Vexpress depends on motherboard firmware + spc for getting opp table. This patch
adds Vexpress glue driver for ARM big LITTLE parent driver.
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
This patch adds spc clock controller. In Vexpress cpu cluster clock is controlled via
spc controller and so it must be present in clk framework.
vexpress_clk_[of_]register_spc() registers cluster clocks with and without DT.
These are added as root clocks without any parents and their names are
"cluster[0|1|..]".
Now, platform must add clocks of all the cpus below these clusters. cpufreq
driver would get cpu clock and will do clk_get[set]_rate() on cpu clock, which
will then pass it to cluster clocks. And finally spc will get programmed.
This patch doesn't add non-DT clocks for clusters and cpus as i don't see a user
of that for now.
Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
This patch enables ARCH_HAS_CPUFREQ for Versatile Express platforms in order
to support CPU frequency scaling.
Signed-off-by: Sudeep KarkadaNagesha <Sudeep.KarkadaNagesha@arm.com>
This patch adds a psci device node to allow the ospm subsystems on the
TC2 to work with a psci backend implemented in the secure firmware. The
function offsets start from 1 instead of 0 as thats whats the current
secure firmware implements.
Signed-off-by: Achin Gupta <achin.gupta@arm.com>
Signed-off-by: Liviu Dudau <Liviu.Dudau@arm.com>
This patch introduces a shim layer for the TC2 platform which converts
'bL_platform_power_ops' routines to their psci counterparts. The psci
counterparts are implemented by the secure firmware. The shim layer
is used only when Linux is running in non-secure world and the secure
firmware implements psci.
It also introduces the use of a reference count to allow a power up call
to go ahead of a power down call.
Signed-off-by: Achin Gupta <achin.gupta@arm.com>
Signed-off-by: Liviu Dudau <Liviu.Dudau@arm.com>
This patch allows the vexpress 'tc2' native backend to probe
the dt for presence of the psci backend. If present then the native
implementation of the 'bL_platform_power_ops' is not used.
Signed-off-by: Achin Gupta <achin.gupta@arm.com>
Signed-off-by: Liviu Dudau <Liviu.Dudau@arm.com>
This patch adds constants in a tc2 specific header file to prevent
use of hard coded values for specifying the number of cpus and
clusters.
Signed-off-by: Achin Gupta <achin.gupta@arm.com>
Signed-off-by: Liviu Dudau <Liviu.Dudau@arm.com>
On TC2, all CPUs in a cluster are woken up when an IRQ event triggers for a
CPU in a cluster in shutdown state.
This patch puts spuriously woken CPUs back in reset by checking the
pending IRQ status in the SPC wake-up interrupt status register; if the
CPU has no pending IRQ routed to it, the core reexecutes wfi and it is put
in reset by FW straight away.
Tested-by: Viresh Kumar <viresh.kumar2@arm.com>
Signed-off-by: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com>
Change the init code for cpuidle-tc2 to check for a
compatible node in the devicetree of "arm,generic"
in preparation for moving it to driver/cpuidle.
Rename functions / variable from tc2_ to bl_.
Signed-off-by: mark hambleton <mahamble@broadcom.com>
Use the bL_cpu_suspend method instead of bL_cpu_power_down.
This allows for the driver to become usable on non SPC based platform
such as RTSM if vexpress_spc_check_loaded() is removed.
Signed-off-by: Nicolas Pitre <nico@linaro.org>
The functions in mcpm_entry.c are mostly intended for use during
scary cache and coherency disabling sequences, or do other things
which confuse trace ... like powering a CPU down and not
returning. Similarly for the backend code.
For simplicity, this patch just makes whole files notrace.
There should be more than enough traceable points on the paths to
these functions, but we can be more fine-grained later if there is
a need for it.
Signed-off-by: Dave Martin <dave.martin@linaro.org>
This is simplistic for the moment as the expected residency is used to
prevent the shutting down of L2 and the cluster if the residency for
the last man is lower than 5 ms. To make this right, the residency
end time for each CPU would need to be recorded and taken into account.
On a suspend, the firmware mailbox address has to be set prior entering
low power mode.
Signed-off-by: Nicolas Pitre <nico@linaro.org>
On TC2 testchip the GIC CPU IF must be disabled before powering down a
core since a pending IRQ might cause wfi completion and the processor
would exit wfi state while power controller is taking action to reset or
power up the CPU upon IRQ reception.
This patch adds code that disables the GIC CPU IF in TC2 specific
power API methods.
Signed-off-by: Lorenzo Pieralisi <lorenzo.pieralisi@arm.com>
This should be queued right before 'Revert "ARM: common: add GIC bybass disable
on GIC CPU IF save function"'.
Signed-off-by: Nicolas Pitre <nico@linaro.org>
To deal with the I-cache discrepancy between Cortex-A15 and Cortex-A7,
let's assume aliasing I-cache in both cases.
Note: this might need to be refined i.e. detect a big.LITTLE system
somehow by probing all CPUs not only the boot one.
Signed-off-by: Nicolas Pitre <nico@linaro.org>
The cpuidle code requires SPC hardware, so check for its presence
before initialising. This enables the cpuidle code to safely exist
in kernels run on hardware without SPC support.
Signed-off-by: Jon Medhurst <tixy@linaro.org>
TC2 test-chip integrates power management circuitry and firmware that
allows to remove voltage from both (A7 and A15) clusters when they are
idle or more generically when the system is forced into shutdown mode.
All CPUs in a cluster share the same voltage source so they cannot be
shutdown independently. In order to take advantage of TC2 power
management capabilities this patch implements a multi-cluster aware
CPU idle implementation. It is based on coupled C-state concept provided
by this code:
http://lists.infradead.org/pipermail/linux-arm-kernel/2012-April/097084.html
CPUs that are part of the same cluster are coupled using the mask
provided by the MPIDR at boot. Once all CPUs hit the coupled barrier the
primary CPU in the cluster (the one with MPIDR[7:0] == 0) waits for
secondaries to clean their L1 and enter wfi. Then it cleans all cache
levels, exits cluster coherency and starts the procedure to shutdown the
respective cluster. All wake-up IRQ sources are enabled by default.
Deep shutdown states for clusters are not enabled by default.
To enabled them:
A15 cluster
echo 0 > /sys/kernel/debug/idle_debug/enable_idle
A7 cluster
echo 1 > /sys/kernel/debug/idle_debug/enable_idle
Tested thoroughly using lookbusy to modulate system load and trigger idle
states entry/exit.
The patch "ARM: kernel: fix MPIDR cpu_{suspend}/{resume} usage"
uses the BFC assembler instruction but this isn't available
on ARMv6 CPUs, which breaks compilation when building kernels which
support both SMP and ARMv6, e.g. omap2plus_defconifg.
Fix this by using a BIC instruction instead.
Signed-off-by: Jon Medhurst <tixy@linaro.org>
The current version of cpu_{suspend}/{resume} relies on the 8 LSBs of
the MPIDR register to index the context pointer saved and restored on
CPU shutdown. This approach breaks as soon as platforms with populated
MPIDR affinity levels 1 and 2 are deployed, since the MPIDR cannot be
considered a linear index anymore.
There are multiple solutions to this problem, each with pros and cons.
This patch changes cpu_{suspend}/{resume} so that the CPU logical id
is used to retrieve an index into the context pointers array.
Performance is impacted on both save and restore paths. On save path
the CPU logical id has to be retrieved from thread_info; since caches
are on, the performance hit should be neglectable. In the resume code
path the MMU is off and so are the caches. The patch adds a trivial for
loop that polls the cpu_logical_map array scanning the present MPIDRs and
retrieves the actual CPU logical index. Since everything runs out of
strongly ordered memory the perfomance hit in the resume code path must
be measured and thought over; it worsens as the number of CPUs increases
since it is a linear search (but can be improved).
On the up side, the logical index approach is by far the easiest solution in
terms of coding and make dynamic changes to the cpu mapping trivial at
run-time.
Any change to the cpu_logical_map (ie in-kernel switcher) at run time must be
cleaned from the caches since this data has to be retrieved with the MMU
off, when caches are not searched.
Tested on TC2 and fast models.
The functions in mcpm_entry.c are mostly intended for use during
scary cache and coherency disabling sequences, or do other things
which confuse trace ... like powering a CPU down and not
returning. Similarly for the backend code.
For simplicity, this patch just makes whole files notrace.
There should be more than enough traceable points on the paths to
these functions, but we can be more fine-grained later if there is
a need for it.
Signed-off-by: Dave Martin <dave.martin@linaro.org>
This patch allows the vexpress 'rtsm' native backend to probe
the dt for presence of the psci backend. If present then the native
implementation of the 'bL_platform_power_ops' is not used.
Signed-off-by: Achin Gupta <achin.gupta@arm.com>
Signed-off-by: Liviu Dudau <Liviu.Dudau@arm.com>
Signed-off-by: Jon Medhurst <tixy@linaro.org>
