convert them to always return 1
so we get the desired behavior of the driver always succeeding to load
the driver and then properly tearing down the partially loaded driver.

v2 (Tvrtko Ursulin):
- Guard init_funcs[i].exit with GEM_BUG_ON(i >= ARRAY_SIZE(init_funcs))
v2 (Daniel Vetter):
- Update the docstring for i915.mock_selftests

Signed-off-by: Jason Ekstrand <jason@jlekstrand.net>
Reviewed-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Link: https://patchwork.freedesktop.org/patch/msgid/20210721152358.2893314-4-jason@jlekstrand.net

drm/i915/pmu: Deprecate I915_PMU_LAST and optimize state tracking

Adding any kinds of "last" abi markers is usually a mistake which I
repeated when implementing the PMU because it felt convenient at the time.

This patch marks I915_PMU_LAST as deprecated and stops the internal
implementation using it for sizing the event status bitmask and array.

New way of sizing the fields is a bit less elegant, but it omits reserving
slots for tracking events we are not interested in, and as such saves some
runtime space. Adding sampling events is likely to be a special event and
the new plumbing needed will be easily detected in testing. Existing
asserts against the bitfield and array sizes are keeping the code safe.

First event which gets the new treatment in this new scheme are the
interrupts - which neither needs any tracking in i915 pmu nor needs
waking up the GPU to read it.

v2:
* Streamline helper names. (Chris)

v3:
* Comment which events need tracking. (Chris)

Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Link: https://patchwork.freedesktop.org/patch/msgid/20201201131757.206367-1-tvrtko.ursulin@linux.intel.com

drm/i915/pmu: Replace open coded kstat_irqs() copy

Driver code has no business with the internals of the irq descriptor.

Aside of that the count is per interrupt line and therefore takes
interrupts from other devices into account which share the interrupt line
and are not handled by the graphics driver.

Replace it with a pmu private count which only counts interrupts which
originate from the graphics card.

To avoid atomics or heuristics of some sort make the counter field
'unsigned long'. That limits the count to 4e9 on 32bit which is a lot and
postprocessing can easily deal with the occasional wraparound.

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Acked-by: Jani Nikula <jani.nikula@intel.com>
Cc: Tvrtko Ursulin <tvrtko.ursulin@linux.intel.com>
Link: https://lore.kernel.org/r/20201210194043.957046529@linutronix.de

drm/i915/pmu: Fix CPU hotplug with multiple GPUs

Since we keep a driver global mask of online CPUs and base the decision
whether PMU needs to be migrated upon it, we need to make sure the
migration is done for all registered PMUs (so GPUs).

To do this we need to track the current CPU for each PMU and base the
decision on whether to migrate on a comparison between global and local
state.

At the same time, since dynamic CPU hotplug notification slots are a
scarce resource and given how we already register the multi instance type
state, we can and should add multiple instance of the i915 PMU to this
same state and not allocate a new one for every GPU.

v2:
* Use pr_notice. (Chris)

v3:
* Handle a nasty interaction where unregistration which triggers a false
CPU offline event. (Chris)

Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Suggested-by: Daniel Vetter <daniel.vetter@intel.com> # dynamic slot optimisation
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Link: https://patchwork.freedesktop.org/patch/msgid/20201020161144.678668-1-tvrtko.ursulin@linux.intel.com

drm/i915/pmu: Handle PCI unbind

Mark the device as closed and keep references to driver data alive to
allow for safe driver unbind with active PMU clients. Perf core does not
otherwise handle this case so we have to do it manually like this.

Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Link: https://patchwork.freedesktop.org/patch/msgid/20201020100822.543332-1-tvrtko.ursulin@linux.intel.com

drm/i915: significantly reduce the use of <drm/i915_drm.h>

The #include has been splattered all over the place, but there are
precious few places, all .c files, that actually need it.

v2: remove leftover double newlines

Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20200225133131.3301-1-jani.nikula@intel.com

drm/i915/pmu: Avoid using globals for PMU events

Attempting to bind / unbind module from devices where we have both
integrated and discreete GPU handled by i915, will cause us to try and
double free the global state, hitting null ptr deref in free_event_attributes.

Let's move it to i915_pmu.

Fixes: 05488673a4d4 ("drm/i915/pmu: Support multiple GPUs")
Signed-off-by: Michał Winiarski <michal.winiarski@intel.com>
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Michal Wajdeczko <michal.wajdeczko@intel.com>
Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Link: https://patchwork.freedesktop.org/patch/msgid/20200219161822.24592-2-michal.winiarski@intel.com
(cherry picked from commit 46129dc10f47c5c2b51c93a82b7b2aca46574ae0)
Signed-off-by: Jani Nikula <jani.nikula@intel.com>

drm/i915/pmu: Avoid using globals for CPU hotplug state

Attempting to bind / unbind module from devices where we have both
integrated and discreete GPU handled by i915 can lead to leaks and
warnings from cpuhp:
Error: Removing state XXX which has instances left.

Let's move the state to i915_pmu.

Fixes: 05488673a4d4 ("drm/i915/pmu: Support multiple GPUs")
Signed-off-by: Michał Winiarski <michal.winiarski@intel.com>
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Michal Wajdeczko <michal.wajdeczko@intel.com>
Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Link: https://patchwork.freedesktop.org/patch/msgid/20200219161822.24592-1-michal.winiarski@intel.com
(cherry picked from commit f5a179d4687d4e7bfadd7cbda7ee5d0bad76761f)
Signed-off-by: Jani Nikula <jani.nikula@intel.com>

drm/i915/pmu: Avoid using globals for PMU events

Attempting to bind / unbind module from devices where we have both
integrated and discreete GPU handled by i915, will cause us to try and
double free the global state, hitting null ptr deref in free_event_attributes.

Let's move it to i915_pmu.

Fixes: 05488673a4d4 ("drm/i915/pmu: Support multiple GPUs")
Signed-off-by: Michał Winiarski <michal.winiarski@intel.com>
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Michal Wajdeczko <michal.wajdeczko@intel.com>
Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Link: https://patchwork.freedesktop.org/patch/msgid/20200219161822.24592-2-michal.winiarski@intel.com

drm/i915/pmu: Avoid using globals for CPU hotplug state

Attempting to bind / unbind module from devices where we have both
integrated and discreete GPU handled by i915 can lead to leaks and
warnings from cpuhp:
Error: Removing state XXX which has instances left.

Let's move the state to i915_pmu.

Fixes: 05488673a4d4 ("drm/i915/pmu: Support multiple GPUs")
Signed-off-by: Michał Winiarski <michal.winiarski@intel.com>
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Michal Wajdeczko <michal.wajdeczko@intel.com>
Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Link: https://patchwork.freedesktop.org/patch/msgid/20200219161822.24592-1-michal.winiarski@intel.com

drm/i915/pmu: Ensure monotonic rc6

Avoid rc6 counter going backward in close to 0% RC6 scenarios like:

15.005477996 114,246,613 ns i915/rc6-residency/
16.005876662 667,657 ns i915/rc6-residency/
17.006131417 7,286 ns i915/rc6-residency/
18.006615031 18,446,744,073,708,914,688 ns i915/rc6-residency/
19.007158361 18,446,744,073,709,447,168 ns i915/rc6-residency/
20.007806498 0 ns i915/rc6-residency/
21.008227495 1,440,403 ns i915/rc6-residency/

There are two aspects to this fix.

First is not assuming rc6 value zero means GT is asleep since that can
also mean GPU is fully busy and we do not want to enter the estimation
path in that case.

Second is ensuring monotonicity on the estimation path itself. I suspect
what is happening is with extremely rapid park/unpark cycles we get no
updates on the real rc6 and therefore have to careful not to
unconditionally trust use last known real rc6 when creating a new
estimation.

v2:
* Simplify logic by not tracking the estimate but last reported value.

Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Fixes: 16ffe73c186b ("drm/i915/pmu: Use GT parked for estimating RC6 while asleep")
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> # v1
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Link: https://patchwork.freedesktop.org/patch/msgid/20191217142057.1000-1-tvrtko.ursulin@linux.intel.com
(cherry picked from commit df6a42053513846475ae1fbd224dfbdbcd0c7010)
Signed-off-by: Jani Nikula <jani.nikula@intel.com>

drm/i915/pmu: Ensure monotonic rc6

Avoid rc6 counter going backward in close to 0% RC6 scenarios like:

15.005477996 114,246,613 ns i915/rc6-residency/
16.005876662 667,657 ns i915/rc6-residency/
17.006131417 7,286 ns i915/rc6-residency/
18.006615031 18,446,744,073,708,914,688 ns i915/rc6-residency/
19.007158361 18,446,744,073,709,447,168 ns i915/rc6-residency/
20.007806498 0 ns i915/rc6-residency/
21.008227495 1,440,403 ns i915/rc6-residency/

There are two aspects to this fix.

First is not assuming rc6 value zero means GT is asleep since that can
also mean GPU is fully busy and we do not want to enter the estimation
path in that case.

Second is ensuring monotonicity on the estimation path itself. I suspect
what is happening is with extremely rapid park/unpark cycles we get no
updates on the real rc6 and therefore have to careful not to
unconditionally trust use last known real rc6 when creating a new
estimation.

v2:
* Simplify logic by not tracking the estimate but last reported value.

Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Fixes: 16ffe73c186b ("drm/i915/pmu: Use GT parked for estimating RC6 while asleep")
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk> # v1
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Link: https://patchwork.freedesktop.org/patch/msgid/20191217142057.1000-1-tvrtko.ursulin@linux.intel.com

drm/i915/pmu: Support multiple GPUs

With discrete graphics system can have both integrated and discrete GPU
handled by i915.

Currently we use a fixed name ("i915") when registering as the uncore PMU
provider which stops working in this case.

To fix this we add the PCI device name string to non-integrated devices
handled by us. Integrated devices keep the legacy name preserving
backward compatibility.

v2:
* Detect IGP and keep legacy name. (Michal)
* Use PCI device name as suffix. (Michal, Chris)

v3:
* Constify the name. (Chris)
* Use pci_domain_nr. (Chris)

v4:
* Fix kfree_const usage. (Chris)

v5:
* kfree_const does not work for modules. (Chris)
* Changed is_igp helper to take i915.

Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Michal Wajdeczko <michal.wajdeczko@intel.com>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Link: https://patchwork.freedesktop.org/patch/msgid/20191016093802.12483-1-tvrtko.ursulin@linux.intel.com

drm/i915/pmu: Use GT parked for estimating RC6 while asleep

As we track when we put the GT device to sleep upon idling, we can use
that callback to sample the current rc6 counters and record the
timestamp for estimating samples after that point while asleep.

v2: Stick to using ktime_t
v3: Track user_wakerefs that interfere with the new
intel_gt_pm_wait_for_idle
v4: No need for parked/unparked estimation if !CONFIG_PM
v5: Keep timer park/unpark logic as was
v6: Refactor duplicated estimate/update rc6 logic
v7: Pull intel_get_pm_get_if_awake() out from the pmu->lock.

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=105010
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20190912124813.19225-1-chris@chris-wilson.co.uk

drm/i915/pmu: Fix enable count array size and bounds checking

Enable count array is supposed to have one counter for each possible
engine sampler. As such, array sizing and bounds checking is not correct
and would blow up the asserts if more samplers were added.

No ill-effect in the current code base but lets fix it for correctness.

At the same time tidy the assert for readability and robustness.

v2:
* One check per assert. (Chris Wilson)

Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Fixes: b46a33e271ed ("drm/i915/pmu: Expose a PMU interface for perf queries")
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Link: https://patchwork.freedesktop.org/patch/msgid/20190205130353.21105-1-tvrtko.ursulin@linux.intel.com
(cherry picked from commit 26a11deea685b41a43edb513194718aa1f461c9a)
Signed-off-by: Jani Nikula <jani.nikula@intel.com>

drm/i915/pmu: Fix enable count array size and bounds checking

Enable count array is supposed to have one counter for each possible
engine sampler. As such, array sizing and bounds checking is not correct
and would blow up the asserts if more samplers were added.

No ill-effect in the current code base but lets fix it for correctness.

At the same time tidy the assert for readability and robustness.

v2:
* One check per assert. (Chris Wilson)

Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Fixes: b46a33e271ed ("drm/i915/pmu: Expose a PMU interface for perf queries")
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Link: https://patchwork.freedesktop.org/patch/msgid/20190205130353.21105-1-tvrtko.ursulin@linux.intel.com

drm/i915: Move on the new pm runtime interface

Use the new PM-runtime interface to get the accounted suspended time:
pm_runtime_suspended_time().

This new interface helps to simplify and cleanup the code that computes
__I915_SAMPLE_RC6_ESTIMATED and to remove direct access to internals of
PM-runtime.

Signed-off-by: Vincent Guittot <vincent.guittot@linaro.org>
Reviewed-by: Ulf Hansson <ulf.hansson@linaro.org>
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>

drm/i915/pmu: Do not assume fixed hrtimer period

As Chris has discovered on his Ivybridge, and later automated test runs
have confirmed, on most of our platforms hrtimer faced with heavy GPU load
can occasionally become sufficiently imprecise to affect PMU sampling
calculations.

This means we cannot assume sampling frequency is what we asked for, but
we need to measure the interval ourselves.

This patch is similar to Chris' original proposal for per-engine counters,
but instead of introducing a new set to work around the problem with
frequency sampling, it swaps around the way internal frequency accounting
is done. Instead of accumulating current frequency and dividing by
sampling frequency on readout, it accumulates frequency scaled by each
period.

v2:
* Typo in commit message, comment on period calculation and USEC_PER_SEC.
(Chris Wilson)

Testcase: igt/perf_pmu/*busy* # snb, ivb, hsw
Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Suggested-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Link: https://patchwork.freedesktop.org/patch/msgid/20180605140253.3541-1-tvrtko.ursulin@linux.intel.com

drm/i915: Make header i915_pmu.h more robust

Definitions in i915_pmu.h header depend on other types and
declarations that were not explicitly included. Fix that by
adding related headers and forward declarations.
While here, change license text to SPDX format.

v2: don't drop "intel_ringbuffer.h" (Tvrtko)

Signed-off-by: Michal Wajdeczko <michal.wajdeczko@intel.com>
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Link: https://patchwork.freedesktop.org/patch/msgid/20180308095037.18264-4-michal.wajdeczko@intel.com

drm/i915/pmu: Fix sleep under atomic in RC6 readout

We are not allowed to call intel_runtime_pm_get from the PMU counter read
callback since the former can sleep, and the latter is running under IRQ
context.

To workaround this, we record the last known RC6 and while runtime
suspended estimate its increase by querying the runtime PM core
timestamps.

Downside of this approach is that we can temporarily lose a chunk of RC6
time, from the last PMU read-out to runtime suspend entry, but that will
eventually catch up, once device comes back online and in the presence of
PMU queries.

Also, we have to be careful not to overshoot the RC6 estimate, so once
resumed after a period of approximation, we only update the counter once
it catches up. With the observation that RC6 is increasing while the
device is suspended, this should not pose a problem and can only cause
slight inaccuracies due clock base differences.

v2: Simplify by estimating on top of PM core counters. (Imre)

Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=104943
Fixes: 6060b6aec03c ("drm/i915/pmu: Add RC6 residency metrics")
Testcase: igt/perf_pmu/rc6-runtime-pm
Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Imre Deak <imre.deak@intel.com>
Cc: Jani Nikula <jani.nikula@linux.intel.com>
Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Cc: Rodrigo Vivi <rodrigo.vivi@intel.com>
Cc: David Airlie <airlied@linux.ie>
Cc: intel-gfx@lists.freedesktop.org
Cc: dri-devel@lists.freedesktop.org
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Link: https://patchwork.freedesktop.org/patch/msgid/20180206183311.17924-1-tvrtko.ursulin@linux.intel.com
(cherry picked from commit 1fe699e30113ed6f6e853ff44710d256072ea627)
Signed-off-by: Rodrigo Vivi <rodrigo.vivi@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20180213095747.2424-3-tvrtko.ursulin@linux.intel.com

drm/i915/pmu: Fix sleep under atomic in RC6 readout

We are not allowed to call intel_runtime_pm_get from the PMU counter read
callback since the former can sleep, and the latter is running under IRQ
context.

To workaround this, we record the last known RC6 and while runtime
suspended estimate its increase by querying the runtime PM core
timestamps.

Downside of this approach is that we can temporarily lose a chunk of RC6
time, from the last PMU read-out to runtime suspend entry, but that will
eventually catch up, once device comes back online and in the presence of
PMU queries.

Also, we have to be careful not to overshoot the RC6 estimate, so once
resumed after a period of approximation, we only update the counter once
it catches up. With the observation that RC6 is increasing while the
device is suspended, this should not pose a problem and can only cause
slight inaccuracies due clock base differences.

v2: Simplify by estimating on top of PM core counters. (Imre)

Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=104943
Fixes: 6060b6aec03c ("drm/i915/pmu: Add RC6 residency metrics")
Testcase: igt/perf_pmu/rc6-runtime-pm
Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Imre Deak <imre.deak@intel.com>
Cc: Jani Nikula <jani.nikula@linux.intel.com>
Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Cc: Rodrigo Vivi <rodrigo.vivi@intel.com>
Cc: David Airlie <airlied@linux.ie>
Cc: intel-gfx@lists.freedesktop.org
Cc: dri-devel@lists.freedesktop.org
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Link: https://patchwork.freedesktop.org/patch/msgid/20180206183311.17924-1-tvrtko.ursulin@linux.intel.com

drm/i915/pmu: Only enumerate available counters in sysfs

Switch over to dynamically creating device attributes, which are in turn
used by the perf core to expose available counters in sysfs.

This way we do not expose counters which are not avaiable on the current
platform, and are so more consistent between what we reply to open
attempts via the perf_event_open(2), and what is discoverable in sysfs.

v2:
* Simplify attribute pointer freeing loop.
* Changed attr init from macro to function.
* More common error unwind. (Chris Wilson)
* Rename some locals. (Chris Wilson)

v3:
* Fixed double semi-colon. (Chris Wilson)

Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Link: https://patchwork.freedesktop.org/patch/msgid/20180111083525.32394-1-tvrtko.ursulin@linux.intel.com

drm/i915/pmu: Stop averaging with the previous sample

Averaging with the previous sample brings a small statistical improvement
to sampling counters, but can leek a little bit of state from a current
client to the next which mulls the border between past and present for
observing clients.

This is because on event enable clients record the current counter value
and use it as reference, but with rapid off-on event cycles, and due the
delayed nature of sampling timer self-disarm, previous sample value does
not get cleared under these circumstances.

Solution is to stop averaging with the previous sample. This has a small
downside of losing some precision with short and spiky signals, but the
alternatives look too complicated for the benefit.

Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Sagar Arun Kamble <sagar.a.kamble@intel.com>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Tested-by: Chris Wilson <chris@chris-wilson.co.uk>
Link: https://patchwork.freedesktop.org/patch/msgid/20171124094959.10725-1-tvrtko.ursulin@linux.intel.com

drm/i915/pmu: Suspend sampling when GPU is idle

If only a subset of events is enabled we can afford to suspend
the sampling timer when the GPU is idle and so save some cycles
and power.

v2: Rebase and limit timer even more.
v3: Rebase.
v4: Rebase.
v5: Skip action if perf PMU failed to register.
v6: Checkpatch cleanup.
v7:
* Add a common helper to start the timer if needed. (Chris Wilson)
* Add comment explaining bitwise logic in pmu_needs_timer.
v8: Fix some comments styles. (Chris Wilson)
v9: Rebase.
v10: Move function declarations to i915_pmu.h.
v11: Rename functions to i915_pmu_gt_(un)parked. (Chris Wilson)

Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Link: https://patchwork.freedesktop.org/patch/msgid/20171121181852.16128-3-tvrtko.ursulin@linux.intel.com

drm/i915/pmu: Expose a PMU interface for perf queries

From: Chris Wilson <chris@chris-wilson.co.uk>
From: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
From: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com>

The first goal is to be able to measure GPU (and invidual ring) busyness
without having to poll registers from userspace. (Which not only incurs
holding the forcewake lock indefinitely, perturbing the system, but also
runs the risk of hanging the machine.) As an alternative we can use the
perf event counter interface to sample the ring registers periodically
and send those results to userspace.

Functionality we are exporting to userspace is via the existing perf PMU
API and can be exercised via the existing tools. For example:

perf stat -a -e i915/rcs0-busy/ -I 1000

Will print the render engine busynnes once per second. All the performance
counters can be enumerated (perf list) and have their unit of measure
correctly reported in sysfs.

v1-v2 (Chris Wilson):

v2: Use a common timer for the ring sampling.

v3: (Tvrtko Ursulin)
* Decouple uAPI from i915 engine ids.
* Complete uAPI defines.
* Refactor some code to helpers for clarity.
* Skip sampling disabled engines.
* Expose counters in sysfs.
* Pass in fake regs to avoid null ptr deref in perf core.
* Convert to class/instance uAPI.
* Use shared driver code for rc6 residency, power and frequency.

v4: (Dmitry Rogozhkin)
* Register PMU with .task_ctx_nr=perf_invalid_context
* Expose cpumask for the PMU with the single CPU in the mask
* Properly support pmu->stop(): it should call pmu->read()
* Properly support pmu->del(): it should call stop(event, PERF_EF_UPDATE)
* Introduce refcounting of event subscriptions.
* Make pmu.busy_stats a refcounter to avoid busy stats going away
with some deleted event.
* Expose cpumask for i915 PMU to avoid multiple events creation of
the same type followed by counter aggregation by perf-stat.
* Track CPUs getting online/offline to migrate perf context. If (likely)
cpumask will initially set CPU0, CONFIG_BOOTPARAM_HOTPLUG_CPU0 will be
needed to see effect of CPU status tracking.
* End result is that only global events are supported and perf stat
works correctly.
* Deny perf driver level sampling - it is prohibited for uncore PMU.

v5: (Tvrtko Ursulin)

* Don't hardcode number of engine samplers.
* Rewrite event ref-counting for correctness and simplicity.
* Store initial counter value when starting already enabled events
to correctly report values to all listeners.
* Fix RC6 residency readout.
* Comments, GPL header.

v6:
* Add missing entry to v4 changelog.
* Fix accounting in CPU hotplug case by copying the approach from
arch/x86/events/intel/cstate.c. (Dmitry Rogozhkin)

v7:
* Log failure message only on failure.
* Remove CPU hotplug notification state on unregister.

v8:
* Fix error unwind on failed registration.
* Checkpatch cleanup.

v9:
* Drop the energy metric, it is available via intel_rapl_perf.
(Ville Syrjälä)
* Use HAS_RC6(p). (Chris Wilson)
* Handle unsupported non-engine events. (Dmitry Rogozhkin)
* Rebase for intel_rc6_residency_ns needing caller managed
runtime pm.
* Drop HAS_RC6 checks from the read callback since creating those
events will be rejected at init time already.
* Add counter units to sysfs so perf stat output is nicer.
* Cleanup the attribute tables for brevity and readability.

v10:
* Fixed queued accounting.

v11:
* Move intel_engine_lookup_user to intel_engine_cs.c
* Commit update. (Joonas Lahtinen)

v12:
* More accurate sampling. (Chris Wilson)
* Store and report frequency in MHz for better usability from
perf stat.
* Removed metrics: queued, interrupts, rc6 counters.
* Sample engine busyness based on seqno difference only
for less MMIO (and forcewake) on all platforms. (Chris Wilson)

v13:
* Comment spelling, use mul_u32_u32 to work around potential GCC
issue and somne code alignment changes. (Chris Wilson)

v14:
* Rebase.

v15:
* Rebase for RPS refactoring.

v16:
* Use the dynamic slot in the CPU hotplug state machine so that we are
free to setup our state as multi-instance. Previously we were re-using
the CPUHP_AP_PERF_X86_UNCORE_ONLINE slot which is neither used as
multi-instance, nor owned by our driver to start with.
* Register the CPU hotplug handlers after the PMU, otherwise the callback
will get called before the PMU is initialized which can end up in
perf_pmu_migrate_context with an un-initialized base.
* Added workaround for a probable bug in cpuhp core.

v17:
* Remove workaround for the cpuhp bug.

v18:
* Rebase for drm_i915_gem_engine_class getting upstream before us.

v19:
* Rebase. (trivial)

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Signed-off-by: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com>
Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Dmitry Rogozhkin <dmitry.v.rogozhkin@intel.com>
Cc: Peter Zijlstra <peterz@infradead.org>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20171121181852.16128-2-tvrtko.ursulin@linux.intel.com