There's a HW race condition between OA unit tail pointer register
updates and writes to memory whereby the tail pointer can sometimes get
ahead of what's been written out to the OA buffer so far (in terms of
what's visible to the CPU).
Although this can be observed explicitly while copying reports to
userspace by checking for a zeroed report-id field in tail reports, we
want to account for this earlier, as part of the _oa_buffer_check to
avoid lots of redundant read() attempts.
Previously the driver used to define an effective tail pointer that
lagged the real pointer by a 'tail margin' measured in bytes derived
from OA_TAIL_MARGIN_NSEC and the configured sampling frequency.
Unfortunately this was flawed considering that the OA unit may also
automatically generate non-periodic reports (such as on context switch)
or the OA unit may be enabled without any periodic sampling.
This improves how we define a tail pointer for reading that lags the
real tail pointer by at least %OA_TAIL_MARGIN_NSEC nanoseconds, which
gives enough time for the corresponding reports to become visible to the
CPU.
The driver now maintains two tail pointers:
1) An 'aging' tail with an associated timestamp that is tracked until we
can trust the corresponding data is visible to the CPU; at which point
it is considered 'aged'.
2) An 'aged' tail that can be used for read()ing.
The two separate pointers let us decouple read()s from tail pointer aging.
The tail pointers are checked and updated at a limited rate within a
hrtimer callback (the same callback that is used for delivering POLLIN
events) and since we're now measuring the wall clock time elapsed since
a given tail pointer was read the mechanism no longer cares about
the OA unit's periodic sampling frequency.
The natural place to handle the tail pointer updates was in
gen7_oa_buffer_is_empty() which is called as part of blocking reads and
the hrtimer callback used for polling, and so this was renamed to
oa_buffer_check() considering the added side effect while checking
whether the buffer contains data.
Signed-off-by: Robert Bragg <robert@sixbynine.org>
Reviewed-by: Matthew Auld <matthew.auld@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/20170511154345.962-6-lionel.g.landwerlin@intel.com
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
This avoids redundantly passing an (inout) head and tail pointer to
gen7_append_oa_reports() from gen7_oa_read which doesn't need to
reference either itself.
Moving the head/tail reads and writes into gen7_append_oa_reports should
have no functional effect except to avoid some redundant head pointer
writes in cases where nothing was copied to userspace.
This is a stepping stone towards updating how the head and tail pointer
state is managed to improve the workaround for the OA unit's tail
pointer race. It reduces the number of places we need to read/write the
head and tail pointers.
Signed-off-by: Robert Bragg <robert@sixbynine.org>
Reviewed-by: Matthew Auld <matthew.auld@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/20170511154345.962-5-lionel.g.landwerlin@intel.com
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
There's no need for the driver to keep reading back the head pointer
from hardware since the hardware doesn't update it automatically. This
way we can treat any invalid head pointer value as a software/driver
bug instead of spurious hardware behaviour.
This change is also a small stepping stone towards re-working how
the head and tail state is managed as part of an improved workaround
for the tail register race condition.
Signed-off-by: Robert Bragg <robert@sixbynine.org>
Reviewed-by: Matthew Auld <matthew.auld@intel.com>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Link: http://patchwork.freedesktop.org/patch/msgid/20170511154345.962-4-lionel.g.landwerlin@intel.com
If the function for checking whether there is OA buffer data available
(during a poll or blocking read) has false positives then we want to
avoid a situation where the subsequent read() returns EAGAIN (after
a more accurate check) followed by a poll() immediately reporting
the same false positive POLLIN event and effectively maintaining a
busy loop until there really is data.
This makes sure that we clear the .pollin event status whenever we
return EAGAIN to userspace which will throttle subsequent POLLIN events
and repeated attempts to read to the 5ms intervals of the hrtimer
callback we have.
Signed-off-by: Robert Bragg <robert@sixbynine.org>
Reviewed-by: Matthew Auld <matthew.auld@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/20170511154345.962-3-lionel.g.landwerlin@intel.com
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Since unifying ringbuffer/execlist submission to use
engine->pin_context, we ensure that the intel_ring is available before
we start constructing the request. We can therefore move the assignment
of the request->ring to the central i915_gem_request_alloc() and not
require it in every engine->request_alloc() callback. Another small step
towards simplification (of the core, but at a cost of handling error
pointers in less important callers of engine->pin_context).
v2: Rearrange a few branches to reduce impact of PTR_ERR() on gcc's code
generation.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Oscar Mateo <oscar.mateo@intel.com>
Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Reviewed-by: Oscar Mateo <oscar.mateo@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/20170504093308.4137-1-chris@chris-wilson.co.uk
Don't throw a warning if we are given an invalid property id. While
here let's also bring back Robert' original idea of catching unhandled
enumeration values at compile time.
Fixes: eec688e142 ("drm/i915: Add i915 perf infrastructure")
Signed-off-by: Matthew Auld <matthew.auld@intel.com>
Cc: Robert Bragg <robert@sixbynine.org>
Reviewed-by: Robert Bragg <robert@sixbynine.org>
Signed-off-by: Mika Kuoppala <mika.kuoppala@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/20170327203236.18276-1-matthew.auld@intel.com
If we were to ever encounter a sample_flags mismatch we need to ensure
we destroy the stream when we bail.
Fixes: d79651522e ("drm/i915: Enable i915 perf stream for Haswell OA unit")
Signed-off-by: Matthew Auld <matthew.auld@intel.com>
Cc: Robert Bragg <robert@sixbynine.org>
Reviewed-by: Mika Kuoppala <mika.kuoppala@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/20170327203459.18398-1-matthew.auld@intel.com
assert_spin_locked() becomes an unconditionally compiled BUG_ON(),
adding debug code right into the heart of critical routines like
interrupt handlers.
text data bss dec hex
1296480 19944 2272 1318696 141f28 before (lockdep disabled)
1295984 19944 2272 1318200 141d38 after
1336261 21139 3208 1360608 14c2e0 before (lockdep enabled)
1339920 21139 3208 1364267 14d12b after
Small saving for release; hopefully more instructive in debug.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Cc: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
Cc: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Cc: Ville Syrjälä <ville.syrjala@linux.intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/20170302132801.599-1-chris@chris-wilson.co.uk
Reviewed-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com>
A few users only take the struct_mutex in order to release a reference
to a context. We can expose a kref_put_mutex() wrapper in order to
simplify these users, and optimise taking of the mutex to the final
unref.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/20161218153724.8439-4-chris@chris-wilson.co.uk
The requests conversion introduced a nasty bug where we could generate a
new request in the middle of constructing a request if we needed to idle
the system in order to evict space for a context. The request to idle
would be executed (and waited upon) before the current one, creating a
minor havoc in the seqno accounting, as we will consider the current
request to already be completed (prior to deferred seqno assignment) but
ring->last_retired_head would have been updated and still could allow
us to overwrite the current request before execution.
We also employed two different mechanisms to track the active context
until it was switched out. The legacy method allowed for waiting upon an
active context (it could forcibly evict any vma, including context's),
but the execlists method took a step backwards by pinning the vma for
the entire active lifespan of the context (the only way to evict was to
idle the entire GPU, not individual contexts). However, to circumvent
the tricky issue of locking (i.e. we cannot take struct_mutex at the
time of i915_gem_request_submit(), where we would want to move the
previous context onto the active tracker and unpin it), we take the
execlists approach and keep the contexts pinned until retirement.
The benefit of the execlists approach, more important for execlists than
legacy, was the reduction in work in pinning the context for each
request - as the context was kept pinned until idle, it could short
circuit the pinning for all active contexts.
We introduce new engine vfuncs to pin and unpin the context
respectively. The context is pinned at the start of the request, and
only unpinned when the following request is retired (this ensures that
the context is idle and coherent in main memory before we unpin it). We
move the engine->last_context tracking into the retirement itself
(rather than during request submission) in order to allow the submission
to be reordered or unwound without undue difficultly.
And finally an ulterior motive for unifying context handling was to
prepare for mock requests.
v2: Rename to last_retired_context, split out legacy_context tracking
for MI_SET_CONTEXT.
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Link: http://patchwork.freedesktop.org/patch/msgid/20161218153724.8439-3-chris@chris-wilson.co.uk
This adds a 'Perf' section to i915.rst with the following sub sections:
- Overview
- Comparison with Core Perf
- i915 Driver Entry Points
- i915 Perf Stream
- i915 Perf Observation Architecture Stream
- All i915 Perf Internals
v2:
section headers in i915.rst (Daniel Vetter)
missing symbol docs + other fixups (Matthew Auld)
Signed-off-by: Robert Bragg <robert@sixbynine.org>
Reviewed-by: Matthew Auld <matthew.auld@intel.com>
Cc: Daniel Vetter <daniel.vetter@ffwll.ch>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Link: http://patchwork.freedesktop.org/patch/msgid/20161207214033.3581-1-robert@sixbynine.org
Avoid using DRM_ERROR for conditions userspace can trigger with a bad
config when opening a stream or from not reading data in a timely
fashion (whereby the OA buffer fills up). These conditions are tested
by i-g-t which treats error messages as failures if using the test
runner. This wasn't an issue while the i915-perf igt tests were being
run in isolation.
One message relating to seeing a spurious zeroed report was changed to
use DRM_NOTE instead of DRM_ERROR. Ideally this warning shouldn't be
seen, but it's not a serious problem if it is. Considering that the
tail margin mechanism is only a heuristic it's possible we might see
this from time to time.
Signed-off-by: Robert Bragg <robert@sixbynine.org:
Cc: Daniel Vetter <daniel.vetter@ffwll.ch>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Link: http://patchwork.freedesktop.org/patch/msgid/20161201172152.10893-1-robert@sixbynine.org
Makes all GEM object constructors consistent.
v2: Fix compilation in GVT code.
Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Joonas Lahtinen <joonas.lahtinen@linux.intel.com> (v1)
Just a couple of naked 64bit divides causing link errors on 32bit
builds, with:
ERROR: "__udivdi3" [drivers/gpu/drm/i915/i915.ko] undefined!
v2: do_div() is only u64/u32, we need a u32/u64!
v3: div_u64() == u64/u32, div64_u64() == u64/u64
Reported-by: kbuild test robot <fengguang.wu@intel.com>
Fixes: d79651522e ("drm/i915: Enable i915 perf stream for Haswell OA unit")
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Robert Bragg <robert@sixbynine.org>
Link: http://patchwork.freedesktop.org/patch/msgid/20161123150714.24449-1-chris@chris-wilson.co.uk
Reviewed-by: Robert Bragg <robert@sixbynine.org>
In particular this tries to capture for posterity some of the early
challenges we had with using the core perf infrastructure in case we
ever want to revisit adapting perf for device metrics.
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Signed-off-by: Robert Bragg <robert@sixbynine.org>
Reviewed-by: Matthew Auld <matthew.auld@intel.com>
Reviewed-by: Sourab Gupta <sourab.gupta@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Link: http://patchwork.freedesktop.org/patch/msgid/20161107194957.3385-12-robert@sixbynine.org
The maximum OA sampling frequency is now configurable via a
dev.i915.oa_max_sample_rate sysctl parameter.
Following the precedent set by perf's similar
kernel.perf_event_max_sample_rate the default maximum rate is 100000Hz
Signed-off-by: Robert Bragg <robert@sixbynine.org>
Reviewed-by: Matthew Auld <matthew.auld@intel.com>
Reviewed-by: Sourab Gupta <sourab.gupta@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Link: http://patchwork.freedesktop.org/patch/msgid/20161107194957.3385-10-robert@sixbynine.org
Consistent with the kernel.perf_event_paranoid sysctl option that can
allow non-root users to access system wide cpu metrics, this can
optionally allow non-root users to access system wide OA counter metrics
from Gen graphics hardware.
Signed-off-by: Robert Bragg <robert@sixbynine.org>
Reviewed-by: Matthew Auld <matthew.auld@intel.com>
Reviewed-by: Sourab Gupta <sourab.gupta@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Link: http://patchwork.freedesktop.org/patch/msgid/20161107194957.3385-9-robert@sixbynine.org
Each metric set is given a sysfs entry like:
/sys/class/drm/card0/metrics/<guid>/id
This allows userspace to enumerate the specific sets that are available
for the current system. The 'id' file contains an unsigned integer that
can be used to open the associated metric set via
DRM_IOCTL_I915_PERF_OPEN. The <guid> is a globally unique ID for a
specific OA unit register configuration that can be reliably used by
userspace as a key to lookup corresponding counter meta data and
normalization equations.
The guid registry is currently maintained as part of gputop along with
the XML metric set descriptions and code generation scripts, ref:
https://github.com/rib/gputop
> gputop-data/guids.xml
> scripts/update-guids.py
> gputop-data/oa-*.xml
> scripts/i915-perf-kernelgen.py
$ make -C gputop-data -f Makefile.xml SYSFS=1 WHITELIST=RenderBasic
Signed-off-by: Robert Bragg <robert@sixbynine.org>
Reviewed-by: Matthew Auld <matthew.auld@intel.com>
Reviewed-by: Sourab Gupta <sourab.gupta@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Link: http://patchwork.freedesktop.org/patch/msgid/20161107194957.3385-8-robert@sixbynine.org
Gen graphics hardware can be set up to periodically write snapshots of
performance counters into a circular buffer via its Observation
Architecture and this patch exposes that capability to userspace via the
i915 perf interface.
v2:
Make sure to initialize ->specific_ctx_id when opening, without
relying on _pin_notify hook, in case ctx already pinned.
v3:
Revert back to pinning ctx upfront when opening stream, removing
need to hook in to pinning and to update OACONTROL on the fly.
Signed-off-by: Robert Bragg <robert@sixbynine.org>
Signed-off-by: Zhenyu Wang <zhenyuw@linux.intel.com>
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Reviewed-by: Matthew Auld <matthew.auld@intel.com>
Reviewed-by: Sourab Gupta <sourab.gupta@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Link: http://patchwork.freedesktop.org/patch/msgid/20161107194957.3385-7-robert@sixbynine.org
Adds base i915 perf infrastructure for Gen performance metrics.
This adds a DRM_IOCTL_I915_PERF_OPEN ioctl that takes an array of uint64
properties to configure a stream of metrics and returns a new fd usable
with standard VFS system calls including read() to read typed and sized
records; ioctl() to enable or disable capture and poll() to wait for
data.
A stream is opened something like:
uint64_t properties[] = {
/* Single context sampling */
DRM_I915_PERF_PROP_CTX_HANDLE, ctx_handle,
/* Include OA reports in samples */
DRM_I915_PERF_PROP_SAMPLE_OA, true,
/* OA unit configuration */
DRM_I915_PERF_PROP_OA_METRICS_SET, metrics_set_id,
DRM_I915_PERF_PROP_OA_FORMAT, report_format,
DRM_I915_PERF_PROP_OA_EXPONENT, period_exponent,
};
struct drm_i915_perf_open_param parm = {
.flags = I915_PERF_FLAG_FD_CLOEXEC |
I915_PERF_FLAG_FD_NONBLOCK |
I915_PERF_FLAG_DISABLED,
.properties_ptr = (uint64_t)properties,
.num_properties = sizeof(properties) / 16,
};
int fd = drmIoctl(drm_fd, DRM_IOCTL_I915_PERF_OPEN, ¶m);
Records read all start with a common { type, size } header with
DRM_I915_PERF_RECORD_SAMPLE being of most interest. Sample records
contain an extensible number of fields and it's the
DRM_I915_PERF_PROP_SAMPLE_xyz properties given when opening that
determine what's included in every sample.
No specific streams are supported yet so any attempt to open a stream
will return an error.
v2:
use i915_gem_context_get() - Chris Wilson
v3:
update read() interface to avoid passing state struct - Chris Wilson
fix some rebase fallout, with i915-perf init/deinit
v4:
s/DRM_IORW/DRM_IOW/ - Emil Velikov
Signed-off-by: Robert Bragg <robert@sixbynine.org>
Reviewed-by: Matthew Auld <matthew.auld@intel.com>
Reviewed-by: Sourab Gupta <sourab.gupta@intel.com>
Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch>
Link: http://patchwork.freedesktop.org/patch/msgid/20161107194957.3385-2-robert@sixbynine.org
