During load_balancing, a group with spare capacity will try to pull some
utilizations from an overloaded group. In such case, the load balance
looks for the runqueue with the highest utilization. Nevertheless, it
should also ensure that there are some pending tasks to pull otherwise
the load balance will fail to pull a task and the spread of the load will
be delayed.
This situation is quite transient but it's possible to highlight the
effect with a short run of sysbench test so the time to spread task impacts
the global result significantly.
Below are the average results for 15 iterations on an arm64 octo core:
sysbench --test=cpu --num-threads=8 --max-requests=1000 run
tip/sched/core +patchset
total time: 172ms 158ms
per-request statistics:
avg: 1.337ms 1.244ms
max: 21.191ms 10.753ms
The average max doesn't fully reflect the wide spread of the value which
ranges from 1.350ms to more than 41ms for the tip/sched/core and from
1.350ms to 21ms with the patch.
Other factors like waiting for an idle load balance or cache hotness
can delay the spreading of the tasks which explains why we can still
have up to 21ms with the patch.
Signed-off-by: Vincent Guittot <vincent.guittot@linaro.org>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lkml.kernel.org/r/20200312165429.990-1-vincent.guittot@linaro.org
During our testing, we found a case that shares no longer
working correctly, the cgroup topology is like:
/sys/fs/cgroup/cpu/A (shares=102400)
/sys/fs/cgroup/cpu/A/B (shares=2)
/sys/fs/cgroup/cpu/A/B/C (shares=1024)
/sys/fs/cgroup/cpu/D (shares=1024)
/sys/fs/cgroup/cpu/D/E (shares=1024)
/sys/fs/cgroup/cpu/D/E/F (shares=1024)
The same benchmark is running in group C & F, no other tasks are
running, the benchmark is capable to consumed all the CPUs.
We suppose the group C will win more CPU resources since it could
enjoy all the shares of group A, but it's F who wins much more.
The reason is because we have group B with shares as 2, since
A->cfs_rq.load.weight == B->se.load.weight == B->shares/nr_cpus,
so A->cfs_rq.load.weight become very small.
And in calc_group_shares() we calculate shares as:
load = max(scale_load_down(cfs_rq->load.weight), cfs_rq->avg.load_avg);
shares = (tg_shares * load) / tg_weight;
Since the 'cfs_rq->load.weight' is too small, the load become 0
after scale down, although 'tg_shares' is 102400, shares of the se
which stand for group A on root cfs_rq become 2.
While the se of D on root cfs_rq is far more bigger than 2, so it
wins the battle.
Thus when scale_load_down() scale real weight down to 0, it's no
longer telling the real story, the caller will have the wrong
information and the calculation will be buggy.
This patch add check in scale_load_down(), so the real weight will
be >= MIN_SHARES after scale, after applied the group C wins as
expected.
Suggested-by: Peter Zijlstra <peterz@infradead.org>
Signed-off-by: Michael Wang <yun.wang@linux.alibaba.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Vincent Guittot <vincent.guittot@linaro.org>
Link: https://lkml.kernel.org/r/38e8e212-59a1-64b2-b247-b6d0b52d8dc1@linux.alibaba.com
The task->flags is a 32-bits flag, in which 31 bits have already been
consumed. So it is hardly to introduce other new per process flag.
Currently there're still enough spaces in the bit-field section of
task_struct, so we can define the memstall state as a single bit in
task_struct instead.
This patch also removes an out-of-date comment pointed by Matthew.
Suggested-by: Johannes Weiner <hannes@cmpxchg.org>
Signed-off-by: Yafang Shao <laoar.shao@gmail.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Acked-by: Johannes Weiner <hannes@cmpxchg.org>
Link: https://lkml.kernel.org/r/1584408485-1921-1-git-send-email-laoar.shao@gmail.com
Add a maintainer section for psi, as it's a user-visible, configurable
kernel feature.
The patches are still routed through the scheduler tree due to the
close integration with that code, but get_maintainers.pl does the
right thing and makes sure everybody gets CCd:
$ ./scripts/get_maintainer.pl -f kernel/sched/psi.c
Johannes Weiner <hannes@cmpxchg.org> (maintainer:PRESSURE STALL INFORMATION (PSI))
Ingo Molnar <mingo@redhat.com> (maintainer:SCHEDULER)
Peter Zijlstra <peterz@infradead.org> (maintainer:SCHEDULER)
...
Reported-by: Ivan Babrou <ivan@cloudflare.com>
Signed-off-by: Johannes Weiner <hannes@cmpxchg.org>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lkml.kernel.org/r/20200316191333.115523-4-hannes@cmpxchg.org
When switching tasks running on a CPU, the psi state of a cgroup
containing both of these tasks does not change. Right now, we don't
exploit that, and can perform many unnecessary state changes in nested
hierarchies, especially when most activity comes from one leaf cgroup.
This patch implements an optimization where we only update cgroups
whose state actually changes during a task switch. These are all
cgroups that contain one task but not the other, up to the first
shared ancestor. When both tasks are in the same group, we don't need
to update anything at all.
We can identify the first shared ancestor by walking the groups of the
incoming task until we see TSK_ONCPU set on the local CPU; that's the
first group that also contains the outgoing task.
The new psi_task_switch() is similar to psi_task_change(). To allow
code reuse, move the task flag maintenance code into a new function
and the poll/avg worker wakeups into the shared psi_group_change().
Suggested-by: Peter Zijlstra <peterz@infradead.org>
Signed-off-by: Johannes Weiner <hannes@cmpxchg.org>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lkml.kernel.org/r/20200316191333.115523-3-hannes@cmpxchg.org
For simplicity, cpu pressure is defined as having more than one
runnable task on a given CPU. This works on the system-level, but it
has limitations in a cgrouped reality: When cpu.max is in use, it
doesn't capture the time in which a task is not executing on the CPU
due to throttling. Likewise, it doesn't capture the time in which a
competing cgroup is occupying the CPU - meaning it only reflects
cgroup-internal competitive pressure, not outside pressure.
Enable tracking of currently executing tasks, and then change the
definition of cpu pressure in a cgroup from
NR_RUNNING > 1
to
NR_RUNNING > ON_CPU
which will capture the effects of cpu.max as well as competition from
outside the cgroup.
After this patch, a cgroup running `stress -c 1` with a cpu.max
setting of 5000 10000 shows ~50% continuous CPU pressure.
Signed-off-by: Johannes Weiner <hannes@cmpxchg.org>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Link: https://lkml.kernel.org/r/20200316191333.115523-2-hannes@cmpxchg.org
Currently, when updating the affinity of tasks via either cpusets.cpus,
or, sched_setaffinity(); tasks not currently running within the newly
specified mask will be arbitrarily assigned to the first CPU within the
mask.
This (particularly in the case that we are restricting masks) can
result in many tasks being assigned to the first CPUs of their new
masks.
This:
1) Can induce scheduling delays while the load-balancer has a chance to
spread them between their new CPUs.
2) Can antogonize a poor load-balancer behavior where it has a
difficult time recognizing that a cross-socket imbalance has been
forced by an affinity mask.
This change adds a new cpumask interface to allow iterated calls to
distribute within the intersection of the provided masks.
The cases that this mainly affects are:
- modifying cpuset.cpus
- when tasks join a cpuset
- when modifying a task's affinity via sched_setaffinity(2)
Signed-off-by: Paul Turner <pjt@google.com>
Signed-off-by: Josh Don <joshdon@google.com>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Qais Yousef <qais.yousef@arm.com>
Tested-by: Qais Yousef <qais.yousef@arm.com>
Link: https://lkml.kernel.org/r/20200311010113.136465-1-joshdon@google.com
When a cfs rq is throttled, the latter and its child are removed from the
leaf list but their nr_running is not changed which includes staying higher
than 1. When a task is enqueued in this throttled branch, the cfs rqs must
be added back in order to ensure correct ordering in the list but this can
only happens if nr_running == 1.
When cfs bandwidth is used, we call unconditionnaly list_add_leaf_cfs_rq()
when enqueuing an entity to make sure that the complete branch will be
added.
Similarly unthrottle_cfs_rq() can stop adding cfs in the list when a parent
is throttled. Iterate the remaining entity to ensure that the complete
branch will be added in the list.
Reported-by: Christian Borntraeger <borntraeger@de.ibm.com>
Signed-off-by: Vincent Guittot <vincent.guittot@linaro.org>
Signed-off-by: Peter Zijlstra (Intel) <peterz@infradead.org>
Reviewed-by: Dietmar Eggemann <dietmar.eggemann@arm.com>
Tested-by: Christian Borntraeger <borntraeger@de.ibm.com>
Tested-by: Dietmar Eggemann <dietmar.eggemann@arm.com>
Cc: stable@vger.kernel.org
Cc: stable@vger.kernel.org #v5.1+
Link: https://lkml.kernel.org/r/20200306135257.25044-1-vincent.guittot@linaro.org
Cooling-maps doesn't have to be a required property because there may
be no cooling device on system, or there are no enough cooling devices for
each thermal zone in multiple thermal zone cases since cooling devices
can't be shared.
So make this property optional to remove such limitations.
For thermal zones with no cooling-maps, there could be critic trips
that can trigger CPU reset or shutdown. So they still can take actions.
Signed-off-by: Yuantian Tang <andy.tang@nxp.com>
Reviewed-by: Amit Kucheria <amit.kucheria@linaro.org>
Signed-off-by: Daniel Lezcano <daniel.lezcano@linaro.org>
Link: https://lore.kernel.org/r/20200309045411.21859-1-andy.tang@nxp.com
There's no need to specify 'maxItems' with the same value as the number
of entries in 'items'. A meta-schema update will catch future cases.
Cc: Andy Gross <agross@kernel.org>
Cc: Bjorn Andersson <bjorn.andersson@linaro.org>
Cc: Amit Kucheria <amit.kucheria@linaro.org>
Cc: Zhang Rui <rui.zhang@intel.com>
Cc: Daniel Lezcano <daniel.lezcano@linaro.org>
Cc: linux-arm-msm@vger.kernel.org
Cc: linux-pm@vger.kernel.org
Cc: devicetree@vger.kernel.org
Signed-off-by: Rob Herring <robh@kernel.org>
Reviewed-by: Amit Kucheria <amit.kucheria@linaro.org>
Signed-off-by: Daniel Lezcano <daniel.lezcano@linaro.org>
Link: https://lore.kernel.org/r/20200313214552.845-2-robh@kernel.org
There's no need to specify 'maxItems' with the same value as the number
of entries in 'items'. A meta-schema update will catch future cases.
Cc: Orson Zhai <orsonzhai@gmail.com>
Cc: Baolin Wang <baolin.wang7@gmail.com>
Cc: Chunyan Zhang <zhang.lyra@gmail.com>
Cc: Zhang Rui <rui.zhang@intel.com>
Cc: Daniel Lezcano <daniel.lezcano@linaro.org>
Cc: Amit Kucheria <amit.kucheria@verdurent.com>
Cc: linux-pm@vger.kernel.org
Cc: devicetree@vger.kernel.org
Signed-off-by: Rob Herring <robh@kernel.org>
Reviewed-by: Baolin Wang <baolin.wang7@gmail.com>
Signed-off-by: Daniel Lezcano <daniel.lezcano@linaro.org>
Link: https://lore.kernel.org/r/20200313214552.845-1-robh@kernel.org
imx_thermal_unregister_legacy_cooling() should be used for handling
legacy cpufreq cooling cleanups in .remove callback instead of
calling cpufreq_cooling_unregister() and cpufreq_cpu_put() directly,
especially for !CONFIG_CPU_FREQ scenario, no operation needed for
handling legacy cpufreq cooling cleanups at all.
Signed-off-by: Anson Huang <Anson.Huang@nxp.com>
Signed-off-by: Daniel Lezcano <daniel.lezcano@linaro.org>
Link: https://lore.kernel.org/r/1584088094-24857-1-git-send-email-Anson.Huang@nxp.com
Sort includes alphabetically for consistency, and take this chance
to remove unused include of of_address.h.
Signed-off-by: Anson Huang <Anson.Huang@nxp.com>
Reviewed-by: Amit Kucheria <amit.kucheria@linaro.org>
Signed-off-by: Daniel Lezcano <daniel.lezcano@linaro.org>
Link: https://lore.kernel.org/r/1583903252-2058-2-git-send-email-Anson.Huang@nxp.com
Use devm_add_action_or_reset() to handle all cleanups of failure in
.probe and .remove, then .remove callback can be dropped.
Signed-off-by: Anson Huang <Anson.Huang@nxp.com>
Reviewed-by: Amit Kucheria <amit.kucheria@linaro.org>
Signed-off-by: Daniel Lezcano <daniel.lezcano@linaro.org>
Link: https://lore.kernel.org/r/1583903252-2058-1-git-send-email-Anson.Huang@nxp.com
With the ctemp value returned instead of cached in the private data
structure their is no need to take the lock when translating ctemp into
a temperature.
Signed-off-by: Niklas Söderlund <niklas.soderlund+renesas@ragnatech.se>
Reviewed-by: Geert Uytterhoeven <geert+renesas@glider.be>
Signed-off-by: Daniel Lezcano <daniel.lezcano@linaro.org>
Link: https://lore.kernel.org/r/20200310170029.1648996-4-niklas.soderlund+renesas@ragnatech.se
There is no need to cache the ctemp value in the private data structure
as it's always prefetched before it's used. Remove it from the structure
and have rcar_thermal_update_temp return the value instead of storing
it.
Signed-off-by: Niklas Söderlund <niklas.soderlund+renesas@ragnatech.se>
Reviewed-by: Geert Uytterhoeven <geert+renesas@glider.be>
Signed-off-by: Daniel Lezcano <daniel.lezcano@linaro.org>
Link: https://lore.kernel.org/r/20200310170029.1648996-3-niklas.soderlund+renesas@ragnatech.se
Since commit a1ade56538 ("thermal: rcar: check every
rcar_thermal_update_temp() return value") the temperature is always read
in rcar_thermal_get_current_temp() so comparing it before and after
enabling interrupts have little effect. Remove the check and always
update the thermal zone when we get an interrupt that the temperature
have changed.
Signed-off-by: Niklas Söderlund <niklas.soderlund+renesas@ragnatech.se>
Reviewed-by: Geert Uytterhoeven <geert+renesas@glider.be>
Signed-off-by: Daniel Lezcano <daniel.lezcano@linaro.org>
Link: https://lore.kernel.org/r/20200310170029.1648996-2-niklas.soderlund+renesas@ragnatech.se
IRQF_TRIGGER_HIGH is already specified through devicetree interrupts
property. Remove it from code.
Signed-off-by: Amit Kucheria <amit.kucheria@linaro.org>
Reviewed-by: Bjorn Andersson <bjorn.andersson@linaro.org>
Reviewed-by: Stephen Boyd <swboyd@chromium.org>
Signed-off-by: Daniel Lezcano <daniel.lezcano@linaro.org>
Link: https://lore.kernel.org/r/8ac92e45b65fe411f4aaf70dcde4e7e7c3169b2d.1584015867.git.amit.kucheria@linaro.org
TSENS IP v2.3 onwards adds support for a watchdog to detect if the TSENS
HW FSM is stuck. Add support to detect and restart the FSM in the
driver. The watchdog is configured by the bootloader, we just enable the
watchdog bark as a debug feature in the kernel.
Signed-off-by: Amit Kucheria <amit.kucheria@linaro.org>
Reviewed-by: Bjorn Andersson <bjorn.andersson@linaro.org>
Reviewed-by: Stephen Boyd <swboyd@chromium.org>
Signed-off-by: Daniel Lezcano <daniel.lezcano@linaro.org>
Link: https://lore.kernel.org/r/a314747664a065db592ad77da7beae68128a5b6e.1584015867.git.amit.kucheria@linaro.org
TSENS IP v2.x adds critical threshold interrupt support for each sensor
in addition to the upper/lower threshold interrupt. Add support in the
driver.
While the critical interrupts themselves aren't currently used by Linux,
the HW line is also used by the TSENS watchdog. So this patch acts as
infrastructure to enable watchdog functionality for the TSENS IP.
Signed-off-by: Amit Kucheria <amit.kucheria@linaro.org>
Signed-off-by: Daniel Lezcano <daniel.lezcano@linaro.org>
Link: https://lore.kernel.org/r/51b22461d4b5f85a817274568459db4579fd4298.1584015867.git.amit.kucheria@linaro.org
We don't currently call put_device in case of successfully initialising
the device. So we hold the reference and keep the device pinned forever.
Allow control to fall through so we can use same code for success and
error paths to put_device.
As a part of this fixup, change devm_ioremap_resource to act on the same
device pointer as that used to allocate regmap memory. That ensures that
we are free to release op->dev after examining its resources.
Signed-off-by: Amit Kucheria <amit.kucheria@linaro.org>
Reviewed-by: Bjorn Andersson <bjorn.andersson@linaro.org>
Signed-off-by: Daniel Lezcano <daniel.lezcano@linaro.org>
Link: https://lore.kernel.org/r/d3996667e9f976bb30e97e301585cb1023be422e.1584015867.git.amit.kucheria@linaro.org
We already dereference the sensor and save it into a variable. Use the
variable directly to make the code easier to read.
Signed-off-by: Amit Kucheria <amit.kucheria@linaro.org>
Reviewed-by: Bjorn Andersson <bjorn.andersson@linaro.org>
Reviewed-by: Stephen Boyd <swboyd@chromium.org>
Signed-off-by: Daniel Lezcano <daniel.lezcano@linaro.org>
Link: https://lore.kernel.org/r/5dc4356edfb8dffa377fb561359bf41a6f1fdf17.1584015867.git.amit.kucheria@linaro.org
All the sensor data is initialised at init time. Lock it down by passing
it to functions as a constant.
Signed-off-by: Amit Kucheria <amit.kucheria@linaro.org>
Reviewed-by: Bjorn Andersson <bjorn.andersson@linaro.org>
Reviewed-by: Stephen Boyd <swboyd@chromium.org>
Signed-off-by: Daniel Lezcano <daniel.lezcano@linaro.org>
Link: https://lore.kernel.org/r/740f9254484c08d65869df578628eb523c0049ff.1584015867.git.amit.kucheria@linaro.org
struct tsens_features is currently initialized as part of platform data
at compile-time and not modifiable. We now have some usecases in feature
detection across IP versions where it is more flexible to update the
features after probing registers.
Remove const qualifier from tsens_features and the encapsulating
tsens_plat_data.
Signed-off-by: Amit Kucheria <amit.kucheria@linaro.org>
Reviewed-by: Stephen Boyd <swboyd@chromium.org>
Signed-off-by: Daniel Lezcano <daniel.lezcano@linaro.org>
Link: https://lore.kernel.org/r/2919a72220470366ae11e0bb5330a4ea39838f71.1584015867.git.amit.kucheria@linaro.org
If the common register memory resource is not available the driver needs
to fail gracefully to disable PM. Instead of returning the error
directly store it in ret and use the already existing error path.
Signed-off-by: Niklas Söderlund <niklas.soderlund+renesas@ragnatech.se>
Reviewed-by: Geert Uytterhoeven <geert+renesas@glider.be>
Signed-off-by: Daniel Lezcano <daniel.lezcano@linaro.org>
Link: https://lore.kernel.org/r/20200310114709.1483860-1-niklas.soderlund+renesas@ragnatech.se
Add support for the Thermal Sensor/Chip Internal Voltage Monitor in the
R-Car M3-W+ (R8A77961) SoC.
According to the R-Car Gen3 Hardware Manual Errata for Revision 2.00 of
Jan 31, 2020, the thermal parameters for R-Car M3-W+ are the same as for
R-Car M3-W.
Signed-off-by: Geert Uytterhoeven <geert+renesas@glider.be>
Reviewed-by: Niklas Söderlund <niklas.soderlund+renesas@ragnatech.se>
Signed-off-by: Daniel Lezcano <daniel.lezcano@linaro.org>
Link: https://lore.kernel.org/r/20200306105503.24267-3-geert+renesas@glider.be
Merge __dma_supported into its only caller, and move the resulting
function so that it doesn't need a forward declaration. Also mark
it static as there are no callers outside of dma-mapping.c.
Signed-off-by: Christoph Hellwig <hch@lst.de>
The DMA coherent allocator needs to take bus limits into account for
picking the zone that the memory is allocated from.
Reported-by: Roger Quadros <rogerq@ti.com>
Signed-off-by: Christoph Hellwig <hch@lst.de>
Tested-by: Peter Ujfalusi <peter.ujfalusi@ti.com>
Tested-by: Roger Quadros <rogerq@ti.com>
The core DMA code already checks for valid DMA masks earlier on, and
doesn't allow NULL struct device pointers. Remove the now not required
checks.
Signed-off-by: Christoph Hellwig <hch@lst.de>
test_vdso would try to call a NULL pointer if the vDSO was missing.
vdso_restorer_32 hit a genuine failure: trying to use the
kernel-provided signal restorer doesn't work if the vDSO is missing.
Skip the test if the vDSO is missing, since the test adds no particular
value in that case.
Reported-by: kbuild test robot <lkp@intel.com>
Signed-off-by: Andy Lutomirski <luto@kernel.org>
Signed-off-by: Borislav Petkov <bp@suse.de>
Link: https://lkml.kernel.org/r/618ea7b8c55b10d08b1cb139e9a3a957934b8647.1584653439.git.luto@kernel.org
Building an arm64 defconfig with clang's integrated assembler, this error
occurs:
<instantiation>:2:2: error: unrecognized instruction mnemonic
_ASM_EXTABLE 9999b, 9f
^
arch/arm64/mm/cache.S:50:1: note: while in macro instantiation
user_alt 9f, "dc cvau, x4", "dc civac, x4", 0
^
While GNU as seems fine with case-sensitive macro instantiations, clang
doesn't, so use the actual macro name (_asm_extable) as in the rest of
the file.
Also checked that the generated assembly matches the GCC output.
Reviewed-by: Nick Desaulniers <ndesaulniers@google.com>
Tested-by: Nick Desaulniers <ndesaulniers@google.com>
Fixes: 290622efc7 ("arm64: fix "dc cvau" cache operation on errata-affected core")
Link: https://github.com/ClangBuiltLinux/linux/issues/924
Signed-off-by: Ilie Halip <ilie.halip@gmail.com>
Signed-off-by: Will Deacon <will@kernel.org>
Include Cadence core DT schema and define the Cadence platform DT schema
for both Host and Endpoint mode. Note: The Cadence core DT schema could
be included for other platforms using Cadence PCIe core.
Signed-off-by: Kishon Vijay Abraham I <kishon@ti.com>
Signed-off-by: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com>
Reviewed-by: Rob Herring <robh@kernel.org>
Add PCIe Host (RC) and Endpoint (EP) device tree schema for Cadence
PCIe core library. Platforms using Cadence PCIe core can include the
schemas added here in the platform specific schemas.
Signed-off-by: Kishon Vijay Abraham I <kishon@ti.com>
Signed-off-by: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com>
Reviewed-by: Rob Herring <robh@kernel.org>
Define a common schema for PCI Endpoint Controllers.
Signed-off-by: Kishon Vijay Abraham I <kishon@ti.com>
Signed-off-by: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com>
Reviewed-by: Rob Herring <robh@kernel.org>
These two devres functions devm_gpiochip_[add|remove]()
were in the wrong file. They should be in gpiolib-devres.c
not gpiolib.c.
Link: https://lore.kernel.org/r/20200313081522.35143-1-linus.walleij@linaro.org
Reviewed-by: Bartosz Golaszewski <bgolaszewski@baylibre.com>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
Add a new struct for the channel parameters that is contained in the
CREATE_CHANNEL message. This is in preparation for newer firmwares that
pass the channel parameters in a dedicated buffer instead of embedding
the parameters into the CREATE_CHANNEL message.
Signed-off-by: Michael Tretter <m.tretter@pengutronix.de>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
Move the mail definitions from the driver core to a dedicated file.
The mails that are exchanged between driver and firmware are not stable
across firmware versions. This is in preparation to make the driver able
to handle multiple firmware version by having dedicated code for
handling mails.
Signed-off-by: Michael Tretter <m.tretter@pengutronix.de>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
As we know which buffers are processed by the codec from the address in
the ENCODE_FRAME response, we can queue multiple buffers in the firmware
and retrieve the buffer from the response of the firmware. This enables
the firmware to use the internal scheduling the codec and avoids round
trips through the driver when fetching the next frame.
Remove buffers that have been passed to the firmware from the m2m buffer
queue and put them into a shadow queue for tracking the buffer in the
driver. When we receive a ENCODE_FRAME response from the firmware, get
the buffer from the shadow queue and finish the buffer.
Furthermore, it is necessary to finish the job straight after passing
the buffer to the firmware to allow the V4L2 framework to send further
buffers to the driver.
Signed-off-by: Michael Tretter <m.tretter@pengutronix.de>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
The PUT_STREAM_BUFFER and ENCODE_FRAME request have fields that allow to
pass arbitrary 64 bit values through the firmware to the ENCODE_FRAME
response. Use these values to verify that the buffers when finishing the
frame are actually the same buffers that have been sent for encoding a
frame.
Signed-off-by: Michael Tretter <m.tretter@pengutronix.de>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
The peak bitrate must not be smaller than the configured bitrate. Update
the other control whenever one of the controls changes to reflect this
dependency.
Signed-off-by: Michael Tretter <m.tretter@pengutronix.de>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
There is no need to copy the bitrate mode to a field in the channel and
the value can be read directly from the control.
Signed-off-by: Michael Tretter <m.tretter@pengutronix.de>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
The V4L2_CID_MPEG_VIDEO_FRAME_RC_ENABLE control allows to enable/disable
rate control on a channel. When rate control is disabled, the driver
shall use constant QP, which are set by the application. Also implement
the controls for configuring the QP.
Signed-off-by: Michael Tretter <m.tretter@pengutronix.de>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
The allegro dvt codec adjust the encoding speed according to a
configured frame rate. Furthermore, the frame rate is written into the
coded stream.
Ensure that the coded video data has the correct frame rate by
implementing s_parm for setting the frame rate from user space.
Signed-off-by: Michael Tretter <m.tretter@pengutronix.de>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
The driver instructs the firmware to leave some space space in front of
the coded frame data for SPS/PPS data. If the driver receives an IDR, it
writes the SPS/PPS into that free space and fills the rest with filler
data. However, if there is no additional data, the driver can use the
plane offset to skip this space instead of adding filler data.
As the size of the SPS/PPS is only available after writing it, keep the
filler data between the SPS/PPS and the coded frame data.
Signed-off-by: Michael Tretter <m.tretter@pengutronix.de>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
The driver uses structs to parse the responses from the VCU and expects
a certain size of the responses. However, the size and format of the
mails is not stable across firmware versions. Therefore, print a warning
if the size does not match the expected size to warn the user that
strange things might happen.
Signed-off-by: Michael Tretter <m.tretter@pengutronix.de>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
The mcu and the codec use 32 bit addresses which are sent via the
firmware messages. Add helper functions for this address calculation to
make it obvious which address is used in the message.
As the mcu and the codec have a limited address space, print warnings if
the addresses are outside the respective address space.
Signed-off-by: Michael Tretter <m.tretter@pengutronix.de>
Signed-off-by: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Signed-off-by: Mauro Carvalho Chehab <mchehab+huawei@kernel.org>
