This patch updates to use register distance member instead of hardcode
in UVD7/VCE4. They are for the same ASIC.
Signed-off-by: Huang Rui <ray.huang@amd.com>
Tested-by: AnZhong Huang <anzhong.huang@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
This patch updates to use register distance member instead of
hardcode in VCN1/JEPG1. They are for the same ASIC.
Signed-off-by: Huang Rui <ray.huang@amd.com>
Tested-by: AnZhong Huang <anzhong.huang@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
This patch updates to use register distance member instead of hardcode
in GMC10.
Signed-off-by: Huang Rui <ray.huang@amd.com>
Tested-by: AnZhong Huang <anzhong.huang@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
This patch updates to use register distance member instead of hardcode
in GMC9.
Signed-off-by: Huang Rui <ray.huang@amd.com>
Tested-by: AnZhong Huang <anzhong.huang@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
This patch is to abstract register distances between two continuous
context domains and invalidation engines. In different ip headers, these
distances may be differences.
Signed-off-by: Huang Rui <ray.huang@amd.com>
Tested-by: AnZhong Huang <anzhong.huang@amd.com>
Reviewed-by: Christian König <christian.koenig@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
kfd_pasid.c isn't using the kfd2kgd interface any more. Remove redundant
code trying to look up a device for finding that interface.
Signed-off-by: Felix Kuehling <Felix.Kuehling@amd.com>
Reviewed-by: Kent Russell <kent.russell@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Move the mutext lock/unlock outside of the if(),
as the mutex is always taken: either in the if()
branch or in the else branch.
Signed-off-by: Alex Jivin <alex.jivin@amd.com>
Suggested-By: Luben Tukov <luben.tuikov@amd.com>
Reviewed-by: Luben Tuikov <luben.tuikov@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
This field is missing for dcn3 so add it from dcn20.
enc2_set_dynamic_metadata() trys to sets this and we get a
generic_reg warning since the mask is not defined.
Signed-off-by: Bhawanpreet Lakha <Bhawanpreet.Lakha@amd.com>
Reviewed-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
[WHY]
We want to be able to see SMU messages sent and their responses
Signed-off-by: Wesley Chalmers <Wesley.Chalmers@amd.com>
Reviewed-by: Joshua Aberback <Joshua.Aberback@amd.com>
Acked-by: Eryk Brol <eryk.brol@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
[Why]
Currently we don't explicitly send a request for a minimum PHYCLK, and
we hope that the dependencies other clocks have will raise PHYCLK when
needed.
[How]
- new clk_mgr function to keep track of PHYCLK requirements
- request maximum requirement across all links
- remove PHYCLK from clock state comparator, as it doesn't come from DML
Signed-off-by: Joshua Aberback <joshua.aberback@amd.com>
Reviewed-by: Jun Lei <Jun.Lei@amd.com>
Acked-by: Eryk Brol <eryk.brol@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
[Why]
VSC infoframe is needed for PSR. Previously only DMCU controller
supported PSR. Now DMUB also implements PSR.
[How]
Remove VSC infoframe dependency on DMCU.
Signed-off-by: Roman Li <roman.li@amd.com>
Reviewed-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Acked-by: Eryk Brol <eryk.brol@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
[Why]
Current DAL behaviour is to not send VSIF if mode does not support VRR
(ie. FS range is < 10Hz). However, we should still set FS Native Color
Active bit in some unsupported mode cases.
[How]
Remove check for if VRR is supported before building infopacket.
Signed-off-by: Jaehyun Chung <jaehyun.chung@amd.com>
Reviewed-by: Anthony Koo <Anthony.Koo@amd.com>
Acked-by: Eryk Brol <eryk.brol@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
[Why]
Driver re-initialize registers already set in FW
[How]
Transfer init to FW
Signed-off-by: Igor Kravchenko <Igor.Kravchenko@amd.com>
Reviewed-by: Tony Cheng <Tony.Cheng@amd.com>
Acked-by: Eryk Brol <eryk.brol@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Print scaling parameters as they are calculated in diags.
Signed-off-by: Dmytro Laktyushkin <Dmytro.Laktyushkin@amd.com>
Reviewed-by: Eric Bernstein <Eric.Bernstein@amd.com>
Acked-by: Eryk Brol <eryk.brol@amd.com>
Acked-by: Rodrigo Siqueira <Rodrigo.Siqueira@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
MMHUB powergating should be disabled on navi12 and enabled on sienna
cichlid.
Signed-off-by: Likun Gao <Likun.Gao@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Use RREG32_KIQ to read gfx register when try to get gfx/sdma
clockgating state instead of RREG32, as it will result
to system hard hang when GPU is enter into GFXOFF state.
Signed-off-by: Likun Gao <Likun.Gao@amd.com>
Reviewed-by: Hawking Zhang <Hawking.Zhang@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
add support for loading ucode with ta_firmware_header_v2_0
Reviewed-by: Bhawanpreet Lakha <Bhawanpreet.Lakha@amd.com>
Signed-off-by: John Clements <john.clements@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
TMR is required to be destoried with GFX_CMD_ID_DESTROY_TMR while the
system goes to suspend. Otherwise, PSP may return the failure state
(0xFFFF007) on Gfx-2-PSP command GFX_CMD_ID_SETUP_TMR after do multiple
times suspend/resume.
Signed-off-by: Huang Rui <ray.huang@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
Unload ASD function in suspend phase.
Signed-off-by: Huang Rui <ray.huang@amd.com>
Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Signed-off-by: Alex Deucher <alexander.deucher@amd.com>
This is identical with change for Intel platforms done with
commit 8c05246c0b ("ASoC: SOF: Intel: add min/max channels for SSP on Baytrail/Broadwell")
and fixes a regression on i.MX8/i.MX8M:
[ 25.705750] esai-Codec: ASoC: no backend playback stream
[ 27.923378] esai-Codec: ASoC: no users playback at close - state
This is root-caused to the introduction of the DAI capability checks
with snd_soc_dai_stream_valid(). Its use in soc-pcm.c makes it a
requirement for all DAIs to report at least a non-zero min_channels
field.
Fixes: 9b5db05936 ("ASoC: soc-pcm: dpcm: Only allow playback/capture if supported")
Signed-off-by: Daniel Baluta <daniel.baluta@nxp.com>
Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Reviewed-by: Ranjani Sridharan <ranjani.sridharan@linux.intel.com>
Link: https://lore.kernel.org/r/20200707210439.115300-4-pierre-louis.bossart@linux.intel.com
Signed-off-by: Mark Brown <broonie@kernel.org>
When SOF is used, the normal links are converted into DPCM ones. This
generates an error
[ 58.276668] bdw-rt5677 bdw-rt5677: CPU DAI spi-RT5677AA:00 for rtd
Wake on Voice does not support playback
[ 58.276676] bdw-rt5677 bdw-rt5677: ASoC: can't create pcm Wake on
Voice :-22
Fix by forcing the capture direction.
Fixes: b73287f0b0 ('ASoC: soc-pcm: dpcm: fix playback/capture checks')
Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Reviewed-by: Guennadi Liakhovetski <guennadi.liakhovetski@linux.intel.com>
Reviewed-by: Curtis Malainey <curtis@malainey.com>
Link: https://lore.kernel.org/r/20200707210439.115300-3-pierre-louis.bossart@linux.intel.com
Signed-off-by: Mark Brown <broonie@kernel.org>
Add a helper to walk through all the DAIs and set dpcm_playback and
dpcm_capture flags based on the DAIs capabilities, and use this helper
to avoid setting these flags arbitrarily in generic cards.
The commit referenced in the Fixes tag did not introduce the
configuration issue but will prevent the card from probing when
detecting invalid configurations.
Fixes: b73287f0b0 ('ASoC: soc-pcm: dpcm: fix playback/capture checks')
Signed-off-by: Pierre-Louis Bossart <pierre-louis.bossart@linux.intel.com>
Reviewed-by: Kai Vehmanen <kai.vehmanen@linux.intel.com>
Reviewed-by: Guennadi Liakhovetski <guennadi.liakhovetski@linux.intel.com>
Link: https://lore.kernel.org/r/20200707210439.115300-2-pierre-louis.bossart@linux.intel.com
Signed-off-by: Mark Brown <broonie@kernel.org>
Move the initialization of the vendor_part_id to be before calling
ib_register_device(), this is needed because the query_device() callback
is called from the context of ib_register_device() before initializing the
vendor_part_id, so the reported value is wrong.
Fixes: bdcf26bf9b ("rdma/siw: network and RDMA core interface")
Link: https://lore.kernel.org/r/20200707130931.444724-1-kamalheib1@gmail.com
Signed-off-by: Kamal Heib <kamalheib1@gmail.com>
Reviewed-by: Bernard Metzler <bmt@zurich.ibm.com>
Signed-off-by: Jason Gunthorpe <jgg@nvidia.com>
An extra count on ebb_state.stats.pmc_count[PMC_INDEX(pmc)] is being per-
formed when count_pmc() is used to reset PMCs on a few selftests. This
extra pmc_count can occasionally invalidate results, such as the ones from
cycles_test shown hereafter. The ebb_check_count() failed with an above
the upper limit error due to the extra value on ebb_state.stats.pmc_count.
Furthermore, this extra count is also indicated by extra PMC1 trace_log on
the output of the cycle test (as well as on pmc56_overflow_test):
==========
...
[21]: counter = 8
[22]: register SPRN_MMCR0 = 0x0000000080000080
[23]: register SPRN_PMC1 = 0x0000000080000004
[24]: counter = 9
[25]: register SPRN_MMCR0 = 0x0000000080000080
[26]: register SPRN_PMC1 = 0x0000000080000004
[27]: counter = 10
[28]: register SPRN_MMCR0 = 0x0000000080000080
[29]: register SPRN_PMC1 = 0x0000000080000004
>> [30]: register SPRN_PMC1 = 0x000000004000051e
PMC1 count (0x280000546) above upper limit 0x2800003e8 (+0x15e)
[FAIL] Test FAILED on line 52
failure: cycles
==========
Signed-off-by: Desnes A. Nunes do Rosario <desnesn@linux.ibm.com>
Tested-by: Sachin Sant <sachinp@linux.vnet.ibm.com>
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20200626164737.21943-1-desnesn@linux.ibm.com
Firmware "Selected" state is a transient state - we don't expect to see
it after finishing driver probe, we even have asserts sprinkled over
i915 to confirm whether that's the case.
Unfortunately - we don't handle the transition out of "Selected" in case
of GuC fetch error, leading those asserts to fire when calling
"intel_huc_is_used()".
v2: Add dbg print when moving HuC into error state (Daniele)
Reported-by: Marcin Bernatowicz <marcin.bernatowicz@intel.com>
Signed-off-by: Michał Winiarski <michal.winiarski@intel.com>
Cc: Chris Wilson <chris@chris-wilson.co.uk>
Cc: Daniele Ceraolo Spurio <daniele.ceraolospurio@intel.com>
Cc: Marcin Bernatowicz <marcin.bernatowicz@intel.com>
Cc: Michal Wajdeczko <michal.wajdeczko@intel.com>
Reviewed-by: Daniele Ceraolo Spurio <daniele.ceraolospurio@intel.com>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Link: https://patchwork.freedesktop.org/patch/msgid/20200708100843.297655-2-michal@hardline.pl
It has been pointed out that information about HuC usage doesn't belong
in guc_info debugfs. Let's move "supported/used/wanted" matrix to a
separate debugfs file, keeping guc_info strictly about GuC.
Suggested-by: Daniele Ceraolo Spurio <daniele.ceraolospurio@intel.com>
Signed-off-by: Michał Winiarski <michal.winiarski@intel.com>
Cc: Daniele Ceraolo Spurio <daniele.ceraolospurio@intel.com>
Cc: Lukasz Fiedorowicz <lukasz.fiedorowicz@intel.com>
Cc: Michal Wajdeczko <michal.wajdeczko@intel.com>
Reviewed-by: Daniele Ceraolo Spurio <daniele.ceraolospurio@intel.com>
Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
Link: https://patchwork.freedesktop.org/patch/msgid/20200708100843.297655-1-michal@hardline.pl
Even with one advertisement monitor in place, the scan policy should use
the whitelist while the system is going to suspend to prevent waking by
random advertisement.
The following test was performed.
- With a paired device, register one advertisement monitor, suspend
the system and verify that the host was not awaken by random
advertisements.
Signed-off-by: Miao-chen Chou <mcchou@chromium.org>
Reviewed-by: Abhishek Pandit-Subedi <abhishekpandit@chromium.org>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
sun6i_dphy_ops and sun6i_dphy_regmap_config are not modified so make them
const structs to allow the compiler to put them in read-only memory.
Before:
text data bss dec hex filename
4407 1944 64 6415 190f drivers/phy/allwinner/phy-sun6i-mipi-dphy.o
After:
text data bss dec hex filename
4835 1496 64 6395 18fb drivers/phy/allwinner/phy-sun6i-mipi-dphy.o
Signed-off-by: Rikard Falkeborn <rikard.falkeborn@gmail.com>
Acked-by: Maxime Ripard <mripard@kernel.org>
Link: https://lore.kernel.org/r/20200629195727.9717-1-rikard.falkeborn@gmail.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
ti_pipe3_power_on() sets variable 'ret' but never uses it, so remove it.
drivers/phy/ti/phy-ti-pipe3.c:340:6: warning: variable ‘ret’ set but not
used [-Wunused-but-set-variable]
Link: https://lore.kernel.org/r/20200629145010.122675-4-vkoul@kernel.org
Signed-off-by: Vinod Koul <vkoul@kernel.org>
dm816x_usb_phy_init() sets variable 'error' but never uses it, so remove
it.
drivers/phy/ti/phy-dm816x-usb.c:85:6: warning: variable ‘error’ set but
not used [-Wunused-but-set-variable]
Link: https://lore.kernel.org/r/20200629145010.122675-3-vkoul@kernel.org
Signed-off-by: Vinod Koul <vkoul@kernel.org>
Some function arguments are missing from documentation prompting
validation kernel doc script to complain:
drivers/phy/phy-core.c:1078: warning: Function parameter or member
'children' not described in '__devm_of_phy_provider_register'
drivers/phy/phy-core.c:1125: warning: Function parameter or member
'phy_provider' not described in 'devm_of_phy_provider_unregister'
Add the documentation for these
Link: https://lore.kernel.org/r/20200629145010.122675-2-vkoul@kernel.org
Signed-off-by: Vinod Koul <vkoul@kernel.org>
Documentation/process/coding-style.rst says:
"functions: they have the opening brace at the beginning of the next
line"
devm_of_phy_provider_unregister() function has opening brace at same
line, so fix it up.
Link: https://lore.kernel.org/r/20200629145010.122675-1-vkoul@kernel.org
Signed-off-by: Vinod Koul <vkoul@kernel.org>
The UFS specific QMP PHY driver started off supporting the 14nm and
20nm hardware. With the 20nm support marked broken for a long time and
the 14nm support added to the common QMP PHY, this driver has not been
used in a while. So delete it
Reviewed-by: Bjorn Andersson <bjorn.andersson@linaro.org>
Link: https://lore.kernel.org/r/20200629145452.123035-1-vkoul@kernel.org
Signed-off-by: Vinod Koul <vkoul@kernel.org>
A typo caused the interrupt handler to branch immediately to the
common "unknown interrupt" handler and skip the special case test for
denormal cause.
This does not affect KVM softpatch handling (e.g., for POWER9 TM
assist) because the KVM test was moved to common code by commit
9600f261ac ("powerpc/64s/exception: Move KVM test to common code")
just before this bug was introduced.
Fixes: 3f7fbd97d0 ("powerpc/64s/exception: Clean up SRR specifiers")
Reported-by: Paul Menzel <pmenzel@molgen.mpg.de>
Signed-off-by: Nicholas Piggin <npiggin@gmail.com>
Tested-by: Paul Menzel <pmenzel@molgen.mpg.de>
[mpe: Split selftest into a separate patch]
Signed-off-by: Michael Ellerman <mpe@ellerman.id.au>
Link: https://lore.kernel.org/r/20200708074942.1713396-1-npiggin@gmail.com
Convert the argument to the newer fwnode_handle instead a device tree
node. Fortunately, there are no users for now. So this is an easy
change.
Signed-off-by: Michael Walle <michael@walle.cc>
Link: https://lore.kernel.org/r/20200706175353.16404-2-michael@walle.cc
Signed-off-by: Mark Brown <broonie@kernel.org>
drm_hdcp_check_ksvs_revoked() returns the number of revoked keys and
error codes when the SRM parsing is failed.
Errors in SRM parsing can't affect the HDCP auth, hence with this patch,
I915 will look out for revoked key count alone.
Signed-off-by: Ramalingam C <ramalingam.c@intel.com>
cc: Sean Paul <seanpaul@chromium.org>
Reviewed-by: Sean Paul <sean@poorly.run>
Link: https://patchwork.freedesktop.org/patch/msgid/20200429134555.22106-1-ramalingam.c@intel.com
Add missing pm_runtime_get_sync() into ltdc_crtc_atomic_enable() to
match pm_runtime_put_sync() in ltdc_crtc_atomic_disable(), otherwise
the LTDC might suspend via runtime PM, disable clock, and then fail
to resume later on.
The test which triggers it is roughly -- run qt5 application which
uses eglfs platform and etnaviv, stop the application, sleep for 15
minutes, run the application again. This leads to a timeout waiting
for vsync, because the LTDC has suspended, but did not resume.
Fixes: 35ab6cfbf2 ("drm/stm: support runtime power management")
Signed-off-by: Marek Vasut <marex@denx.de>
Cc: Yannick Fertré <yannick.fertre@st.com>
Cc: Philippe Cornu <philippe.cornu@st.com>
Cc: Benjamin Gaignard <benjamin.gaignard@linaro.org>
Cc: Vincent Abriou <vincent.abriou@st.com>
Cc: Maxime Coquelin <mcoquelin.stm32@gmail.com>
Cc: Alexandre Torgue <alexandre.torgue@st.com>
To: dri-devel@lists.freedesktop.org
Cc: linux-stm32@st-md-mailman.stormreply.com
Cc: linux-arm-kernel@lists.infradead.org
Acked-by: Philippe Cornu <philippe.cornu@st.com>
Tested-by: Yannick Fertre <yannick.fertre@st.com>
Signed-off-by: Benjamin Gaignard <benjamin.gaignard@st.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20200229221649.90813-1-marex@denx.de
Content Protection property should be updated as per the kernel
internal state. Let's say if Content protection is disabled
by userspace, CP property should be set to UNDESIRED so that
reauthentication will not happen until userspace request it again,
but when kernel disables the HDCP due to any DDI disabling sequences
like modeset/DPMS operation, kernel should set the property to
DESIRED, so that when opportunity arises, kernel will start the
HDCP authentication on its own.
Somewhere in the line, state machine to set content protection to
DESIRED from kernel was broken and IGT coverage was missing for it.
This patch fixes it.
v2:
- Fixing hdcp CP state in connector atomic check function
intel_hdcp_atomic_check(). [Maarten]
This will require to check hdcp->value in intel_hdcp_update_pipe()
in order to avoid enabling hdcp, if it was already enabled.
v3:
- Rebased.
Cc: Ramalingam C <ramalingam.c@intel.com>
Cc: Maarten Lankhorst <maarten.lankhorst@linux.intel.com>
Reviewed-by: Uma Shankar <uma.shankar@intel.com>
Signed-off-by: Anshuman Gupta <anshuman.gupta@intel.com>
Link: https://patchwork.freedesktop.org/patch/350962/?series=72664&rev=2 #v1
Link: https://patchwork.freedesktop.org/patch/359396/?series=72251&rev=3 #v2
Reviewed-by: Ramalingam C <ramalingam.c@intel.com>
Signed-off-by: Ramalingam C <ramalingam.c@intel.com>
Link: https://patchwork.freedesktop.org/patch/msgid/20200630082048.22308-1-anshuman.gupta@intel.com
Add a bare tracepoint trace_sched_update_nr_running_tp which tracks
->nr_running CPU's rq. This is used to accurately trace this data and
provide a visualization of scheduler imbalances in, for example, the
form of a heat map. The tracepoint is accessed by loading an external
kernel module. An example module (forked from Qais' module and including
the pelt related tracepoints) can be found at:
https://github.com/auldp/tracepoints-helpers.git
A script to turn the trace-cmd report output into a heatmap plot can be
found at:
https://github.com/jirvoz/plot-nr-running
The tracepoints are added to add_nr_running() and sub_nr_running() which
are in kernel/sched/sched.h. In order to avoid CREATE_TRACE_POINTS in
the header a wrapper call is used and the trace/events/sched.h include
is moved before sched.h in kernel/sched/core.
Signed-off-by: Phil Auld <pauld@redhat.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lkml.kernel.org/r/20200629192303.GC120228@lorien.usersys.redhat.com
With the existing implementation of store_rps_map(), packets are queued
in the receive path on the backlog queues of other CPUs irrespective of
whether they are isolated or not. This could add a latency overhead to
any RT workload that is running on the same CPU.
Ensure that store_rps_map() only uses available housekeeping CPUs for
storing the rps_map.
Signed-off-by: Alex Belits <abelits@marvell.com>
Signed-off-by: Nitesh Narayan Lal <nitesh@redhat.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lkml.kernel.org/r/20200625223443.2684-4-nitesh@redhat.com
pci_call_probe() prevents the nesting of work_on_cpu() for a scenario
where a VF device is probed from work_on_cpu() of the PF.
Replace the cpumask used in pci_call_probe() from all online CPUs to only
housekeeping CPUs. This is to ensure that there are no additional latency
overheads caused due to the pinning of jobs on isolated CPUs.
Signed-off-by: Alex Belits <abelits@marvell.com>
Signed-off-by: Nitesh Narayan Lal <nitesh@redhat.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Frederic Weisbecker <frederic@kernel.org>
Acked-by: Bjorn Helgaas <bhelgaas@google.com>
Link: https://lkml.kernel.org/r/20200625223443.2684-3-nitesh@redhat.com
The current implementation of cpumask_local_spread() does not respect the
isolated CPUs, i.e., even if a CPU has been isolated for Real-Time task,
it will return it to the caller for pinning of its IRQ threads. Having
these unwanted IRQ threads on an isolated CPU adds up to a latency
overhead.
Restrict the CPUs that are returned for spreading IRQs only to the
available housekeeping CPUs.
Signed-off-by: Alex Belits <abelits@marvell.com>
Signed-off-by: Nitesh Narayan Lal <nitesh@redhat.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lkml.kernel.org/r/20200625223443.2684-2-nitesh@redhat.com
There is a report that when uclamp is enabled, a netperf UDP test
regresses compared to a kernel compiled without uclamp.
https://lore.kernel.org/lkml/20200529100806.GA3070@suse.de/
While investigating the root cause, there were no sign that the uclamp
code is doing anything particularly expensive but could suffer from bad
cache behavior under certain circumstances that are yet to be
understood.
https://lore.kernel.org/lkml/20200616110824.dgkkbyapn3io6wik@e107158-lin/
To reduce the pressure on the fast path anyway, add a static key that is
by default will skip executing uclamp logic in the
enqueue/dequeue_task() fast path until it's needed.
As soon as the user start using util clamp by:
1. Changing uclamp value of a task with sched_setattr()
2. Modifying the default sysctl_sched_util_clamp_{min, max}
3. Modifying the default cpu.uclamp.{min, max} value in cgroup
We flip the static key now that the user has opted to use util clamp.
Effectively re-introducing uclamp logic in the enqueue/dequeue_task()
fast path. It stays on from that point forward until the next reboot.
This should help minimize the effect of util clamp on workloads that
don't need it but still allow distros to ship their kernels with uclamp
compiled in by default.
SCHED_WARN_ON() in uclamp_rq_dec_id() was removed since now we can end
up with unbalanced call to uclamp_rq_dec_id() if we flip the key while
a task is running in the rq. Since we know it is harmless we just
quietly return if we attempt a uclamp_rq_dec_id() when
rq->uclamp[].bucket[].tasks is 0.
In schedutil, we introduce a new uclamp_is_enabled() helper which takes
the static key into account to ensure RT boosting behavior is retained.
The following results demonstrates how this helps on 2 Sockets Xeon E5
2x10-Cores system.
nouclamp uclamp uclamp-static-key
Hmean send-64 162.43 ( 0.00%) 157.84 * -2.82%* 163.39 * 0.59%*
Hmean send-128 324.71 ( 0.00%) 314.78 * -3.06%* 326.18 * 0.45%*
Hmean send-256 641.55 ( 0.00%) 628.67 * -2.01%* 648.12 * 1.02%*
Hmean send-1024 2525.28 ( 0.00%) 2448.26 * -3.05%* 2543.73 * 0.73%*
Hmean send-2048 4836.14 ( 0.00%) 4712.08 * -2.57%* 4867.69 * 0.65%*
Hmean send-3312 7540.83 ( 0.00%) 7425.45 * -1.53%* 7621.06 * 1.06%*
Hmean send-4096 9124.53 ( 0.00%) 8948.82 * -1.93%* 9276.25 * 1.66%*
Hmean send-8192 15589.67 ( 0.00%) 15486.35 * -0.66%* 15819.98 * 1.48%*
Hmean send-16384 26386.47 ( 0.00%) 25752.25 * -2.40%* 26773.74 * 1.47%*
The perf diff between nouclamp and uclamp-static-key when uclamp is
disabled in the fast path:
8.73% -1.55% [kernel.kallsyms] [k] try_to_wake_up
0.07% +0.04% [kernel.kallsyms] [k] deactivate_task
0.13% -0.02% [kernel.kallsyms] [k] activate_task
The diff between nouclamp and uclamp-static-key when uclamp is enabled
in the fast path:
8.73% -0.72% [kernel.kallsyms] [k] try_to_wake_up
0.13% +0.39% [kernel.kallsyms] [k] activate_task
0.07% +0.38% [kernel.kallsyms] [k] deactivate_task
Fixes: 69842cba9a ("sched/uclamp: Add CPU's clamp buckets refcounting")
Reported-by: Mel Gorman <mgorman@suse.de>
Signed-off-by: Qais Yousef <qais.yousef@arm.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Tested-by: Lukasz Luba <lukasz.luba@arm.com>
Link: https://lkml.kernel.org/r/20200630112123.12076-3-qais.yousef@arm.com
struct uclamp_rq was zeroed out entirely in assumption that in the first
call to uclamp_rq_inc() they'd be initialized correctly in accordance to
default settings.
But when next patch introduces a static key to skip
uclamp_rq_{inc,dec}() until userspace opts in to use uclamp, schedutil
will fail to perform any frequency changes because the
rq->uclamp[UCLAMP_MAX].value is zeroed at init and stays as such. Which
means all rqs are capped to 0 by default.
Fix it by making sure we do proper initialization at init without
relying on uclamp_rq_inc() doing it later.
Fixes: 69842cba9a ("sched/uclamp: Add CPU's clamp buckets refcounting")
Signed-off-by: Qais Yousef <qais.yousef@arm.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Valentin Schneider <valentin.schneider@arm.com>
Tested-by: Lukasz Luba <lukasz.luba@arm.com>
Link: https://lkml.kernel.org/r/20200630112123.12076-2-qais.yousef@arm.com
