If the MAX_MSIX_TABLE_NUM is specified in scenario.xml. Return the
largest number from count of MSI, table_size of MSIX or
MAX_MSIX_TABLE_NUM of scenario.xml.
Tracked-On: #6235
Signed-off-by: Yang,Yu-chu <yu-chu.yang@intel.com>
The ACPI specification allows both assigning to buffers and indexing to a
certain byte of a buffer using the Index operator. This patch adds the
implementation of these two operations in the interpreter.
Tracked-On: #6287
Signed-off-by: Junjie Mao <junjie.mao@intel.com>
The PackageElementList builder takes variadic arguments, each of which is
an element of the package to be created, not a single argument being the
list of the elements. This patch fix the call to PackageElementList in
build_value() where the wrong type of argument was passed.
Tracked-On: #6287
Signed-off-by: Junjie Mao <junjie.mao@intel.com>
It is typical in AML resource descriptors to have 0-length region
descriptors which are typically templates of resources that are not
assigned on the current platform. For such regions, the `base + length - 1`
formula does not calculate the max of the region properly.
This patch updates the resource descriptor parsers to use max = min when
the length of the region is 0.
Tracked-On: #6287
Signed-off-by: Junjie Mao <junjie.mao@intel.com>
When parsing an AML object representing a host bridge, the current board
inspector may encounter the following issues:
1. The host DSDT may contain multiple host bridge instances, with some of
them not being present. In this case the _BBN of these instances may
evaluate to the same value that coincide with the bus assigned to an
existing host bridge, leading to multiple PCI bus nodes with the same
bus number and thus confusion in later information extraction phases.
2. Methods of a host bridge may refer to the PCI configuration space of
itself (which is typically Device 0, Function 0 under that
bus). However, such objects may not have an _ADR object as the bus
number is encoded by the _BBN object instead.
This patch fixes the issues above.
Tracked-On: #6287
Signed-off-by: Junjie Mao <junjie.mao@intel.com>
This patch enables the interpretation of the following AML objects.
* OnesOp. A OnesOp object always evaluates to a 64-bit integer with all bits
set to 1. It is assumed that the host DSDT is always revision 2 or above,
which is typically the case on modern platforms.
* DefMatch. A DefMatch object evaluates to the index of the first
element (starting from a given index) in a package that matches the given
two predicates. If a match is not found, the constant Ones is returned.
* DefSizeOf. A DefSizeOf object evaluates to the byte length of a buffer, the
length of a string (without the terminating NUL character) or the number of
elements in a package.
Tracked-On: #6287
Signed-off-by: Junjie Mao <junjie.mao@intel.com>
In commit e5ba06cbe8 ("board_inspector: a workaround to an incorrect
interpretation") a workaround is introduced to check the data type of
predicate operands. That commit assumes that both operands must be exactly
integers, which is not usually the case as operation fields or strings can
also be used in predicates.
This patch applies the following conversions on both operands when
evaluating a predicate:
1. Try converting both operands to integers
2. If either conversion in step 1 fails, try regarding both operands as
strings.
3. If either operand is not a string, return the default
result (i.e. False).
Fixes: e5ba06cbe8 ("board_inspector: a workaround to an incorrect interpretation")
Tracked-On: #6287
Signed-off-by: Junjie Mao <junjie.mao@intel.com>
The method `enter_scope` of class `Context` is a reduced implementation of
`change_scope` which assumes that the given name is simply a NameSeg. This
method is currently only used when a new named scope is opened by a
DefDevice object for historical reasons.
As the other named-scope-opening objects all use `change_scope` which can
handle arbitrary NameString, this patch unifies the code by removing
`enter_scope` and replacing the only occurrence with `change_scope`. This
also resolves the parsing of AML templates in board XMLs where device names
can be more than a simple NameSeg.
Tracked-On: #6287
Signed-off-by: Junjie Mao <junjie.mao@intel.com>
According to the specification, the _MAX of memory address space resource
descriptors may not necessarily be _MIN + _LEN - 1. Typically the resource
locates at the low end of the region specified by _MIN and _MAX. This patch
makes the ACPI extractor always calculating the end of a memory region
using _MIN and _LEN instead.
Tracked-On: #6287
Signed-off-by: Junjie Mao <junjie.mao@intel.com>
The current design of AML parsing, objects are first defined in the
namespace and later dropped if they are in a False branch. This leads to
incorrect interpretation of the AML code where:
1. A name T is defined in the root scope as an integer.
2. A method M in an inner scope S references T.
3. The name T is defined as a device, power resource or other named
objects in scope S under conditions where M will not be called.
As a workaround, check if both the left and right hand sides are integers
first. If either is not the case,
Tracked-On: #6287
Signed-off-by: Junjie Mao <junjie.mao@intel.com>
The current vACPI image layout reserves 512 bytes for vDSDT. Given the fact
that the size of vDSDT of a pre-launched VM grows when more devices are
assigned to the VM, this size limit can be easily exceeded. As an example,
a single pass-through TSN NIC requires 291 bytes in vDSDT to define device
objects representing its PCS (Physical Coding Sublayer), which means the
current reserved space for vDSDT does not allow two TSN NICs to be assigned
to the same VM.
This patch enlarges the reserved space for vDSDT to 2432 bytes by moving
the MCFG and MADT spaces. 768 bytes are still reserved for MADT which
is sufficient to encode the LAPIC information for more than 64 vCPUs.
This patch is added in v2 of the series.
Tracked-On: #6287
Signed-off-by: Junjie Mao <junjie.mao@intel.com>
This patch generates vDSDT for pre-launched VMs based on the scenario
configuration and static allocation results. The vDSDT contains the
following objects.
* A device object representing the root PCI bus which has the same
location and name as the object in physical DSDT. This object contains
device identification objects such as _HID, _CID, _BBN and _UID, the
_CRS object encoding the bus numbers and PCI hole regions, and the _PRT
object encoding the routing from vPCI device pins to virtual interrupt
lines.
* For each pass-through device assigned to the VM, a device object is
created based on the AML template of that device in board XML (if
exists). The _ADR object is rewritten with the vBDF..
v1 -> v2:
* In order to replace the current static vDSDT, the AML templates in
board XMLs now include objects that are referenced by other devices as
well.
* When TPM is assigned, the corresponding device object in DSDT (if
exists) will be copied.
* Add the _S5 control method to vDSDT, as is done in the static DSDT
sources.
* The old vDSDT generation routine and the static DSDT sources it uses
are removed.
v2 -> v3:
* Only follow 'uses', 'is used by' or 'consumes resources by'
dependencies.
* Organize the device objects in vDSDT according to the scopes they are
in.
Tracked-On: #6287
Signed-off-by: Junjie Mao <junjie.mao@intel.com>
This patch allocates interrupt lines among VMs according to the PCI devices
assigned to them.
v1 -> v2:
* Remove the usage of VMx_PT_INTX_NUM macro in vm_configuration.c; use the
concrete numbers directly.
* The static allocator will also complain if any interrupt line is allocated to
a VM with LAPIC_PASSTHROUGH.
v2 -> v3:
* Fix a minor coding style issue.
Tracked-On: #6287
Signed-off-by: Junjie Mao <junjie.mao@intel.com>
It is a common practice to parse PCI BDF in the static allocators. This
patch moves the BusDevFunc class (which is a named tuple encoding a BDF) to
lib.py and uses it for BDF parsing throughout the static allocators.
Tracked-On: #6287
Signed-off-by: Junjie Mao <junjie.mao@intel.com>
Customized environment variables are not inherited to child processes
created by the subprocess module. As a result the legacy board parser may
not be able to locate the prerequisite utilities if they can be found only
with the customized PATH.
This patch passes the PATH of cli.py to the legacy parser so that both
scripts use the same PATH to search for utilities.
This patch is added in v2 of the series.
Tracked-On: #6287
Signed-off-by: Junjie Mao <junjie.mao@intel.com>
With a growing demand on host ACPI namespace for devices pass-through, it
is now important to parse the ACPI namespace when generating board
XMLs. This patch makes ACPI namespace parsing enabled by default by
replacing the `--advanced` option, which is designed to enable the parsing,
to `--basic` which disables it.
The option provides a reliable way to disable ACPI namespace parsing
completely in case the parsing blocks the generation of board XMLs, while
the ACPI namespace parser will gracefully stop without polluting the board
XML when it fails.
This patch is added in v2 of the patch series.
Tracked-On: #6287
Signed-off-by: Junjie Mao <junjie.mao@intel.com>
This patch updates the board inspector to collect the bits of physical and
linear addresses of the processors and generate this information to the
board XML for further uses at configuration phase.
Tracked-On: #6292
Signed-off-by: Junjie Mao <junjie.mao@intel.com>
It is witnessed on some boards that the device status (as is reported by
the _STA object) returns 0 while the device object is still useful for
pass-through devices. The original implementation, however, assumes that
only a device is a non-zero status is useful as long as the _STA object
exists.
This patch makes this filtering disabled by default and adds a command-line
argument `--check-device-status` to enable this filtering. As disabled or
non-present devices can have empty resources, the sorting algorithm is
updated accordingly to gracefully handle such descriptors.
This patch is added in v3 of the series.
Tracked-On: #6287
Signed-off-by: Junjie Mao <junjie.mao@intel.com>
Occasionally buffer fields (which are typically integers) are converted to
hexadecimal strings for debugging purposes. This patch adds the conversion
to suppress interpretation errors of these debugging calls.
This patch is added in v3 of the series.
Tracked-On: #6287
Signed-off-by: Junjie Mao <junjie.mao@intel.com>
Servers typically have multiple peer host bridges that need to be scanned
separately. This patch extends the PCI information extractor to visit all
PCI host bridges by enumerating /sys/devices/pci* directories.
Tracked-On: #6292
Signed-off-by: Junjie Mao <junjie.mao@intel.com>
While not allowed by ACPI specification, using a ConstObj (e.g. OneOp) as a
term in a TermList IS witnessed in the DSDT of some BIOS. This patch allows
ConstObj to act as a TermObj so that a TermList can contain a ConstObj as a
statement (which is essentially no-op).
This patch is added in v2 of the series.
Tracked-On: #6287
Signed-off-by: Junjie Mao <junjie.mao@intel.com>
It is seen that the BAR in the PCI configuration space of a device can be
cleared when the device is put to D3. This patch resumes a device not in D0
before parsing its configuration space in order to collect accurate
information.
This patch is added in v2 of the series.
Tracked-On: #6287
Signed-off-by: Junjie Mao <junjie.mao@intel.com>
AML allows devices defined in an ACPI namespace to have inter-dependency,
i.e. a method defined in one device can refer to objects in other
devices. While such inter-dependency is common in device manipulation
methods, device identification and configuration methods, such as _CRS, may
depend on other devices as well.
An example we have already met is a PCS (Physical Coding Sublayer) which
calculates resource descriptors by accessing the PCI configuration space of
the accompanying Ethernet controller. Without the ACPI object describing
the PCS, a driver of the Ethernet controller may refuse to initialize.
This patch adds a preliminary dependency analyzer to detect such
inter-device dependency. The analyzer walks through the reference chains of
an object, identifying whether the referenced objects are operation fields
of a device. Depending on the result of this analysis, the board XML is
refined as follows.
* When an object (probably a method) references such fields, the original
object definition in host DSDT/SSDTs will be copied in the AML template
so that they still work in VMs where the operation fields may be
virtualized. Such objects will be referred to as "copied objects"
hereinafter.
* The objects that are **directly** referenced by a copied object is
added in the AML template as well. Such objects still belong to devices
where they are originally defined in the host ACPI namespace. Their
definition, however, may be copied or replaced with constant values,
depending on the dependency analysis on these objects.
* Nodes with the "dependency" tag are added under "device" nodes in the
board XML, allowing the configuration tools to follow the device
dependency chain when generating vACPI tables. These nodes only
represent direct dependencies; indirect dependencies can be inferred by
following those direct ones.
The current implementation does not allow objects being added to AML
templates if they refer to any of the following.
* Global objects, i.e. objects not belonging to any device. Such objects
tend to encode system-wide information, such as the ACPI
NVS (Non-Volatile Storage) or its fields.
* Methods with parameters.
Objects with such references are thus being hidden from guest software,
just like how they are invisible in the current implementation.
This patch is added in v2 of the series.
v2 -> v3:
* Also collect dependencies due to providing or consuming resources.
* Refactor the dependency detection logic for clarity.
Tracked-On: #6287
Signed-off-by: Junjie Mao <junjie.mao@intel.com>
GPIO and generic serial connection resources in ACPI resource descriptors
usually encode resource sources which are important in detecting
cross-device dependencies. This patch adds parsers for GPIO and generic
serial connection descriptors.
Tracked-On: #6287
Signed-off-by: Junjie Mao <junjie.mao@intel.com>
This patch adds AML template (in XML hexBinary format) for each device to
board XMLs. For now these templates contain the following objects if they
exists in the physical DSDT:
- Device identification objects: _ADR, _HID, _UID and _STR
- _CRS which encodes the current resources consumed by the device
- _STA which encodes the status of the device
An AML template is always a DefScope with a single DefDevice so that they
can be appended anywhere in the vDSDT.
v1 -> v2:
* Remove the temporary visitor that collects cross-device dependencies. Such
check will be replaced with another visitor introduced in the next patch.
v2 -> v3:
* The AML templates are now DefDevice objects with their names being the
full namepath. The vDSDT generator will take care of this and organize
the objects properly.
Tracked-On: #6287
Signed-off-by: Junjie Mao <junjie.mao@intel.com>
Object type codes are used to identify, as the name suggests, type of
objects. Typical object types in AML include integers, strings, methods,
devices, buffers, packages and operation regions. In DefExternal terms the
object type codes help specify the type of the external objects so that an
AML parser can parse the code without knowing the concrete definition of
these objects.
The per-device AML templates in board XMLs need DefExternal terms to
declare the objects in other devices, as these templates are meant to be
parsed and integrated separately. This patch adds a static method to object
declaration classes to make it easier to generate such DefExternal terms
for a given declaration.
A complete definition of object type codes can be found in section 19.6.96
of ACPI specification 6.4.
v1 -> v2:
* Remove the object_type of FieldDecl and OperationFieldDecl as 0x5 is
not a proper object type for buffer fields.
Tracked-On: #6287
Signed-off-by: Junjie Mao <junjie.mao@intel.com>
The current implementation of the ConditionallyUnregisterSymbolVisitor
exits upon visiting a DefMethod node without unregistering that method. As
a result, methods in False branches in DSDT/SSDTs are not removed from the
parsed namespace, which can lead to further confusions when these methods
are referenced (e.g. a _CRS method visited by the board inspector).
This patch fixes this by always visiting a DefMethod node but stops
traversing its children.
Tracked-On: #6287
Signed-off-by: Junjie Mao <junjie.mao@intel.com>
/usr/share/pci.ids.gz is another typical path to the pci.ids file of the
lspci tool which is used in Yocto-based systems. This patch adds this path
as another candidate when searching for pci.ids. The builtin gzip module is
used to open this file.
Tracked-On: #6287
Signed-off-by: Junjie Mao <junjie.mao@intel.com>
Address space resource descriptors have an optional field to encode the
resource source, which is not commonly used when creating new resource
descriptors.
For modules which want to create a class to parse address space resource
descriptors without resource source, this patch sets the length of such
descriptors as the default value of the `_len` factory parameter so that
callers do not need to care about these lengths.
Tracked-On: #6287
Signed-off-by: Junjie Mao <junjie.mao@intel.com>
With AML templates for devices in the board XML, the parser now needs to be able
to parse a stream as an arbitrary object. This patch adds the `parse_tree`
method to the acpiparser.aml.parser module for this purpose.
Tracked-On: #6287
Signed-off-by: Junjie Mao <junjie.mao@intel.com>
In addition to the mandatory _HID (Hardware ID), the ACPI spec also defines
an optional _CID (Compatible ID) object for device identification.
This patch enhances the ACPI extractor by parsing the _CID objects of devices as
well.
Tracked-On: #6320
Signed-off-by: Junjie Mao <junjie.mao@intel.com>
It is quite common to meet permissions errors when opening a specific
region of /dev/mem due to kernel configurations. This patch adds a bit more
logs on this for eaiser debugging.
Tracked-On: #6287
Signed-off-by: Junjie Mao <junjie.mao@intel.com>
This patch adds interrupt pin related information into the board XML,
including:
* The PCI routing table in ACPI DSDT/SSDT are parsed and generated into
the board XML as "interrupt_pin_routing" nodes.
* IRQs encoded in _CRS directly are represented as resources of type
"irq".
* Interrupt lines (i.e. INTx#) of PCI devices are represented as
resources of type "interrupt_pin". When the PCI routing table is
available, the corresponding interrupt line is identified and
represented as the "source" attribute of the resource node.
Due to the existence of vIOAPIC in ACRN VMs, the board inspector interprets
the \_PIC method with parameter 1 to inform the ACPI namespace that the
interrupt model should be in APIC mode.
v1 -> v2:
* Remove the msi_enable variable which is defined but never used.
Tracked-On: #6287
Signed-off-by: Junjie Mao <junjie.mao@intel.com>
This patch adds the acpiparser.aml.builder module which provides methods to
construct AML trees from scratch in Python. Similar to how parsers and
binary generators are implemented, this module constructs most builder
methods from the AML grammar defined in the acpiparser.aml.grammar
module. AML objects whose grammar are not present in the grammar module
require special treatment and their builders are implemented
explicitly. The methods have the same name as the AML tree labels defined
in the grammar.
In addition, this module also provides the method `build_value` which
converts plain integers, strings or interpreter values (which are defined
in the datatypes module) to AML trees.
With the builders, the `interpret_method_call` method in the
ConcreteInterpreter is refined to build the (fake) MethodInvocation node
using the builders and handle the actual parameters as well.
Tracked-On: #6287
Signed-off-by: Junjie Mao <junjie.mao@intel.com>
Tree visitors usually have a fixed direction (either top-down or bottom-up)
and invoking a visitor with a wrong direction typically leads to unintended
behavior. However, the current implementation exposes both `visit_topdown`
and `visit_bottomup` methods to users, allowing callers to invoke the
visitors in an undesigned way. The same issue exists in the implementation
of transformers.
This patch refactors the base classes of visitors and transformers so that
they require an explicit direction from their sub-classes to
initialize. Callers of the visitors and transformers are now expected to
invoke the `visit` or `transform` methods without assuming the correct
direction of the visitor or transformer. The original `visit_topdown` or
`visit_bottomup` methods (or their transformer counterparts) are now
used to initialize the `visit` method and can be used by the subclasses in
case those subclasses visits the tree in a customized manner.
Tracked-On: #6287
Signed-off-by: Junjie Mao <junjie.mao@intel.com>
This patch introduces a visitor that converts an arbitrary AML tree to an
AML binary. Most nodes can be converted in a straightforward way by
following the defined grammar, but the following nodes require some
additional effort:
- NameStrings can be formatted as either a NameSeg (i.e. four upper case
characters), a DualNamePath, a MultiNamePath or a NullName.
- PkgLengths are recalculated according to the actual length of the
following object (in case they are changed dynamically after being
generated by the parser) and generated following the AML encoding of
such lengths.
The visitor works in a bottom-up manner, i.e. the children are visited and
converted to binary before the parent.
The whole trees parsed from DSDT/SSDTs are now also stored in the Context
for further reference.
Tracked-On: #6287
Signed-off-by: Junjie Mao <junjie.mao@intel.com>
_STR is another device identification object defined in ACPI spec that
describes a device. This patch collects this string (when available) into
board XML as well.
Tracked-On: #6287
Signed-off-by: Junjie Mao <junjie.mao@intel.com>
This patch adds the property `irqs` to the class SmallResourceitemIRQ so
that the list of IRQs encoded in this resource item can be retrieved
easily.
Tracked-On: #6287
Signed-off-by: Junjie Mao <junjie.mao@intel.com>
This patch adds a parser to the PCI routing tables returned by _PRT objects
of platform devices. The parsed result is a list of PRTMappingPackage
instances, each of which is a named tuple with the following fields:
* address: a dword with higher 16 bits being the function number and
lower 16 bits all 1's.
* pin: a byte representing the mapped pin.
* source: either a DeviceDecl of the device that allocates the interrupt
line, or the byte 0.
* source_index:
- If `source` is a DeviceDecl, this is the index of the interrupt
source within that device.
- If `source` is 0, this is the interrupt line connected to the pin.
Tracked-On: #6287
Signed-off-by: Junjie Mao <junjie.mao@intel.com>
Local variables can be assigned with formal arguments in AML. As a result
when interpreting a DefReturn node, the interpreter shall unwrap multiple
layers of argument/local variable wrappings until a concrete value is
found. This patch implements this logic.
Tracked-On: #6287
Signed-off-by: Junjie Mao <junjie.mao@intel.com>
Resource template buffers always end with an end tag. Concatenation of two
resource buffers thus requires that the end tag of the first buffer is
stripped. This patch adds this logic to the interpretation of DefConcatRes
AML nodes.
Tracked-On: #6287
Signed-off-by: Junjie Mao <junjie.mao@intel.com>
The warning, info and debug logging levels are intended to be used in the
following way.
* Warnings are used when users are expected to be aware of a certain
failure.
* Info messages are used to track parsing process and major internal
errors for development.
* Debug messages are used to collect verbose debug logs.
To align the current usage of logs to the above guidelines, this patch
adjusts the logging level of the following messages:
* DSDT/SSDT interpretation failures are now warnings, not information
* Failures of parsing deferred AML blocks are now information, not debug
messages
The default log level when running `cli.py` is adjusted to WARNING as well,
as INFO is primarily used for development. A new command line option
`loglevel` is added to adjust the log level per user needs.
v2 -> v3:
* Make address collisions in ACPI namespace as an info rather than a
warning.
Tracked-On: #6287
Signed-off-by: Junjie Mao <junjie.mao@intel.com>
A DualNamePath clause is a NamePath that only follows rootchar or
prefixpath. Thus, it is never necessary to check if a dot is necessary for
separating segments before a DualNamePath. This patch removes the code that
conduct that check.
Tracked-On: #6287
Signed-off-by: Junjie Mao <junjie.mao@intel.com>
It is seen occasionally that a memory/port BAR of a PCI device is
programmed with the address 0 which is clearly invalid. This patch
gracefully handles this case by printing an error to warn the users that
this device cannot be passed through to any VM.
Tracked-On: #6298
Signed-off-by: Junjie Mao <junjie.mao@intel.com>
According to section 19 of ACPI spec 6.4, the following clauses open name
scopes (in addition to the Scope clauses).
- Function
- Device
- Method
- Power Resource
- Thermal Zone
The current AML parser only opens a scope when parsing DefMethod and
DefDevice, however. This patch fixes the AML parsing by opening a scope on
visiting a DefPowerRes or DefThermalZone clause.
Note: Functions in ASL are equivalent to Methods at AML level.
Tracked-On: #6298
Signed-off-by: Junjie Mao <junjie.mao@intel.com>
The current ConditionallyUnregisterSymbolVisitor has the following two
issues.
1. The visitor will crash when a DefIfElse node is not fully parsed due
to failed deferred expansion.
2. Nested DefIfElse of disabled blocks are still checked and one of its
branch may still take effect.
This patch fixes those issues by checking the predicates of a DefIfElse
block only when conditionally_hidden is False and check existence of
TermList and DefElse clauses.
Tracked-On: #6298
Signed-off-by: Junjie Mao <junjie.mao@intel.com>
When parsing a sequence of clauses, it is not necessary to peek an opcode
from the current stream unless that sequence starts with one. Peeking an
opcode is even an error when the actual clause is empty (e.g. as a
TermList).
This patch makes the SequenceFactory only peeking at the next opcode when
the grammar expects one.
Tracked-On: #6298
Signed-off-by: Junjie Mao <junjie.mao@intel.com>
This patch refines the AML parser to improve its readability and
performance in the following ways.
1. A Tree object now has the parts of the corresponding object being
member fields. As an example, a Tree with label `DefMethod` now has
members `NameString`, `MethodFlags` and `TermList`.
2. It is now possible to assign names each part of an object. The grammar
is updated to assign different names to the parts with the same type
in the same object.
3. Invocation to intermediate factories is now skipped. As an example,
when parsing a ByteConst that acts as a ByteIndex, the original
implementation invokes the following factories in sequence:
ByteIndex -> TermArg -> DataObject -> ComputationalData -> ByteConst
The factories TermArg, DataObject and ComputationalData does nothing
but forward the parsing to the next-level factory according to the
opcode of the next object. With this patch, the invocation sequence
becomes:
ByteIndex -> ByteConst
i.e. ByteIndex directly passes to ByteConst which can parse the next
opcode.
Tracked-On: #6298
Signed-off-by: Junjie Mao <junjie.mao@intel.com>
The current ACPI AML parser can generate incorrect AST if a DSDT/SSDT
satisfies the following:
1. The body of a method invokes a NameString that is defined later.
2. Before that method the same NameString is also defined but in an outer
scope and with a different number of parameter.
Since method bodies hardly define any further symbol that is referenced
outside the method itself, this patch forces the parsing of method bodies
to be deferred to the second pass when all symbols have been declared.
Tracked-On: #6298
Signed-off-by: Junjie Mao <junjie.mao@intel.com>
The current implementation of I/O buffers have the following issues.
1. I/O buffers are filled with values on creation. This may be fine for
memory-mapped I/O regions, but could be a problem to port I/O regions
and indexed I/O regions.
2. While not commonly seen, it IS witnessed that some devices only allow
its MMIO registers to be accessed with certain width. Accessing such
registers with a larger width will not be handled by the device,
causing SW to get all 1's rather than the actual values in these
registers.
This patch resolves the issues above as follows:
1. I/O buffers now do not access any register on creation. Instead, the
register is accessed only upon requests.
2. The access width of these registers are followed to ensure that the
registers are accessed properly.
The classes that represents buffers when interpreting AML is also
refactored to abstract the common code that manages fields within
buffers. The class hierarchy now looks like this:
BufferBase: Implement methods that registers, reads or writes fields
Buffer(BufferBase): Implement memory buffer
StreamIOBuffer(BufferBase): Implement I/Os available via /dev files
IndexedIOBuffer(BufferBase): Implement I/Os via index/data registers
Tracked-On: #6298
Signed-off-by: Junjie Mao <junjie.mao@intel.com>
DefDevide is now enountered when interpreting host DSDT/SSDT. This patch
implements the interpretation of the integer division operation.
Tracked-On: #6298
Signed-off-by: Junjie Mao <junjie.mao@intel.com>
The current implementation of the AML interpreter continues interpreting a
method after meeting a DefReturn object, which is incorrect. This patch
fixes this issue by raising a dedicated exception on return and catching
that exception on the caller side.
Tracked-On: #6298
Signed-off-by: Junjie Mao <junjie.mao@intel.com>
Update generic_board/generic_code folder with compile result
on the nuc11tnbi5 platform.
Tracked-On: #6292
Signed-off-by: Kunhui-Li <kunhuix.li@intel.com>
config editor creates the default scenario xmls and launch xmls
based on generic configs when users import a new board.
Tracked-On: #6208
Signed-off-by: Shuang Zheng <shuang.zheng@intel.com>
add interface to save scenario xmls and launch xmls to user defined
path in config editor; move all config xmls and generated scripts
out of acrn-hypervisor.
Tracked-On: #6208
Signed-off-by: Shuang Zheng <shuang.zheng@intel.com>
Reviewed-by: Victor Sun <victor.sun@intel.com>
While running in a nested environment, such as qemu, parse the board
information should be allowed even it is not in a native environment.
Replace the error with warning message and does not exit the program.
Tracked-On: #6208
Signed-off-by: Yang,Yu-chu <yu-chu.yang@intel.com>
Enlarge the max size to store the PTCT/RTCT table to 1k bytes
because the size of RTCT table exceeded the original max size
0x1100 - 0xF00 which makes RTCT table overlap other ACPI tables.
Tracked-On: #6303
Signed-off-by: Kunhui-Li <kunhuix.li@intel.com>
1. Update the vaule of the tag CLOS_MASK to 0xfffff according to board.xml
in all scenario xml files.
2. Replace industry_launch_2uos.xml launch file with industry_launch_6uos.xml.
3. Update logical_partition.xml file.
4. Remove hybrid_rt.xml file, then add a second POST_STD_VM in hybrid.xml and
add hybrid_launch_2uos.xml launch file correspondingly.
Tracked-On: #6244
Signed-off-by: Kunhui-Li <kunhuix.li@intel.com>
Add the MSI-X capability structure nodes under <capability
id="MSI-X"> in board.xml.
Example:
<capability id="MSI-X">
<table_size>16</table_size>
<table_bir>1</table_bir>
<table_offset>0x1000000</table_offset>
<pba_bir>1</pba_bir>
<pba_offset>0x0</pba_offset>
</capability>
Fix the MSI <count> nodes when there is only one vector.
Tracked-On: #6235
Signed-off-by: Yang,Yu-chu <yu-chu.yang@intel.com>
The PR 6236 has modified the board.xml format for MAX_MSIX_TABLE_NUM fix.
To compromise this PR, updates all the source file board.xmls.
Tracked-On: #6235
Signed-off-by: Kunhui-Li <kunhuix.li@intel.com>
IC_ASSIGN_MMIODEV -> ACRN_IOCTL_ASSIGN_MMIODEV
IC_DEASSIGN_MMIODEV -> ACRN_IOCTL_DEASSIGN_MMIODEV
struct acrn_mmiodev has slight change. Move struct acrn_mmiodev into
acrn_common.h because it is used by both DM and HV.
Tracked-On: #6282
Signed-off-by: Shuo A Liu <shuo.a.liu@intel.com>
In launch script, update cpu offline method to fix the issue
that it isn't offline cpu on ADL-S board.
Tracked-On: #6266
Signed-off-by: Kunhui-Li <kunhuix.li@intel.com>
1. Update the necessary libraries to consistent with the "Build ACRN From Source"
document in the "Getting Started Guide" document.
2. Delete the related introduction with acrngt.conf and launch_uos_id1.sh files in
"Getting Started Guide" document.
3. Update WHL-IPC-I7 board's processor in Supported HW document.
4. Add cpu_affinity element's description in ACRN Configuration Data.
5. Update the description for shm_region in Launch XML format.
6. Update configurable/readonly attributes values.
7. Update the description for hv.CAPACITIES.MAX_MSIX_TABLE_NUM in schema/config.xsd.
Tracked-On: #5692
Signed-off-by: Kunhui-Li <kunhuix.li@intel.com>
1. Update the value of the tag MAX_MSIX_TABLE_NUM from 64 to empty
for all scenario xml except ehl-crb-b board.
2. Update the value of the tag MAX_MSIX_TABLE_NUM to 96 for the
scenario xml on the ehl-crb-b board.
Tracked-On: #6186
Signed-off-by: Kunhui-Li <kunhuix.li@intel.com>
Update the value of the tag MAX_PT_IRQ_ENTRIES from 64 to 128
in TGL-RVP scenario xml.
Tracked-On: #6185
Signed-off-by: Kunhui-Li <kunhuix.li@intel.com>
A vmsix supported passthrough device expects the first unused bar region
for vmsix. Pop the first unused_bar_index in gpa.py instead.
Reference code: init_vmsix_on_msi of hypervisor\dm\vpci\vmsix_on_msi.c
Tracked-On: #6192
Signed-off-by: Yang,Yu-chu <yu-chu.yang@intel.com>
Without this, the GUEST_FLAG_NVMX_ENABLED doesn't show up in the
drop-down list of "guest_flags" in the ACRN config GUI.
Tracked-On: #5923
Signed-off-by: Zide Chen <zide.chen@intel.com>
Update the content about getting board xml from native
enviroment in acrn_configuration_tool.rst and README.
Tracked-On: #6134
Signed-off-by: Kunhui Li <kunhuix.li@intel.com>
Refine get_pt_intx_table. The method parse the <pt_inx> of scenario and
append the pair of phys_gsi and virt_gsi if there is any.
Refine check_pt_intx. Add a condition that the method returns if the
phys_gsi and virt_gsi are empty dictionary.
Tracked-On: #6178
Signed-off-by: Yang,Yu-chu <yu-chu.yang@intel.com>
Add flag "allow_trigger_s5" to launch script xmls. If this flag sets to
'y' and the poweroff_channel sets to "vuart1(pty)" or "vuart1(tty)", the
"allow_trigger_s5" will appends to the end of "--pm_notify_channel
uart".
Tracked-On: #6138
Signed-off-by: Yang,Yu-chu <yu-chu.yang@intel.com>
Make up qemu.xml to compromise the static allocators which use the
new xpath based on board inspector.
Tracked-On: #6102
Signed-off-by: Yang,Yu-chu <yu-chu.yang@intel.com>
For illegal characters, replace original characters with escaped characters in board.xml.
Tracked-On: #6113
Signed-off-by: Kunhui Li <kunhuix.li@intel.com>
Modify the initial value of PT_SLOT variable and
update the get slot logic that all device call the virtual_dev_slot function to get slot number directly.
Copy the launch_uos_id1.sh to launch_win.sh.
Tracked-On: #6072
Signed-off-by: Kunhui Li <kunhuix.li@intel.com>
Update the severity from "warning" to "error" for hybrid cores check.
Tracked-On: #5918
Reviewed-by: Junjie Mao <junjie.mao@intel.com>
Signed-off-by: Kunhui Li <kunhuix.li@intel.com>
Retrieve physical APIC IDs from board xml file and use them to fill in the ACPI MADT table
for pre-Launched VMs.
Note that the config-tool will throw an error if the processors/die/core/thread tags are absent.
User needs to run board_inspector.py to regenerate the board xml file when this commit is merged,
if the processors/die/core/thread tags are missing in the board xml file.
Tracked-On: #6020
Signed-off-by: dongshen <dongsheng.x.zhang@intel.com>
commit 873ed75 ("misc: sanity check VM config for nested virtualization")
requires that the guest_flag tag can't be empty, or it will fail to build.
This patch changes adl instances of "<guest_flag></guest_flag>" to
"<guest_flag>0</guest_flag>".
Tracked-On: #5923
Signed-off-by: Jiang, Yanting <yanting.jiang@intel.com>
Configure PTM in post-launched VM using <PTM> element. If the //vm/PTM
sets to 'y', pci_dev.c.xsl appends the virtual root port to
corresponding struct acrn_vm_pci_dev_config of that VM. Currently it
supports only post-launched VMs.
Configure enable_ptm for dm argument. If a uos/enable_ptm with uos id
= 'vm_id 'sets to 'y' and the vm/PTM with the same vm_id sets to 'y',
append an "enable_ptm" flag to the end of passthrough ethernet devices.
Currently there is only ethernet card can support the "enable_ptm"flag.
For the schema validation, the <PTM> can only be ['y', 'n'].
For the launched script validation, the <enable_ptm> can only be ['y',
'n']. If the <enable_ptm> sets to 'y' but the corresponding <PTM> sets
to 'n', the launch script will fail to generate.
Tracked-On: #6054
Signed-off-by: Yang,Yu-chu <yu-chu.yang@intel.com>
Add an xslt file "board_info.h.xsl". This file is used to
generate board_info.h which is used by hypervisor.
Tracked-On: #6024
Signed-off-by: Yang,Yu-chu <yu-chu.yang@intel.com>
Reviewed-by: Junjie Mao <junjie.mao@intel.com>
Add an xslt file "pci_dev.c.xsl". This file is used to
generate pci_dev.c which is used by hypervisor.
Tracked-On: #6024
Signed-off-by: Yang,Yu-chu <yu-chu.yang@intel.com>
Reviewed-by: Junjie Mao <junjie.mao@intel.com>
acrn:get-vbdf: get the virtual bdf from allocation.xml based on vmid and device name
acrn:get-pbdf: get physical bdf from <pci_dev>
acrn:ptdev-name-suffix: fix the name to look up allocation.xml
acrn:get-hidden-device-num: get the number of hidden devices based on
board name
acrn:is-vmsix-supported-device: check if a device is a vmsix supported
device based on the vendor and identifier
Tracked-On: #6024
Signed-off-by: Yang,Yu-chu <yu-chu.yang@intel.com>
Assign bdf to pci emulated and passthrough devices.
For pre-launched VM, assigns unique bdf to passthrough devices, inter-vm
shared memory, pci vuart(console and communication vuarts).
For SOS vm, assigns unique bdf to inter-vm shared memory and pci
vuart(console and communication vuarts).
The bdf follows the rules below:
- the bdf 00:00.0 is reserved for pci hostbridge
- the assigned bdf range: bus is 0x00, dev is in range [0x1, 0x20)
and the fuc is 0x00
- the bdf must be unique, which means any vm's emulated devices cannot
share the same bdf with existing devices
- some devices's bdf is hardcoded, modify its bdf would leads the
device cannot be dicoverd by os. A HARDCODED_BDF_LIST in bdf.py documents
them
- the passthrough devices' bdf can be reused in SOS vm
Tracked-On: #6024
Signed-off-by: Yang,Yu-chu <yu-chu.yang@intel.com>
Reviewed-by: Junjie Mao <junjie.mao@intel.com>
Add methods allocates the mmio bar base to console vuart,
communication vuarts, inter-vm shared memory and passthrough pci
devices.
For SOS:
- get low mem by parsing board xml.
- get high mem by parsing board xml, if the high mem is not enabled,
the high mem start address would be ~0UL and the end address is 0UL
- get the occupied mmio windows by parsing board.xml
- for each console vuart, communication vuart and inter-vm shared memory
devices, assign unused mmio windows to them
- all the assigned mmio windows must be unique and should not overlay
with any devices' mmio window
- the passthrough devices mmio windows can be reused in SOS vm
- each allocated mmio start address must be 4k alignment if the length
of bar is smaller than 4k
- each allocated mmio start address must be aligned with the bar length
if its length is greater than 4k
- the 32bits bar will fall in low mem range only
- 64bits bar will look for free mmio in low mem rage first, if the high
mem is enabled, the 64bits bar will look for free mmio in high mem
range if there is not enough space in low mem range
- allocator raises an error if there is not enough mmio space
For pre-launched VM:
- the high mem range is [256G, 512G)
- the low mem range is [2G, 3.5G)
- there is no used mmio window initially
- for each console vuart, communication vuart, inter-vm shared memory
devices and passthrough devices, assign unused mmio windows to them
- all the assigned mmio windows must be unique and should not overlay
with any devices' mmio window
- the 32bits bar will fall in low mem range only
- 64bits bar will look for free mmio in low mem rage first and then
look for free mmio in high mem range if there is not enough space in
low mem range
- each allocated mmio start address must be 4k alignment if the length
of bar is smaller than 4k
- each allocated mmio start address must be aligned with the bar length
if its lenght is greater than 4k
- allocator raises an error if there is not enough mmio space
Tracked-On: #6024
Signed-off-by: Yang,Yu-chu <yu-chu.yang@intel.com>
Reviewed-by: Junjie Mao <junjie.mao@intel.com>
Add a common method "get_shmem_regions":
This method get <IVSHMEM_REGION> and extracts the region size, region
position in xml and and vm ids which share this regions. Returns a
dictionary:
{'vm_id':{'region_name':{'id': region position,'size': region size,}}}
Add vm type checking methods:
is_pre_launched_vm, is_post_launched_vm and is_sos_vm.
Tracked-On: #6024
Signed-off-by: Yang,Yu-chu <yu-chu.yang@intel.com>
Reviewed-by: Junjie Mao <junjie.mao@intel.com>
If there is hugepage support from board xml, config tool will
add hugepagesz=1G hugepages=[size] into sos kernel cmdline,
the size is calculated by memory size in G minusing 3.
The reason for reducing 3 is that it is reserved for SOS VM use.
Tracked-On: #5815
Signed-off-by: Kunhui Li <kunhuix.li@intel.com>
Reviewed-By: Junjie Mao <junjie.mao@intel.com>
This patch applies the latest board inspector on ehl-crb-b and tgl-rvp to
generate additional information to the board XMLs.
Tracked-On: #5922
Signed-off-by: Junjie Mao <junjie.mao@intel.com>
