Hexadecimal integer literals are signed if they can fit into a signed int,
which causes undefined behavior.
This happens here because 0xFF can fit into a signed int and then gets
left-shifted by 24, undefined behavior for signed integers.
Signed-off-by: Daniel Hajjar <daniel.hajjar16@gmail.com>
Information about IOAPIC can be located not in the first
DMAR Hardware Unit Definition subtable. Iterate them all.
Signed-off-by: Andrei Emeltchenko <andrei.emeltchenko@intel.com>
Use DEVICE_DT_INST_DEFINE instead of DEVICE_DEFINE to declare the device
structure. This ensures that the device gets an ordinal and is
initialized before any device depending on it.
Signed-off-by: Fabio Baltieri <fabiobaltieri@google.com>
The old acpi implimentation is replaced with acpica interface
and updated x86 arch porting with the new interface.
Signed-off-by: Najumon B.A <najumon.ba@intel.com>
The cast for the offset variable will make incorrect register
access when it is over 0x80. Change the char cast to unsigned
char to fix it.
Fixes#49803.
Signed-off-by: Enjia Mai <enjia.mai@intel.com>
Initialize the vtd device at compile time. Also removed unnecessary
macro hackery. As a result, vtd pointer can be constified.
Signed-off-by: Gerard Marull-Paretas <gerard.marull@nordicsemi.no>
In order to bring consistency in-tree, migrate all drivers to the new
prefix <zephyr/...>. Note that the conversion has been scripted, refer
to #45388 for more details.
Signed-off-by: Gerard Marull-Paretas <gerard.marull@nordicsemi.no>
Refactors interrupt controller drivers to use the shared driver class
initialization priority configuration, CONFIG_INTC_INIT_PRIORITY, to
allow configuring interrupt controller drivers separately from other
devices. This is similar to other driver classes.
The default is set to CONFIG_KERNEL_INIT_PRIORITY_DEFAULT to preserve
the existing default initialization priority for most drivers.
Signed-off-by: Maureen Helm <maureen.helm@intel.com>
The device instance can be obtained at compile time. Here
DEVICE_DT_GET_OR_NULL is used as the following code seems to accept a
NULL condition, meaning instance is optional.
Signed-off-by: Gerard Marull-Paretas <gerard.marull@nordicsemi.no>
In order to align with macros used to obtain a device reference (e.g.
DEVICE_DT_GET), align the PM macros to use "GET" instead of "REF". This
change should have low impact since no official release has gone out yet
with the "REF" macros.
Signed-off-by: Gerard Marull-Paretas <gerard.marull@nordicsemi.no>
For functions returning nothing, there is no need to document
with @return, as Doxgen complains about "documented empty
return type of ...".
Signed-off-by: Daniel Leung <daniel.leung@intel.com>
If VT-D's interrupt remapping is in place, all IOAPIC RTEs need to get
remapped as well (or then they will be simply blocked).
Signed-off-by: Tomasz Bursztyka <tomasz.bursztyka@linux.intel.com>
The device PM subsystem _depends_ on device, not vice-versa. Devices
only hold a reference to struct pm_device now, and initialize this
reference with the value provided in Z_DEVICE_DEFINE. This requirement
can be solved with a forward struct declaration, meaning there is no
need to include device PM headers.
Signed-off-by: Gerard Marull-Paretas <gerard.marull@nordicsemi.no>
Port some drivers to the recently introduced macros to showcase its
usage and be able to do some initial testing (nRF52840).
Signed-off-by: Gerard Marull-Paretas <gerard.marull@nordicsemi.no>
Define the device using DEVICE_DEFINE macro, so that a single option can
be used regardless of PM being enabled or not.
Signed-off-by: Gerard Marull-Paretas <gerard.marull@nordicsemi.no>
The PM callback is no longer referenced as "pm_control" but
"pm_action_cb", so reflect this new naming on the callbacks.
Signed-off-by: Gerard Marull-Paretas <gerard.marull@nordicsemi.no>
Instead of passing target states, use actions for device PM control.
Actions represent better the meaning of the callback argument.
Furthermore, they are more future proof as they can be suitable for
other PM actions that have no direct mapping to a state. If we compare
with Linux, we could have a multi-stage suspend/resume. Such scenario
would not have a good mapping when using target states.
Signed-off-by: Gerard Marull-Paretas <gerard.marull@nordicsemi.no>
According to the documentation the OFF state has to be used when the
devices is fully turned off, ie, power removed. Most drivers were using
a sort of fall-through for all non-active states, leading to behaviors
not following the specifications.
Signed-off-by: Gerard Marull-Paretas <gerard.marull@nordicsemi.no>
- Return -ENOTSUP if the requested state is not supported
- Remove redundant "noop style" functions.
- Use switch everywhere to handle requested state (not necessary in all
drivers, but better take off with consistency in place after current
changes).
Signed-off-by: Gerard Marull-Paretas <gerard.marull@nordicsemi.no>
The verb tense for the suspended state was not consistent with other
states. The likely reason: state was being used as a command/action.
Signed-off-by: Gerard Marull-Paretas <gerard.marull@nordicsemi.no>
The difference between low power and suspend states is a thin blur line
that is is not clear and most drivers have used indistinctly. This patch
converges to the usage of the suspend state for low power, since
contrary to the low power state, it is used by both system and runtime
device PM. The low power state is still kept, but its future is unclear
and needs some discussion.
Signed-off-by: Gerard Marull-Paretas <gerard.marull@nordicsemi.no>
Some devices are using PM_DEVICE_STATE_FORCE_SUSPEND as a sort of low
power state, something that is not correct. In fact, this state is not
an actual state and will be eventually moved, if found necessary, to an
action or command.
Signed-off-by: Gerard Marull-Paretas <gerard.marull@nordicsemi.no>
The device PM control function will only be called if the requested
state is different from the current one. A significant amount of drivers
were checking for state changes, now unnecessary. This patch removes all
this redundant logic.
Signed-off-by: Gerard Marull-Paretas <gerard.marull@nordicsemi.no>
Since the state is no longer modified by the device PM callback, just
use the state value.
Signed-off-by: Gerard Marull-Paretas <gerard.marull@nordicsemi.no>
The callback is now invoked to set the device PM state in all cases, so
the usage of ctrl_command is redundant.
Signed-off-by: Gerard Marull-Paretas <gerard.marull@nordicsemi.no>
The device PM subsystem already holds the device state, so there is no
need to keep duplicates inside the device. The pm_device_state_get has
been refactored to just return the device state. Note that this is still
not safe, but the same applied to the previous implementation. This
problem will be addressed later.
Signed-off-by: Gerard Marull-Paretas <gerard.marull@nordicsemi.no>
The callback is not used anymore, so just delete it from the pm_control
callback signature.
Signed-off-by: Gerard Marull-Paretas <gerard.marull@nordicsemi.no>
the device PM callback is not used anymore by the device PM subsystem,
so remove it from all drivers/tests using it.
Signed-off-by: Gerard Marull-Paretas <gerard.marull@nordicsemi.no>
Move all PM_DEVICE_STATE_* definitions to an enum. The
PM_DEVICE_STATE_SET and PM_DEVICE_STATE_GET definitions have been kept
out of the enum since they do not represent any state. However, their
name has not been changed since they will be removed soon.
All drivers and tests have been adjusted accordingly.
Signed-off-by: Gerard Marull-Paretas <gerard.marull@nordicsemi.no>
The context parameter used across device power management is
actually the power state. Just use it and avoid a lot of
unnecessary casts.
Signed-off-by: Flavio Ceolin <flavio.ceolin@intel.com>
The fixed number of 24 RTEs is a legacy thing, and long gone by now.
IOAPICs expose the maximum number of RTEs they have via the version
register, so let's use it.
This avoids to manually tweak a Kconfig option (which is now removed)
and fixes the RTE number for all x86 targets relevantly.
Signed-off-by: Tomasz Bursztyka <tomasz.bursztyka@linux.intel.com>
- Remove SYS_ prefix
- shorten POWER_MANAGEMENT to just PM
- DEVICE_POWER_MANAGEMENT -> PM_DEVICE
and use PM_ as the prefix for all PM related Kconfigs
Signed-off-by: Anas Nashif <anas.nashif@intel.com>
Now that device_api attribute is unmodified at runtime, as well as all
the other attributes, it is possible to switch all device driver
instance to be constant.
A coccinelle rule is used for this:
@r_const_dev_1
disable optional_qualifier
@
@@
-struct device *
+const struct device *
@r_const_dev_2
disable optional_qualifier
@
@@
-struct device * const
+const struct device *
Fixes#27399
Signed-off-by: Tomasz Bursztyka <tomasz.bursztyka@linux.intel.com>
There is nothing wrong with instance numbers and they are
recommended for use whenever possible, but this is an API
design problem because it's not always possible to get nodes
by instance number; in some cases, drivers need to get node
identifiers from node labels, for example.
Change these APIs (which are not yet in any Zephyr release)
to take node IDs instead of instance IDs.
Fixes: #26984
Signed-off-by: Andrew Boie <andrew.p.boie@intel.com>
A hack was required for the loapic code due to the address
range not being in DTS. A bug was filed.
Signed-off-by: Andrew Boie <andrew.p.boie@intel.com>
Currently all IO APIC interrupts are configured at fixed delivery mode,
which is good for HEPT timer interrupt but it imposes burdens to
device drivers to properly handle the repeated interrupt sent to all
processors.
This patch makes it more flexible so that device drivers can specify
the delivery mode it desires in the IRQ connect APIs.
- Don't hard code IOAPIC_FIXED in z_ioapic_irq_set(), meaning the
IRQ delivery mode is passed in from the 'flags' argument and
individual device driver needs to choose delivery mode for its own
IO APIC interrupt.
- To support different delivery mode in different IO APIC interrupts,
need to save and restore RTE[10:8] during IOAPIC suspend and resume.
If device driver doesn't pass either IOAPIC_FIXED or IOAPIC_LOWEST
in IRQ_CONNECT()/irq_connect_dynamic() alike APIs, the delivery mode
bit fields in the target RTE register are '0' which implies fixed mode.
If the device driver wants the interrupt to be delivered to one CPU
only, it needs to explicitly specify IOAPIC_LOWEST in one of the IRQ
connect APIs.
Signed-off-by: Zide Chen <zide.chen@intel.com>
Currently IO APIC is working in physical destination mode, which
doesn't support interrupt to be delivered to multiple local APICs.
By definition only 4 bits [59:63] in IO APIC IOREDTBL register are
available for destination addresses and it contains an APIC ID only.
This patch changes it to logical destination mode so that IOREDTBL
can potentially define a set of processors and it's posible to deliver
interrupts to multiple APICs.
Signed-off-by: Zide Chen <zide.chen@intel.com>
Daniel Hajjar
Andrei Emeltchenko
Flavio Ceolin
Fabio Baltieri
Najumon B.A
Enjia Mai
Gerard Marull-Paretas
Maureen Helm
Daniel Leung
Tomasz Bursztyka
Dong Wang
Anas Nashif
Andrew Boie
Kumar Gala
Zide Chen