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: 16ffe73c18 ("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
We used to report the minimum possible frequency as both requested and
active while GPU was in sleep state. This was a consequence of sampling
the value from the "current frequency" field in our software tracking.
This was strictly speaking wrong, but given that until recently the
current frequency in sleeping state used to be equal to minimum, it did
not stand out sufficiently to be noticed as such.
After some recent changes have made the current frequency be reported
as last active before GPU went to sleep, meaning both requested and active
frequencies could end up being reported at their maximum values for the
duration of the GPU idle state, it became much more obvious that this does
not make sense.
To fix this we will now sample the frequency counters only when the GPU is
awake. As a consequence reported frequencies could be reported as below
the GPU reported minimum but that should be much less confusing that the
current situation.
v2:
* Split out early exit conditions for readability. (Chris)
Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Closes: https://gitlab.freedesktop.org/drm/intel/issues/675
Link: https://patchwork.freedesktop.org/patch/msgid/20191129105436.20100-1-tvrtko.ursulin@linux.intel.com
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
Since d0aa694b92 ("drm/i915/pmu: Always sample an active ringbuffer")
the cost of sampling the engine state on execlists platforms became a
little bit higher when both engine busyness and one of the wait states are
being monitored. (Previously the busyness sampling on legacy platforms was
done via seqno comparison so there was no cost of mmio read.)
We can avoid that by skipping busyness sampling when engine supports
software busy stats and so avoid the cost of potential mmio read and
sample accumulation.
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/20190911160730.22687-1-tvrtko.ursulin@linux.intel.com
Pull drm updates from Dave Airlie:
"This is the main drm pull request for the 5.1 merge window.
The big changes I'd highlight are:
- nouveau has HMM support now, there is finally an in-tree user so we
can quieten down the rip it out people.
- i915 now enables fastboot by default on Skylake+
- Displayport Multistream support has been refactored and should
hopefully be more reliable.
Core:
- header cleanups aiming towards removing drmP.h
- dma-buf fence seqnos to 64-bits
- common helper for DP mst hotplug for radeon,i915,amdgpu + new
refcounting scheme
- MST i2c improvements
- drm_syncobj_cb removal
- ARM FB compression fourcc
- P010 + P016 fourcc
- allwinner tiled format modifier
- i2c over aux I2C_M_STOP support
- DRM_AUTH handling fixes
TTM:
- ref/unref renaming
New driver:
- ARM komeda display driver
scheduler:
- refactor mirror list handling
- rework hw fence processing
- 0 run queue entity fix
bridge:
- TI DS90C185 LVDS bridge
- thc631lvdm83d bridge improvements
- cadence + allwinner DSI ported to generic phy
panels:
- Sitronix ST7701 panel
- Kingdisplay KD097D04
- LeMaker BL035-RGB-002
- PDA 91-00156-A0
- Innolux EE101IA-01D
i915:
- Enable fastboot by default on SKL+/VLV/CHV
- Export RPCS configuration for ICL media driver
- Coffelake PCI ID
- CNL clocks setup fixes
- ACPI/PMIC support for MIPI/DSI
- Per-engine WA init for all engines
- Shrinker locking fixes
- Kerneldoc updates
- Lots of ring improvements and reset fixes
- Coffeelake GVT Support
- VFIO GVT EDID Region support
- runtime PM wakeref tracking
- ILK->IVB primary plane enable delays
- userptr mutex locking fixes
- DSI fixes
- LVDS/TV cleanups
- HW readout fixes
- LUT robustness fixes
- ICL display and watermark fixes
- gem mmap race fix
amdgpu:
- add scheduled dependencies interface
- DCC on scanout surfaces
- vega10/20 BACO support
- Multiple IH rings on soc15
- XGMI locking fixes
- DC i2c/aux cleanups
- runtime SMU debug interface
- Kexec improvmeents
- SR-IOV fixes
- DC freesync + ABM fixes
- GDS fixes
- GPUVM fixes
- vega20 PCIE DPM switching fixes
- Context priority handling fixes
radeon:
- fix missing break in evergreen parser
nouveau:
- SVM support via HMM
msm:
- QCOM Compressed modifier support
exynos:
- s5pv210 rotator support
imx:
- zpos property support
- pending update fixes
v3d:
- cache flush improvments
vc4:
- reflection support
- HDMI overscan support
tegra:
- CEC refactoring
- HDMI audio fixes
- Tegra186 prep work
- SOR crossbar device tree fixes
sun4i:
- implicit fencing support
- YUV and scalar support improvements
- A23 support
- tiling fixes
atmel-hlcdc:
- clipping and rotation property fixes
qxl:
- BO and PRIME improvements
- generic fbdev emulation
dw-hdmi:
- HDMI 2.0 2160p
- YUV420 ouput
rockchip:
- implicit fencing support
- reflection proerties
virtio-gpu:
- use generic fbdev emulation
tilcdc:
- cpufreq vs crtc init fix
rcar-du:
- R8A774C0 support
- D3/E3 RGB output routing fixes and DPAD0 support
- RA87744 LVDS support
bochs:
- atomic and generic fbdev emulation
- ID mismatch error on bochs load
meson:
- remove firmware fbs"
* tag 'drm-next-2019-03-06' of git://anongit.freedesktop.org/drm/drm: (1130 commits)
drm/amd/display: Use vrr friendly pageflip throttling in DC.
drm/imx: only send commit done event when all state has been applied
drm/imx: allow building under COMPILE_TEST
drm/imx: imx-tve: depend on COMMON_CLK
drm/imx: ipuv3-plane: add zpos property
drm/imx: ipuv3-plane: add function to query atomic update status
gpu: ipu-v3: prg: add function to get channel configure status
gpu: ipu-v3: pre: add double buffer status readback
drm/amdgpu: Bump amdgpu version for context priority override.
drm/amdgpu/powerplay: fix typo in BACO header guards
drm/amdgpu/powerplay: fix return codes in BACO code
drm/amdgpu: add missing license on baco files
drm/bochs: Fix the ID mismatch error
drm/nouveau/dmem: use dma addresses during migration copies
drm/nouveau/dmem: use physical vram addresses during migration copies
drm/nouveau/dmem: extend copy function to allow direct use of physical addresses
drm/nouveau/svm: new ioctl to migrate process memory to GPU memory
drm/nouveau/dmem: device memory helpers for SVM
drm/nouveau/svm: initial support for shared virtual memory
drm/nouveau: prepare for enabling svm with existing userspace interfaces
...
In the next patch, we are introducing a broad virtual engine to encompass
multiple physical engines, losing the 1:1 nature of BIT(engine->id). To
reflect the broader set of engines implied by the virtual instance, lets
store the full bitmask.
v2: Use intel_engine_mask_t (s/ring_mask/engine_mask/)
v3: Tvrtko voted for moah churn so teach everyone to not mention ring
and use $class$instance throughout.
v4: Comment upon the disparity in bspec for using VCS1,VCS2 in gen8 and
VCS[0-4] in later gen. We opt to keep the code consistent and use
0-index naming throughout.
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/20190305180332.30900-1-chris@chris-wilson.co.uk
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: b46a33e271 ("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 26a11deea6)
Signed-off-by: Jani Nikula <jani.nikula@intel.com>
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: b46a33e271 ("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
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>
The majority of runtime-pm operations are bounded and scoped within a
function; these are easy to verify that the wakeref are handled
correctly. We can employ the compiler to help us, and reduce the number
of wakerefs tracked when debugging, by passing around cookies provided
by the various rpm_get functions to their rpm_put counterpart. This
makes the pairing explicit, and given the required wakeref cookie the
compiler can verify that we pass an initialised value to the rpm_put
(quite handy for double checking error paths).
For regular builds, the compiler should be able to eliminate the unused
local variables and the program growth should be minimal. Fwiw, it came
out as a net improvement as gcc was able to refactor rpm_get and
rpm_get_if_in_use together,
v2: Just s/rpm_put/rpm_put_unchecked/ everywhere, leaving the manual
mark up for smaller more targeted patches.
v3: Mention the cookie in Returns
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Jani Nikula <jani.nikula@intel.com>
Cc: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Reviewed-by: Mika Kuoppala <mika.kuoppala@linux.intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20190114142129.24398-2-chris@chris-wilson.co.uk
Pull drm updates from Dave Airlie:
"This is the main drm pull request for 4.19.
Rob has some new hardware support for new qualcomm hw that I'll send
along separately. This has the display part of it, the remaining pull
is for the acceleration engine.
This also contains a wound-wait/wait-die mutex rework, Peter has acked
it for merging via my tree.
Otherwise mostly the usual level of activity. Summary:
core:
- Wound-wait/wait-die mutex rework
- Add writeback connector type
- Add "content type" property for HDMI
- Move GEM bo to drm_framebuffer
- Initial gpu scheduler documentation
- GPU scheduler fixes for dying processes
- Console deferred fbcon takeover support
- Displayport support for CEC tunneling over AUX
panel:
- otm8009a panel driver fixes
- Innolux TV123WAM and G070Y2-L01 panel driver
- Ilitek ILI9881c panel driver
- Rocktech RK070ER9427 LCD
- EDT ETM0700G0EDH6 and EDT ETM0700G0BDH6
- DLC DLC0700YZG-1
- BOE HV070WSA-100
- newhaven, nhd-4.3-480272ef-atxl LCD
- DataImage SCF0700C48GGU18
- Sharp LQ035Q7DB03
- p079zca: Refactor to support multiple panels
tinydrm:
- ILI9341 display panel
New driver:
- vkms - virtual kms driver to testing.
i915:
- Icelake:
Display enablement
DSI support
IRQ support
Powerwell support
- GPU reset fixes and improvements
- Full ppgtt support refactoring
- PSR fixes and improvements
- Execlist improvments
- GuC related fixes
amdgpu:
- Initial amdgpu documentation
- JPEG engine support on VCN
- CIK uses powerplay by default
- Move to using core PCIE functionality for gens/lanes
- DC/Powerplay interface rework
- Stutter mode support for RV
- Vega12 Powerplay updates
- GFXOFF fixes
- GPUVM fault debugging
- Vega12 GFXOFF
- DC improvements
- DC i2c/aux changes
- UVD 7.2 fixes
- Powerplay fixes for Polaris12, CZ/ST
- command submission bo_list fixes
amdkfd:
- Raven support
- Power management fixes
udl:
- Cleanups and fixes
nouveau:
- misc fixes and cleanups.
msm:
- DPU1 support display controller in sdm845
- GPU coredump support.
vmwgfx:
- Atomic modesetting validation fixes
- Support for multisample surfaces
armada:
- Atomic modesetting support completed.
exynos:
- IPPv2 fixes
- Move g2d to component framework
- Suspend/resume support cleanups
- Driver cleanups
imx:
- CSI configuration improvements
- Driver cleanups
- Use atomic suspend/resume helpers
- ipu-v3 V4L2 XRGB32/XBGR32 support
pl111:
- Add Nomadik LCDC variant
v3d:
- GPU scheduler jobs management
sun4i:
- R40 display engine support
- TCON TOP driver
mediatek:
- MT2712 SoC support
rockchip:
- vop fixes
omapdrm:
- Workaround for DRA7 errata i932
- Fix mm_list locking
mali-dp:
- Writeback implementation
PM improvements
- Internal error reporting debugfs
tilcdc:
- Single fix for deferred probing
hdlcd:
- Teardown fixes
tda998x:
- Converted to a bridge driver.
etnaviv:
- Misc fixes"
* tag 'drm-next-2018-08-15' of git://anongit.freedesktop.org/drm/drm: (1506 commits)
drm/amdgpu/sriov: give 8s for recover vram under RUNTIME
drm/scheduler: fix param documentation
drm/i2c: tda998x: correct PLL divider calculation
drm/i2c: tda998x: get rid of private fill_modes function
drm/i2c: tda998x: move mode_valid() to bridge
drm/i2c: tda998x: register bridge outside of component helper
drm/i2c: tda998x: cleanup from previous changes
drm/i2c: tda998x: allocate tda998x_priv inside tda998x_create()
drm/i2c: tda998x: convert to bridge driver
drm/scheduler: fix timeout worker setup for out of order job completions
drm/amd/display: display connected to dp-1 does not light up
drm/amd/display: update clk for various HDMI color depths
drm/amd/display: program display clock on cache match
drm/amd/display: Add NULL check for enabling dp ss
drm/amd/display: add vbios table check for enabling dp ss
drm/amd/display: Don't share clk source between DP and HDMI
drm/amd/display: Fix DP HBR2 Eye Diagram Pattern on Carrizo
drm/amd/display: Use calculated disp_clk_khz value for dce110
drm/amd/display: Implement custom degamma lut on dcn
drm/amd/display: Destroy aux_engines only once
...
The next patch in this series will have to make the definition of
irq_cpustat_t available to entering_irq().
Inclusion of asm/hardirq.h into asm/apic.h would cause circular header
dependencies like
asm/smp.h
asm/apic.h
asm/hardirq.h
linux/irq.h
linux/topology.h
linux/smp.h
asm/smp.h
or
linux/gfp.h
linux/mmzone.h
asm/mmzone.h
asm/mmzone_64.h
asm/smp.h
asm/apic.h
asm/hardirq.h
linux/irq.h
linux/irqdesc.h
linux/kobject.h
linux/sysfs.h
linux/kernfs.h
linux/idr.h
linux/gfp.h
and others.
This causes compilation errors because of the header guards becoming
effective in the second inclusion: symbols/macros that had been defined
before wouldn't be available to intermediate headers in the #include chain
anymore.
A possible workaround would be to move the definition of irq_cpustat_t
into its own header and include that from both, asm/hardirq.h and
asm/apic.h.
However, this wouldn't solve the real problem, namely asm/harirq.h
unnecessarily pulling in all the linux/irq.h cruft: nothing in
asm/hardirq.h itself requires it. Also, note that there are some other
archs, like e.g. arm64, which don't have that #include in their
asm/hardirq.h.
Remove the linux/irq.h #include from x86' asm/hardirq.h.
Fix resulting compilation errors by adding appropriate #includes to *.c
files as needed.
Note that some of these *.c files could be cleaned up a bit wrt. to their
set of #includes, but that should better be done from separate patches, if
at all.
Signed-off-by: Nicolai Stange <nstange@suse.de>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
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
While thinking about sporadic failures of perf_pmu/rc6-runtime-pm* tests
on some CI machines I have concluded that: a) the PMU readout of RC6 can
race against runtime PM transitions, and b) there are other reasons than
being runtime suspended which can cause intel_runtime_pm_get_if_in_use to
fail.
Therefore when estimating RC6 the code needs to assert we are indeed in
suspended state, and if not, the best we can do is return the last known
RC6 value.
Without this check we can calculate the estimated value based on un-
initialized or inappropriate internal state, which can result in over-
estimation, or in any case incorrect value being returned.
v2:
* Re-arrange the code a bit to avoid second unlock and return branch.
(Chris Wilson)
v3:
* Insert some strategic blank lines and improve commit msg.
(Chris Wilson)
Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Fixes: 1fe699e301 ("drm/i915/pmu: Fix sleep under atomic in RC6 readout")
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=105010
Cc: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Imre Deak <imre.deak@intel.com>
Reviewed-by: Chris Wilson <chris@chris-wilson.co.uk>
Link: https://patchwork.freedesktop.org/patch/msgid/20180410112704.24462-1-tvrtko.ursulin@linux.intel.com
Arnd Bergman reports:
"""
The conditional spinlock confuses gcc into thinking the 'flags' value
might contain uninitialized data:
drivers/gpu/drm/i915/i915_pmu.c: In function '__i915_pmu_event_read':
arch/x86/include/asm/paravirt_types.h:573:3: error: 'flags' may be used uninitialized in this function [-Werror=maybe-uninitialized]
The code is correct, but it's easy to see how the compiler gets confused
here. This avoids the problem by pulling the lock outside of the function
into its only caller.
"""
On deeper look it seems this is caused by paravirt spinlocks
implementation when CONFIG_PARAVIRT_DEBUG is set, which by being
complicated, manages to convince gcc locked parameter can be changed
externally (impossible).
Work around it by removing the conditional locking parameters altogether.
(It was never the most elegant code anyway.)
Slight penalty we now pay is an additional irqsave spin lock/unlock cycle
on the event enable path. But since enable is not a fast path, that is
preferrable to the alternative solution which was doing MMIO under irqsave
spinlock.
Signed-off-by: Tvrtko Ursulin <tvrtko.ursulin@intel.com>
Reported-by: Arnd Bergmann <arnd@arndb.de>
Fixes: 1fe699e301 ("drm/i915/pmu: Fix sleep under atomic in RC6 readout")
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/20180314080535.17490-1-tvrtko.ursulin@linux.intel.com
