In irq_safe_dev_in_sleep_domain() we correctly skip the dev_warn_once() if
the corresponding genpd for the device, has the GENPD_FLAG_ALWAYS_ON flag
being set. For the same reason (the genpd is always-on in runtime), let's
also skip the warning if the GENPD_FLAG_RPM_ALWAYS_ON flag is set for the
genpd.
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
The name "irq_safe_dev_in_no_sleep_domain", doesn't really match the
conditions that are being checked in the function, hence the code becomes a
bit confusing to read.
Let's clarify this by renaming it into "irq_safe_dev_in_sleep_domain" and
let's also take the opportunity to clarify a corresponding comment in the
code.
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Back in the days when genpd supported intermediate power states of its
devices, it made sense to check the PM_QOS_FLAG_NO_POWER_OFF in
genpd_power_off(). This because the attached devices were all being put
into low power state together when the PM domain was also being powered
off.
At this point, the flag PM_QOS_FLAG_NO_POWER_OFF is better checked by
drivers from their ->runtime_suspend() callbacks, like in the
usb_port_runtime_suspend(), for example. Or perhaps an even better option
is to set the QoS resume latency constraint for the device to zero, which
informs the runtime PM core to prevent the device from being runtime
suspended.
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Due to recent changes, the always-on governor is always used with a genpd
that has the GENPD_FLAG_RPM_ALWAYS_ON flag being set. This means genpd,
doesn't invoke the governor's ->power_down_ok() callback, which makes the
code in the governor redundant, so let's drop it.
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Rather than relying on the genpd provider to set the corresponding flag,
GENPD_FLAG_RPM_ALWAYS_ON, when the always-on governor is being used, let's
add it in pm_genpd_init(). In this way, it starts to benefits all genpd
providers immediately.
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
When calling idxd_wq_enable() and wq is already enabled, code should return 0
and indicate function is successful instead of return error code and fail.
This should also put idxd_wq_enable() in sync with idxd_wq_disable() where
it returns 0 if wq is already disabled.
Signed-off-by: Dave Jiang <dave.jiang@intel.com>
Link: https://lore.kernel.org/r/165090980906.1378449.1939401700832432886.stgit@djiang5-desk3.ch.intel.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
So far it appears to match the configuration of the A100 variant.
Since D1 is a RISC-V chip, it does not meet any of the existing
dependencies for this driver, so relax the dependency somewhat.
Acked-by: Maxime Ripard <maxime@cerno.tech>
Signed-off-by: Samuel Holland <samuel@sholland.org>
Link: https://lore.kernel.org/r/20220424172759.33383-5-samuel@sholland.org
Signed-off-by: Vinod Koul <vkoul@kernel.org>
Recent Allwinner SoCs support >4 GiB of DRAM, so those variants of the
DMA engine support >32 bit physical addresses. This is accomplished by
placing the high bits in the "para" word in the DMA descriptor.
DMA descriptors themselves can be located at >32 bit addresses by
putting the high bits in the LSBs of the descriptor address register,
taking advantage of the required DMA descriptor alignment. However,
support for this is not really necessary, so we can avoid the
complication by allocating them from the DMA_32 zone.
Signed-off-by: Samuel Holland <samuel@sholland.org>
Acked-by: Jernej Skrabec <jernej.skrabec@gmail.com>
Link: https://lore.kernel.org/r/20220424172759.33383-4-samuel@sholland.org
Signed-off-by: Vinod Koul <vkoul@kernel.org>
This breaks on RISC-V, because dma_pool_alloc returns addresses which
are not in the linear map. Instead, plumb through the physical address
which is already known anyway.
Acked-by: Maxime Ripard <maxime@cerno.tech>
Signed-off-by: Samuel Holland <samuel@sholland.org>
Link: https://lore.kernel.org/r/20220424172759.33383-3-samuel@sholland.org
Signed-off-by: Vinod Koul <vkoul@kernel.org>
D1 has a DMA controller similar to the one in other Allwinner SoCs.
Add its compatible, and include it in the list of variants with a
separate MBUS clock gate.
Acked-by: Rob Herring <robh@kernel.org>
Acked-by: Maxime Ripard <maxime@cerno.tech>
Signed-off-by: Samuel Holland <samuel@sholland.org>
Link: https://lore.kernel.org/r/20220424172759.33383-2-samuel@sholland.org
Signed-off-by: Vinod Koul <vkoul@kernel.org>
Initialize slave_bw in dma_prep*() functions as the parameter is not
set for DMA_MEM_TO_MEM case in get_transfer_param(). Though the case
may never occur, initializing it avoids warning from certain static
checkers
Fixes: ee17028009 ("dmaengine: tegra: Add tegra gpcdma driver")
Reported-by: Dan Carpenter <dan.carpenter@oracle.com>
Signed-off-by: Akhil R <akhilrajeev@nvidia.com>
Link: https://lore.kernel.org/r/20220426101913.43335-2-akhilrajeev@nvidia.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
At any time, a DMA transfer can be suspended to be restarted later before
the end of the DMA transfer.
In order to restart from the point where the transfer was stopped,
DMA_SxNDTR has to be read after disabling the channel by clearing the EN
bit in DMA_SxCR register, to know the number of data items already
collected.
Peripheral and/or memory addresses have to be updated in order to adjust
the address pointers.
SxNDTR register has to be updated with the remaining number of data items
to be transferred (the value read when the channel was disabled).
Then the channel can be re-enabled to resume the transfer from the point
it was suspended.
If the channel was configured in circular or double-buffer mode, the
circular or double-buffer mode must be disabled before re-enabling the
channel to be able to reconfigure SxNDTR register and re-activate circular
or double-buffer mode on next Transfer Complete interrupt where channel
will be disabled by HW. This is due to the fact that on resume, re-writing
SxNDTR register value updates internal HW auto-reload data counter, and
then it truncates all next transfers after a pause/resume sequence.
Signed-off-by: Amelie Delaunay <amelie.delaunay@foss.st.com>
Link: https://lore.kernel.org/r/20220505115611.38845-5-amelie.delaunay@foss.st.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
dmaengine framework offers device_pause and device_resume ops to pause an
on-going transfer and resume it later.
To avoid any misunderstanding with system sleep pm ops, rename pm ops into
stm32_dma_pm_suspend and stm32_dma_pm_resume.
Signed-off-by: Amelie Delaunay <amelie.delaunay@foss.st.com>
Link: https://lore.kernel.org/r/20220505115611.38845-4-amelie.delaunay@foss.st.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
stm32_dma_handle_chan_done() is called on Transfer Complete interrupt.
As DMA_SxSCR register is read in interrupt handler, pass the value as
parameter of stm32_dma_handle_chan_done(). Also return directly if
chan->desc is null to remove one ident level.
Then, stm32_dma_configure_next_sg() is doing something only if
Double-Buffer Mode (DBM) is enabled, so, check it is enabled prior calling
stm32_dma_configure_next_sg(), to remove one ident level in
stm32_dma_configure_next_sg().
Signed-off-by: Amelie Delaunay <amelie.delaunay@foss.st.com>
Link: https://lore.kernel.org/r/20220505115611.38845-3-amelie.delaunay@foss.st.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
chan->next_sg is used to know which transfer will start after the ongoing
one. It is incremented for each new transfer, either on transfer start for
non-cyclic transfers, or on transfer complete interrupt for cyclic
transfers.
For cyclic transfer, when the last item is reached, chan->next_sg must be
reinitialized to the first item.
Signed-off-by: Amelie Delaunay <amelie.delaunay@foss.st.com>
Link: https://lore.kernel.org/r/20220505115611.38845-2-amelie.delaunay@foss.st.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
One of the two DMA controllers managed by the DMAMUX can be used by the
coprocessor. It is defined in the device tree with dma-masters.
When the two DMA controllers are used by the main CPU,
dma-masters = <&dma1, &dma2>; is specified in the device tree.
When one of the controllers is used by coprocessor (so not managed by
Linux), dma-masters = <&dma1>; is specified in the device tree.
In this case, Linux driver must not reset the DMAMUX, because it could have
been configured by the coprocessor to use the second DMA controller.
count is the number of DMA controllers defined in dma-masters property.
Reset only if resets property is found and valid in device tree, and if
the two DMA controllers are under Linux control.
Signed-off-by: Amelie Delaunay <amelie.delaunay@foss.st.com>
Link: https://lore.kernel.org/r/20220504161724.123180-1-amelie.delaunay@foss.st.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
Add compatible and driver_data for GPI DMA engines found in Qualcomm
SC7280. The driver_data contains ee_offset of 0x10000.
Signed-off-by: Vinod Koul <vkoul@kernel.org>
Link: https://lore.kernel.org/r/20220421121733.1829350-1-vkoul@kernel.org
Signed-off-by: Vinod Koul <vkoul@kernel.org>
This adds helpers for accessing and appending service data (0x16) ad
type.
Signed-off-by: Luiz Augusto von Dentz <luiz.von.dentz@intel.com>
Signed-off-by: Marcel Holtmann <marcel@holtmann.org>
PM notifier callbacks should check for supported events rather than filter
out the unsupported events. So that it won't break when a new event is
introduced.
No functional change in this patch.
Signed-off-by: Zhang Rui <rui.zhang@intel.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
The _DSM evaluation warning in its current form is not very helpful, as
it lacks any specific information:
ACPI: \: failed to evaluate _DSM (0x1001)
Thus, include the UUID of the missing _DSM:
ACPI: \: failed to evaluate _DSM bf0212f2-... (0x1001)
Signed-off-by: Michael Niewöhner <linux@mniewoehner.de>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
The pl330 DMA controller on Exynos SoC (e.g. dma-controller@3880000 in
Exynos5420) belongs to power domain, so allow such property.
Reported-by: Rob Herring <robh@kernel.org>
Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>
Acked-by: Rob Herring <robh@kernel.org>
Link: https://lore.kernel.org/r/20220427064048.86635-1-krzysztof.kozlowski@linaro.org
Signed-off-by: Vinod Koul <vkoul@kernel.org>
If interrupt occurs while !chan->busy, it means channel has been disabled
between the raise of the interruption and the read of status and ien, so,
spurious interrupt can be silently discarded.
Signed-off-by: Amelie Delaunay <amelie.delaunay@foss.st.com>
Link: https://lore.kernel.org/r/20220504155322.121431-4-amelie.delaunay@foss.st.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
The parameter to pass back to the handler function when irq has been
requested is a struct stm32_mdma_device pointer, not a struct
stm32_mdma_chan pointer.
Even if chan is reinit later in the function, remove this wrong
initialization.
Fixes: a4ffb13c89 ("dmaengine: Add STM32 MDMA driver")
Signed-off-by: Amelie Delaunay <amelie.delaunay@foss.st.com>
Link: https://lore.kernel.org/r/20220504155322.121431-3-amelie.delaunay@foss.st.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
GISR1 was described in a not up-to-date documentation when the stm32-mdma
driver has been developed. This register has not been added in reference
manual of STM32 SoC with MDMA, which have only 32 MDMA channels.
So remove it from stm32-mdma driver.
Fixes: a4ffb13c89 ("dmaengine: Add STM32 MDMA driver")
Signed-off-by: Amelie Delaunay <amelie.delaunay@foss.st.com>
Link: https://lore.kernel.org/r/20220504155322.121431-2-amelie.delaunay@foss.st.com
Signed-off-by: Vinod Koul <vkoul@kernel.org>
There is no need to store the result of the multiply back to variable value
after the multiplication. The store is redundant, replace *= with just *.
Cleans up clang scan build warning:
warning: Although the value stored to 'value' is used in the enclosing
expression, the value is never actually read from 'value'
[deadcode.DeadStores]
Signed-off-by: Colin Ian King <colin.i.king@gmail.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
The driver is using functions from a compilation unit which is enabled
by CONFIG_CPU_SUP_INTEL. Add that dependency to Kconfig explicitly
otherwise:
drivers/platform/x86/intel/ifs/load.o: in function `ifs_load_firmware':
load.c:(.text+0x3b8): undefined reference to `intel_cpu_collect_info'
Reported-by: Randy Dunlap <rdunlap@infradead.org>
Signed-off-by: Borislav Petkov <bp@suse.de>
Link: https://lore.kernel.org/r/YoZay8YR0zRGyVu+@zn.tnic
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
The communication mean of the _CPC desired performance can be
PCC, System Memory, System IO, or Functional Fixed Hardware (FFH).
PCC, SystemMemory and SystemIo address spaces are available from any
CPU. Thus, dvfs_possible_from_any_cpu should be enabled in such case.
For FFH, let the FFH implementation do smp_call_function_*() calls.
Signed-off-by: Pierre Gondois <pierre.gondois@arm.com>
Reviewed-by: Sudeep Holla <sudeep.holla@arm.com>
Acked-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
The communication mean of the _CPC desired performance can be
PCC, System Memory, System IO, or Functional Fixed Hardware.
commit b7898fda5b ("cpufreq: Support for fast frequency switching")
fast_switching is 'for switching CPU frequencies from interrupt
context'.
Writes to SystemMemory and SystemIo are fast and suitable this.
This is not the case for PCC and might not be the case for FFH.
Enable fast_switching for the cppc_cpufreq driver in above cases.
Add cppc_allow_fast_switch() to check the desired performance
register address space and set fast_switching accordingly.
Signed-off-by: Pierre Gondois <pierre.gondois@arm.com>
Reviewed-by: Sudeep Holla <sudeep.holla@arm.com>
Acked-by: Viresh Kumar <viresh.kumar@linaro.org>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
The transition_delay_us (struct cpufreq_policy) is currently defined
as:
Preferred average time interval between consecutive invocations of
the driver to set the frequency for this policy. To be set by the
scaling driver (0, which is the default, means no preference).
The transition_latency represents the amount of time necessary for a
CPU to change its frequency.
A PCCT table advertises mutliple values:
- pcc_nominal: Expected latency to process a command, in microseconds
- pcc_mpar: The maximum number of periodic requests that the subspace
channel can support, reported in commands per minute. 0 indicates no
limitation.
- pcc_mrtt: The minimum amount of time that OSPM must wait after the
completion of a command before issuing the next command,
in microseconds.
cppc_get_transition_latency() allows to get the max of them.
commit d4f3388afd ("cpufreq / CPPC: Set platform specific
transition_delay_us") allows to select transition_delay_us based on
the platform, and fallbacks to cppc_get_transition_latency()
otherwise.
If _CPC objects are not using PCC channels (no PPCT table), the
transition_delay_us is set to CPUFREQ_ETERNAL, leading to really long
periods between frequency updates (~4s).
If the desired_reg, where performance requests are written, is in
SystemMemory or SystemIo ACPI address space, there is no delay
in requests. So return 0 instead of CPUFREQ_ETERNAL, leading to
transition_delay_us being set to LATENCY_MULTIPLIER us (1000 us).
This patch also adds two macros to check the address spaces.
Signed-off-by: Pierre Gondois <pierre.gondois@arm.com>
Reviewed-by: Sudeep Holla <sudeep.holla@arm.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
The _OSC method allows the OS and firmware to communicate about
supported features/capabitlities. It also allows the OS to take
control of some features.
In ACPI 6.4, s6.2.11.2 Platform-Wide OSPM Capabilities, the CPPC
(resp. v2) bit should be set by the OS if it 'supports controlling
processor performance via the interfaces described in the _CPC
object'.
The OS supports CPPC and parses the _CPC object only if
CONFIG_ACPI_CPPC_LIB is set. Replace the x86 specific
boot_cpu_has(X86_FEATURE_HWP) dynamic check with an arch
generic CONFIG_ACPI_CPPC_LIB build-time check.
Note:
CONFIG_X86_INTEL_PSTATE selects CONFIG_ACPI_CPPC_LIB.
Signed-off-by: Pierre Gondois <pierre.gondois@arm.com>
Reviewed-by: Sudeep Holla <sudeep.holla@arm.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
ACPI 6.2 Section 6.2.11.2 'Platform-Wide OSPM Capabilities':
Starting with ACPI Specification 6.2, all _CPC registers can be in
PCC, System Memory, System IO, or Functional Fixed Hardware address
spaces. OSPM support for this more flexible register space scheme is
indicated by the “Flexible Address Space for CPPC Registers” _OSC bit
Otherwise (cf ACPI 6.1, s8.4.7.1.1.X), _CPC registers must be in:
- PCC or Functional Fixed Hardware address space if defined
- SystemMemory address space (NULL register) if not defined
Add the corresponding _OSC bit and check it when parsing _CPC objects.
Signed-off-by: Pierre Gondois <pierre.gondois@arm.com>
Reviewed-by: Sudeep Holla <sudeep.holla@arm.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Module removal fails because cht_int33fe_typec_remove()
tries to access driver data that does not exist. Fixing by
assigning the data at the end of probe.
Fixes: 915623a80b ("platform/x86: intel_cht_int33fe: Switch to DMI modalias based loading")
Signed-off-by: Heikki Krogerus <heikki.krogerus@linux.intel.com>
Link: https://lore.kernel.org/r/20220519122103.78546-1-heikki.krogerus@linux.intel.com
Reviewed-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
Previously, during suspend, intel_pch_thermal driver logs for every
cooling iteration, about the current PCH temperature and number of cooling
iterations that have been tried, like below
[ 100.955526] intel_pch_thermal 0000:00:14.2: CPU-PCH current temp [53C] higher than the threshold temp [50C], sleep 1 times for 100 ms duration
[ 101.064156] intel_pch_thermal 0000:00:14.2: CPU-PCH current temp [53C] higher than the threshold temp [50C], sleep 2 times for 100 ms duration
After changing the default delay_cnt to 600, in practice, it is common to
see tens of the above messages if the system is suspended when PCH
overheats. Thus, change this log message from dev_warn to dev_dbg because
it is only useful when we want to check the temperature trend.
At the same time, there is always a one-line message given by the driver
with the patch applied, with below four possibilities.
1. PCH is cool, no cooling delay needed
[ 1791.902853] intel_pch_thermal 0000:00:12.0: CPU-PCH is cool [48C]
2. PCH overheats and becomes cool after the cooling delays
[ 1475.511617] intel_pch_thermal 0000:00:12.0: CPU-PCH is cool [49C] after 30700 ms delay
3. PCH still overheats after the overall cooling timeout
[ 2250.157487] intel_pch_thermal 0000:00:12.0: CPU-PCH is hot [60C] after 60000 ms delay. S0ix might fail
4. PCH aborts cooling because of wakeup event detected during the delay
[ 1933.639509] intel_pch_thermal 0000:00:12.0: Wakeup event detected, abort cooling
Signed-off-by: Zhang Rui <rui.zhang@intel.com>
Tested-by: Sumeet Pawnikar <sumeet.r.pawnikar@intel.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Commit ef63b043ac ("thermal: intel: pch: fix S0ix failure due to PCH
temperature above threshold") introduces delay loop mechanism that allows
PCH temperature to go down below threshold during suspend so it won't
block S0ix. And the default overall delay timeout is 1 second.
However, in practice, we found that the time it takes to cool the PCH down
below threshold highly depends on the initial PCH temperature when the
delay starts, as well as the ambient temperature.
And in some cases, the 1 second delay is not sufficient. As a result, the
system stays in a shallower power state like PCx instead of S0ix, and
drains the battery power, without user' notice.
To make sure S0ix is not blocked by the PCH overheating, we
1. expand the default overall timeout to 60 seconds.
2. make sure the temperature is below threshold rather than equal to it.
At the same time, as the cooling delay can be much longer and many wakeup
events (ACPI Power Button press, USB mouse move, etc) becomes valid in the
suspend_noirq phase, add detection of wakeup event so that the driver
does not delay blindly when the system suspend is likely to abort soon.
This patch may introduce longer suspend time, but only in the cases when
the system overheats and Linux used to enter a shallower S2idle state,
say, PCx instead of S0ix.
Signed-off-by: Zhang Rui <rui.zhang@intel.com>
Tested-by: Sumeet Pawnikar <sumeet.r.pawnikar@intel.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Move the PCH Thermal driver suspend callback to suspend_noirq to do
cooling while the system is more quiescent.
Signed-off-by: Zhang Rui <rui.zhang@intel.com>
Tested-by: Sumeet Pawnikar <sumeet.r.pawnikar@intel.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
intel_pch_thermal driver needs a long delay to cool itself (60 seconds
in maximum) during suspend. When a wakeup event occures during the
delay, it is better for the intel_pch_thermal driver to detect this and
quit cooling because the suspend is likely to abort anyway.
Thus expose pm_wakeup_pending to modules so that intel_pch_thermal
driver can be aware of the wakeup events.
Signed-off-by: Zhang Rui <rui.zhang@intel.com>
Tested-by: Sumeet Pawnikar <sumeet.r.pawnikar@intel.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
intel_hid_dsm_fn_mask is a bit mask containing one bit for each function
index. Fix the function index check in intel_hid_evaluate_method
accordingly, which was missed in commit 97ab451620 ("platform/x86:
intel-hid: fix _DSM function index handling").
Fixes: 97ab451620 ("platform/x86: intel-hid: fix _DSM function index handling")
Cc: stable@vger.kernel.org
Signed-off-by: Michael Niewöhner <linux@mniewoehner.de>
Link: https://lore.kernel.org/r/66f813f5bcc724a0f6dd5adefe6a9728dbe509e3.camel@mniewoehner.de
Reviewed-by: Hans de Goede <hdegoede@redhat.com>
Signed-off-by: Hans de Goede <hdegoede@redhat.com>
In __device_attach function, The lock holding logic is as follows:
...
__device_attach
device_lock(dev) // get lock dev
async_schedule_dev(__device_attach_async_helper, dev); // func
async_schedule_node
async_schedule_node_domain(func)
entry = kzalloc(sizeof(struct async_entry), GFP_ATOMIC);
/* when fail or work limit, sync to execute func, but
__device_attach_async_helper will get lock dev as
well, which will lead to A-A deadlock. */
if (!entry || atomic_read(&entry_count) > MAX_WORK) {
func;
else
queue_work_node(node, system_unbound_wq, &entry->work)
device_unlock(dev)
As shown above, when it is allowed to do async probes, because of
out of memory or work limit, async work is not allowed, to do
sync execute instead. it will lead to A-A deadlock because of
__device_attach_async_helper getting lock dev.
To fix the deadlock, move the async_schedule_dev outside device_lock,
as we can see, in async_schedule_node_domain, the parameter of
queue_work_node is system_unbound_wq, so it can accept concurrent
operations. which will also not change the code logic, and will
not lead to deadlock.
Fixes: 765230b5f0 ("driver-core: add asynchronous probing support for drivers")
Signed-off-by: Zhang Wensheng <zhangwensheng5@huawei.com>
Link: https://lore.kernel.org/r/20220518074516.1225580-1-zhangwensheng5@huawei.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Previously the protection of kernfs_pr_cont_buf was piggy backed by
rename_lock, which means that pr_cont() needs to be protected under
rename_lock. This can cause potential circular lock dependencies.
If there is an OOM, we have the following call hierarchy:
-> cpuset_print_current_mems_allowed()
-> pr_cont_cgroup_name()
-> pr_cont_kernfs_name()
pr_cont_kernfs_name() will grab rename_lock and call printk. So we have
the following lock dependencies:
kernfs_rename_lock -> console_sem
Sometimes, printk does a wakeup before releasing console_sem, which has
the dependence chain:
console_sem -> p->pi_lock -> rq->lock
Now, imagine one wants to read cgroup_name under rq->lock, for example,
printing cgroup_name in a tracepoint in the scheduler code. They will
be holding rq->lock and take rename_lock:
rq->lock -> kernfs_rename_lock
Now they will deadlock.
A prevention to this circular lock dependency is to separate the
protection of pr_cont_buf from rename_lock. In principle, rename_lock
is to protect the integrity of cgroup name when copying to buf. Once
pr_cont_buf has got its content, rename_lock can be dropped. So it's
safe to drop rename_lock after kernfs_name_locked (and
kernfs_path_from_node_locked) and rely on a dedicated pr_cont_lock
to protect pr_cont_buf.
Acked-by: Tejun Heo <tj@kernel.org>
Signed-off-by: Hao Luo <haoluo@google.com>
Link: https://lore.kernel.org/r/20220516190951.3144144-1-haoluo@google.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The intent is to use a genpd governor when there are some states that needs
to be managed. Although, the current code ends up to never assign a
governor, let's fix this.
Fixes: 6abf32f1d9 ("cpuidle: Add RISC-V SBI CPU idle driver")
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Reviewed-by: Anup Patel <anup@brainfault.org>
Tested-by: Anup Patel <anup@brainfault.org>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
While factoring out the PM domain related code from PSCI domain driver into
a set of library functions, a regression when initializing the genpds got
introduced. More precisely, we fail to assign a genpd governor, so let's
fix this.
Fixes: 9d976d6721 ("cpuidle: Factor-out power domain related code from PSCI domain driver")
Signed-off-by: Ulf Hansson <ulf.hansson@linaro.org>
Reviewed-by: Anup Patel <anup@brainfault.org>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
This code has a check to prevent read overflow but it needs another
check to prevent writing beyond the end of the ->Ssid[] array.
Fixes: 2b42bd58b3 ("staging: r8188eu: introduce new os_dep dir for RTL8188eu driver")
Cc: stable <stable@vger.kernel.org>
Signed-off-by: Denis Efremov <denis.e.efremov@oracle.com>
Link: https://lore.kernel.org/r/20220518070052.108287-1-denis.e.efremov@oracle.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
default_topology[] uses cpu_clustergroup_mask() for the CLS level
(guarded by CONFIG_SCHED_CLUSTER) which is currently provided by x86
(arch/x86/kernel/smpboot.c) and arm64 (drivers/base/arch_topology.c).
Fixes: 778c558f49 ("sched: Add cluster scheduler level in core and related Kconfig for ARM64")
Acked-by: Barry Song <baohua@kernel.org>
Signed-off-by: Dietmar Eggemann <dietmar.eggemann@arm.com>
Link: https://lore.kernel.org/r/20220513093433.425163-1-dietmar.eggemann@arm.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
The deferred probe timer that's used for this currently starts at
late_initcall and runs for driver_deferred_probe_timeout seconds. The
assumption being that all available drivers would be loaded and
registered before the timer expires. This means, the
driver_deferred_probe_timeout has to be pretty large for it to cover the
worst case. But if we set the default value for it to cover the worst
case, it would significantly slow down the average case. For this
reason, the default value is set to 0.
Also, with CONFIG_MODULES=y and the current default values of
driver_deferred_probe_timeout=0 and fw_devlink=on, devices with missing
drivers will cause their consumer devices to always defer their probes.
This is because device links created by fw_devlink defer the probe even
before the consumer driver's probe() is called.
Instead of a fixed timeout, if we extend an unexpired deferred probe
timer on every successful driver registration, with the expectation more
modules would be loaded in the near future, then the default value of
driver_deferred_probe_timeout only needs to be as long as the worst case
time difference between two consecutive module loads.
So let's implement that and set the default value to 10 seconds when
CONFIG_MODULES=y.
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Cc: "Rafael J. Wysocki" <rjw@rjwysocki.net>
Cc: Rob Herring <robh@kernel.org>
Cc: Linus Walleij <linus.walleij@linaro.org>
Cc: Will Deacon <will@kernel.org>
Cc: Ulf Hansson <ulf.hansson@linaro.org>
Cc: Kevin Hilman <khilman@kernel.org>
Cc: Thierry Reding <treding@nvidia.com>
Cc: Mark Brown <broonie@kernel.org>
Cc: Pavel Machek <pavel@ucw.cz>
Cc: Geert Uytterhoeven <geert@linux-m68k.org>
Cc: Yoshihiro Shimoda <yoshihiro.shimoda.uh@renesas.com>
Cc: Paul Kocialkowski <paul.kocialkowski@bootlin.com>
Cc: linux-gpio@vger.kernel.org
Cc: linux-pm@vger.kernel.org
Cc: iommu@lists.linux-foundation.org
Reviewed-by: Mark Brown <broonie@kernel.org>
Acked-by: Rob Herring <robh@kernel.org>
Signed-off-by: Saravana Kannan <saravanak@google.com>
Link: https://lore.kernel.org/r/20220429220933.1350374-1-saravanak@google.com
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Each devfreq governor specifies the supported governor event
such as GOV_START and GOV_STOP. When not-supported event is required,
just return non-error. But, commit ce9a0d88d97a ("PM / devfreq: Add
cpu based scaling support to passive governor") returned the error
value. So that return non-error value when not-supported event is required.
Fixes: ce9a0d88d97a ("PM / devfreq: Add cpu based scaling support to passive governor")
Reported-by: Marek Szyprowski <m.szyprowski@samsung.com>
Signed-off-by: Chanwoo Choi <cw00.choi@samsung.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Add devm_register_restart_handler() helper that registers sys-off
handler using restart mode and with a default priority. Most drivers
will want to register restart handler with a default priority, so this
helper will reduce the boilerplate code and make code easier to read and
follow.
Signed-off-by: Dmitry Osipenko <dmitry.osipenko@collabora.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Add devm_register_power_off_handler() helper that registers sys-off
handler using power-off mode and with a default priority. Most drivers
will want to register power-off handler with a default priority, so this
helper will reduce the boilerplate code and make code easier to read and
follow.
Signed-off-by: Dmitry Osipenko <dmitry.osipenko@collabora.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Nexus 7 Android tablet can be turned off using a special bootloader
command which is conveyed to bootloader by putting magic value into the
special scratch register and then rebooting normally. This power-off
method should be invoked if USB cable is connected. Bootloader then will
display battery status and power off the device. This behaviour is
borrowed from downstream kernel and matches user expectations, otherwise
it looks like device got hung during power-off and it may wake up on
USB disconnect.
Switch PMC driver to sys-off handler API, which provides drivers with
chained power-off callbacks functionality that is required for powering-off
devices properly. It also brings resource-managed API for the restart
handler registration that makes PMC driver code cleaner.
Signed-off-by: Dmitry Osipenko <dmitry.osipenko@collabora.com>
Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
