GNUBoot/coreboot-kgpe-d16
2
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity
coreboot-kgpe-d16/src/arch/x86/acpi_device.c

962 lines
22 KiB
C
Raw Normal View History

src/arch/x86: Convert to SPDX license header This also drops individual copyright notices, all mentioned authors in that part of the tree are listed in AUTHORS. Change-Id: Ib5a92bb46ff2b9d2928aae3763daec71747044c2 Signed-off-by: Patrick Georgi <pgeorgi@google.com> Reviewed-on: https://review.coreboot.org/c/coreboot/+/39284 Tested-by: build bot (Jenkins) <no-reply@coreboot.org> Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net> Reviewed-by: HAOUAS Elyes <ehaouas@noos.fr>
2020-03-04 15:10:45 +01:00
/* SPDX-License-Identifier: GPL-2.0-only */
/* This file is part of the coreboot project. */
device: Add an ACPI device name and path concept to devices Add a function to "struct device_operations" to return the ACPI name for the device, and helper functions to find this name (either from the device or its parent) and to build a fully qualified ACPI path from the root device. This addition will allow device drivers to generate their ACPI AML in the SSDT at boot, with customization supplied by devicetree.cb, instead of needing custom DSDT ASL for every mainboard. The root device acpi_name is defined as "\\_SB" and is used to start the path when building a fully qualified name. This requires SOC support to provide handlers for returning the ACPI name for devices that it owns, and those names must match the objects declared in the DSDT. The handler can be done either in each device driver or with a global handler for the entire SOC. Simplified example of how this can be used for an i2c device declared in devicetree.cb with: chip soc/intel/skylake # "\_SB" (from root device) device domain 0 on # "PCI0" device pci 19.2 on # "I2C4" chip drivers/i2c/test0 device i2c 1a.0 on end # "TST0" end end end end And basic SSDT generating code in the device driver: acpigen_write_scope(acpi_device_scope(dev)); acpigen_write_device(acpi_device_name(dev)); acpigen_write_string("_HID", "TEST0000"); acpigen_write_byte("_UID", 0); acpigen_pop_len(); /* device */ acpigen_pop_len(); /* scope */ Will produce this ACPI code: Scope (\_SB.PCI0.I2C4) { Device (TST0) { Name (_HID, "TEST0000") Name (_UID, 0) } } Change-Id: Ie149595aeab96266fa5f006e7934339f0119ac54 Signed-off-by: Duncan Laurie <dlaurie@chromium.org> Reviewed-on: https://review.coreboot.org/14840 Tested-by: build bot (Jenkins) Reviewed-by: Aaron Durbin <adurbin@chromium.org>
2016-05-09 03:15:25 +02:00
arch/x86/acpi_device: Add a helper function to write PCI device This change adds a helper function to write a PCI device with _ADR and _STA defined for it. BUG=b:153858769 Signed-off-by: Furquan Shaikh <furquan@google.com> Change-Id: I932af917d91198876fe8e90af9bb7a2531bd8960 Reviewed-on: https://review.coreboot.org/c/coreboot/+/40674 Reviewed-by: Aaron Durbin <adurbin@chromium.org> Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
2020-04-23 21:51:42 +02:00
#include <assert.h>
device: Add an ACPI device name and path concept to devices Add a function to "struct device_operations" to return the ACPI name for the device, and helper functions to find this name (either from the device or its parent) and to build a fully qualified ACPI path from the root device. This addition will allow device drivers to generate their ACPI AML in the SSDT at boot, with customization supplied by devicetree.cb, instead of needing custom DSDT ASL for every mainboard. The root device acpi_name is defined as "\\_SB" and is used to start the path when building a fully qualified name. This requires SOC support to provide handlers for returning the ACPI name for devices that it owns, and those names must match the objects declared in the DSDT. The handler can be done either in each device driver or with a global handler for the entire SOC. Simplified example of how this can be used for an i2c device declared in devicetree.cb with: chip soc/intel/skylake # "\_SB" (from root device) device domain 0 on # "PCI0" device pci 19.2 on # "I2C4" chip drivers/i2c/test0 device i2c 1a.0 on end # "TST0" end end end end And basic SSDT generating code in the device driver: acpigen_write_scope(acpi_device_scope(dev)); acpigen_write_device(acpi_device_name(dev)); acpigen_write_string("_HID", "TEST0000"); acpigen_write_byte("_UID", 0); acpigen_pop_len(); /* device */ acpigen_pop_len(); /* scope */ Will produce this ACPI code: Scope (\_SB.PCI0.I2C4) { Device (TST0) { Name (_HID, "TEST0000") Name (_UID, 0) } } Change-Id: Ie149595aeab96266fa5f006e7934339f0119ac54 Signed-off-by: Duncan Laurie <dlaurie@chromium.org> Reviewed-on: https://review.coreboot.org/14840 Tested-by: build bot (Jenkins) Reviewed-by: Aaron Durbin <adurbin@chromium.org>
2016-05-09 03:15:25 +02:00
#include <string.h>
#include <arch/acpi.h>
#include <arch/acpi_device.h>
#include <arch/acpigen.h>
#include <device/device.h>
#include <device/path.h>
src: Add missing include <stdlib.h> Change-Id: I17dc2fed6c6518daf5af286788c98c049088911e Signed-off-by: Elyes HAOUAS <ehaouas@noos.fr> Reviewed-on: https://review.coreboot.org/c/coreboot/+/37366 Tested-by: build bot (Jenkins) <no-reply@coreboot.org> Reviewed-by: Werner Zeh <werner.zeh@siemens.com>
2019-12-02 20:47:50 +01:00
#include <stdlib.h>
acpi: Be more ACPI compliant when generating _UID * Add function to generate unique _UID using CRC32 * Add function to write the _UID based on a device's ACPI path ACPI devices that have the same _HID must use different _UID. Linux doesn't care about _UID if it's not used. Windows 10 verifies the ACPI code on boot and BSODs if two devices with the same _HID share the same _UID. Fixes BSOD seen on Windows 10. Change-Id: I47cd5396060d325f9ce338afced6af021e7ff2b4 Signed-off-by: Patrick Rudolph <patrick.rudolph@9elements.com> Reviewed-on: https://review.coreboot.org/c/coreboot/+/37695 Tested-by: build bot (Jenkins) <no-reply@coreboot.org> Reviewed-by: Philipp Deppenwiese <zaolin.daisuki@gmail.com>
2019-12-13 12:16:06 +01:00
#include <crc_byte.h>
coreboot: Replace all IS_ENABLED(CONFIG_XXX) with CONFIG(XXX) This patch is a raw application of find src/ -type f | xargs sed -i -e 's/IS_ENABLED\s*(CONFIG_/CONFIG(/g' Change-Id: I6262d6d5c23cabe23c242b4f38d446b74fe16b88 Signed-off-by: Julius Werner <jwerner@chromium.org> Reviewed-on: https://review.coreboot.org/c/coreboot/+/31774 Tested-by: build bot (Jenkins) <no-reply@coreboot.org> Reviewed-by: Patrick Georgi <pgeorgi@google.com>
2019-03-06 01:53:33 +01:00
#if CONFIG(GENERIC_GPIO_LIB)
acpi_device: Add support for writing ACPI GPIO descriptors Add definitions to describe GPIOs in generated ACPI objects and a method to write a GpioIo() or GpioInt() descriptor to the SSDT. ACPI GPIOs have many possible configuration options and a structure is created to describe it accurately in ACPI terms. There are many shared descriptor fields between GpioIo() and GpioInt() so the same function can write both types. GpioInt shares many properties with ACPI Interrupts and the same types are re-used here where possible. One addition is that GpioInt can be configured to trigger on both low and high edge transitions. One descriptor can describe multiple GPIO pins (limited to 8 in this implementation) that all share configuration and controller and are used by the same device scope. Accurately referring to the GPIO controller that this pin is connected to requires the SoC/board to implement a function handler for acpi_gpio_path(), or for the caller to provide this directly as a string in the acpi_gpio->reference variable. This will get used by device drivers to describe their resources in the SSDT. Here is a sample for a Maxim 98357A I2S codec which has a GPIO for power and channel selection called "sdmode". chip.h: struct drivers_generic_max98357a_config { struct acpi_gpio sdmode_gpio; }; max98357a.c: void acpi_fill_ssdt_generator(struct device *dev) { struct drivers_generic_max98357a_config *config = dev->chip_info; ... acpi_device_write_gpio(&config->sdmode_gpio); ... } devicetree.cb: device pci 1f.3 on chip drivers/generic/max98357a register "sdmode_gpio" = "ACPI_GPIO_OUTPUT(GPP_C5)" device generic 0 on end end end SSDT.dsl: GpioIo (Exclusive, PullDefault, 0, 0, IoRestrictionOutputOnly, "\\_SB.PCI0.GPIO", 0, ResourceConsumer, ,) { 53 } Signed-off-by: Duncan Laurie <dlaurie@chromium.org> Change-Id: Ibf5bab9c4bf6f21252373fb013e78f872550b167 Reviewed-on: https://review.coreboot.org/14934 Tested-by: build bot (Jenkins) Reviewed-by: Aaron Durbin <adurbin@chromium.org>
2016-05-10 02:08:38 +02:00
#include <gpio.h>
#endif
device: Add an ACPI device name and path concept to devices Add a function to "struct device_operations" to return the ACPI name for the device, and helper functions to find this name (either from the device or its parent) and to build a fully qualified ACPI path from the root device. This addition will allow device drivers to generate their ACPI AML in the SSDT at boot, with customization supplied by devicetree.cb, instead of needing custom DSDT ASL for every mainboard. The root device acpi_name is defined as "\\_SB" and is used to start the path when building a fully qualified name. This requires SOC support to provide handlers for returning the ACPI name for devices that it owns, and those names must match the objects declared in the DSDT. The handler can be done either in each device driver or with a global handler for the entire SOC. Simplified example of how this can be used for an i2c device declared in devicetree.cb with: chip soc/intel/skylake # "\_SB" (from root device) device domain 0 on # "PCI0" device pci 19.2 on # "I2C4" chip drivers/i2c/test0 device i2c 1a.0 on end # "TST0" end end end end And basic SSDT generating code in the device driver: acpigen_write_scope(acpi_device_scope(dev)); acpigen_write_device(acpi_device_name(dev)); acpigen_write_string("_HID", "TEST0000"); acpigen_write_byte("_UID", 0); acpigen_pop_len(); /* device */ acpigen_pop_len(); /* scope */ Will produce this ACPI code: Scope (\_SB.PCI0.I2C4) { Device (TST0) { Name (_HID, "TEST0000") Name (_UID, 0) } } Change-Id: Ie149595aeab96266fa5f006e7934339f0119ac54 Signed-off-by: Duncan Laurie <dlaurie@chromium.org> Reviewed-on: https://review.coreboot.org/14840 Tested-by: build bot (Jenkins) Reviewed-by: Aaron Durbin <adurbin@chromium.org>
2016-05-09 03:15:25 +02:00
acpi: Change device properties to work as a tree There is a second ACPI _DSD document from the UEFI Forum that details how _DSD style tables can be nested, creating a tree of similarly formatted tables. This document is linked from acpi_device.h. In order to support this the device property interface needs to be more flexible and build up a tree of properties to write all entries at once instead of writing each entry as it is generated. In the end this is a more flexible solution that can support drivers that need child tables like the DA7219 codec, while only requiring minor changes to the existing drivers that use the device property interface. This was tested on reef (apollolake) and chell (skylake) boards to ensure that there was no change in the generated SSDT AML. Change-Id: Ia22e3a5fd3982ffa7c324bee1a8d190d49f853dd Signed-off-by: Duncan Laurie <dlaurie@chromium.org> Reviewed-on: https://review.coreboot.org/15537 Tested-by: build bot (Jenkins) Reviewed-by: Aaron Durbin <adurbin@chromium.org>
2016-07-03 04:56:06 +02:00
#define ACPI_DP_UUID "daffd814-6eba-4d8c-8a91-bc9bbf4aa301"
#define ACPI_DP_CHILD_UUID "dbb8e3e6-5886-4ba6-8795-1319f52a966b"
acpi_device: Add support for writing ACPI Interrupt descriptors Add definitions for ACPI device extended interrupts and a method to write an Interrupt() descriptor to the SSDT output stream. Interrupts are often tied together with other resources and some configuration items are shared (though not always compatibly) with other constructs like GPIOs and GPEs. These will get used by device drivers to write _CRS sections for devices into the SSDT. One usage is to include a "struct acpi_irq" inside a config struct for a device so it can be initialized based on settings in devicetree. Example usage: chip.h: struct drivers_i2c_generic_config { struct acpi_irq irq; }; generic.c: void acpi_fill_ssdt_generator(struct device *dev) { struct drivers_i2c_generic_config *config = dev->chip_info; ... acpi_device_write_interrupt(&config->irq); ... } devicetree.cb: device pci 15.0 on chip drivers/i2c/generic register "irq" = "IRQ_EDGE_LOW(GPP_E7_IRQ)" device i2c 10 on end end end SSDT.dsl: Interrupt (ResourceConsumer, Edge, ActiveLow, Exclusive,,,) { 31 } Change-Id: I3b64170cc2ebac178e7a17df479eda7670a42703 Signed-off-by: Duncan Laurie <dlaurie@chromium.org> Reviewed-on: https://review.coreboot.org/14933 Tested-by: build bot (Jenkins) Reviewed-by: Aaron Durbin <adurbin@chromium.org>
2016-05-10 00:38:36 +02:00
/* Write empty word value and return pointer to it */
static void *acpi_device_write_zero_len(void)
{
char *p = acpigen_get_current();
acpigen_emit_word(0);
return p;
}
/* Fill in length value from start to current at specified location */
static void acpi_device_fill_from_len(char *ptr, char *start)
{
uint16_t len = acpigen_get_current() - start;
ptr[0] = len & 0xff;
ptr[1] = (len >> 8) & 0xff;
}
/*
* Fill in the length field with the value calculated from after
* the 16bit field to acpigen current as this length value does
* not include the length field itself.
*/
static void acpi_device_fill_len(void *ptr)
{
acpi_device_fill_from_len(ptr, ptr + sizeof(uint16_t));
}
device: Add an ACPI device name and path concept to devices Add a function to "struct device_operations" to return the ACPI name for the device, and helper functions to find this name (either from the device or its parent) and to build a fully qualified ACPI path from the root device. This addition will allow device drivers to generate their ACPI AML in the SSDT at boot, with customization supplied by devicetree.cb, instead of needing custom DSDT ASL for every mainboard. The root device acpi_name is defined as "\\_SB" and is used to start the path when building a fully qualified name. This requires SOC support to provide handlers for returning the ACPI name for devices that it owns, and those names must match the objects declared in the DSDT. The handler can be done either in each device driver or with a global handler for the entire SOC. Simplified example of how this can be used for an i2c device declared in devicetree.cb with: chip soc/intel/skylake # "\_SB" (from root device) device domain 0 on # "PCI0" device pci 19.2 on # "I2C4" chip drivers/i2c/test0 device i2c 1a.0 on end # "TST0" end end end end And basic SSDT generating code in the device driver: acpigen_write_scope(acpi_device_scope(dev)); acpigen_write_device(acpi_device_name(dev)); acpigen_write_string("_HID", "TEST0000"); acpigen_write_byte("_UID", 0); acpigen_pop_len(); /* device */ acpigen_pop_len(); /* scope */ Will produce this ACPI code: Scope (\_SB.PCI0.I2C4) { Device (TST0) { Name (_HID, "TEST0000") Name (_UID, 0) } } Change-Id: Ie149595aeab96266fa5f006e7934339f0119ac54 Signed-off-by: Duncan Laurie <dlaurie@chromium.org> Reviewed-on: https://review.coreboot.org/14840 Tested-by: build bot (Jenkins) Reviewed-by: Aaron Durbin <adurbin@chromium.org>
2016-05-09 03:15:25 +02:00
/* Locate and return the ACPI name for this device */
src/arch/x86/acpi: Constify struct device instances Constify the struct device arguments in below APIs: > acpi_device_name > acpi_device_path_fill > acpi_device_path > acpi_write_dbg2_pci_uart > acpi_device_scope > acpi_device_path_join The APIs do not seem to modify the argument and are using device argument as reference to device only. Change-Id: Ic2ce045f17efa288eb41503795723d0ad5ec78bd Signed-off-by: Aamir Bohra <aamir.bohra@intel.com> Reviewed-on: https://review.coreboot.org/c/coreboot/+/34625 Reviewed-by: Subrata Banik <subrata.banik@intel.com> Reviewed-by: Patrick Rudolph <patrick.rudolph@9elements.com> Reviewed-by: Nico Huber <nico.h@gmx.de> Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
2019-07-30 06:41:41 +02:00
const char *acpi_device_name(const struct device *dev)
device: Add an ACPI device name and path concept to devices Add a function to "struct device_operations" to return the ACPI name for the device, and helper functions to find this name (either from the device or its parent) and to build a fully qualified ACPI path from the root device. This addition will allow device drivers to generate their ACPI AML in the SSDT at boot, with customization supplied by devicetree.cb, instead of needing custom DSDT ASL for every mainboard. The root device acpi_name is defined as "\\_SB" and is used to start the path when building a fully qualified name. This requires SOC support to provide handlers for returning the ACPI name for devices that it owns, and those names must match the objects declared in the DSDT. The handler can be done either in each device driver or with a global handler for the entire SOC. Simplified example of how this can be used for an i2c device declared in devicetree.cb with: chip soc/intel/skylake # "\_SB" (from root device) device domain 0 on # "PCI0" device pci 19.2 on # "I2C4" chip drivers/i2c/test0 device i2c 1a.0 on end # "TST0" end end end end And basic SSDT generating code in the device driver: acpigen_write_scope(acpi_device_scope(dev)); acpigen_write_device(acpi_device_name(dev)); acpigen_write_string("_HID", "TEST0000"); acpigen_write_byte("_UID", 0); acpigen_pop_len(); /* device */ acpigen_pop_len(); /* scope */ Will produce this ACPI code: Scope (\_SB.PCI0.I2C4) { Device (TST0) { Name (_HID, "TEST0000") Name (_UID, 0) } } Change-Id: Ie149595aeab96266fa5f006e7934339f0119ac54 Signed-off-by: Duncan Laurie <dlaurie@chromium.org> Reviewed-on: https://review.coreboot.org/14840 Tested-by: build bot (Jenkins) Reviewed-by: Aaron Durbin <adurbin@chromium.org>
2016-05-09 03:15:25 +02:00
{
src/arch/x86/acpi: Constify struct device instances Constify the struct device arguments in below APIs: > acpi_device_name > acpi_device_path_fill > acpi_device_path > acpi_write_dbg2_pci_uart > acpi_device_scope > acpi_device_path_join The APIs do not seem to modify the argument and are using device argument as reference to device only. Change-Id: Ic2ce045f17efa288eb41503795723d0ad5ec78bd Signed-off-by: Aamir Bohra <aamir.bohra@intel.com> Reviewed-on: https://review.coreboot.org/c/coreboot/+/34625 Reviewed-by: Subrata Banik <subrata.banik@intel.com> Reviewed-by: Patrick Rudolph <patrick.rudolph@9elements.com> Reviewed-by: Nico Huber <nico.h@gmx.de> Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
2019-07-30 06:41:41 +02:00
const struct device *pdev = dev;
acpi: device: Walk up the tree to find identifier Instead of just checking the immediate parent for an device name, walk up the tree to check if any parent can identify the device. This allows devices to be nested more than one level deep and still have them identified in one place by the SOC. The recursive method calling this function has been changed to handle a null return from acpi_device_name and abort instead of continuing and perhaps forming an invalid ACPI path. Change-Id: Ic17c5b6facdcb1a0ac696912867d62652b2bf18e Signed-off-by: Duncan Laurie <dlaurie@google.com> Reviewed-on: https://review.coreboot.org/26487 Tested-by: build bot (Jenkins) <no-reply@coreboot.org> Reviewed-by: Marc Jones <marc@marcjonesconsulting.com> Reviewed-by: Aaron Durbin <adurbin@chromium.org> Reviewed-by: Furquan Shaikh <furquan@google.com>
2018-05-07 23:28:53 +02:00
const char *name = NULL;
device: Add an ACPI device name and path concept to devices Add a function to "struct device_operations" to return the ACPI name for the device, and helper functions to find this name (either from the device or its parent) and to build a fully qualified ACPI path from the root device. This addition will allow device drivers to generate their ACPI AML in the SSDT at boot, with customization supplied by devicetree.cb, instead of needing custom DSDT ASL for every mainboard. The root device acpi_name is defined as "\\_SB" and is used to start the path when building a fully qualified name. This requires SOC support to provide handlers for returning the ACPI name for devices that it owns, and those names must match the objects declared in the DSDT. The handler can be done either in each device driver or with a global handler for the entire SOC. Simplified example of how this can be used for an i2c device declared in devicetree.cb with: chip soc/intel/skylake # "\_SB" (from root device) device domain 0 on # "PCI0" device pci 19.2 on # "I2C4" chip drivers/i2c/test0 device i2c 1a.0 on end # "TST0" end end end end And basic SSDT generating code in the device driver: acpigen_write_scope(acpi_device_scope(dev)); acpigen_write_device(acpi_device_name(dev)); acpigen_write_string("_HID", "TEST0000"); acpigen_write_byte("_UID", 0); acpigen_pop_len(); /* device */ acpigen_pop_len(); /* scope */ Will produce this ACPI code: Scope (\_SB.PCI0.I2C4) { Device (TST0) { Name (_HID, "TEST0000") Name (_UID, 0) } } Change-Id: Ie149595aeab96266fa5f006e7934339f0119ac54 Signed-off-by: Duncan Laurie <dlaurie@chromium.org> Reviewed-on: https://review.coreboot.org/14840 Tested-by: build bot (Jenkins) Reviewed-by: Aaron Durbin <adurbin@chromium.org>
2016-05-09 03:15:25 +02:00
if (!dev)
return NULL;
/* Check for device specific handler */
if (dev->ops->acpi_name)
return dev->ops->acpi_name(dev);
acpi: device: Walk up the tree to find identifier Instead of just checking the immediate parent for an device name, walk up the tree to check if any parent can identify the device. This allows devices to be nested more than one level deep and still have them identified in one place by the SOC. The recursive method calling this function has been changed to handle a null return from acpi_device_name and abort instead of continuing and perhaps forming an invalid ACPI path. Change-Id: Ic17c5b6facdcb1a0ac696912867d62652b2bf18e Signed-off-by: Duncan Laurie <dlaurie@google.com> Reviewed-on: https://review.coreboot.org/26487 Tested-by: build bot (Jenkins) <no-reply@coreboot.org> Reviewed-by: Marc Jones <marc@marcjonesconsulting.com> Reviewed-by: Aaron Durbin <adurbin@chromium.org> Reviewed-by: Furquan Shaikh <furquan@google.com>
2018-05-07 23:28:53 +02:00
/* Walk up the tree to find if any parent can identify this device */
while (pdev->bus) {
pdev = pdev->bus->dev;
if (!pdev)
break;
if (pdev->path.type == DEVICE_PATH_ROOT)
break;
if (pdev->ops && pdev->ops->acpi_name)
name = pdev->ops->acpi_name(dev);
if (name)
return name;
}
device: Add an ACPI device name and path concept to devices Add a function to "struct device_operations" to return the ACPI name for the device, and helper functions to find this name (either from the device or its parent) and to build a fully qualified ACPI path from the root device. This addition will allow device drivers to generate their ACPI AML in the SSDT at boot, with customization supplied by devicetree.cb, instead of needing custom DSDT ASL for every mainboard. The root device acpi_name is defined as "\\_SB" and is used to start the path when building a fully qualified name. This requires SOC support to provide handlers for returning the ACPI name for devices that it owns, and those names must match the objects declared in the DSDT. The handler can be done either in each device driver or with a global handler for the entire SOC. Simplified example of how this can be used for an i2c device declared in devicetree.cb with: chip soc/intel/skylake # "\_SB" (from root device) device domain 0 on # "PCI0" device pci 19.2 on # "I2C4" chip drivers/i2c/test0 device i2c 1a.0 on end # "TST0" end end end end And basic SSDT generating code in the device driver: acpigen_write_scope(acpi_device_scope(dev)); acpigen_write_device(acpi_device_name(dev)); acpigen_write_string("_HID", "TEST0000"); acpigen_write_byte("_UID", 0); acpigen_pop_len(); /* device */ acpigen_pop_len(); /* scope */ Will produce this ACPI code: Scope (\_SB.PCI0.I2C4) { Device (TST0) { Name (_HID, "TEST0000") Name (_UID, 0) } } Change-Id: Ie149595aeab96266fa5f006e7934339f0119ac54 Signed-off-by: Duncan Laurie <dlaurie@chromium.org> Reviewed-on: https://review.coreboot.org/14840 Tested-by: build bot (Jenkins) Reviewed-by: Aaron Durbin <adurbin@chromium.org>
2016-05-09 03:15:25 +02:00
return NULL;
}
arch/x86/acpi: Add acpi_device_hid Allow a driver to return device specific _HID, which will be consumed by acpigen in order to generate proper SSDTs. Change-Id: Ibb79eb00c008a3c3cdc12ad2a48b88a055a9216f Signed-off-by: Patrick Rudolph <patrick.rudolph@9elements.com> Reviewed-on: https://review.coreboot.org/c/coreboot/+/35006 Reviewed-by: Philipp Deppenwiese <zaolin.daisuki@gmail.com> Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
2019-08-20 08:20:01 +02:00
/* Locate and return the ACPI _HID (Hardware ID) for this device */
const char *acpi_device_hid(const struct device *dev)
{
if (!dev)
return NULL;
/* Check for device specific handler */
if (dev->ops->acpi_hid)
return dev->ops->acpi_hid(dev);
/*
* Don't walk up the tree to find any parent that can identify this device, as
* PNP devices are hard to identify.
*/
return NULL;
}
acpi: Be more ACPI compliant when generating _UID * Add function to generate unique _UID using CRC32 * Add function to write the _UID based on a device's ACPI path ACPI devices that have the same _HID must use different _UID. Linux doesn't care about _UID if it's not used. Windows 10 verifies the ACPI code on boot and BSODs if two devices with the same _HID share the same _UID. Fixes BSOD seen on Windows 10. Change-Id: I47cd5396060d325f9ce338afced6af021e7ff2b4 Signed-off-by: Patrick Rudolph <patrick.rudolph@9elements.com> Reviewed-on: https://review.coreboot.org/c/coreboot/+/37695 Tested-by: build bot (Jenkins) <no-reply@coreboot.org> Reviewed-by: Philipp Deppenwiese <zaolin.daisuki@gmail.com>
2019-12-13 12:16:06 +01:00
/*
* Generate unique ID based on the ACPI path.
* Collisions on the same _HID are possible but very unlikely.
*/
arch/x86/acpi_device: Constify struct device * parameter to UID functions acpi_device_uid() and acpi_device_write_uid() do not need to make changes to the device structure. Thus, this change marks struct device * parameter to these functions as const. Change-Id: I3755223766c78f93c57ac80caf392985cfd5c5e5 Signed-off-by: Furquan Shaikh <furquan@google.com> Reviewed-on: https://review.coreboot.org/c/coreboot/+/40702 Reviewed-by: HAOUAS Elyes <ehaouas@noos.fr> Reviewed-by: Aaron Durbin <adurbin@chromium.org> Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
2020-04-25 06:27:29 +02:00
uint32_t acpi_device_uid(const struct device *dev)
acpi: Be more ACPI compliant when generating _UID * Add function to generate unique _UID using CRC32 * Add function to write the _UID based on a device's ACPI path ACPI devices that have the same _HID must use different _UID. Linux doesn't care about _UID if it's not used. Windows 10 verifies the ACPI code on boot and BSODs if two devices with the same _HID share the same _UID. Fixes BSOD seen on Windows 10. Change-Id: I47cd5396060d325f9ce338afced6af021e7ff2b4 Signed-off-by: Patrick Rudolph <patrick.rudolph@9elements.com> Reviewed-on: https://review.coreboot.org/c/coreboot/+/37695 Tested-by: build bot (Jenkins) <no-reply@coreboot.org> Reviewed-by: Philipp Deppenwiese <zaolin.daisuki@gmail.com>
2019-12-13 12:16:06 +01:00
{
const char *path = acpi_device_path(dev);
if (!path)
return 0;
return CRC(path, strlen(path), crc32_byte);
}
device: Add an ACPI device name and path concept to devices Add a function to "struct device_operations" to return the ACPI name for the device, and helper functions to find this name (either from the device or its parent) and to build a fully qualified ACPI path from the root device. This addition will allow device drivers to generate their ACPI AML in the SSDT at boot, with customization supplied by devicetree.cb, instead of needing custom DSDT ASL for every mainboard. The root device acpi_name is defined as "\\_SB" and is used to start the path when building a fully qualified name. This requires SOC support to provide handlers for returning the ACPI name for devices that it owns, and those names must match the objects declared in the DSDT. The handler can be done either in each device driver or with a global handler for the entire SOC. Simplified example of how this can be used for an i2c device declared in devicetree.cb with: chip soc/intel/skylake # "\_SB" (from root device) device domain 0 on # "PCI0" device pci 19.2 on # "I2C4" chip drivers/i2c/test0 device i2c 1a.0 on end # "TST0" end end end end And basic SSDT generating code in the device driver: acpigen_write_scope(acpi_device_scope(dev)); acpigen_write_device(acpi_device_name(dev)); acpigen_write_string("_HID", "TEST0000"); acpigen_write_byte("_UID", 0); acpigen_pop_len(); /* device */ acpigen_pop_len(); /* scope */ Will produce this ACPI code: Scope (\_SB.PCI0.I2C4) { Device (TST0) { Name (_HID, "TEST0000") Name (_UID, 0) } } Change-Id: Ie149595aeab96266fa5f006e7934339f0119ac54 Signed-off-by: Duncan Laurie <dlaurie@chromium.org> Reviewed-on: https://review.coreboot.org/14840 Tested-by: build bot (Jenkins) Reviewed-by: Aaron Durbin <adurbin@chromium.org>
2016-05-09 03:15:25 +02:00
/* Recursive function to find the root device and print a path from there */
src/arch/x86/acpi: Constify struct device instances Constify the struct device arguments in below APIs: > acpi_device_name > acpi_device_path_fill > acpi_device_path > acpi_write_dbg2_pci_uart > acpi_device_scope > acpi_device_path_join The APIs do not seem to modify the argument and are using device argument as reference to device only. Change-Id: Ic2ce045f17efa288eb41503795723d0ad5ec78bd Signed-off-by: Aamir Bohra <aamir.bohra@intel.com> Reviewed-on: https://review.coreboot.org/c/coreboot/+/34625 Reviewed-by: Subrata Banik <subrata.banik@intel.com> Reviewed-by: Patrick Rudolph <patrick.rudolph@9elements.com> Reviewed-by: Nico Huber <nico.h@gmx.de> Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
2019-07-30 06:41:41 +02:00
static ssize_t acpi_device_path_fill(const struct device *dev, char *buf,
acpi: device: Walk up the tree to find identifier Instead of just checking the immediate parent for an device name, walk up the tree to check if any parent can identify the device. This allows devices to be nested more than one level deep and still have them identified in one place by the SOC. The recursive method calling this function has been changed to handle a null return from acpi_device_name and abort instead of continuing and perhaps forming an invalid ACPI path. Change-Id: Ic17c5b6facdcb1a0ac696912867d62652b2bf18e Signed-off-by: Duncan Laurie <dlaurie@google.com> Reviewed-on: https://review.coreboot.org/26487 Tested-by: build bot (Jenkins) <no-reply@coreboot.org> Reviewed-by: Marc Jones <marc@marcjonesconsulting.com> Reviewed-by: Aaron Durbin <adurbin@chromium.org> Reviewed-by: Furquan Shaikh <furquan@google.com>
2018-05-07 23:28:53 +02:00
size_t buf_len, size_t cur)
device: Add an ACPI device name and path concept to devices Add a function to "struct device_operations" to return the ACPI name for the device, and helper functions to find this name (either from the device or its parent) and to build a fully qualified ACPI path from the root device. This addition will allow device drivers to generate their ACPI AML in the SSDT at boot, with customization supplied by devicetree.cb, instead of needing custom DSDT ASL for every mainboard. The root device acpi_name is defined as "\\_SB" and is used to start the path when building a fully qualified name. This requires SOC support to provide handlers for returning the ACPI name for devices that it owns, and those names must match the objects declared in the DSDT. The handler can be done either in each device driver or with a global handler for the entire SOC. Simplified example of how this can be used for an i2c device declared in devicetree.cb with: chip soc/intel/skylake # "\_SB" (from root device) device domain 0 on # "PCI0" device pci 19.2 on # "I2C4" chip drivers/i2c/test0 device i2c 1a.0 on end # "TST0" end end end end And basic SSDT generating code in the device driver: acpigen_write_scope(acpi_device_scope(dev)); acpigen_write_device(acpi_device_name(dev)); acpigen_write_string("_HID", "TEST0000"); acpigen_write_byte("_UID", 0); acpigen_pop_len(); /* device */ acpigen_pop_len(); /* scope */ Will produce this ACPI code: Scope (\_SB.PCI0.I2C4) { Device (TST0) { Name (_HID, "TEST0000") Name (_UID, 0) } } Change-Id: Ie149595aeab96266fa5f006e7934339f0119ac54 Signed-off-by: Duncan Laurie <dlaurie@chromium.org> Reviewed-on: https://review.coreboot.org/14840 Tested-by: build bot (Jenkins) Reviewed-by: Aaron Durbin <adurbin@chromium.org>
2016-05-09 03:15:25 +02:00
{
const char *name = acpi_device_name(dev);
acpi: device: Walk up the tree to find identifier Instead of just checking the immediate parent for an device name, walk up the tree to check if any parent can identify the device. This allows devices to be nested more than one level deep and still have them identified in one place by the SOC. The recursive method calling this function has been changed to handle a null return from acpi_device_name and abort instead of continuing and perhaps forming an invalid ACPI path. Change-Id: Ic17c5b6facdcb1a0ac696912867d62652b2bf18e Signed-off-by: Duncan Laurie <dlaurie@google.com> Reviewed-on: https://review.coreboot.org/26487 Tested-by: build bot (Jenkins) <no-reply@coreboot.org> Reviewed-by: Marc Jones <marc@marcjonesconsulting.com> Reviewed-by: Aaron Durbin <adurbin@chromium.org> Reviewed-by: Furquan Shaikh <furquan@google.com>
2018-05-07 23:28:53 +02:00
ssize_t next = 0;
if (!name)
return -1;
device: Add an ACPI device name and path concept to devices Add a function to "struct device_operations" to return the ACPI name for the device, and helper functions to find this name (either from the device or its parent) and to build a fully qualified ACPI path from the root device. This addition will allow device drivers to generate their ACPI AML in the SSDT at boot, with customization supplied by devicetree.cb, instead of needing custom DSDT ASL for every mainboard. The root device acpi_name is defined as "\\_SB" and is used to start the path when building a fully qualified name. This requires SOC support to provide handlers for returning the ACPI name for devices that it owns, and those names must match the objects declared in the DSDT. The handler can be done either in each device driver or with a global handler for the entire SOC. Simplified example of how this can be used for an i2c device declared in devicetree.cb with: chip soc/intel/skylake # "\_SB" (from root device) device domain 0 on # "PCI0" device pci 19.2 on # "I2C4" chip drivers/i2c/test0 device i2c 1a.0 on end # "TST0" end end end end And basic SSDT generating code in the device driver: acpigen_write_scope(acpi_device_scope(dev)); acpigen_write_device(acpi_device_name(dev)); acpigen_write_string("_HID", "TEST0000"); acpigen_write_byte("_UID", 0); acpigen_pop_len(); /* device */ acpigen_pop_len(); /* scope */ Will produce this ACPI code: Scope (\_SB.PCI0.I2C4) { Device (TST0) { Name (_HID, "TEST0000") Name (_UID, 0) } } Change-Id: Ie149595aeab96266fa5f006e7934339f0119ac54 Signed-off-by: Duncan Laurie <dlaurie@chromium.org> Reviewed-on: https://review.coreboot.org/14840 Tested-by: build bot (Jenkins) Reviewed-by: Aaron Durbin <adurbin@chromium.org>
2016-05-09 03:15:25 +02:00
/*
* Make sure this name segment will fit, including the path segment
* separator and possible NUL terminator if this is the last segment.
*/
acpi: device: Walk up the tree to find identifier Instead of just checking the immediate parent for an device name, walk up the tree to check if any parent can identify the device. This allows devices to be nested more than one level deep and still have them identified in one place by the SOC. The recursive method calling this function has been changed to handle a null return from acpi_device_name and abort instead of continuing and perhaps forming an invalid ACPI path. Change-Id: Ic17c5b6facdcb1a0ac696912867d62652b2bf18e Signed-off-by: Duncan Laurie <dlaurie@google.com> Reviewed-on: https://review.coreboot.org/26487 Tested-by: build bot (Jenkins) <no-reply@coreboot.org> Reviewed-by: Marc Jones <marc@marcjonesconsulting.com> Reviewed-by: Aaron Durbin <adurbin@chromium.org> Reviewed-by: Furquan Shaikh <furquan@google.com>
2018-05-07 23:28:53 +02:00
if (!dev || (cur + strlen(name) + 2) > buf_len)
device: Add an ACPI device name and path concept to devices Add a function to "struct device_operations" to return the ACPI name for the device, and helper functions to find this name (either from the device or its parent) and to build a fully qualified ACPI path from the root device. This addition will allow device drivers to generate their ACPI AML in the SSDT at boot, with customization supplied by devicetree.cb, instead of needing custom DSDT ASL for every mainboard. The root device acpi_name is defined as "\\_SB" and is used to start the path when building a fully qualified name. This requires SOC support to provide handlers for returning the ACPI name for devices that it owns, and those names must match the objects declared in the DSDT. The handler can be done either in each device driver or with a global handler for the entire SOC. Simplified example of how this can be used for an i2c device declared in devicetree.cb with: chip soc/intel/skylake # "\_SB" (from root device) device domain 0 on # "PCI0" device pci 19.2 on # "I2C4" chip drivers/i2c/test0 device i2c 1a.0 on end # "TST0" end end end end And basic SSDT generating code in the device driver: acpigen_write_scope(acpi_device_scope(dev)); acpigen_write_device(acpi_device_name(dev)); acpigen_write_string("_HID", "TEST0000"); acpigen_write_byte("_UID", 0); acpigen_pop_len(); /* device */ acpigen_pop_len(); /* scope */ Will produce this ACPI code: Scope (\_SB.PCI0.I2C4) { Device (TST0) { Name (_HID, "TEST0000") Name (_UID, 0) } } Change-Id: Ie149595aeab96266fa5f006e7934339f0119ac54 Signed-off-by: Duncan Laurie <dlaurie@chromium.org> Reviewed-on: https://review.coreboot.org/14840 Tested-by: build bot (Jenkins) Reviewed-by: Aaron Durbin <adurbin@chromium.org>
2016-05-09 03:15:25 +02:00
return cur;
/* Walk up the tree to the root device */
if (dev->path.type != DEVICE_PATH_ROOT && dev->bus && dev->bus->dev)
next = acpi_device_path_fill(dev->bus->dev, buf, buf_len, cur);
acpi: device: Walk up the tree to find identifier Instead of just checking the immediate parent for an device name, walk up the tree to check if any parent can identify the device. This allows devices to be nested more than one level deep and still have them identified in one place by the SOC. The recursive method calling this function has been changed to handle a null return from acpi_device_name and abort instead of continuing and perhaps forming an invalid ACPI path. Change-Id: Ic17c5b6facdcb1a0ac696912867d62652b2bf18e Signed-off-by: Duncan Laurie <dlaurie@google.com> Reviewed-on: https://review.coreboot.org/26487 Tested-by: build bot (Jenkins) <no-reply@coreboot.org> Reviewed-by: Marc Jones <marc@marcjonesconsulting.com> Reviewed-by: Aaron Durbin <adurbin@chromium.org> Reviewed-by: Furquan Shaikh <furquan@google.com>
2018-05-07 23:28:53 +02:00
if (next < 0)
return next;
device: Add an ACPI device name and path concept to devices Add a function to "struct device_operations" to return the ACPI name for the device, and helper functions to find this name (either from the device or its parent) and to build a fully qualified ACPI path from the root device. This addition will allow device drivers to generate their ACPI AML in the SSDT at boot, with customization supplied by devicetree.cb, instead of needing custom DSDT ASL for every mainboard. The root device acpi_name is defined as "\\_SB" and is used to start the path when building a fully qualified name. This requires SOC support to provide handlers for returning the ACPI name for devices that it owns, and those names must match the objects declared in the DSDT. The handler can be done either in each device driver or with a global handler for the entire SOC. Simplified example of how this can be used for an i2c device declared in devicetree.cb with: chip soc/intel/skylake # "\_SB" (from root device) device domain 0 on # "PCI0" device pci 19.2 on # "I2C4" chip drivers/i2c/test0 device i2c 1a.0 on end # "TST0" end end end end And basic SSDT generating code in the device driver: acpigen_write_scope(acpi_device_scope(dev)); acpigen_write_device(acpi_device_name(dev)); acpigen_write_string("_HID", "TEST0000"); acpigen_write_byte("_UID", 0); acpigen_pop_len(); /* device */ acpigen_pop_len(); /* scope */ Will produce this ACPI code: Scope (\_SB.PCI0.I2C4) { Device (TST0) { Name (_HID, "TEST0000") Name (_UID, 0) } } Change-Id: Ie149595aeab96266fa5f006e7934339f0119ac54 Signed-off-by: Duncan Laurie <dlaurie@chromium.org> Reviewed-on: https://review.coreboot.org/14840 Tested-by: build bot (Jenkins) Reviewed-by: Aaron Durbin <adurbin@chromium.org>
2016-05-09 03:15:25 +02:00
/* Fill in the path from the root device */
next += snprintf(buf + next, buf_len - next, "%s%s",
arch/x86/acpi: Allow "transparent" ACPI device names Certain devices, such as the northbridge on AMD Opteron systems, do not require a node in the ACPI device path. Allow such devices to be passed over by the ACPI path generator if the device-specific ACPI name function returns a zero-length (non-NULL) string. Change-Id: Iffffc9a30b395b0bd6d60e411439a437e89f554e Signed-off-by: Timothy Pearson <tpearson@raptorengineering.com> Reviewed-on: https://review.coreboot.org/19281 Tested-by: build bot (Jenkins) Tested-by: Raptor Engineering Automated Test Stand <noreply@raptorengineeringinc.com> Reviewed-by: Duncan Laurie <dlaurie@chromium.org> Reviewed-by: Aaron Durbin <adurbin@chromium.org>
2017-04-14 00:06:16 +02:00
(dev->path.type == DEVICE_PATH_ROOT
|| (strlen(name) == 0)) ?
"" : ".", name);
device: Add an ACPI device name and path concept to devices Add a function to "struct device_operations" to return the ACPI name for the device, and helper functions to find this name (either from the device or its parent) and to build a fully qualified ACPI path from the root device. This addition will allow device drivers to generate their ACPI AML in the SSDT at boot, with customization supplied by devicetree.cb, instead of needing custom DSDT ASL for every mainboard. The root device acpi_name is defined as "\\_SB" and is used to start the path when building a fully qualified name. This requires SOC support to provide handlers for returning the ACPI name for devices that it owns, and those names must match the objects declared in the DSDT. The handler can be done either in each device driver or with a global handler for the entire SOC. Simplified example of how this can be used for an i2c device declared in devicetree.cb with: chip soc/intel/skylake # "\_SB" (from root device) device domain 0 on # "PCI0" device pci 19.2 on # "I2C4" chip drivers/i2c/test0 device i2c 1a.0 on end # "TST0" end end end end And basic SSDT generating code in the device driver: acpigen_write_scope(acpi_device_scope(dev)); acpigen_write_device(acpi_device_name(dev)); acpigen_write_string("_HID", "TEST0000"); acpigen_write_byte("_UID", 0); acpigen_pop_len(); /* device */ acpigen_pop_len(); /* scope */ Will produce this ACPI code: Scope (\_SB.PCI0.I2C4) { Device (TST0) { Name (_HID, "TEST0000") Name (_UID, 0) } } Change-Id: Ie149595aeab96266fa5f006e7934339f0119ac54 Signed-off-by: Duncan Laurie <dlaurie@chromium.org> Reviewed-on: https://review.coreboot.org/14840 Tested-by: build bot (Jenkins) Reviewed-by: Aaron Durbin <adurbin@chromium.org>
2016-05-09 03:15:25 +02:00
return next;
}
/*
* Warning: just as with dev_path() this uses a static buffer
* so should not be called mulitple times in one statement
*/
src/arch/x86/acpi: Constify struct device instances Constify the struct device arguments in below APIs: > acpi_device_name > acpi_device_path_fill > acpi_device_path > acpi_write_dbg2_pci_uart > acpi_device_scope > acpi_device_path_join The APIs do not seem to modify the argument and are using device argument as reference to device only. Change-Id: Ic2ce045f17efa288eb41503795723d0ad5ec78bd Signed-off-by: Aamir Bohra <aamir.bohra@intel.com> Reviewed-on: https://review.coreboot.org/c/coreboot/+/34625 Reviewed-by: Subrata Banik <subrata.banik@intel.com> Reviewed-by: Patrick Rudolph <patrick.rudolph@9elements.com> Reviewed-by: Nico Huber <nico.h@gmx.de> Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
2019-07-30 06:41:41 +02:00
const char *acpi_device_path(const struct device *dev)
device: Add an ACPI device name and path concept to devices Add a function to "struct device_operations" to return the ACPI name for the device, and helper functions to find this name (either from the device or its parent) and to build a fully qualified ACPI path from the root device. This addition will allow device drivers to generate their ACPI AML in the SSDT at boot, with customization supplied by devicetree.cb, instead of needing custom DSDT ASL for every mainboard. The root device acpi_name is defined as "\\_SB" and is used to start the path when building a fully qualified name. This requires SOC support to provide handlers for returning the ACPI name for devices that it owns, and those names must match the objects declared in the DSDT. The handler can be done either in each device driver or with a global handler for the entire SOC. Simplified example of how this can be used for an i2c device declared in devicetree.cb with: chip soc/intel/skylake # "\_SB" (from root device) device domain 0 on # "PCI0" device pci 19.2 on # "I2C4" chip drivers/i2c/test0 device i2c 1a.0 on end # "TST0" end end end end And basic SSDT generating code in the device driver: acpigen_write_scope(acpi_device_scope(dev)); acpigen_write_device(acpi_device_name(dev)); acpigen_write_string("_HID", "TEST0000"); acpigen_write_byte("_UID", 0); acpigen_pop_len(); /* device */ acpigen_pop_len(); /* scope */ Will produce this ACPI code: Scope (\_SB.PCI0.I2C4) { Device (TST0) { Name (_HID, "TEST0000") Name (_UID, 0) } } Change-Id: Ie149595aeab96266fa5f006e7934339f0119ac54 Signed-off-by: Duncan Laurie <dlaurie@chromium.org> Reviewed-on: https://review.coreboot.org/14840 Tested-by: build bot (Jenkins) Reviewed-by: Aaron Durbin <adurbin@chromium.org>
2016-05-09 03:15:25 +02:00
{
static char buf[DEVICE_PATH_MAX] = {};
if (!dev)
return NULL;
if (acpi_device_path_fill(dev, buf, sizeof(buf), 0) <= 0)
return NULL;
return buf;
}
/* Return the path of the parent device as the ACPI Scope for this device */
src/arch/x86/acpi: Constify struct device instances Constify the struct device arguments in below APIs: > acpi_device_name > acpi_device_path_fill > acpi_device_path > acpi_write_dbg2_pci_uart > acpi_device_scope > acpi_device_path_join The APIs do not seem to modify the argument and are using device argument as reference to device only. Change-Id: Ic2ce045f17efa288eb41503795723d0ad5ec78bd Signed-off-by: Aamir Bohra <aamir.bohra@intel.com> Reviewed-on: https://review.coreboot.org/c/coreboot/+/34625 Reviewed-by: Subrata Banik <subrata.banik@intel.com> Reviewed-by: Patrick Rudolph <patrick.rudolph@9elements.com> Reviewed-by: Nico Huber <nico.h@gmx.de> Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
2019-07-30 06:41:41 +02:00
const char *acpi_device_scope(const struct device *dev)
device: Add an ACPI device name and path concept to devices Add a function to "struct device_operations" to return the ACPI name for the device, and helper functions to find this name (either from the device or its parent) and to build a fully qualified ACPI path from the root device. This addition will allow device drivers to generate their ACPI AML in the SSDT at boot, with customization supplied by devicetree.cb, instead of needing custom DSDT ASL for every mainboard. The root device acpi_name is defined as "\\_SB" and is used to start the path when building a fully qualified name. This requires SOC support to provide handlers for returning the ACPI name for devices that it owns, and those names must match the objects declared in the DSDT. The handler can be done either in each device driver or with a global handler for the entire SOC. Simplified example of how this can be used for an i2c device declared in devicetree.cb with: chip soc/intel/skylake # "\_SB" (from root device) device domain 0 on # "PCI0" device pci 19.2 on # "I2C4" chip drivers/i2c/test0 device i2c 1a.0 on end # "TST0" end end end end And basic SSDT generating code in the device driver: acpigen_write_scope(acpi_device_scope(dev)); acpigen_write_device(acpi_device_name(dev)); acpigen_write_string("_HID", "TEST0000"); acpigen_write_byte("_UID", 0); acpigen_pop_len(); /* device */ acpigen_pop_len(); /* scope */ Will produce this ACPI code: Scope (\_SB.PCI0.I2C4) { Device (TST0) { Name (_HID, "TEST0000") Name (_UID, 0) } } Change-Id: Ie149595aeab96266fa5f006e7934339f0119ac54 Signed-off-by: Duncan Laurie <dlaurie@chromium.org> Reviewed-on: https://review.coreboot.org/14840 Tested-by: build bot (Jenkins) Reviewed-by: Aaron Durbin <adurbin@chromium.org>
2016-05-09 03:15:25 +02:00
{
acpi_device: Have acpi_device_scope() use a separate buffer Have the different acpi_device_ path functions use a different static buffer so they can be called interchangeably. Change-Id: I270a80f66880861d5847bd586a16a73f8f1e2511 Signed-off-by: Duncan Laurie <dlaurie@chromium.org> Reviewed-on: https://review.coreboot.org/15521 Tested-by: build bot (Jenkins) Reviewed-by: Aaron Durbin <adurbin@chromium.org> Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
2016-06-30 23:37:37 +02:00
static char buf[DEVICE_PATH_MAX] = {};
device: Add an ACPI device name and path concept to devices Add a function to "struct device_operations" to return the ACPI name for the device, and helper functions to find this name (either from the device or its parent) and to build a fully qualified ACPI path from the root device. This addition will allow device drivers to generate their ACPI AML in the SSDT at boot, with customization supplied by devicetree.cb, instead of needing custom DSDT ASL for every mainboard. The root device acpi_name is defined as "\\_SB" and is used to start the path when building a fully qualified name. This requires SOC support to provide handlers for returning the ACPI name for devices that it owns, and those names must match the objects declared in the DSDT. The handler can be done either in each device driver or with a global handler for the entire SOC. Simplified example of how this can be used for an i2c device declared in devicetree.cb with: chip soc/intel/skylake # "\_SB" (from root device) device domain 0 on # "PCI0" device pci 19.2 on # "I2C4" chip drivers/i2c/test0 device i2c 1a.0 on end # "TST0" end end end end And basic SSDT generating code in the device driver: acpigen_write_scope(acpi_device_scope(dev)); acpigen_write_device(acpi_device_name(dev)); acpigen_write_string("_HID", "TEST0000"); acpigen_write_byte("_UID", 0); acpigen_pop_len(); /* device */ acpigen_pop_len(); /* scope */ Will produce this ACPI code: Scope (\_SB.PCI0.I2C4) { Device (TST0) { Name (_HID, "TEST0000") Name (_UID, 0) } } Change-Id: Ie149595aeab96266fa5f006e7934339f0119ac54 Signed-off-by: Duncan Laurie <dlaurie@chromium.org> Reviewed-on: https://review.coreboot.org/14840 Tested-by: build bot (Jenkins) Reviewed-by: Aaron Durbin <adurbin@chromium.org>
2016-05-09 03:15:25 +02:00
if (!dev || !dev->bus || !dev->bus->dev)
return NULL;
acpi_device: Have acpi_device_scope() use a separate buffer Have the different acpi_device_ path functions use a different static buffer so they can be called interchangeably. Change-Id: I270a80f66880861d5847bd586a16a73f8f1e2511 Signed-off-by: Duncan Laurie <dlaurie@chromium.org> Reviewed-on: https://review.coreboot.org/15521 Tested-by: build bot (Jenkins) Reviewed-by: Aaron Durbin <adurbin@chromium.org> Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
2016-06-30 23:37:37 +02:00
if (acpi_device_path_fill(dev->bus->dev, buf, sizeof(buf), 0) <= 0)
return NULL;
return buf;
device: Add an ACPI device name and path concept to devices Add a function to "struct device_operations" to return the ACPI name for the device, and helper functions to find this name (either from the device or its parent) and to build a fully qualified ACPI path from the root device. This addition will allow device drivers to generate their ACPI AML in the SSDT at boot, with customization supplied by devicetree.cb, instead of needing custom DSDT ASL for every mainboard. The root device acpi_name is defined as "\\_SB" and is used to start the path when building a fully qualified name. This requires SOC support to provide handlers for returning the ACPI name for devices that it owns, and those names must match the objects declared in the DSDT. The handler can be done either in each device driver or with a global handler for the entire SOC. Simplified example of how this can be used for an i2c device declared in devicetree.cb with: chip soc/intel/skylake # "\_SB" (from root device) device domain 0 on # "PCI0" device pci 19.2 on # "I2C4" chip drivers/i2c/test0 device i2c 1a.0 on end # "TST0" end end end end And basic SSDT generating code in the device driver: acpigen_write_scope(acpi_device_scope(dev)); acpigen_write_device(acpi_device_name(dev)); acpigen_write_string("_HID", "TEST0000"); acpigen_write_byte("_UID", 0); acpigen_pop_len(); /* device */ acpigen_pop_len(); /* scope */ Will produce this ACPI code: Scope (\_SB.PCI0.I2C4) { Device (TST0) { Name (_HID, "TEST0000") Name (_UID, 0) } } Change-Id: Ie149595aeab96266fa5f006e7934339f0119ac54 Signed-off-by: Duncan Laurie <dlaurie@chromium.org> Reviewed-on: https://review.coreboot.org/14840 Tested-by: build bot (Jenkins) Reviewed-by: Aaron Durbin <adurbin@chromium.org>
2016-05-09 03:15:25 +02:00
}
arch/x86: Fix typos Change-Id: I944b9bf8f518eff9b539769825174bf1544e6b34 Signed-off-by: Elyes HAOUAS <ehaouas@noos.fr> Reviewed-on: https://review.coreboot.org/c/coreboot/+/38310 Tested-by: build bot (Jenkins) <no-reply@coreboot.org> Reviewed-by: Angel Pons <th3fanbus@gmail.com>
2020-01-09 21:29:25 +01:00
/* Concatenate the device path and provided name suffix */
src/arch/x86/acpi: Constify struct device instances Constify the struct device arguments in below APIs: > acpi_device_name > acpi_device_path_fill > acpi_device_path > acpi_write_dbg2_pci_uart > acpi_device_scope > acpi_device_path_join The APIs do not seem to modify the argument and are using device argument as reference to device only. Change-Id: Ic2ce045f17efa288eb41503795723d0ad5ec78bd Signed-off-by: Aamir Bohra <aamir.bohra@intel.com> Reviewed-on: https://review.coreboot.org/c/coreboot/+/34625 Reviewed-by: Subrata Banik <subrata.banik@intel.com> Reviewed-by: Patrick Rudolph <patrick.rudolph@9elements.com> Reviewed-by: Nico Huber <nico.h@gmx.de> Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
2019-07-30 06:41:41 +02:00
const char *acpi_device_path_join(const struct device *dev, const char *name)
device: Add an ACPI device name and path concept to devices Add a function to "struct device_operations" to return the ACPI name for the device, and helper functions to find this name (either from the device or its parent) and to build a fully qualified ACPI path from the root device. This addition will allow device drivers to generate their ACPI AML in the SSDT at boot, with customization supplied by devicetree.cb, instead of needing custom DSDT ASL for every mainboard. The root device acpi_name is defined as "\\_SB" and is used to start the path when building a fully qualified name. This requires SOC support to provide handlers for returning the ACPI name for devices that it owns, and those names must match the objects declared in the DSDT. The handler can be done either in each device driver or with a global handler for the entire SOC. Simplified example of how this can be used for an i2c device declared in devicetree.cb with: chip soc/intel/skylake # "\_SB" (from root device) device domain 0 on # "PCI0" device pci 19.2 on # "I2C4" chip drivers/i2c/test0 device i2c 1a.0 on end # "TST0" end end end end And basic SSDT generating code in the device driver: acpigen_write_scope(acpi_device_scope(dev)); acpigen_write_device(acpi_device_name(dev)); acpigen_write_string("_HID", "TEST0000"); acpigen_write_byte("_UID", 0); acpigen_pop_len(); /* device */ acpigen_pop_len(); /* scope */ Will produce this ACPI code: Scope (\_SB.PCI0.I2C4) { Device (TST0) { Name (_HID, "TEST0000") Name (_UID, 0) } } Change-Id: Ie149595aeab96266fa5f006e7934339f0119ac54 Signed-off-by: Duncan Laurie <dlaurie@chromium.org> Reviewed-on: https://review.coreboot.org/14840 Tested-by: build bot (Jenkins) Reviewed-by: Aaron Durbin <adurbin@chromium.org>
2016-05-09 03:15:25 +02:00
{
static char buf[DEVICE_PATH_MAX] = {};
arch/x86: Use ssize_t to store length size_t is the wrong type to store the return value of acpi_device_path_fill(), since any negative error values will be converted to a very large unsigned integer and potentially cause buffer overflow. Change-Id: Ia8ed62ecfac8eaa18a61545bd203b3c7a7cd9ca5 Signed-off-by: Jacob Garber <jgarber1@ualberta.ca> Found-by: Coverity CID 1402095 Reviewed-on: https://review.coreboot.org/c/coreboot/+/33962 Tested-by: build bot (Jenkins) <no-reply@coreboot.org> Reviewed-by: HAOUAS Elyes <ehaouas@noos.fr> Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net>
2019-07-02 18:50:10 +02:00
ssize_t len;
device: Add an ACPI device name and path concept to devices Add a function to "struct device_operations" to return the ACPI name for the device, and helper functions to find this name (either from the device or its parent) and to build a fully qualified ACPI path from the root device. This addition will allow device drivers to generate their ACPI AML in the SSDT at boot, with customization supplied by devicetree.cb, instead of needing custom DSDT ASL for every mainboard. The root device acpi_name is defined as "\\_SB" and is used to start the path when building a fully qualified name. This requires SOC support to provide handlers for returning the ACPI name for devices that it owns, and those names must match the objects declared in the DSDT. The handler can be done either in each device driver or with a global handler for the entire SOC. Simplified example of how this can be used for an i2c device declared in devicetree.cb with: chip soc/intel/skylake # "\_SB" (from root device) device domain 0 on # "PCI0" device pci 19.2 on # "I2C4" chip drivers/i2c/test0 device i2c 1a.0 on end # "TST0" end end end end And basic SSDT generating code in the device driver: acpigen_write_scope(acpi_device_scope(dev)); acpigen_write_device(acpi_device_name(dev)); acpigen_write_string("_HID", "TEST0000"); acpigen_write_byte("_UID", 0); acpigen_pop_len(); /* device */ acpigen_pop_len(); /* scope */ Will produce this ACPI code: Scope (\_SB.PCI0.I2C4) { Device (TST0) { Name (_HID, "TEST0000") Name (_UID, 0) } } Change-Id: Ie149595aeab96266fa5f006e7934339f0119ac54 Signed-off-by: Duncan Laurie <dlaurie@chromium.org> Reviewed-on: https://review.coreboot.org/14840 Tested-by: build bot (Jenkins) Reviewed-by: Aaron Durbin <adurbin@chromium.org>
2016-05-09 03:15:25 +02:00
if (!dev)
return NULL;
/* Build the path of this device */
len = acpi_device_path_fill(dev, buf, sizeof(buf), 0);
if (len <= 0)
return NULL;
/* Ensure there is room for the added name, separator, and NUL */
if ((len + strlen(name) + 2) > sizeof(buf))
return NULL;
snprintf(buf + len, sizeof(buf) - len, ".%s", name);
return buf;
}
acpi_device: Add support for writing ACPI Interrupt descriptors Add definitions for ACPI device extended interrupts and a method to write an Interrupt() descriptor to the SSDT output stream. Interrupts are often tied together with other resources and some configuration items are shared (though not always compatibly) with other constructs like GPIOs and GPEs. These will get used by device drivers to write _CRS sections for devices into the SSDT. One usage is to include a "struct acpi_irq" inside a config struct for a device so it can be initialized based on settings in devicetree. Example usage: chip.h: struct drivers_i2c_generic_config { struct acpi_irq irq; }; generic.c: void acpi_fill_ssdt_generator(struct device *dev) { struct drivers_i2c_generic_config *config = dev->chip_info; ... acpi_device_write_interrupt(&config->irq); ... } devicetree.cb: device pci 15.0 on chip drivers/i2c/generic register "irq" = "IRQ_EDGE_LOW(GPP_E7_IRQ)" device i2c 10 on end end end SSDT.dsl: Interrupt (ResourceConsumer, Edge, ActiveLow, Exclusive,,,) { 31 } Change-Id: I3b64170cc2ebac178e7a17df479eda7670a42703 Signed-off-by: Duncan Laurie <dlaurie@chromium.org> Reviewed-on: https://review.coreboot.org/14933 Tested-by: build bot (Jenkins) Reviewed-by: Aaron Durbin <adurbin@chromium.org>
2016-05-10 00:38:36 +02:00
acpi: Call acpi_gen_writeSTA by status from device tree The device tree now supports 'hidden' and the status can be found in `struct device.hidden`. A new acpi_device_status() will return the expected setting of STA from a `struct device`. BUG=b:72200466 BRANCH=eve TEST=Builds and boots properly on device eve Change-Id: I6dc62aff63cc3cb950739398a4dcac21836c9766 Signed-off-by: Hung-Te Lin <hungte@chromium.org> Reviewed-on: https://review.coreboot.org/28567 Reviewed-by: Duncan Laurie <dlaurie@chromium.org> Reviewed-by: Furquan Shaikh <furquan@google.com> Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
2018-09-10 04:55:49 +02:00
int acpi_device_status(const struct device *dev)
{
if (!dev->enabled)
return ACPI_STATUS_DEVICE_ALL_OFF;
if (dev->hidden)
return ACPI_STATUS_DEVICE_HIDDEN_ON;
return ACPI_STATUS_DEVICE_ALL_ON;
}
acpi: Be more ACPI compliant when generating _UID * Add function to generate unique _UID using CRC32 * Add function to write the _UID based on a device's ACPI path ACPI devices that have the same _HID must use different _UID. Linux doesn't care about _UID if it's not used. Windows 10 verifies the ACPI code on boot and BSODs if two devices with the same _HID share the same _UID. Fixes BSOD seen on Windows 10. Change-Id: I47cd5396060d325f9ce338afced6af021e7ff2b4 Signed-off-by: Patrick Rudolph <patrick.rudolph@9elements.com> Reviewed-on: https://review.coreboot.org/c/coreboot/+/37695 Tested-by: build bot (Jenkins) <no-reply@coreboot.org> Reviewed-by: Philipp Deppenwiese <zaolin.daisuki@gmail.com>
2019-12-13 12:16:06 +01:00
/* Write the unique _UID based on ACPI device path. */
arch/x86/acpi_device: Constify struct device * parameter to UID functions acpi_device_uid() and acpi_device_write_uid() do not need to make changes to the device structure. Thus, this change marks struct device * parameter to these functions as const. Change-Id: I3755223766c78f93c57ac80caf392985cfd5c5e5 Signed-off-by: Furquan Shaikh <furquan@google.com> Reviewed-on: https://review.coreboot.org/c/coreboot/+/40702 Reviewed-by: HAOUAS Elyes <ehaouas@noos.fr> Reviewed-by: Aaron Durbin <adurbin@chromium.org> Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
2020-04-25 06:27:29 +02:00
void acpi_device_write_uid(const struct device *dev)
acpi: Be more ACPI compliant when generating _UID * Add function to generate unique _UID using CRC32 * Add function to write the _UID based on a device's ACPI path ACPI devices that have the same _HID must use different _UID. Linux doesn't care about _UID if it's not used. Windows 10 verifies the ACPI code on boot and BSODs if two devices with the same _HID share the same _UID. Fixes BSOD seen on Windows 10. Change-Id: I47cd5396060d325f9ce338afced6af021e7ff2b4 Signed-off-by: Patrick Rudolph <patrick.rudolph@9elements.com> Reviewed-on: https://review.coreboot.org/c/coreboot/+/37695 Tested-by: build bot (Jenkins) <no-reply@coreboot.org> Reviewed-by: Philipp Deppenwiese <zaolin.daisuki@gmail.com>
2019-12-13 12:16:06 +01:00
{
acpigen_write_name_integer("_UID", acpi_device_uid(dev));
}
acpi_device: Add support for writing ACPI Interrupt descriptors Add definitions for ACPI device extended interrupts and a method to write an Interrupt() descriptor to the SSDT output stream. Interrupts are often tied together with other resources and some configuration items are shared (though not always compatibly) with other constructs like GPIOs and GPEs. These will get used by device drivers to write _CRS sections for devices into the SSDT. One usage is to include a "struct acpi_irq" inside a config struct for a device so it can be initialized based on settings in devicetree. Example usage: chip.h: struct drivers_i2c_generic_config { struct acpi_irq irq; }; generic.c: void acpi_fill_ssdt_generator(struct device *dev) { struct drivers_i2c_generic_config *config = dev->chip_info; ... acpi_device_write_interrupt(&config->irq); ... } devicetree.cb: device pci 15.0 on chip drivers/i2c/generic register "irq" = "IRQ_EDGE_LOW(GPP_E7_IRQ)" device i2c 10 on end end end SSDT.dsl: Interrupt (ResourceConsumer, Edge, ActiveLow, Exclusive,,,) { 31 } Change-Id: I3b64170cc2ebac178e7a17df479eda7670a42703 Signed-off-by: Duncan Laurie <dlaurie@chromium.org> Reviewed-on: https://review.coreboot.org/14933 Tested-by: build bot (Jenkins) Reviewed-by: Aaron Durbin <adurbin@chromium.org>
2016-05-10 00:38:36 +02:00
/* ACPI 6.1 section 6.4.3.6: Extended Interrupt Descriptor */
void acpi_device_write_interrupt(const struct acpi_irq *irq)
{
void *desc_length;
uint8_t flags;
if (!irq || !irq->pin)
return;
/* This is supported by GpioInt() but not Interrupt() */
acpi: Add ACPI_ prefix to IRQ enum and struct names This is done to avoid any conflicts with same IRQ enums defined by other drivers. BUG=None BRANCH=None TEST=Compiles successfully Change-Id: I539831d853286ca45f6c36c3812a6fa9602df24c Signed-off-by: Furquan Shaikh <furquan@chromium.org> Reviewed-on: https://review.coreboot.org/18444 Tested-by: build bot (Jenkins) Reviewed-by: Aaron Durbin <adurbin@chromium.org>
2017-02-21 22:16:30 +01:00
if (irq->polarity == ACPI_IRQ_ACTIVE_BOTH)
acpi_device: Add support for writing ACPI Interrupt descriptors Add definitions for ACPI device extended interrupts and a method to write an Interrupt() descriptor to the SSDT output stream. Interrupts are often tied together with other resources and some configuration items are shared (though not always compatibly) with other constructs like GPIOs and GPEs. These will get used by device drivers to write _CRS sections for devices into the SSDT. One usage is to include a "struct acpi_irq" inside a config struct for a device so it can be initialized based on settings in devicetree. Example usage: chip.h: struct drivers_i2c_generic_config { struct acpi_irq irq; }; generic.c: void acpi_fill_ssdt_generator(struct device *dev) { struct drivers_i2c_generic_config *config = dev->chip_info; ... acpi_device_write_interrupt(&config->irq); ... } devicetree.cb: device pci 15.0 on chip drivers/i2c/generic register "irq" = "IRQ_EDGE_LOW(GPP_E7_IRQ)" device i2c 10 on end end end SSDT.dsl: Interrupt (ResourceConsumer, Edge, ActiveLow, Exclusive,,,) { 31 } Change-Id: I3b64170cc2ebac178e7a17df479eda7670a42703 Signed-off-by: Duncan Laurie <dlaurie@chromium.org> Reviewed-on: https://review.coreboot.org/14933 Tested-by: build bot (Jenkins) Reviewed-by: Aaron Durbin <adurbin@chromium.org>
2016-05-10 00:38:36 +02:00
return;
/* Byte 0: Descriptor Type */
acpigen_emit_byte(ACPI_DESCRIPTOR_INTERRUPT);
/* Byte 1-2: Length (filled in later) */
desc_length = acpi_device_write_zero_len();
/*
* Byte 3: Flags
* [7:5]: Reserved
* [4]: Wake (0=NO_WAKE 1=WAKE)
* [3]: Sharing (0=EXCLUSIVE 1=SHARED)
* [2]: Polarity (0=HIGH 1=LOW)
* [1]: Mode (0=LEVEL 1=EDGE)
* [0]: Resource (0=PRODUCER 1=CONSUMER)
*/
flags = 1 << 0; /* ResourceConsumer */
acpi: Add ACPI_ prefix to IRQ enum and struct names This is done to avoid any conflicts with same IRQ enums defined by other drivers. BUG=None BRANCH=None TEST=Compiles successfully Change-Id: I539831d853286ca45f6c36c3812a6fa9602df24c Signed-off-by: Furquan Shaikh <furquan@chromium.org> Reviewed-on: https://review.coreboot.org/18444 Tested-by: build bot (Jenkins) Reviewed-by: Aaron Durbin <adurbin@chromium.org>
2017-02-21 22:16:30 +01:00
if (irq->mode == ACPI_IRQ_EDGE_TRIGGERED)
acpi_device: Add support for writing ACPI Interrupt descriptors Add definitions for ACPI device extended interrupts and a method to write an Interrupt() descriptor to the SSDT output stream. Interrupts are often tied together with other resources and some configuration items are shared (though not always compatibly) with other constructs like GPIOs and GPEs. These will get used by device drivers to write _CRS sections for devices into the SSDT. One usage is to include a "struct acpi_irq" inside a config struct for a device so it can be initialized based on settings in devicetree. Example usage: chip.h: struct drivers_i2c_generic_config { struct acpi_irq irq; }; generic.c: void acpi_fill_ssdt_generator(struct device *dev) { struct drivers_i2c_generic_config *config = dev->chip_info; ... acpi_device_write_interrupt(&config->irq); ... } devicetree.cb: device pci 15.0 on chip drivers/i2c/generic register "irq" = "IRQ_EDGE_LOW(GPP_E7_IRQ)" device i2c 10 on end end end SSDT.dsl: Interrupt (ResourceConsumer, Edge, ActiveLow, Exclusive,,,) { 31 } Change-Id: I3b64170cc2ebac178e7a17df479eda7670a42703 Signed-off-by: Duncan Laurie <dlaurie@chromium.org> Reviewed-on: https://review.coreboot.org/14933 Tested-by: build bot (Jenkins) Reviewed-by: Aaron Durbin <adurbin@chromium.org>
2016-05-10 00:38:36 +02:00
flags |= 1 << 1;
acpi: Add ACPI_ prefix to IRQ enum and struct names This is done to avoid any conflicts with same IRQ enums defined by other drivers. BUG=None BRANCH=None TEST=Compiles successfully Change-Id: I539831d853286ca45f6c36c3812a6fa9602df24c Signed-off-by: Furquan Shaikh <furquan@chromium.org> Reviewed-on: https://review.coreboot.org/18444 Tested-by: build bot (Jenkins) Reviewed-by: Aaron Durbin <adurbin@chromium.org>
2017-02-21 22:16:30 +01:00
if (irq->polarity == ACPI_IRQ_ACTIVE_LOW)
acpi_device: Add support for writing ACPI Interrupt descriptors Add definitions for ACPI device extended interrupts and a method to write an Interrupt() descriptor to the SSDT output stream. Interrupts are often tied together with other resources and some configuration items are shared (though not always compatibly) with other constructs like GPIOs and GPEs. These will get used by device drivers to write _CRS sections for devices into the SSDT. One usage is to include a "struct acpi_irq" inside a config struct for a device so it can be initialized based on settings in devicetree. Example usage: chip.h: struct drivers_i2c_generic_config { struct acpi_irq irq; }; generic.c: void acpi_fill_ssdt_generator(struct device *dev) { struct drivers_i2c_generic_config *config = dev->chip_info; ... acpi_device_write_interrupt(&config->irq); ... } devicetree.cb: device pci 15.0 on chip drivers/i2c/generic register "irq" = "IRQ_EDGE_LOW(GPP_E7_IRQ)" device i2c 10 on end end end SSDT.dsl: Interrupt (ResourceConsumer, Edge, ActiveLow, Exclusive,,,) { 31 } Change-Id: I3b64170cc2ebac178e7a17df479eda7670a42703 Signed-off-by: Duncan Laurie <dlaurie@chromium.org> Reviewed-on: https://review.coreboot.org/14933 Tested-by: build bot (Jenkins) Reviewed-by: Aaron Durbin <adurbin@chromium.org>
2016-05-10 00:38:36 +02:00
flags |= 1 << 2;
acpi: Add ACPI_ prefix to IRQ enum and struct names This is done to avoid any conflicts with same IRQ enums defined by other drivers. BUG=None BRANCH=None TEST=Compiles successfully Change-Id: I539831d853286ca45f6c36c3812a6fa9602df24c Signed-off-by: Furquan Shaikh <furquan@chromium.org> Reviewed-on: https://review.coreboot.org/18444 Tested-by: build bot (Jenkins) Reviewed-by: Aaron Durbin <adurbin@chromium.org>
2017-02-21 22:16:30 +01:00
if (irq->shared == ACPI_IRQ_SHARED)
acpi_device: Add support for writing ACPI Interrupt descriptors Add definitions for ACPI device extended interrupts and a method to write an Interrupt() descriptor to the SSDT output stream. Interrupts are often tied together with other resources and some configuration items are shared (though not always compatibly) with other constructs like GPIOs and GPEs. These will get used by device drivers to write _CRS sections for devices into the SSDT. One usage is to include a "struct acpi_irq" inside a config struct for a device so it can be initialized based on settings in devicetree. Example usage: chip.h: struct drivers_i2c_generic_config { struct acpi_irq irq; }; generic.c: void acpi_fill_ssdt_generator(struct device *dev) { struct drivers_i2c_generic_config *config = dev->chip_info; ... acpi_device_write_interrupt(&config->irq); ... } devicetree.cb: device pci 15.0 on chip drivers/i2c/generic register "irq" = "IRQ_EDGE_LOW(GPP_E7_IRQ)" device i2c 10 on end end end SSDT.dsl: Interrupt (ResourceConsumer, Edge, ActiveLow, Exclusive,,,) { 31 } Change-Id: I3b64170cc2ebac178e7a17df479eda7670a42703 Signed-off-by: Duncan Laurie <dlaurie@chromium.org> Reviewed-on: https://review.coreboot.org/14933 Tested-by: build bot (Jenkins) Reviewed-by: Aaron Durbin <adurbin@chromium.org>
2016-05-10 00:38:36 +02:00
flags |= 1 << 3;
acpi: Add ACPI_ prefix to IRQ enum and struct names This is done to avoid any conflicts with same IRQ enums defined by other drivers. BUG=None BRANCH=None TEST=Compiles successfully Change-Id: I539831d853286ca45f6c36c3812a6fa9602df24c Signed-off-by: Furquan Shaikh <furquan@chromium.org> Reviewed-on: https://review.coreboot.org/18444 Tested-by: build bot (Jenkins) Reviewed-by: Aaron Durbin <adurbin@chromium.org>
2017-02-21 22:16:30 +01:00
if (irq->wake == ACPI_IRQ_WAKE)
acpi_device: Add support for writing ACPI Interrupt descriptors Add definitions for ACPI device extended interrupts and a method to write an Interrupt() descriptor to the SSDT output stream. Interrupts are often tied together with other resources and some configuration items are shared (though not always compatibly) with other constructs like GPIOs and GPEs. These will get used by device drivers to write _CRS sections for devices into the SSDT. One usage is to include a "struct acpi_irq" inside a config struct for a device so it can be initialized based on settings in devicetree. Example usage: chip.h: struct drivers_i2c_generic_config { struct acpi_irq irq; }; generic.c: void acpi_fill_ssdt_generator(struct device *dev) { struct drivers_i2c_generic_config *config = dev->chip_info; ... acpi_device_write_interrupt(&config->irq); ... } devicetree.cb: device pci 15.0 on chip drivers/i2c/generic register "irq" = "IRQ_EDGE_LOW(GPP_E7_IRQ)" device i2c 10 on end end end SSDT.dsl: Interrupt (ResourceConsumer, Edge, ActiveLow, Exclusive,,,) { 31 } Change-Id: I3b64170cc2ebac178e7a17df479eda7670a42703 Signed-off-by: Duncan Laurie <dlaurie@chromium.org> Reviewed-on: https://review.coreboot.org/14933 Tested-by: build bot (Jenkins) Reviewed-by: Aaron Durbin <adurbin@chromium.org>
2016-05-10 00:38:36 +02:00
flags |= 1 << 4;
acpigen_emit_byte(flags);
/* Byte 4: Interrupt Table Entry Count */
acpigen_emit_byte(1);
/* Byte 5-8: Interrupt Number */
acpigen_emit_dword(irq->pin);
/* Fill in Descriptor Length (account for len word) */
acpi_device_fill_len(desc_length);
}
acpi_device: Add support for writing ACPI GPIO descriptors Add definitions to describe GPIOs in generated ACPI objects and a method to write a GpioIo() or GpioInt() descriptor to the SSDT. ACPI GPIOs have many possible configuration options and a structure is created to describe it accurately in ACPI terms. There are many shared descriptor fields between GpioIo() and GpioInt() so the same function can write both types. GpioInt shares many properties with ACPI Interrupts and the same types are re-used here where possible. One addition is that GpioInt can be configured to trigger on both low and high edge transitions. One descriptor can describe multiple GPIO pins (limited to 8 in this implementation) that all share configuration and controller and are used by the same device scope. Accurately referring to the GPIO controller that this pin is connected to requires the SoC/board to implement a function handler for acpi_gpio_path(), or for the caller to provide this directly as a string in the acpi_gpio->reference variable. This will get used by device drivers to describe their resources in the SSDT. Here is a sample for a Maxim 98357A I2S codec which has a GPIO for power and channel selection called "sdmode". chip.h: struct drivers_generic_max98357a_config { struct acpi_gpio sdmode_gpio; }; max98357a.c: void acpi_fill_ssdt_generator(struct device *dev) { struct drivers_generic_max98357a_config *config = dev->chip_info; ... acpi_device_write_gpio(&config->sdmode_gpio); ... } devicetree.cb: device pci 1f.3 on chip drivers/generic/max98357a register "sdmode_gpio" = "ACPI_GPIO_OUTPUT(GPP_C5)" device generic 0 on end end end SSDT.dsl: GpioIo (Exclusive, PullDefault, 0, 0, IoRestrictionOutputOnly, "\\_SB.PCI0.GPIO", 0, ResourceConsumer, ,) { 53 } Signed-off-by: Duncan Laurie <dlaurie@chromium.org> Change-Id: Ibf5bab9c4bf6f21252373fb013e78f872550b167 Reviewed-on: https://review.coreboot.org/14934 Tested-by: build bot (Jenkins) Reviewed-by: Aaron Durbin <adurbin@chromium.org>
2016-05-10 02:08:38 +02:00
/* ACPI 6.1 section 6.4.3.8.1 - GPIO Interrupt or I/O */
void acpi_device_write_gpio(const struct acpi_gpio *gpio)
{
void *start, *desc_length;
void *pin_table_offset, *vendor_data_offset, *resource_offset;
uint16_t flags = 0;
int pin;
if (!gpio || gpio->type > ACPI_GPIO_TYPE_IO)
return;
start = acpigen_get_current();
/* Byte 0: Descriptor Type */
acpigen_emit_byte(ACPI_DESCRIPTOR_GPIO);
/* Byte 1-2: Length (fill in later) */
desc_length = acpi_device_write_zero_len();
/* Byte 3: Revision ID */
acpigen_emit_byte(ACPI_GPIO_REVISION_ID);
/* Byte 4: GpioIo or GpioInt */
acpigen_emit_byte(gpio->type);
/*
* Byte 5-6: General Flags
* [15:1]: 0 => Reserved
* [0]: 1 => ResourceConsumer
*/
acpigen_emit_word(1 << 0);
switch (gpio->type) {
case ACPI_GPIO_TYPE_INTERRUPT:
/*
* Byte 7-8: GPIO Interrupt Flags
* [15:5]: 0 => Reserved
* [4]: Wake (0=NO_WAKE 1=WAKE)
* [3]: Sharing (0=EXCLUSIVE 1=SHARED)
* [2:1]: Polarity (0=HIGH 1=LOW 2=BOTH)
* [0]: Mode (0=LEVEL 1=EDGE)
*/
acpi: Add ACPI_ prefix to IRQ enum and struct names This is done to avoid any conflicts with same IRQ enums defined by other drivers. BUG=None BRANCH=None TEST=Compiles successfully Change-Id: I539831d853286ca45f6c36c3812a6fa9602df24c Signed-off-by: Furquan Shaikh <furquan@chromium.org> Reviewed-on: https://review.coreboot.org/18444 Tested-by: build bot (Jenkins) Reviewed-by: Aaron Durbin <adurbin@chromium.org>
2017-02-21 22:16:30 +01:00
if (gpio->irq.mode == ACPI_IRQ_EDGE_TRIGGERED)
acpi_device: Add support for writing ACPI GPIO descriptors Add definitions to describe GPIOs in generated ACPI objects and a method to write a GpioIo() or GpioInt() descriptor to the SSDT. ACPI GPIOs have many possible configuration options and a structure is created to describe it accurately in ACPI terms. There are many shared descriptor fields between GpioIo() and GpioInt() so the same function can write both types. GpioInt shares many properties with ACPI Interrupts and the same types are re-used here where possible. One addition is that GpioInt can be configured to trigger on both low and high edge transitions. One descriptor can describe multiple GPIO pins (limited to 8 in this implementation) that all share configuration and controller and are used by the same device scope. Accurately referring to the GPIO controller that this pin is connected to requires the SoC/board to implement a function handler for acpi_gpio_path(), or for the caller to provide this directly as a string in the acpi_gpio->reference variable. This will get used by device drivers to describe their resources in the SSDT. Here is a sample for a Maxim 98357A I2S codec which has a GPIO for power and channel selection called "sdmode". chip.h: struct drivers_generic_max98357a_config { struct acpi_gpio sdmode_gpio; }; max98357a.c: void acpi_fill_ssdt_generator(struct device *dev) { struct drivers_generic_max98357a_config *config = dev->chip_info; ... acpi_device_write_gpio(&config->sdmode_gpio); ... } devicetree.cb: device pci 1f.3 on chip drivers/generic/max98357a register "sdmode_gpio" = "ACPI_GPIO_OUTPUT(GPP_C5)" device generic 0 on end end end SSDT.dsl: GpioIo (Exclusive, PullDefault, 0, 0, IoRestrictionOutputOnly, "\\_SB.PCI0.GPIO", 0, ResourceConsumer, ,) { 53 } Signed-off-by: Duncan Laurie <dlaurie@chromium.org> Change-Id: Ibf5bab9c4bf6f21252373fb013e78f872550b167 Reviewed-on: https://review.coreboot.org/14934 Tested-by: build bot (Jenkins) Reviewed-by: Aaron Durbin <adurbin@chromium.org>
2016-05-10 02:08:38 +02:00
flags |= 1 << 0;
acpi: Add ACPI_ prefix to IRQ enum and struct names This is done to avoid any conflicts with same IRQ enums defined by other drivers. BUG=None BRANCH=None TEST=Compiles successfully Change-Id: I539831d853286ca45f6c36c3812a6fa9602df24c Signed-off-by: Furquan Shaikh <furquan@chromium.org> Reviewed-on: https://review.coreboot.org/18444 Tested-by: build bot (Jenkins) Reviewed-by: Aaron Durbin <adurbin@chromium.org>
2017-02-21 22:16:30 +01:00
if (gpio->irq.shared == ACPI_IRQ_SHARED)
acpi_device: Add support for writing ACPI GPIO descriptors Add definitions to describe GPIOs in generated ACPI objects and a method to write a GpioIo() or GpioInt() descriptor to the SSDT. ACPI GPIOs have many possible configuration options and a structure is created to describe it accurately in ACPI terms. There are many shared descriptor fields between GpioIo() and GpioInt() so the same function can write both types. GpioInt shares many properties with ACPI Interrupts and the same types are re-used here where possible. One addition is that GpioInt can be configured to trigger on both low and high edge transitions. One descriptor can describe multiple GPIO pins (limited to 8 in this implementation) that all share configuration and controller and are used by the same device scope. Accurately referring to the GPIO controller that this pin is connected to requires the SoC/board to implement a function handler for acpi_gpio_path(), or for the caller to provide this directly as a string in the acpi_gpio->reference variable. This will get used by device drivers to describe their resources in the SSDT. Here is a sample for a Maxim 98357A I2S codec which has a GPIO for power and channel selection called "sdmode". chip.h: struct drivers_generic_max98357a_config { struct acpi_gpio sdmode_gpio; }; max98357a.c: void acpi_fill_ssdt_generator(struct device *dev) { struct drivers_generic_max98357a_config *config = dev->chip_info; ... acpi_device_write_gpio(&config->sdmode_gpio); ... } devicetree.cb: device pci 1f.3 on chip drivers/generic/max98357a register "sdmode_gpio" = "ACPI_GPIO_OUTPUT(GPP_C5)" device generic 0 on end end end SSDT.dsl: GpioIo (Exclusive, PullDefault, 0, 0, IoRestrictionOutputOnly, "\\_SB.PCI0.GPIO", 0, ResourceConsumer, ,) { 53 } Signed-off-by: Duncan Laurie <dlaurie@chromium.org> Change-Id: Ibf5bab9c4bf6f21252373fb013e78f872550b167 Reviewed-on: https://review.coreboot.org/14934 Tested-by: build bot (Jenkins) Reviewed-by: Aaron Durbin <adurbin@chromium.org>
2016-05-10 02:08:38 +02:00
flags |= 1 << 3;
acpi: Add ACPI_ prefix to IRQ enum and struct names This is done to avoid any conflicts with same IRQ enums defined by other drivers. BUG=None BRANCH=None TEST=Compiles successfully Change-Id: I539831d853286ca45f6c36c3812a6fa9602df24c Signed-off-by: Furquan Shaikh <furquan@chromium.org> Reviewed-on: https://review.coreboot.org/18444 Tested-by: build bot (Jenkins) Reviewed-by: Aaron Durbin <adurbin@chromium.org>
2017-02-21 22:16:30 +01:00
if (gpio->irq.wake == ACPI_IRQ_WAKE)
acpi_device: Add support for writing ACPI GPIO descriptors Add definitions to describe GPIOs in generated ACPI objects and a method to write a GpioIo() or GpioInt() descriptor to the SSDT. ACPI GPIOs have many possible configuration options and a structure is created to describe it accurately in ACPI terms. There are many shared descriptor fields between GpioIo() and GpioInt() so the same function can write both types. GpioInt shares many properties with ACPI Interrupts and the same types are re-used here where possible. One addition is that GpioInt can be configured to trigger on both low and high edge transitions. One descriptor can describe multiple GPIO pins (limited to 8 in this implementation) that all share configuration and controller and are used by the same device scope. Accurately referring to the GPIO controller that this pin is connected to requires the SoC/board to implement a function handler for acpi_gpio_path(), or for the caller to provide this directly as a string in the acpi_gpio->reference variable. This will get used by device drivers to describe their resources in the SSDT. Here is a sample for a Maxim 98357A I2S codec which has a GPIO for power and channel selection called "sdmode". chip.h: struct drivers_generic_max98357a_config { struct acpi_gpio sdmode_gpio; }; max98357a.c: void acpi_fill_ssdt_generator(struct device *dev) { struct drivers_generic_max98357a_config *config = dev->chip_info; ... acpi_device_write_gpio(&config->sdmode_gpio); ... } devicetree.cb: device pci 1f.3 on chip drivers/generic/max98357a register "sdmode_gpio" = "ACPI_GPIO_OUTPUT(GPP_C5)" device generic 0 on end end end SSDT.dsl: GpioIo (Exclusive, PullDefault, 0, 0, IoRestrictionOutputOnly, "\\_SB.PCI0.GPIO", 0, ResourceConsumer, ,) { 53 } Signed-off-by: Duncan Laurie <dlaurie@chromium.org> Change-Id: Ibf5bab9c4bf6f21252373fb013e78f872550b167 Reviewed-on: https://review.coreboot.org/14934 Tested-by: build bot (Jenkins) Reviewed-by: Aaron Durbin <adurbin@chromium.org>
2016-05-10 02:08:38 +02:00
flags |= 1 << 4;
switch (gpio->irq.polarity) {
acpi: Add ACPI_ prefix to IRQ enum and struct names This is done to avoid any conflicts with same IRQ enums defined by other drivers. BUG=None BRANCH=None TEST=Compiles successfully Change-Id: I539831d853286ca45f6c36c3812a6fa9602df24c Signed-off-by: Furquan Shaikh <furquan@chromium.org> Reviewed-on: https://review.coreboot.org/18444 Tested-by: build bot (Jenkins) Reviewed-by: Aaron Durbin <adurbin@chromium.org>
2017-02-21 22:16:30 +01:00
case ACPI_IRQ_ACTIVE_HIGH:
acpi_device: Add support for writing ACPI GPIO descriptors Add definitions to describe GPIOs in generated ACPI objects and a method to write a GpioIo() or GpioInt() descriptor to the SSDT. ACPI GPIOs have many possible configuration options and a structure is created to describe it accurately in ACPI terms. There are many shared descriptor fields between GpioIo() and GpioInt() so the same function can write both types. GpioInt shares many properties with ACPI Interrupts and the same types are re-used here where possible. One addition is that GpioInt can be configured to trigger on both low and high edge transitions. One descriptor can describe multiple GPIO pins (limited to 8 in this implementation) that all share configuration and controller and are used by the same device scope. Accurately referring to the GPIO controller that this pin is connected to requires the SoC/board to implement a function handler for acpi_gpio_path(), or for the caller to provide this directly as a string in the acpi_gpio->reference variable. This will get used by device drivers to describe their resources in the SSDT. Here is a sample for a Maxim 98357A I2S codec which has a GPIO for power and channel selection called "sdmode". chip.h: struct drivers_generic_max98357a_config { struct acpi_gpio sdmode_gpio; }; max98357a.c: void acpi_fill_ssdt_generator(struct device *dev) { struct drivers_generic_max98357a_config *config = dev->chip_info; ... acpi_device_write_gpio(&config->sdmode_gpio); ... } devicetree.cb: device pci 1f.3 on chip drivers/generic/max98357a register "sdmode_gpio" = "ACPI_GPIO_OUTPUT(GPP_C5)" device generic 0 on end end end SSDT.dsl: GpioIo (Exclusive, PullDefault, 0, 0, IoRestrictionOutputOnly, "\\_SB.PCI0.GPIO", 0, ResourceConsumer, ,) { 53 } Signed-off-by: Duncan Laurie <dlaurie@chromium.org> Change-Id: Ibf5bab9c4bf6f21252373fb013e78f872550b167 Reviewed-on: https://review.coreboot.org/14934 Tested-by: build bot (Jenkins) Reviewed-by: Aaron Durbin <adurbin@chromium.org>
2016-05-10 02:08:38 +02:00
flags |= 0 << 1;
break;
acpi: Add ACPI_ prefix to IRQ enum and struct names This is done to avoid any conflicts with same IRQ enums defined by other drivers. BUG=None BRANCH=None TEST=Compiles successfully Change-Id: I539831d853286ca45f6c36c3812a6fa9602df24c Signed-off-by: Furquan Shaikh <furquan@chromium.org> Reviewed-on: https://review.coreboot.org/18444 Tested-by: build bot (Jenkins) Reviewed-by: Aaron Durbin <adurbin@chromium.org>
2017-02-21 22:16:30 +01:00
case ACPI_IRQ_ACTIVE_LOW:
acpi_device: Add support for writing ACPI GPIO descriptors Add definitions to describe GPIOs in generated ACPI objects and a method to write a GpioIo() or GpioInt() descriptor to the SSDT. ACPI GPIOs have many possible configuration options and a structure is created to describe it accurately in ACPI terms. There are many shared descriptor fields between GpioIo() and GpioInt() so the same function can write both types. GpioInt shares many properties with ACPI Interrupts and the same types are re-used here where possible. One addition is that GpioInt can be configured to trigger on both low and high edge transitions. One descriptor can describe multiple GPIO pins (limited to 8 in this implementation) that all share configuration and controller and are used by the same device scope. Accurately referring to the GPIO controller that this pin is connected to requires the SoC/board to implement a function handler for acpi_gpio_path(), or for the caller to provide this directly as a string in the acpi_gpio->reference variable. This will get used by device drivers to describe their resources in the SSDT. Here is a sample for a Maxim 98357A I2S codec which has a GPIO for power and channel selection called "sdmode". chip.h: struct drivers_generic_max98357a_config { struct acpi_gpio sdmode_gpio; }; max98357a.c: void acpi_fill_ssdt_generator(struct device *dev) { struct drivers_generic_max98357a_config *config = dev->chip_info; ... acpi_device_write_gpio(&config->sdmode_gpio); ... } devicetree.cb: device pci 1f.3 on chip drivers/generic/max98357a register "sdmode_gpio" = "ACPI_GPIO_OUTPUT(GPP_C5)" device generic 0 on end end end SSDT.dsl: GpioIo (Exclusive, PullDefault, 0, 0, IoRestrictionOutputOnly, "\\_SB.PCI0.GPIO", 0, ResourceConsumer, ,) { 53 } Signed-off-by: Duncan Laurie <dlaurie@chromium.org> Change-Id: Ibf5bab9c4bf6f21252373fb013e78f872550b167 Reviewed-on: https://review.coreboot.org/14934 Tested-by: build bot (Jenkins) Reviewed-by: Aaron Durbin <adurbin@chromium.org>
2016-05-10 02:08:38 +02:00
flags |= 1 << 1;
break;
acpi: Add ACPI_ prefix to IRQ enum and struct names This is done to avoid any conflicts with same IRQ enums defined by other drivers. BUG=None BRANCH=None TEST=Compiles successfully Change-Id: I539831d853286ca45f6c36c3812a6fa9602df24c Signed-off-by: Furquan Shaikh <furquan@chromium.org> Reviewed-on: https://review.coreboot.org/18444 Tested-by: build bot (Jenkins) Reviewed-by: Aaron Durbin <adurbin@chromium.org>
2017-02-21 22:16:30 +01:00
case ACPI_IRQ_ACTIVE_BOTH:
acpi_device: Add support for writing ACPI GPIO descriptors Add definitions to describe GPIOs in generated ACPI objects and a method to write a GpioIo() or GpioInt() descriptor to the SSDT. ACPI GPIOs have many possible configuration options and a structure is created to describe it accurately in ACPI terms. There are many shared descriptor fields between GpioIo() and GpioInt() so the same function can write both types. GpioInt shares many properties with ACPI Interrupts and the same types are re-used here where possible. One addition is that GpioInt can be configured to trigger on both low and high edge transitions. One descriptor can describe multiple GPIO pins (limited to 8 in this implementation) that all share configuration and controller and are used by the same device scope. Accurately referring to the GPIO controller that this pin is connected to requires the SoC/board to implement a function handler for acpi_gpio_path(), or for the caller to provide this directly as a string in the acpi_gpio->reference variable. This will get used by device drivers to describe their resources in the SSDT. Here is a sample for a Maxim 98357A I2S codec which has a GPIO for power and channel selection called "sdmode". chip.h: struct drivers_generic_max98357a_config { struct acpi_gpio sdmode_gpio; }; max98357a.c: void acpi_fill_ssdt_generator(struct device *dev) { struct drivers_generic_max98357a_config *config = dev->chip_info; ... acpi_device_write_gpio(&config->sdmode_gpio); ... } devicetree.cb: device pci 1f.3 on chip drivers/generic/max98357a register "sdmode_gpio" = "ACPI_GPIO_OUTPUT(GPP_C5)" device generic 0 on end end end SSDT.dsl: GpioIo (Exclusive, PullDefault, 0, 0, IoRestrictionOutputOnly, "\\_SB.PCI0.GPIO", 0, ResourceConsumer, ,) { 53 } Signed-off-by: Duncan Laurie <dlaurie@chromium.org> Change-Id: Ibf5bab9c4bf6f21252373fb013e78f872550b167 Reviewed-on: https://review.coreboot.org/14934 Tested-by: build bot (Jenkins) Reviewed-by: Aaron Durbin <adurbin@chromium.org>
2016-05-10 02:08:38 +02:00
flags |= 2 << 1;
break;
}
break;
case ACPI_GPIO_TYPE_IO:
/*
* Byte 7-8: GPIO IO Flags
* [15:4]: 0 => Reserved
* [3]: Sharing (0=EXCLUSIVE 1=SHARED)
* [2]: 0 => Reserved
* [1:0]: IO Restriction
* 0 => IoRestrictionNone
* 1 => IoRestrictionInputOnly
* 2 => IoRestrictionOutputOnly
* 3 => IoRestrictionNoneAndPreserve
*/
flags |= gpio->io_restrict & 3;
if (gpio->io_shared)
flags |= 1 << 3;
break;
}
acpigen_emit_word(flags);
/*
* Byte 9: Pin Configuration
* 0x01 => Default (no configuration applied)
* 0x02 => Pull-up
* 0x03 => Pull-down
* 0x04-0x7F => Reserved
* 0x80-0xff => Vendor defined
*/
acpigen_emit_byte(gpio->pull);
/* Byte 10-11: Output Drive Strength in 1/100 mA */
acpigen_emit_word(gpio->output_drive_strength);
/* Byte 12-13: Debounce Timeout in 1/100 ms */
acpigen_emit_word(gpio->interrupt_debounce_timeout);
/* Byte 14-15: Pin Table Offset, relative to start */
pin_table_offset = acpi_device_write_zero_len();
/* Byte 16: Reserved */
acpigen_emit_byte(0);
/* Byte 17-18: Resource Source Name Offset, relative to start */
resource_offset = acpi_device_write_zero_len();
/* Byte 19-20: Vendor Data Offset, relative to start */
vendor_data_offset = acpi_device_write_zero_len();
/* Byte 21-22: Vendor Data Length */
acpigen_emit_word(0);
/* Fill in Pin Table Offset */
acpi_device_fill_from_len(pin_table_offset, start);
/* Pin Table, one word for each pin */
gpio: Add support for translating gpio_t into ACPI pin Add a function for an SOC to define that will allow it to map the SOC-specific gpio_t value into an appropriate ACPI pin. The exact behavior depends on the GPIO implementation in the SOC, but it can be used to provide a pin number that is relative to the community or bank that a GPIO resides in. Change-Id: Icb97ccf7d6a9034877614d49166bc9e4fe659bcf Signed-off-by: Duncan Laurie <dlaurie@chromium.org> Reviewed-on: https://review.coreboot.org/15512 Tested-by: build bot (Jenkins) Reviewed-by: Aaron Durbin <adurbin@chromium.org>
2016-06-29 19:47:22 +02:00
for (pin = 0; pin < gpio->pin_count; pin++) {
uint16_t acpi_pin = gpio->pins[pin];
coreboot: Replace all IS_ENABLED(CONFIG_XXX) with CONFIG(XXX) This patch is a raw application of find src/ -type f | xargs sed -i -e 's/IS_ENABLED\s*(CONFIG_/CONFIG(/g' Change-Id: I6262d6d5c23cabe23c242b4f38d446b74fe16b88 Signed-off-by: Julius Werner <jwerner@chromium.org> Reviewed-on: https://review.coreboot.org/c/coreboot/+/31774 Tested-by: build bot (Jenkins) <no-reply@coreboot.org> Reviewed-by: Patrick Georgi <pgeorgi@google.com>
2019-03-06 01:53:33 +01:00
#if CONFIG(GENERIC_GPIO_LIB)
gpio: Add support for translating gpio_t into ACPI pin Add a function for an SOC to define that will allow it to map the SOC-specific gpio_t value into an appropriate ACPI pin. The exact behavior depends on the GPIO implementation in the SOC, but it can be used to provide a pin number that is relative to the community or bank that a GPIO resides in. Change-Id: Icb97ccf7d6a9034877614d49166bc9e4fe659bcf Signed-off-by: Duncan Laurie <dlaurie@chromium.org> Reviewed-on: https://review.coreboot.org/15512 Tested-by: build bot (Jenkins) Reviewed-by: Aaron Durbin <adurbin@chromium.org>
2016-06-29 19:47:22 +02:00
acpi_pin = gpio_acpi_pin(acpi_pin);
#endif
acpigen_emit_word(acpi_pin);
}
acpi_device: Add support for writing ACPI GPIO descriptors Add definitions to describe GPIOs in generated ACPI objects and a method to write a GpioIo() or GpioInt() descriptor to the SSDT. ACPI GPIOs have many possible configuration options and a structure is created to describe it accurately in ACPI terms. There are many shared descriptor fields between GpioIo() and GpioInt() so the same function can write both types. GpioInt shares many properties with ACPI Interrupts and the same types are re-used here where possible. One addition is that GpioInt can be configured to trigger on both low and high edge transitions. One descriptor can describe multiple GPIO pins (limited to 8 in this implementation) that all share configuration and controller and are used by the same device scope. Accurately referring to the GPIO controller that this pin is connected to requires the SoC/board to implement a function handler for acpi_gpio_path(), or for the caller to provide this directly as a string in the acpi_gpio->reference variable. This will get used by device drivers to describe their resources in the SSDT. Here is a sample for a Maxim 98357A I2S codec which has a GPIO for power and channel selection called "sdmode". chip.h: struct drivers_generic_max98357a_config { struct acpi_gpio sdmode_gpio; }; max98357a.c: void acpi_fill_ssdt_generator(struct device *dev) { struct drivers_generic_max98357a_config *config = dev->chip_info; ... acpi_device_write_gpio(&config->sdmode_gpio); ... } devicetree.cb: device pci 1f.3 on chip drivers/generic/max98357a register "sdmode_gpio" = "ACPI_GPIO_OUTPUT(GPP_C5)" device generic 0 on end end end SSDT.dsl: GpioIo (Exclusive, PullDefault, 0, 0, IoRestrictionOutputOnly, "\\_SB.PCI0.GPIO", 0, ResourceConsumer, ,) { 53 } Signed-off-by: Duncan Laurie <dlaurie@chromium.org> Change-Id: Ibf5bab9c4bf6f21252373fb013e78f872550b167 Reviewed-on: https://review.coreboot.org/14934 Tested-by: build bot (Jenkins) Reviewed-by: Aaron Durbin <adurbin@chromium.org>
2016-05-10 02:08:38 +02:00
/* Fill in Resource Source Name Offset */
acpi_device_fill_from_len(resource_offset, start);
/* Resource Source Name String */
coreboot: Replace all IS_ENABLED(CONFIG_XXX) with CONFIG(XXX) This patch is a raw application of find src/ -type f | xargs sed -i -e 's/IS_ENABLED\s*(CONFIG_/CONFIG(/g' Change-Id: I6262d6d5c23cabe23c242b4f38d446b74fe16b88 Signed-off-by: Julius Werner <jwerner@chromium.org> Reviewed-on: https://review.coreboot.org/c/coreboot/+/31774 Tested-by: build bot (Jenkins) <no-reply@coreboot.org> Reviewed-by: Patrick Georgi <pgeorgi@google.com>
2019-03-06 01:53:33 +01:00
#if CONFIG(GENERIC_GPIO_LIB)
acpi_device: Add support for writing ACPI GPIO descriptors Add definitions to describe GPIOs in generated ACPI objects and a method to write a GpioIo() or GpioInt() descriptor to the SSDT. ACPI GPIOs have many possible configuration options and a structure is created to describe it accurately in ACPI terms. There are many shared descriptor fields between GpioIo() and GpioInt() so the same function can write both types. GpioInt shares many properties with ACPI Interrupts and the same types are re-used here where possible. One addition is that GpioInt can be configured to trigger on both low and high edge transitions. One descriptor can describe multiple GPIO pins (limited to 8 in this implementation) that all share configuration and controller and are used by the same device scope. Accurately referring to the GPIO controller that this pin is connected to requires the SoC/board to implement a function handler for acpi_gpio_path(), or for the caller to provide this directly as a string in the acpi_gpio->reference variable. This will get used by device drivers to describe their resources in the SSDT. Here is a sample for a Maxim 98357A I2S codec which has a GPIO for power and channel selection called "sdmode". chip.h: struct drivers_generic_max98357a_config { struct acpi_gpio sdmode_gpio; }; max98357a.c: void acpi_fill_ssdt_generator(struct device *dev) { struct drivers_generic_max98357a_config *config = dev->chip_info; ... acpi_device_write_gpio(&config->sdmode_gpio); ... } devicetree.cb: device pci 1f.3 on chip drivers/generic/max98357a register "sdmode_gpio" = "ACPI_GPIO_OUTPUT(GPP_C5)" device generic 0 on end end end SSDT.dsl: GpioIo (Exclusive, PullDefault, 0, 0, IoRestrictionOutputOnly, "\\_SB.PCI0.GPIO", 0, ResourceConsumer, ,) { 53 } Signed-off-by: Duncan Laurie <dlaurie@chromium.org> Change-Id: Ibf5bab9c4bf6f21252373fb013e78f872550b167 Reviewed-on: https://review.coreboot.org/14934 Tested-by: build bot (Jenkins) Reviewed-by: Aaron Durbin <adurbin@chromium.org>
2016-05-10 02:08:38 +02:00
acpigen_emit_string(gpio->resource ? : gpio_acpi_path(gpio->pins[0]));
#else
acpigen_emit_string(gpio->resource);
#endif
/* Fill in Vendor Data Offset */
acpi_device_fill_from_len(vendor_data_offset, start);
/* Fill in GPIO Descriptor Length (account for len word) */
acpi_device_fill_len(desc_length);
}
acpi_device: Add support for writing ACPI I2C descriptors Add required definitions to describe an ACPI I2C bus and a method to write the I2cSerialBus() descriptor to the SSDT. This will be used by device drivers to describe their I2C resources to the OS. The devicetree i2c device can supply the address and 7 or 10 bit mode as well as indicate the GPIO controller device, and the bus speed can be fixed or configured by the driver. chip.h: struct drivers_i2c_generic_config { enum i2c_speed bus_speed; }; generic.c: void acpi_fill_ssdt_generator(struct device *dev) { struct drivers_i2c_generic_config *config = dev->chip_info; struct acpi_i2c i2c = { .address = dev->path->i2c.device, .mode_10bit = dev->path.i2c.mode_10bit, .speed = config->bus_speed ? : I2C_SPEED_FAST, .resource = acpi_device_path(dev->bus->dev) }; ... acpi_device_write_i2c(&i2c); ... } devicetree.cb: device pci 15.0 on chip drivers/i2c/generic device i2c 10.0 on end end end SSDT.dsl: I2cSerialBus (0x10, ControllerInitiated, 400000, AddressingMode7Bit, "\\_SB.PCI0.I2C0", 0, ResourceConsumer) Change-Id: I598401ac81a92c72f19da0271af1e218580a6c49 Signed-off-by: Duncan Laurie <dlaurie@chromium.org> Reviewed-on: https://review.coreboot.org/14935 Tested-by: build bot (Jenkins) Reviewed-by: Aaron Durbin <adurbin@chromium.org>
2016-05-10 05:10:47 +02:00
/* ACPI 6.1 section 6.4.3.8.2.1 - I2cSerialBus() */
void acpi_device_write_i2c(const struct acpi_i2c *i2c)
{
void *desc_length, *type_length;
/* Byte 0: Descriptor Type */
acpigen_emit_byte(ACPI_DESCRIPTOR_SERIAL_BUS);
/* Byte 1+2: Length (filled in later) */
desc_length = acpi_device_write_zero_len();
/* Byte 3: Revision ID */
ACPI: Clarify serial bus revision and specific revision Serial bus revision [Byte 3] and serial bus specific revision [Byte 9] are not the same. Change-Id: I366f62e6aa0e9c0dfbc1ec17adeebc42a0e777eb Signed-off-by: Elyes HAOUAS <ehaouas@noos.fr> Reviewed-on: https://review.coreboot.org/c/coreboot/+/32329 Tested-by: build bot (Jenkins) <no-reply@coreboot.org> Reviewed-by: Lijian Zhao <lijian.zhao@intel.com>
2019-04-16 08:12:10 +02:00
acpigen_emit_byte(ACPI_I2C_SERIAL_BUS_REVISION_ID);
acpi_device: Add support for writing ACPI I2C descriptors Add required definitions to describe an ACPI I2C bus and a method to write the I2cSerialBus() descriptor to the SSDT. This will be used by device drivers to describe their I2C resources to the OS. The devicetree i2c device can supply the address and 7 or 10 bit mode as well as indicate the GPIO controller device, and the bus speed can be fixed or configured by the driver. chip.h: struct drivers_i2c_generic_config { enum i2c_speed bus_speed; }; generic.c: void acpi_fill_ssdt_generator(struct device *dev) { struct drivers_i2c_generic_config *config = dev->chip_info; struct acpi_i2c i2c = { .address = dev->path->i2c.device, .mode_10bit = dev->path.i2c.mode_10bit, .speed = config->bus_speed ? : I2C_SPEED_FAST, .resource = acpi_device_path(dev->bus->dev) }; ... acpi_device_write_i2c(&i2c); ... } devicetree.cb: device pci 15.0 on chip drivers/i2c/generic device i2c 10.0 on end end end SSDT.dsl: I2cSerialBus (0x10, ControllerInitiated, 400000, AddressingMode7Bit, "\\_SB.PCI0.I2C0", 0, ResourceConsumer) Change-Id: I598401ac81a92c72f19da0271af1e218580a6c49 Signed-off-by: Duncan Laurie <dlaurie@chromium.org> Reviewed-on: https://review.coreboot.org/14935 Tested-by: build bot (Jenkins) Reviewed-by: Aaron Durbin <adurbin@chromium.org>
2016-05-10 05:10:47 +02:00
/* Byte 4: Resource Source Index is Reserved */
acpigen_emit_byte(0);
/* Byte 5: Serial Bus Type is I2C */
acpigen_emit_byte(ACPI_SERIAL_BUS_TYPE_I2C);
/*
* Byte 6: Flags
* [7:2]: 0 => Reserved
* [1]: 1 => ResourceConsumer
* [0]: 0 => ControllerInitiated
*/
acpigen_emit_byte(1 << 1);
/*
* Byte 7-8: Type Specific Flags
* [15:1]: 0 => Reserved
* [0]: 0 => 7bit, 1 => 10bit
*/
acpigen_emit_word(i2c->mode_10bit);
/* Byte 9: Type Specific Revision ID */
ACPI: Clarify serial bus revision and specific revision Serial bus revision [Byte 3] and serial bus specific revision [Byte 9] are not the same. Change-Id: I366f62e6aa0e9c0dfbc1ec17adeebc42a0e777eb Signed-off-by: Elyes HAOUAS <ehaouas@noos.fr> Reviewed-on: https://review.coreboot.org/c/coreboot/+/32329 Tested-by: build bot (Jenkins) <no-reply@coreboot.org> Reviewed-by: Lijian Zhao <lijian.zhao@intel.com>
2019-04-16 08:12:10 +02:00
acpigen_emit_byte(ACPI_I2C_TYPE_SPECIFIC_REVISION_ID);
acpi_device: Add support for writing ACPI I2C descriptors Add required definitions to describe an ACPI I2C bus and a method to write the I2cSerialBus() descriptor to the SSDT. This will be used by device drivers to describe their I2C resources to the OS. The devicetree i2c device can supply the address and 7 or 10 bit mode as well as indicate the GPIO controller device, and the bus speed can be fixed or configured by the driver. chip.h: struct drivers_i2c_generic_config { enum i2c_speed bus_speed; }; generic.c: void acpi_fill_ssdt_generator(struct device *dev) { struct drivers_i2c_generic_config *config = dev->chip_info; struct acpi_i2c i2c = { .address = dev->path->i2c.device, .mode_10bit = dev->path.i2c.mode_10bit, .speed = config->bus_speed ? : I2C_SPEED_FAST, .resource = acpi_device_path(dev->bus->dev) }; ... acpi_device_write_i2c(&i2c); ... } devicetree.cb: device pci 15.0 on chip drivers/i2c/generic device i2c 10.0 on end end end SSDT.dsl: I2cSerialBus (0x10, ControllerInitiated, 400000, AddressingMode7Bit, "\\_SB.PCI0.I2C0", 0, ResourceConsumer) Change-Id: I598401ac81a92c72f19da0271af1e218580a6c49 Signed-off-by: Duncan Laurie <dlaurie@chromium.org> Reviewed-on: https://review.coreboot.org/14935 Tested-by: build bot (Jenkins) Reviewed-by: Aaron Durbin <adurbin@chromium.org>
2016-05-10 05:10:47 +02:00
/* Byte 10-11: I2C Type Data Length */
type_length = acpi_device_write_zero_len();
/* Byte 12-15: I2C Bus Speed */
acpigen_emit_dword(i2c->speed);
/* Byte 16-17: I2C Slave Address */
acpigen_emit_word(i2c->address);
/* Fill in Type Data Length */
acpi_device_fill_len(type_length);
/* Byte 18+: ResourceSource */
acpigen_emit_string(i2c->resource);
/* Fill in I2C Descriptor Length */
acpi_device_fill_len(desc_length);
}
acpi_device: Add support for writing ACPI SPI descriptors Add required definitions to describe an ACPI SPI bus and a method to write the SpiSerialBus() descriptor to the SSDT. This will be used by device drivers to describe their SPI resources to the OS. SPI devices are not currently enumerated in the devicetree but can be enumerated by device drivers directly. generic.c: void acpi_fill_ssdt_generator(struct device *dev) { struct acpi_spi spi = { .device_select = dev->path->generic.device.id, .device_select_polarity = SPI_POLARITY_LOW, .spi_wire_mode = SPI_4_WIRE_MODE, .speed = 1000 * 1000; /* 1 mHz */ .data_bit_length = 8, .clock_phase = SPI_CLOCK_PHASE_FIRST, .clock_polarity = SPI_POLARITY_LOW, .resource = acpi_device_path(dev->bus->dev) }; ... acpi_device_write_spi(&spi); ... } devicetree.cb: device pci 1e.2 on chip drivers/spi/generic device generic 0 on end end end SSDT.dsl: SpiSerialBus (0, PolarityLow, FourWireMode, 8, ControllerInitiated, 1000000, ClockPolarityLow, ClockPhaseFirst, "\\_SB.PCI0.SPI0", 0, ResourceConsumer) Change-Id: I0ef83dc111ac6c19d68872ab64e1e5e3a7756cae Signed-off-by: Duncan Laurie <dlaurie@chromium.org> Reviewed-on: https://review.coreboot.org/14936 Tested-by: build bot (Jenkins) Reviewed-by: Aaron Durbin <adurbin@chromium.org>
2016-05-10 16:26:34 +02:00
/* ACPI 6.1 section 6.4.3.8.2.2 - SpiSerialBus() */
void acpi_device_write_spi(const struct acpi_spi *spi)
{
void *desc_length, *type_length;
uint16_t flags = 0;
/* Byte 0: Descriptor Type */
acpigen_emit_byte(ACPI_DESCRIPTOR_SERIAL_BUS);
/* Byte 1+2: Length (filled in later) */
desc_length = acpi_device_write_zero_len();
/* Byte 3: Revision ID */
ACPI: Clarify serial bus revision and specific revision Serial bus revision [Byte 3] and serial bus specific revision [Byte 9] are not the same. Change-Id: I366f62e6aa0e9c0dfbc1ec17adeebc42a0e777eb Signed-off-by: Elyes HAOUAS <ehaouas@noos.fr> Reviewed-on: https://review.coreboot.org/c/coreboot/+/32329 Tested-by: build bot (Jenkins) <no-reply@coreboot.org> Reviewed-by: Lijian Zhao <lijian.zhao@intel.com>
2019-04-16 08:12:10 +02:00
acpigen_emit_byte(ACPI_SPI_SERIAL_BUS_REVISION_ID);
acpi_device: Add support for writing ACPI SPI descriptors Add required definitions to describe an ACPI SPI bus and a method to write the SpiSerialBus() descriptor to the SSDT. This will be used by device drivers to describe their SPI resources to the OS. SPI devices are not currently enumerated in the devicetree but can be enumerated by device drivers directly. generic.c: void acpi_fill_ssdt_generator(struct device *dev) { struct acpi_spi spi = { .device_select = dev->path->generic.device.id, .device_select_polarity = SPI_POLARITY_LOW, .spi_wire_mode = SPI_4_WIRE_MODE, .speed = 1000 * 1000; /* 1 mHz */ .data_bit_length = 8, .clock_phase = SPI_CLOCK_PHASE_FIRST, .clock_polarity = SPI_POLARITY_LOW, .resource = acpi_device_path(dev->bus->dev) }; ... acpi_device_write_spi(&spi); ... } devicetree.cb: device pci 1e.2 on chip drivers/spi/generic device generic 0 on end end end SSDT.dsl: SpiSerialBus (0, PolarityLow, FourWireMode, 8, ControllerInitiated, 1000000, ClockPolarityLow, ClockPhaseFirst, "\\_SB.PCI0.SPI0", 0, ResourceConsumer) Change-Id: I0ef83dc111ac6c19d68872ab64e1e5e3a7756cae Signed-off-by: Duncan Laurie <dlaurie@chromium.org> Reviewed-on: https://review.coreboot.org/14936 Tested-by: build bot (Jenkins) Reviewed-by: Aaron Durbin <adurbin@chromium.org>
2016-05-10 16:26:34 +02:00
/* Byte 4: Resource Source Index is Reserved */
acpigen_emit_byte(0);
/* Byte 5: Serial Bus Type is SPI */
acpigen_emit_byte(ACPI_SERIAL_BUS_TYPE_SPI);
/*
* Byte 6: Flags
* [7:2]: 0 => Reserved
* [1]: 1 => ResourceConsumer
* [0]: 0 => ControllerInitiated
*/
acpigen_emit_byte(1 << 1);
/*
* Byte 7-8: Type Specific Flags
* [15:2]: 0 => Reserved
* [1]: 0 => ActiveLow, 1 => ActiveHigh
* [0]: 0 => FourWire, 1 => ThreeWire
*/
if (spi->wire_mode == SPI_3_WIRE_MODE)
flags |= 1 << 0;
if (spi->device_select_polarity == SPI_POLARITY_HIGH)
flags |= 1 << 1;
acpigen_emit_word(flags);
/* Byte 9: Type Specific Revision ID */
ACPI: Clarify serial bus revision and specific revision Serial bus revision [Byte 3] and serial bus specific revision [Byte 9] are not the same. Change-Id: I366f62e6aa0e9c0dfbc1ec17adeebc42a0e777eb Signed-off-by: Elyes HAOUAS <ehaouas@noos.fr> Reviewed-on: https://review.coreboot.org/c/coreboot/+/32329 Tested-by: build bot (Jenkins) <no-reply@coreboot.org> Reviewed-by: Lijian Zhao <lijian.zhao@intel.com>
2019-04-16 08:12:10 +02:00
acpigen_emit_byte(ACPI_SPI_TYPE_SPECIFIC_REVISION_ID);
acpi_device: Add support for writing ACPI SPI descriptors Add required definitions to describe an ACPI SPI bus and a method to write the SpiSerialBus() descriptor to the SSDT. This will be used by device drivers to describe their SPI resources to the OS. SPI devices are not currently enumerated in the devicetree but can be enumerated by device drivers directly. generic.c: void acpi_fill_ssdt_generator(struct device *dev) { struct acpi_spi spi = { .device_select = dev->path->generic.device.id, .device_select_polarity = SPI_POLARITY_LOW, .spi_wire_mode = SPI_4_WIRE_MODE, .speed = 1000 * 1000; /* 1 mHz */ .data_bit_length = 8, .clock_phase = SPI_CLOCK_PHASE_FIRST, .clock_polarity = SPI_POLARITY_LOW, .resource = acpi_device_path(dev->bus->dev) }; ... acpi_device_write_spi(&spi); ... } devicetree.cb: device pci 1e.2 on chip drivers/spi/generic device generic 0 on end end end SSDT.dsl: SpiSerialBus (0, PolarityLow, FourWireMode, 8, ControllerInitiated, 1000000, ClockPolarityLow, ClockPhaseFirst, "\\_SB.PCI0.SPI0", 0, ResourceConsumer) Change-Id: I0ef83dc111ac6c19d68872ab64e1e5e3a7756cae Signed-off-by: Duncan Laurie <dlaurie@chromium.org> Reviewed-on: https://review.coreboot.org/14936 Tested-by: build bot (Jenkins) Reviewed-by: Aaron Durbin <adurbin@chromium.org>
2016-05-10 16:26:34 +02:00
/* Byte 10-11: SPI Type Data Length */
type_length = acpi_device_write_zero_len();
/* Byte 12-15: Connection Speed */
acpigen_emit_dword(spi->speed);
/* Byte 16: Data Bit Length */
acpigen_emit_byte(spi->data_bit_length);
/* Byte 17: Clock Phase */
acpigen_emit_byte(spi->clock_phase);
/* Byte 18: Clock Polarity */
acpigen_emit_byte(spi->clock_polarity);
/* Byte 19-20: Device Selection */
acpigen_emit_word(spi->device_select);
/* Fill in Type Data Length */
acpi_device_fill_len(type_length);
/* Byte 21+: ResourceSource String */
acpigen_emit_string(spi->resource);
/* Fill in SPI Descriptor Length */
acpi_device_fill_len(desc_length);
}
acpi_device: Add support for writing ACPI Device Properties The recent ACPI specification extensions have formally defined a method for describing device information with a key=value format that is modeled after the Devicetree/DTS format using a special crafted object named _DSD with a specific UUID for this format. There are three defined Device Property types: Integers, Strings, and References. It is also possible to have arrays of these properties under one key=value pair. Strings and References are both represented as character arrays but result in different generated ACPI OpCodes. Various helpers are provided for writing the Device Property header (to fill in the object name and UUID) and footer (to fill in the property count and device length values) as well as for writing the different Device Property types. A specific helper is provided for writing the defined GPIO binding Device Property that is used to allow GPIOs to be referred to by name rather than resource index. This is all documented in the _DSD Device Properties UUID document: http://uefi.org/sites/default/files/resources/_DSD-device-properties-UUID.pdf This will be used by device drivers to provide device properties that are consumed by the operating system. Devicetree bindings are often described in the linux kernel at Documentation/devicetree/bindings/ A sample driver here has an input GPIO that it needs to describe to the kernel driver: chip.h: struct drivers_generic_sample_config { struct acpi_gpio mode_gpio; }; sample.c: static void acpi_fill_ssdt_generator(struct device *dev) { struct drivers_generic_sample_config *config = dev->chip_info; const char *path = acpi_device_path(dev); ... acpi_device_write_gpio(&config->mode_gpio); ... acpi_dp_write_header(); acpi_dp_write_gpio("mode-gpio", path, 0, 0, 0); acpi_dp_write_footer(); ... } devicetree.cb: device pci 1f.0 on chip drivers/generic/sample register "mode_gpio" = "ACPI_GPIO_INPUT(GPP_B1)" device generic 0 on end end end SSDT.dsl: Name (_CRS, ResourceTemplate () { GpioIo (Exclusive, PullDefault, 0, 0, IoRestrictionInputOnly, "\\_SB.PCI0.GPIO", 0, ResourceConsumer) { 25 } }) Name (_DSD, Package () { ToUUID ("daffd814-6eba-4d8c-8a91-bc9bbf4aa301"), Package () { Package () {"mode-gpio", Package () { \_SB.PCI0.LPCB, 0, 0, 1 }} } }) Change-Id: I93ffd09e59d05c09e38693e221a87085469be3ad Signed-off-by: Duncan Laurie <dlaurie@chromium.org> Reviewed-on: https://review.coreboot.org/14937 Tested-by: build bot (Jenkins) Reviewed-by: Aaron Durbin <adurbin@chromium.org>
2016-05-10 22:18:17 +02:00
acpi_device: Move power resource function to generic code Move the function that adds a power resource block from i2c/generic to the acpi device code at src/arch/x86/acpi_device.c so it can be used by more drivers. BUG=chrome-os-partner:61233 TEST=verify SSDT table generation is unchanged Change-Id: I0ffb61a4f46028cbe912e85c0124d9f5200b9c76 Signed-off-by: Duncan Laurie <dlaurie@chromium.org> Reviewed-on: https://review.coreboot.org/18391 Tested-by: build bot (Jenkins) Reviewed-by: Furquan Shaikh <furquan@google.com>
2017-02-18 02:05:03 +01:00
/* PowerResource() with Enable and/or Reset control */
src/drivers: Add reset/enable/stop_off_ms variables to ACPI devices Some touchscreens need to adhere to certain timings during the power off sequence as well as during the power on sequence. Adding reset_off_delay_ms, enable_off_delay_ms, and stop_off_delay_ms to accommodate these devices. BUG=b:78311818 BRANCH=None TEST=./util/abuild/abuild -p none -t google/poppy -x -a Change-Id: Idb4a5dbe56eee4749d2f2b514e92c28fb2c6078f Signed-off-by: Shelley Chen <shchen@google.com> Reviewed-on: https://review.coreboot.org/25882 Tested-by: build bot (Jenkins) <no-reply@coreboot.org> Reviewed-by: Duncan Laurie <dlaurie@chromium.org> Reviewed-by: Furquan Shaikh <furquan@google.com>
2018-04-26 22:52:30 +02:00
void acpi_device_add_power_res(const struct acpi_power_res_params *params)
acpi_device: Move power resource function to generic code Move the function that adds a power resource block from i2c/generic to the acpi device code at src/arch/x86/acpi_device.c so it can be used by more drivers. BUG=chrome-os-partner:61233 TEST=verify SSDT table generation is unchanged Change-Id: I0ffb61a4f46028cbe912e85c0124d9f5200b9c76 Signed-off-by: Duncan Laurie <dlaurie@chromium.org> Reviewed-on: https://review.coreboot.org/18391 Tested-by: build bot (Jenkins) Reviewed-by: Furquan Shaikh <furquan@google.com>
2017-02-18 02:05:03 +01:00
{
arch/x86: Fix most of remaining issues detected by checkpatch Fix the following errors and warnings detected by checkpatch.pl: ERROR: do not use assignment in if condition ERROR: trailing statements should be on next line ERROR: Macros with complex values should be enclosed in parentheses ERROR: switch and case should be at the same indent WARNING: char * array declaration might be better as static const WARNING: else is not generally useful after a break or return WARNING: storage class should be at the beginning of the declaration WARNING: void function return statements are not generally useful WARNING: break is not useful after a goto or return WARNING: Single statement macros should not use a do {} while (0) loop WARNING: sizeof *t should be sizeof(*t) WARNING: Comparisons should place the constant on the right side of the test TEST=Build and run on Galileo Gen2 Change-Id: I39d49790c5eaeedec5051e1fab0b1279275f6e7f Signed-off-by: Lee Leahy <Leroy.P.Leahy@intel.com> Reviewed-on: https://review.coreboot.org/18865 Tested-by: build bot (Jenkins) Reviewed-by: Martin Roth <martinroth@google.com>
2017-03-17 00:01:40 +01:00
static const char *power_res_dev_states[] = { "_PR0", "_PR3" };
x86/acpi_device: Allow acpi_device_add_power_res params as optional Allow for making both reset_gpio && enable_gpio as optional in the params by fixing a potential NULL deref and defaulting to zero values. BUG=b:147026979 BRANCH=none TEST=builds Change-Id: I8053d7a080dfed898400c0994bcea492c826fe3d Signed-off-by: Edward O'Callaghan <quasisec@google.com> Reviewed-on: https://review.coreboot.org/c/coreboot/+/38522 Tested-by: build bot (Jenkins) <no-reply@coreboot.org> Reviewed-by: Furquan Shaikh <furquan@google.com>
2020-01-23 00:32:33 +01:00
unsigned int reset_gpio = params->reset_gpio ? params->reset_gpio->pins[0] : 0;
unsigned int enable_gpio = params->enable_gpio ? params->enable_gpio->pins[0] : 0;
unsigned int stop_gpio = params->stop_gpio ? params->stop_gpio->pins[0] : 0;
acpi_device: Move power resource function to generic code Move the function that adds a power resource block from i2c/generic to the acpi device code at src/arch/x86/acpi_device.c so it can be used by more drivers. BUG=chrome-os-partner:61233 TEST=verify SSDT table generation is unchanged Change-Id: I0ffb61a4f46028cbe912e85c0124d9f5200b9c76 Signed-off-by: Duncan Laurie <dlaurie@chromium.org> Reviewed-on: https://review.coreboot.org/18391 Tested-by: build bot (Jenkins) Reviewed-by: Furquan Shaikh <furquan@google.com>
2017-02-18 02:05:03 +01:00
acpigen: Add stop gpio control to power resource There is at least one I2C device (being used by Soraka) that has 3 controls -- enable, reset and stop. If the stop gpio is not put into the right state when turning off the device in suspend mode, then it causes leakage. Thus, we need control in power resource to be able to stop the device when entering suspend state. BUG=b:64987428 TEST=Verified on soraka that touchscreen stop is correctly configured on suspend. Change-Id: Iae5ec7eb3972c5c7f80956d60d0d3c321bbefb0f Signed-off-by: Furquan Shaikh <furquan@chromium.org> Reviewed-on: https://review.coreboot.org/21249 Reviewed-by: Duncan Laurie <dlaurie@chromium.org> Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
2017-08-29 02:20:49 +02:00
if (!reset_gpio && !enable_gpio && !stop_gpio)
acpi_device: Move power resource function to generic code Move the function that adds a power resource block from i2c/generic to the acpi device code at src/arch/x86/acpi_device.c so it can be used by more drivers. BUG=chrome-os-partner:61233 TEST=verify SSDT table generation is unchanged Change-Id: I0ffb61a4f46028cbe912e85c0124d9f5200b9c76 Signed-off-by: Duncan Laurie <dlaurie@chromium.org> Reviewed-on: https://review.coreboot.org/18391 Tested-by: build bot (Jenkins) Reviewed-by: Furquan Shaikh <furquan@google.com>
2017-02-18 02:05:03 +01:00
return;
/* PowerResource (PRIC, 0, 0) */
acpigen_write_power_res("PRIC", 0, 0, power_res_dev_states,
ARRAY_SIZE(power_res_dev_states));
/* Method (_STA, 0, NotSerialized) { Return (0x1) } */
acpigen_write_STA(0x1);
/* Method (_ON, 0, Serialized) */
acpigen_write_method_serialized("_ON", 0);
if (reset_gpio)
src/drivers: Add reset/enable/stop_off_ms variables to ACPI devices Some touchscreens need to adhere to certain timings during the power off sequence as well as during the power on sequence. Adding reset_off_delay_ms, enable_off_delay_ms, and stop_off_delay_ms to accommodate these devices. BUG=b:78311818 BRANCH=None TEST=./util/abuild/abuild -p none -t google/poppy -x -a Change-Id: Idb4a5dbe56eee4749d2f2b514e92c28fb2c6078f Signed-off-by: Shelley Chen <shchen@google.com> Reviewed-on: https://review.coreboot.org/25882 Tested-by: build bot (Jenkins) <no-reply@coreboot.org> Reviewed-by: Duncan Laurie <dlaurie@chromium.org> Reviewed-by: Furquan Shaikh <furquan@google.com>
2018-04-26 22:52:30 +02:00
acpigen_enable_tx_gpio(params->reset_gpio);
acpi_device: Move power resource function to generic code Move the function that adds a power resource block from i2c/generic to the acpi device code at src/arch/x86/acpi_device.c so it can be used by more drivers. BUG=chrome-os-partner:61233 TEST=verify SSDT table generation is unchanged Change-Id: I0ffb61a4f46028cbe912e85c0124d9f5200b9c76 Signed-off-by: Duncan Laurie <dlaurie@chromium.org> Reviewed-on: https://review.coreboot.org/18391 Tested-by: build bot (Jenkins) Reviewed-by: Furquan Shaikh <furquan@google.com>
2017-02-18 02:05:03 +01:00
if (enable_gpio) {
src/drivers: Add reset/enable/stop_off_ms variables to ACPI devices Some touchscreens need to adhere to certain timings during the power off sequence as well as during the power on sequence. Adding reset_off_delay_ms, enable_off_delay_ms, and stop_off_delay_ms to accommodate these devices. BUG=b:78311818 BRANCH=None TEST=./util/abuild/abuild -p none -t google/poppy -x -a Change-Id: Idb4a5dbe56eee4749d2f2b514e92c28fb2c6078f Signed-off-by: Shelley Chen <shchen@google.com> Reviewed-on: https://review.coreboot.org/25882 Tested-by: build bot (Jenkins) <no-reply@coreboot.org> Reviewed-by: Duncan Laurie <dlaurie@chromium.org> Reviewed-by: Furquan Shaikh <furquan@google.com>
2018-04-26 22:52:30 +02:00
acpigen_enable_tx_gpio(params->enable_gpio);
if (params->enable_delay_ms)
acpigen_write_sleep(params->enable_delay_ms);
acpi_device: Move power resource function to generic code Move the function that adds a power resource block from i2c/generic to the acpi device code at src/arch/x86/acpi_device.c so it can be used by more drivers. BUG=chrome-os-partner:61233 TEST=verify SSDT table generation is unchanged Change-Id: I0ffb61a4f46028cbe912e85c0124d9f5200b9c76 Signed-off-by: Duncan Laurie <dlaurie@chromium.org> Reviewed-on: https://review.coreboot.org/18391 Tested-by: build bot (Jenkins) Reviewed-by: Furquan Shaikh <furquan@google.com>
2017-02-18 02:05:03 +01:00
}
if (reset_gpio) {
src/drivers: Add reset/enable/stop_off_ms variables to ACPI devices Some touchscreens need to adhere to certain timings during the power off sequence as well as during the power on sequence. Adding reset_off_delay_ms, enable_off_delay_ms, and stop_off_delay_ms to accommodate these devices. BUG=b:78311818 BRANCH=None TEST=./util/abuild/abuild -p none -t google/poppy -x -a Change-Id: Idb4a5dbe56eee4749d2f2b514e92c28fb2c6078f Signed-off-by: Shelley Chen <shchen@google.com> Reviewed-on: https://review.coreboot.org/25882 Tested-by: build bot (Jenkins) <no-reply@coreboot.org> Reviewed-by: Duncan Laurie <dlaurie@chromium.org> Reviewed-by: Furquan Shaikh <furquan@google.com>
2018-04-26 22:52:30 +02:00
acpigen_disable_tx_gpio(params->reset_gpio);
if (params->reset_delay_ms)
acpigen_write_sleep(params->reset_delay_ms);
acpi_device: Move power resource function to generic code Move the function that adds a power resource block from i2c/generic to the acpi device code at src/arch/x86/acpi_device.c so it can be used by more drivers. BUG=chrome-os-partner:61233 TEST=verify SSDT table generation is unchanged Change-Id: I0ffb61a4f46028cbe912e85c0124d9f5200b9c76 Signed-off-by: Duncan Laurie <dlaurie@chromium.org> Reviewed-on: https://review.coreboot.org/18391 Tested-by: build bot (Jenkins) Reviewed-by: Furquan Shaikh <furquan@google.com>
2017-02-18 02:05:03 +01:00
}
acpigen: Add stop gpio control to power resource There is at least one I2C device (being used by Soraka) that has 3 controls -- enable, reset and stop. If the stop gpio is not put into the right state when turning off the device in suspend mode, then it causes leakage. Thus, we need control in power resource to be able to stop the device when entering suspend state. BUG=b:64987428 TEST=Verified on soraka that touchscreen stop is correctly configured on suspend. Change-Id: Iae5ec7eb3972c5c7f80956d60d0d3c321bbefb0f Signed-off-by: Furquan Shaikh <furquan@chromium.org> Reviewed-on: https://review.coreboot.org/21249 Reviewed-by: Duncan Laurie <dlaurie@chromium.org> Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
2017-08-29 02:20:49 +02:00
if (stop_gpio) {
src/drivers: Add reset/enable/stop_off_ms variables to ACPI devices Some touchscreens need to adhere to certain timings during the power off sequence as well as during the power on sequence. Adding reset_off_delay_ms, enable_off_delay_ms, and stop_off_delay_ms to accommodate these devices. BUG=b:78311818 BRANCH=None TEST=./util/abuild/abuild -p none -t google/poppy -x -a Change-Id: Idb4a5dbe56eee4749d2f2b514e92c28fb2c6078f Signed-off-by: Shelley Chen <shchen@google.com> Reviewed-on: https://review.coreboot.org/25882 Tested-by: build bot (Jenkins) <no-reply@coreboot.org> Reviewed-by: Duncan Laurie <dlaurie@chromium.org> Reviewed-by: Furquan Shaikh <furquan@google.com>
2018-04-26 22:52:30 +02:00
acpigen_disable_tx_gpio(params->stop_gpio);
if (params->stop_delay_ms)
acpigen_write_sleep(params->stop_delay_ms);
acpigen: Add stop gpio control to power resource There is at least one I2C device (being used by Soraka) that has 3 controls -- enable, reset and stop. If the stop gpio is not put into the right state when turning off the device in suspend mode, then it causes leakage. Thus, we need control in power resource to be able to stop the device when entering suspend state. BUG=b:64987428 TEST=Verified on soraka that touchscreen stop is correctly configured on suspend. Change-Id: Iae5ec7eb3972c5c7f80956d60d0d3c321bbefb0f Signed-off-by: Furquan Shaikh <furquan@chromium.org> Reviewed-on: https://review.coreboot.org/21249 Reviewed-by: Duncan Laurie <dlaurie@chromium.org> Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
2017-08-29 02:20:49 +02:00
}
acpi_device: Move power resource function to generic code Move the function that adds a power resource block from i2c/generic to the acpi device code at src/arch/x86/acpi_device.c so it can be used by more drivers. BUG=chrome-os-partner:61233 TEST=verify SSDT table generation is unchanged Change-Id: I0ffb61a4f46028cbe912e85c0124d9f5200b9c76 Signed-off-by: Duncan Laurie <dlaurie@chromium.org> Reviewed-on: https://review.coreboot.org/18391 Tested-by: build bot (Jenkins) Reviewed-by: Furquan Shaikh <furquan@google.com>
2017-02-18 02:05:03 +01:00
acpigen_pop_len(); /* _ON method */
/* Method (_OFF, 0, Serialized) */
acpigen_write_method_serialized("_OFF", 0);
src/drivers: Add reset/enable/stop_off_ms variables to ACPI devices Some touchscreens need to adhere to certain timings during the power off sequence as well as during the power on sequence. Adding reset_off_delay_ms, enable_off_delay_ms, and stop_off_delay_ms to accommodate these devices. BUG=b:78311818 BRANCH=None TEST=./util/abuild/abuild -p none -t google/poppy -x -a Change-Id: Idb4a5dbe56eee4749d2f2b514e92c28fb2c6078f Signed-off-by: Shelley Chen <shchen@google.com> Reviewed-on: https://review.coreboot.org/25882 Tested-by: build bot (Jenkins) <no-reply@coreboot.org> Reviewed-by: Duncan Laurie <dlaurie@chromium.org> Reviewed-by: Furquan Shaikh <furquan@google.com>
2018-04-26 22:52:30 +02:00
if (stop_gpio) {
acpigen_enable_tx_gpio(params->stop_gpio);
if (params->stop_off_delay_ms)
acpigen_write_sleep(params->stop_off_delay_ms);
}
if (reset_gpio) {
acpigen_enable_tx_gpio(params->reset_gpio);
if (params->reset_off_delay_ms)
acpigen_write_sleep(params->reset_off_delay_ms);
}
if (enable_gpio) {
acpigen_disable_tx_gpio(params->enable_gpio);
if (params->enable_off_delay_ms)
acpigen_write_sleep(params->enable_off_delay_ms);
}
acpi_device: Move power resource function to generic code Move the function that adds a power resource block from i2c/generic to the acpi device code at src/arch/x86/acpi_device.c so it can be used by more drivers. BUG=chrome-os-partner:61233 TEST=verify SSDT table generation is unchanged Change-Id: I0ffb61a4f46028cbe912e85c0124d9f5200b9c76 Signed-off-by: Duncan Laurie <dlaurie@chromium.org> Reviewed-on: https://review.coreboot.org/18391 Tested-by: build bot (Jenkins) Reviewed-by: Furquan Shaikh <furquan@google.com>
2017-02-18 02:05:03 +01:00
acpigen_pop_len(); /* _OFF method */
acpigen_pop_len(); /* PowerResource PRIC */
}
acpi: Change device properties to work as a tree There is a second ACPI _DSD document from the UEFI Forum that details how _DSD style tables can be nested, creating a tree of similarly formatted tables. This document is linked from acpi_device.h. In order to support this the device property interface needs to be more flexible and build up a tree of properties to write all entries at once instead of writing each entry as it is generated. In the end this is a more flexible solution that can support drivers that need child tables like the DA7219 codec, while only requiring minor changes to the existing drivers that use the device property interface. This was tested on reef (apollolake) and chell (skylake) boards to ensure that there was no change in the generated SSDT AML. Change-Id: Ia22e3a5fd3982ffa7c324bee1a8d190d49f853dd Signed-off-by: Duncan Laurie <dlaurie@chromium.org> Reviewed-on: https://review.coreboot.org/15537 Tested-by: build bot (Jenkins) Reviewed-by: Aaron Durbin <adurbin@chromium.org>
2016-07-03 04:56:06 +02:00
static void acpi_dp_write_array(const struct acpi_dp *array);
static void acpi_dp_write_value(const struct acpi_dp *prop)
acpi_device: Add support for writing ACPI Device Properties The recent ACPI specification extensions have formally defined a method for describing device information with a key=value format that is modeled after the Devicetree/DTS format using a special crafted object named _DSD with a specific UUID for this format. There are three defined Device Property types: Integers, Strings, and References. It is also possible to have arrays of these properties under one key=value pair. Strings and References are both represented as character arrays but result in different generated ACPI OpCodes. Various helpers are provided for writing the Device Property header (to fill in the object name and UUID) and footer (to fill in the property count and device length values) as well as for writing the different Device Property types. A specific helper is provided for writing the defined GPIO binding Device Property that is used to allow GPIOs to be referred to by name rather than resource index. This is all documented in the _DSD Device Properties UUID document: http://uefi.org/sites/default/files/resources/_DSD-device-properties-UUID.pdf This will be used by device drivers to provide device properties that are consumed by the operating system. Devicetree bindings are often described in the linux kernel at Documentation/devicetree/bindings/ A sample driver here has an input GPIO that it needs to describe to the kernel driver: chip.h: struct drivers_generic_sample_config { struct acpi_gpio mode_gpio; }; sample.c: static void acpi_fill_ssdt_generator(struct device *dev) { struct drivers_generic_sample_config *config = dev->chip_info; const char *path = acpi_device_path(dev); ... acpi_device_write_gpio(&config->mode_gpio); ... acpi_dp_write_header(); acpi_dp_write_gpio("mode-gpio", path, 0, 0, 0); acpi_dp_write_footer(); ... } devicetree.cb: device pci 1f.0 on chip drivers/generic/sample register "mode_gpio" = "ACPI_GPIO_INPUT(GPP_B1)" device generic 0 on end end end SSDT.dsl: Name (_CRS, ResourceTemplate () { GpioIo (Exclusive, PullDefault, 0, 0, IoRestrictionInputOnly, "\\_SB.PCI0.GPIO", 0, ResourceConsumer) { 25 } }) Name (_DSD, Package () { ToUUID ("daffd814-6eba-4d8c-8a91-bc9bbf4aa301"), Package () { Package () {"mode-gpio", Package () { \_SB.PCI0.LPCB, 0, 0, 1 }} } }) Change-Id: I93ffd09e59d05c09e38693e221a87085469be3ad Signed-off-by: Duncan Laurie <dlaurie@chromium.org> Reviewed-on: https://review.coreboot.org/14937 Tested-by: build bot (Jenkins) Reviewed-by: Aaron Durbin <adurbin@chromium.org>
2016-05-10 22:18:17 +02:00
{
switch (prop->type) {
case ACPI_DP_TYPE_INTEGER:
acpigen_write_integer(prop->integer);
break;
case ACPI_DP_TYPE_STRING:
acpi: Change API called to write the name for ACPI_DP_TYPE_CHILD The API called to write the name of the child table in the dp entry (type ACPI_DP_TYPE_CHILD) was not including the quotes, e.g., it was DAAD and not "DAAD". Thus, the kernel driver did not get the right information from SSDT. Change the API to acpigen_write_string() to fix the issue. Signed-off-by: Harsha Priya <harshapriya.n@intel.com> Change-Id: Id33ad29e637bf1fe6b02e8a4b0fd9e220e8984e7 Reviewed-on: https://review.coreboot.org/15724 Tested-by: build bot (Jenkins) Reviewed-by: Aaron Durbin <adurbin@chromium.org>
2016-07-16 02:31:43 +02:00
case ACPI_DP_TYPE_CHILD:
acpi_device: Add support for writing ACPI Device Properties The recent ACPI specification extensions have formally defined a method for describing device information with a key=value format that is modeled after the Devicetree/DTS format using a special crafted object named _DSD with a specific UUID for this format. There are three defined Device Property types: Integers, Strings, and References. It is also possible to have arrays of these properties under one key=value pair. Strings and References are both represented as character arrays but result in different generated ACPI OpCodes. Various helpers are provided for writing the Device Property header (to fill in the object name and UUID) and footer (to fill in the property count and device length values) as well as for writing the different Device Property types. A specific helper is provided for writing the defined GPIO binding Device Property that is used to allow GPIOs to be referred to by name rather than resource index. This is all documented in the _DSD Device Properties UUID document: http://uefi.org/sites/default/files/resources/_DSD-device-properties-UUID.pdf This will be used by device drivers to provide device properties that are consumed by the operating system. Devicetree bindings are often described in the linux kernel at Documentation/devicetree/bindings/ A sample driver here has an input GPIO that it needs to describe to the kernel driver: chip.h: struct drivers_generic_sample_config { struct acpi_gpio mode_gpio; }; sample.c: static void acpi_fill_ssdt_generator(struct device *dev) { struct drivers_generic_sample_config *config = dev->chip_info; const char *path = acpi_device_path(dev); ... acpi_device_write_gpio(&config->mode_gpio); ... acpi_dp_write_header(); acpi_dp_write_gpio("mode-gpio", path, 0, 0, 0); acpi_dp_write_footer(); ... } devicetree.cb: device pci 1f.0 on chip drivers/generic/sample register "mode_gpio" = "ACPI_GPIO_INPUT(GPP_B1)" device generic 0 on end end end SSDT.dsl: Name (_CRS, ResourceTemplate () { GpioIo (Exclusive, PullDefault, 0, 0, IoRestrictionInputOnly, "\\_SB.PCI0.GPIO", 0, ResourceConsumer) { 25 } }) Name (_DSD, Package () { ToUUID ("daffd814-6eba-4d8c-8a91-bc9bbf4aa301"), Package () { Package () {"mode-gpio", Package () { \_SB.PCI0.LPCB, 0, 0, 1 }} } }) Change-Id: I93ffd09e59d05c09e38693e221a87085469be3ad Signed-off-by: Duncan Laurie <dlaurie@chromium.org> Reviewed-on: https://review.coreboot.org/14937 Tested-by: build bot (Jenkins) Reviewed-by: Aaron Durbin <adurbin@chromium.org>
2016-05-10 22:18:17 +02:00
acpigen_write_string(prop->string);
break;
case ACPI_DP_TYPE_REFERENCE:
acpigen_emit_namestring(prop->string);
break;
acpi: Change device properties to work as a tree There is a second ACPI _DSD document from the UEFI Forum that details how _DSD style tables can be nested, creating a tree of similarly formatted tables. This document is linked from acpi_device.h. In order to support this the device property interface needs to be more flexible and build up a tree of properties to write all entries at once instead of writing each entry as it is generated. In the end this is a more flexible solution that can support drivers that need child tables like the DA7219 codec, while only requiring minor changes to the existing drivers that use the device property interface. This was tested on reef (apollolake) and chell (skylake) boards to ensure that there was no change in the generated SSDT AML. Change-Id: Ia22e3a5fd3982ffa7c324bee1a8d190d49f853dd Signed-off-by: Duncan Laurie <dlaurie@chromium.org> Reviewed-on: https://review.coreboot.org/15537 Tested-by: build bot (Jenkins) Reviewed-by: Aaron Durbin <adurbin@chromium.org>
2016-07-03 04:56:06 +02:00
case ACPI_DP_TYPE_ARRAY:
acpi_dp_write_array(prop->array);
break;
default:
break;
acpi_device: Add support for writing ACPI Device Properties The recent ACPI specification extensions have formally defined a method for describing device information with a key=value format that is modeled after the Devicetree/DTS format using a special crafted object named _DSD with a specific UUID for this format. There are three defined Device Property types: Integers, Strings, and References. It is also possible to have arrays of these properties under one key=value pair. Strings and References are both represented as character arrays but result in different generated ACPI OpCodes. Various helpers are provided for writing the Device Property header (to fill in the object name and UUID) and footer (to fill in the property count and device length values) as well as for writing the different Device Property types. A specific helper is provided for writing the defined GPIO binding Device Property that is used to allow GPIOs to be referred to by name rather than resource index. This is all documented in the _DSD Device Properties UUID document: http://uefi.org/sites/default/files/resources/_DSD-device-properties-UUID.pdf This will be used by device drivers to provide device properties that are consumed by the operating system. Devicetree bindings are often described in the linux kernel at Documentation/devicetree/bindings/ A sample driver here has an input GPIO that it needs to describe to the kernel driver: chip.h: struct drivers_generic_sample_config { struct acpi_gpio mode_gpio; }; sample.c: static void acpi_fill_ssdt_generator(struct device *dev) { struct drivers_generic_sample_config *config = dev->chip_info; const char *path = acpi_device_path(dev); ... acpi_device_write_gpio(&config->mode_gpio); ... acpi_dp_write_header(); acpi_dp_write_gpio("mode-gpio", path, 0, 0, 0); acpi_dp_write_footer(); ... } devicetree.cb: device pci 1f.0 on chip drivers/generic/sample register "mode_gpio" = "ACPI_GPIO_INPUT(GPP_B1)" device generic 0 on end end end SSDT.dsl: Name (_CRS, ResourceTemplate () { GpioIo (Exclusive, PullDefault, 0, 0, IoRestrictionInputOnly, "\\_SB.PCI0.GPIO", 0, ResourceConsumer) { 25 } }) Name (_DSD, Package () { ToUUID ("daffd814-6eba-4d8c-8a91-bc9bbf4aa301"), Package () { Package () {"mode-gpio", Package () { \_SB.PCI0.LPCB, 0, 0, 1 }} } }) Change-Id: I93ffd09e59d05c09e38693e221a87085469be3ad Signed-off-by: Duncan Laurie <dlaurie@chromium.org> Reviewed-on: https://review.coreboot.org/14937 Tested-by: build bot (Jenkins) Reviewed-by: Aaron Durbin <adurbin@chromium.org>
2016-05-10 22:18:17 +02:00
}
}
acpi: Change device properties to work as a tree There is a second ACPI _DSD document from the UEFI Forum that details how _DSD style tables can be nested, creating a tree of similarly formatted tables. This document is linked from acpi_device.h. In order to support this the device property interface needs to be more flexible and build up a tree of properties to write all entries at once instead of writing each entry as it is generated. In the end this is a more flexible solution that can support drivers that need child tables like the DA7219 codec, while only requiring minor changes to the existing drivers that use the device property interface. This was tested on reef (apollolake) and chell (skylake) boards to ensure that there was no change in the generated SSDT AML. Change-Id: Ia22e3a5fd3982ffa7c324bee1a8d190d49f853dd Signed-off-by: Duncan Laurie <dlaurie@chromium.org> Reviewed-on: https://review.coreboot.org/15537 Tested-by: build bot (Jenkins) Reviewed-by: Aaron Durbin <adurbin@chromium.org>
2016-07-03 04:56:06 +02:00
/* Package (2) { "prop->name", VALUE } */
static void acpi_dp_write_property(const struct acpi_dp *prop)
acpi_device: Add support for writing ACPI Device Properties The recent ACPI specification extensions have formally defined a method for describing device information with a key=value format that is modeled after the Devicetree/DTS format using a special crafted object named _DSD with a specific UUID for this format. There are three defined Device Property types: Integers, Strings, and References. It is also possible to have arrays of these properties under one key=value pair. Strings and References are both represented as character arrays but result in different generated ACPI OpCodes. Various helpers are provided for writing the Device Property header (to fill in the object name and UUID) and footer (to fill in the property count and device length values) as well as for writing the different Device Property types. A specific helper is provided for writing the defined GPIO binding Device Property that is used to allow GPIOs to be referred to by name rather than resource index. This is all documented in the _DSD Device Properties UUID document: http://uefi.org/sites/default/files/resources/_DSD-device-properties-UUID.pdf This will be used by device drivers to provide device properties that are consumed by the operating system. Devicetree bindings are often described in the linux kernel at Documentation/devicetree/bindings/ A sample driver here has an input GPIO that it needs to describe to the kernel driver: chip.h: struct drivers_generic_sample_config { struct acpi_gpio mode_gpio; }; sample.c: static void acpi_fill_ssdt_generator(struct device *dev) { struct drivers_generic_sample_config *config = dev->chip_info; const char *path = acpi_device_path(dev); ... acpi_device_write_gpio(&config->mode_gpio); ... acpi_dp_write_header(); acpi_dp_write_gpio("mode-gpio", path, 0, 0, 0); acpi_dp_write_footer(); ... } devicetree.cb: device pci 1f.0 on chip drivers/generic/sample register "mode_gpio" = "ACPI_GPIO_INPUT(GPP_B1)" device generic 0 on end end end SSDT.dsl: Name (_CRS, ResourceTemplate () { GpioIo (Exclusive, PullDefault, 0, 0, IoRestrictionInputOnly, "\\_SB.PCI0.GPIO", 0, ResourceConsumer) { 25 } }) Name (_DSD, Package () { ToUUID ("daffd814-6eba-4d8c-8a91-bc9bbf4aa301"), Package () { Package () {"mode-gpio", Package () { \_SB.PCI0.LPCB, 0, 0, 1 }} } }) Change-Id: I93ffd09e59d05c09e38693e221a87085469be3ad Signed-off-by: Duncan Laurie <dlaurie@chromium.org> Reviewed-on: https://review.coreboot.org/14937 Tested-by: build bot (Jenkins) Reviewed-by: Aaron Durbin <adurbin@chromium.org>
2016-05-10 22:18:17 +02:00
{
acpigen_write_package(2);
acpi: Change device properties to work as a tree There is a second ACPI _DSD document from the UEFI Forum that details how _DSD style tables can be nested, creating a tree of similarly formatted tables. This document is linked from acpi_device.h. In order to support this the device property interface needs to be more flexible and build up a tree of properties to write all entries at once instead of writing each entry as it is generated. In the end this is a more flexible solution that can support drivers that need child tables like the DA7219 codec, while only requiring minor changes to the existing drivers that use the device property interface. This was tested on reef (apollolake) and chell (skylake) boards to ensure that there was no change in the generated SSDT AML. Change-Id: Ia22e3a5fd3982ffa7c324bee1a8d190d49f853dd Signed-off-by: Duncan Laurie <dlaurie@chromium.org> Reviewed-on: https://review.coreboot.org/15537 Tested-by: build bot (Jenkins) Reviewed-by: Aaron Durbin <adurbin@chromium.org>
2016-07-03 04:56:06 +02:00
acpigen_write_string(prop->name);
acpi_device: Add support for writing ACPI Device Properties The recent ACPI specification extensions have formally defined a method for describing device information with a key=value format that is modeled after the Devicetree/DTS format using a special crafted object named _DSD with a specific UUID for this format. There are three defined Device Property types: Integers, Strings, and References. It is also possible to have arrays of these properties under one key=value pair. Strings and References are both represented as character arrays but result in different generated ACPI OpCodes. Various helpers are provided for writing the Device Property header (to fill in the object name and UUID) and footer (to fill in the property count and device length values) as well as for writing the different Device Property types. A specific helper is provided for writing the defined GPIO binding Device Property that is used to allow GPIOs to be referred to by name rather than resource index. This is all documented in the _DSD Device Properties UUID document: http://uefi.org/sites/default/files/resources/_DSD-device-properties-UUID.pdf This will be used by device drivers to provide device properties that are consumed by the operating system. Devicetree bindings are often described in the linux kernel at Documentation/devicetree/bindings/ A sample driver here has an input GPIO that it needs to describe to the kernel driver: chip.h: struct drivers_generic_sample_config { struct acpi_gpio mode_gpio; }; sample.c: static void acpi_fill_ssdt_generator(struct device *dev) { struct drivers_generic_sample_config *config = dev->chip_info; const char *path = acpi_device_path(dev); ... acpi_device_write_gpio(&config->mode_gpio); ... acpi_dp_write_header(); acpi_dp_write_gpio("mode-gpio", path, 0, 0, 0); acpi_dp_write_footer(); ... } devicetree.cb: device pci 1f.0 on chip drivers/generic/sample register "mode_gpio" = "ACPI_GPIO_INPUT(GPP_B1)" device generic 0 on end end end SSDT.dsl: Name (_CRS, ResourceTemplate () { GpioIo (Exclusive, PullDefault, 0, 0, IoRestrictionInputOnly, "\\_SB.PCI0.GPIO", 0, ResourceConsumer) { 25 } }) Name (_DSD, Package () { ToUUID ("daffd814-6eba-4d8c-8a91-bc9bbf4aa301"), Package () { Package () {"mode-gpio", Package () { \_SB.PCI0.LPCB, 0, 0, 1 }} } }) Change-Id: I93ffd09e59d05c09e38693e221a87085469be3ad Signed-off-by: Duncan Laurie <dlaurie@chromium.org> Reviewed-on: https://review.coreboot.org/14937 Tested-by: build bot (Jenkins) Reviewed-by: Aaron Durbin <adurbin@chromium.org>
2016-05-10 22:18:17 +02:00
acpi_dp_write_value(prop);
acpigen_pop_len();
}
acpi: Change device properties to work as a tree There is a second ACPI _DSD document from the UEFI Forum that details how _DSD style tables can be nested, creating a tree of similarly formatted tables. This document is linked from acpi_device.h. In order to support this the device property interface needs to be more flexible and build up a tree of properties to write all entries at once instead of writing each entry as it is generated. In the end this is a more flexible solution that can support drivers that need child tables like the DA7219 codec, while only requiring minor changes to the existing drivers that use the device property interface. This was tested on reef (apollolake) and chell (skylake) boards to ensure that there was no change in the generated SSDT AML. Change-Id: Ia22e3a5fd3982ffa7c324bee1a8d190d49f853dd Signed-off-by: Duncan Laurie <dlaurie@chromium.org> Reviewed-on: https://review.coreboot.org/15537 Tested-by: build bot (Jenkins) Reviewed-by: Aaron Durbin <adurbin@chromium.org>
2016-07-03 04:56:06 +02:00
/* Write array of Device Properties */
static void acpi_dp_write_array(const struct acpi_dp *array)
acpi_device: Add support for writing ACPI Device Properties The recent ACPI specification extensions have formally defined a method for describing device information with a key=value format that is modeled after the Devicetree/DTS format using a special crafted object named _DSD with a specific UUID for this format. There are three defined Device Property types: Integers, Strings, and References. It is also possible to have arrays of these properties under one key=value pair. Strings and References are both represented as character arrays but result in different generated ACPI OpCodes. Various helpers are provided for writing the Device Property header (to fill in the object name and UUID) and footer (to fill in the property count and device length values) as well as for writing the different Device Property types. A specific helper is provided for writing the defined GPIO binding Device Property that is used to allow GPIOs to be referred to by name rather than resource index. This is all documented in the _DSD Device Properties UUID document: http://uefi.org/sites/default/files/resources/_DSD-device-properties-UUID.pdf This will be used by device drivers to provide device properties that are consumed by the operating system. Devicetree bindings are often described in the linux kernel at Documentation/devicetree/bindings/ A sample driver here has an input GPIO that it needs to describe to the kernel driver: chip.h: struct drivers_generic_sample_config { struct acpi_gpio mode_gpio; }; sample.c: static void acpi_fill_ssdt_generator(struct device *dev) { struct drivers_generic_sample_config *config = dev->chip_info; const char *path = acpi_device_path(dev); ... acpi_device_write_gpio(&config->mode_gpio); ... acpi_dp_write_header(); acpi_dp_write_gpio("mode-gpio", path, 0, 0, 0); acpi_dp_write_footer(); ... } devicetree.cb: device pci 1f.0 on chip drivers/generic/sample register "mode_gpio" = "ACPI_GPIO_INPUT(GPP_B1)" device generic 0 on end end end SSDT.dsl: Name (_CRS, ResourceTemplate () { GpioIo (Exclusive, PullDefault, 0, 0, IoRestrictionInputOnly, "\\_SB.PCI0.GPIO", 0, ResourceConsumer) { 25 } }) Name (_DSD, Package () { ToUUID ("daffd814-6eba-4d8c-8a91-bc9bbf4aa301"), Package () { Package () {"mode-gpio", Package () { \_SB.PCI0.LPCB, 0, 0, 1 }} } }) Change-Id: I93ffd09e59d05c09e38693e221a87085469be3ad Signed-off-by: Duncan Laurie <dlaurie@chromium.org> Reviewed-on: https://review.coreboot.org/14937 Tested-by: build bot (Jenkins) Reviewed-by: Aaron Durbin <adurbin@chromium.org>
2016-05-10 22:18:17 +02:00
{
acpi: Change device properties to work as a tree There is a second ACPI _DSD document from the UEFI Forum that details how _DSD style tables can be nested, creating a tree of similarly formatted tables. This document is linked from acpi_device.h. In order to support this the device property interface needs to be more flexible and build up a tree of properties to write all entries at once instead of writing each entry as it is generated. In the end this is a more flexible solution that can support drivers that need child tables like the DA7219 codec, while only requiring minor changes to the existing drivers that use the device property interface. This was tested on reef (apollolake) and chell (skylake) boards to ensure that there was no change in the generated SSDT AML. Change-Id: Ia22e3a5fd3982ffa7c324bee1a8d190d49f853dd Signed-off-by: Duncan Laurie <dlaurie@chromium.org> Reviewed-on: https://review.coreboot.org/15537 Tested-by: build bot (Jenkins) Reviewed-by: Aaron Durbin <adurbin@chromium.org>
2016-07-03 04:56:06 +02:00
const struct acpi_dp *dp;
char *pkg_count;
acpi_device: Add support for writing ACPI Device Properties The recent ACPI specification extensions have formally defined a method for describing device information with a key=value format that is modeled after the Devicetree/DTS format using a special crafted object named _DSD with a specific UUID for this format. There are three defined Device Property types: Integers, Strings, and References. It is also possible to have arrays of these properties under one key=value pair. Strings and References are both represented as character arrays but result in different generated ACPI OpCodes. Various helpers are provided for writing the Device Property header (to fill in the object name and UUID) and footer (to fill in the property count and device length values) as well as for writing the different Device Property types. A specific helper is provided for writing the defined GPIO binding Device Property that is used to allow GPIOs to be referred to by name rather than resource index. This is all documented in the _DSD Device Properties UUID document: http://uefi.org/sites/default/files/resources/_DSD-device-properties-UUID.pdf This will be used by device drivers to provide device properties that are consumed by the operating system. Devicetree bindings are often described in the linux kernel at Documentation/devicetree/bindings/ A sample driver here has an input GPIO that it needs to describe to the kernel driver: chip.h: struct drivers_generic_sample_config { struct acpi_gpio mode_gpio; }; sample.c: static void acpi_fill_ssdt_generator(struct device *dev) { struct drivers_generic_sample_config *config = dev->chip_info; const char *path = acpi_device_path(dev); ... acpi_device_write_gpio(&config->mode_gpio); ... acpi_dp_write_header(); acpi_dp_write_gpio("mode-gpio", path, 0, 0, 0); acpi_dp_write_footer(); ... } devicetree.cb: device pci 1f.0 on chip drivers/generic/sample register "mode_gpio" = "ACPI_GPIO_INPUT(GPP_B1)" device generic 0 on end end end SSDT.dsl: Name (_CRS, ResourceTemplate () { GpioIo (Exclusive, PullDefault, 0, 0, IoRestrictionInputOnly, "\\_SB.PCI0.GPIO", 0, ResourceConsumer) { 25 } }) Name (_DSD, Package () { ToUUID ("daffd814-6eba-4d8c-8a91-bc9bbf4aa301"), Package () { Package () {"mode-gpio", Package () { \_SB.PCI0.LPCB, 0, 0, 1 }} } }) Change-Id: I93ffd09e59d05c09e38693e221a87085469be3ad Signed-off-by: Duncan Laurie <dlaurie@chromium.org> Reviewed-on: https://review.coreboot.org/14937 Tested-by: build bot (Jenkins) Reviewed-by: Aaron Durbin <adurbin@chromium.org>
2016-05-10 22:18:17 +02:00
acpi: Change device properties to work as a tree There is a second ACPI _DSD document from the UEFI Forum that details how _DSD style tables can be nested, creating a tree of similarly formatted tables. This document is linked from acpi_device.h. In order to support this the device property interface needs to be more flexible and build up a tree of properties to write all entries at once instead of writing each entry as it is generated. In the end this is a more flexible solution that can support drivers that need child tables like the DA7219 codec, while only requiring minor changes to the existing drivers that use the device property interface. This was tested on reef (apollolake) and chell (skylake) boards to ensure that there was no change in the generated SSDT AML. Change-Id: Ia22e3a5fd3982ffa7c324bee1a8d190d49f853dd Signed-off-by: Duncan Laurie <dlaurie@chromium.org> Reviewed-on: https://review.coreboot.org/15537 Tested-by: build bot (Jenkins) Reviewed-by: Aaron Durbin <adurbin@chromium.org>
2016-07-03 04:56:06 +02:00
/* Package element count determined as it is populated */
pkg_count = acpigen_write_package(0);
x86/acpi_device: Fix writing of array property Only acpi_dp of type DP_TYPE_TABLE is allowed to be an array. This DP_TYPE_TABLE does not have a value which is written. Thus, acpi_dp_write_array needs to start counting from the next element type in the array. Fix this by updating the initialization in for loop for writing array elements. BUG=chrome-os-partner:55988 BRANCH=None TEST=Verified that the correct number of elements are passed for add_gpio in maxim sdmode-gpio. Change-Id: I8e1e540d66086971de2edf0bb83494d3b1dbd176 Signed-off-by: Furquan Shaikh <furquan@chromium.org> Reviewed-on: https://review.coreboot.org/16871 Reviewed-by: Aaron Durbin <adurbin@chromium.org> Tested-by: build bot (Jenkins) Reviewed-by: Duncan Laurie <dlaurie@chromium.org>
2016-10-04 08:30:14 +02:00
/*
* Only acpi_dp of type DP_TYPE_TABLE is allowed to be an array.
* DP_TYPE_TABLE does not have a value to be written. Thus, start
* the loop from next type in the array.
*/
for (dp = array->next; dp; dp = dp->next) {
acpi: Change device properties to work as a tree There is a second ACPI _DSD document from the UEFI Forum that details how _DSD style tables can be nested, creating a tree of similarly formatted tables. This document is linked from acpi_device.h. In order to support this the device property interface needs to be more flexible and build up a tree of properties to write all entries at once instead of writing each entry as it is generated. In the end this is a more flexible solution that can support drivers that need child tables like the DA7219 codec, while only requiring minor changes to the existing drivers that use the device property interface. This was tested on reef (apollolake) and chell (skylake) boards to ensure that there was no change in the generated SSDT AML. Change-Id: Ia22e3a5fd3982ffa7c324bee1a8d190d49f853dd Signed-off-by: Duncan Laurie <dlaurie@chromium.org> Reviewed-on: https://review.coreboot.org/15537 Tested-by: build bot (Jenkins) Reviewed-by: Aaron Durbin <adurbin@chromium.org>
2016-07-03 04:56:06 +02:00
acpi_dp_write_value(dp);
(*pkg_count)++;
}
acpigen_pop_len();
acpi_device: Add support for writing ACPI Device Properties The recent ACPI specification extensions have formally defined a method for describing device information with a key=value format that is modeled after the Devicetree/DTS format using a special crafted object named _DSD with a specific UUID for this format. There are three defined Device Property types: Integers, Strings, and References. It is also possible to have arrays of these properties under one key=value pair. Strings and References are both represented as character arrays but result in different generated ACPI OpCodes. Various helpers are provided for writing the Device Property header (to fill in the object name and UUID) and footer (to fill in the property count and device length values) as well as for writing the different Device Property types. A specific helper is provided for writing the defined GPIO binding Device Property that is used to allow GPIOs to be referred to by name rather than resource index. This is all documented in the _DSD Device Properties UUID document: http://uefi.org/sites/default/files/resources/_DSD-device-properties-UUID.pdf This will be used by device drivers to provide device properties that are consumed by the operating system. Devicetree bindings are often described in the linux kernel at Documentation/devicetree/bindings/ A sample driver here has an input GPIO that it needs to describe to the kernel driver: chip.h: struct drivers_generic_sample_config { struct acpi_gpio mode_gpio; }; sample.c: static void acpi_fill_ssdt_generator(struct device *dev) { struct drivers_generic_sample_config *config = dev->chip_info; const char *path = acpi_device_path(dev); ... acpi_device_write_gpio(&config->mode_gpio); ... acpi_dp_write_header(); acpi_dp_write_gpio("mode-gpio", path, 0, 0, 0); acpi_dp_write_footer(); ... } devicetree.cb: device pci 1f.0 on chip drivers/generic/sample register "mode_gpio" = "ACPI_GPIO_INPUT(GPP_B1)" device generic 0 on end end end SSDT.dsl: Name (_CRS, ResourceTemplate () { GpioIo (Exclusive, PullDefault, 0, 0, IoRestrictionInputOnly, "\\_SB.PCI0.GPIO", 0, ResourceConsumer) { 25 } }) Name (_DSD, Package () { ToUUID ("daffd814-6eba-4d8c-8a91-bc9bbf4aa301"), Package () { Package () {"mode-gpio", Package () { \_SB.PCI0.LPCB, 0, 0, 1 }} } }) Change-Id: I93ffd09e59d05c09e38693e221a87085469be3ad Signed-off-by: Duncan Laurie <dlaurie@chromium.org> Reviewed-on: https://review.coreboot.org/14937 Tested-by: build bot (Jenkins) Reviewed-by: Aaron Durbin <adurbin@chromium.org>
2016-05-10 22:18:17 +02:00
}
acpi: Change device properties to work as a tree There is a second ACPI _DSD document from the UEFI Forum that details how _DSD style tables can be nested, creating a tree of similarly formatted tables. This document is linked from acpi_device.h. In order to support this the device property interface needs to be more flexible and build up a tree of properties to write all entries at once instead of writing each entry as it is generated. In the end this is a more flexible solution that can support drivers that need child tables like the DA7219 codec, while only requiring minor changes to the existing drivers that use the device property interface. This was tested on reef (apollolake) and chell (skylake) boards to ensure that there was no change in the generated SSDT AML. Change-Id: Ia22e3a5fd3982ffa7c324bee1a8d190d49f853dd Signed-off-by: Duncan Laurie <dlaurie@chromium.org> Reviewed-on: https://review.coreboot.org/15537 Tested-by: build bot (Jenkins) Reviewed-by: Aaron Durbin <adurbin@chromium.org>
2016-07-03 04:56:06 +02:00
static void acpi_dp_free(struct acpi_dp *dp)
acpi_device: Add support for writing ACPI Device Properties The recent ACPI specification extensions have formally defined a method for describing device information with a key=value format that is modeled after the Devicetree/DTS format using a special crafted object named _DSD with a specific UUID for this format. There are three defined Device Property types: Integers, Strings, and References. It is also possible to have arrays of these properties under one key=value pair. Strings and References are both represented as character arrays but result in different generated ACPI OpCodes. Various helpers are provided for writing the Device Property header (to fill in the object name and UUID) and footer (to fill in the property count and device length values) as well as for writing the different Device Property types. A specific helper is provided for writing the defined GPIO binding Device Property that is used to allow GPIOs to be referred to by name rather than resource index. This is all documented in the _DSD Device Properties UUID document: http://uefi.org/sites/default/files/resources/_DSD-device-properties-UUID.pdf This will be used by device drivers to provide device properties that are consumed by the operating system. Devicetree bindings are often described in the linux kernel at Documentation/devicetree/bindings/ A sample driver here has an input GPIO that it needs to describe to the kernel driver: chip.h: struct drivers_generic_sample_config { struct acpi_gpio mode_gpio; }; sample.c: static void acpi_fill_ssdt_generator(struct device *dev) { struct drivers_generic_sample_config *config = dev->chip_info; const char *path = acpi_device_path(dev); ... acpi_device_write_gpio(&config->mode_gpio); ... acpi_dp_write_header(); acpi_dp_write_gpio("mode-gpio", path, 0, 0, 0); acpi_dp_write_footer(); ... } devicetree.cb: device pci 1f.0 on chip drivers/generic/sample register "mode_gpio" = "ACPI_GPIO_INPUT(GPP_B1)" device generic 0 on end end end SSDT.dsl: Name (_CRS, ResourceTemplate () { GpioIo (Exclusive, PullDefault, 0, 0, IoRestrictionInputOnly, "\\_SB.PCI0.GPIO", 0, ResourceConsumer) { 25 } }) Name (_DSD, Package () { ToUUID ("daffd814-6eba-4d8c-8a91-bc9bbf4aa301"), Package () { Package () {"mode-gpio", Package () { \_SB.PCI0.LPCB, 0, 0, 1 }} } }) Change-Id: I93ffd09e59d05c09e38693e221a87085469be3ad Signed-off-by: Duncan Laurie <dlaurie@chromium.org> Reviewed-on: https://review.coreboot.org/14937 Tested-by: build bot (Jenkins) Reviewed-by: Aaron Durbin <adurbin@chromium.org>
2016-05-10 22:18:17 +02:00
{
acpi: Change device properties to work as a tree There is a second ACPI _DSD document from the UEFI Forum that details how _DSD style tables can be nested, creating a tree of similarly formatted tables. This document is linked from acpi_device.h. In order to support this the device property interface needs to be more flexible and build up a tree of properties to write all entries at once instead of writing each entry as it is generated. In the end this is a more flexible solution that can support drivers that need child tables like the DA7219 codec, while only requiring minor changes to the existing drivers that use the device property interface. This was tested on reef (apollolake) and chell (skylake) boards to ensure that there was no change in the generated SSDT AML. Change-Id: Ia22e3a5fd3982ffa7c324bee1a8d190d49f853dd Signed-off-by: Duncan Laurie <dlaurie@chromium.org> Reviewed-on: https://review.coreboot.org/15537 Tested-by: build bot (Jenkins) Reviewed-by: Aaron Durbin <adurbin@chromium.org>
2016-07-03 04:56:06 +02:00
while (dp) {
struct acpi_dp *p = dp->next;
switch (dp->type) {
case ACPI_DP_TYPE_CHILD:
acpi_dp_free(dp->child);
break;
case ACPI_DP_TYPE_ARRAY:
acpi_dp_free(dp->array);
break;
default:
break;
}
free(dp);
dp = p;
}
acpi_device: Add support for writing ACPI Device Properties The recent ACPI specification extensions have formally defined a method for describing device information with a key=value format that is modeled after the Devicetree/DTS format using a special crafted object named _DSD with a specific UUID for this format. There are three defined Device Property types: Integers, Strings, and References. It is also possible to have arrays of these properties under one key=value pair. Strings and References are both represented as character arrays but result in different generated ACPI OpCodes. Various helpers are provided for writing the Device Property header (to fill in the object name and UUID) and footer (to fill in the property count and device length values) as well as for writing the different Device Property types. A specific helper is provided for writing the defined GPIO binding Device Property that is used to allow GPIOs to be referred to by name rather than resource index. This is all documented in the _DSD Device Properties UUID document: http://uefi.org/sites/default/files/resources/_DSD-device-properties-UUID.pdf This will be used by device drivers to provide device properties that are consumed by the operating system. Devicetree bindings are often described in the linux kernel at Documentation/devicetree/bindings/ A sample driver here has an input GPIO that it needs to describe to the kernel driver: chip.h: struct drivers_generic_sample_config { struct acpi_gpio mode_gpio; }; sample.c: static void acpi_fill_ssdt_generator(struct device *dev) { struct drivers_generic_sample_config *config = dev->chip_info; const char *path = acpi_device_path(dev); ... acpi_device_write_gpio(&config->mode_gpio); ... acpi_dp_write_header(); acpi_dp_write_gpio("mode-gpio", path, 0, 0, 0); acpi_dp_write_footer(); ... } devicetree.cb: device pci 1f.0 on chip drivers/generic/sample register "mode_gpio" = "ACPI_GPIO_INPUT(GPP_B1)" device generic 0 on end end end SSDT.dsl: Name (_CRS, ResourceTemplate () { GpioIo (Exclusive, PullDefault, 0, 0, IoRestrictionInputOnly, "\\_SB.PCI0.GPIO", 0, ResourceConsumer) { 25 } }) Name (_DSD, Package () { ToUUID ("daffd814-6eba-4d8c-8a91-bc9bbf4aa301"), Package () { Package () {"mode-gpio", Package () { \_SB.PCI0.LPCB, 0, 0, 1 }} } }) Change-Id: I93ffd09e59d05c09e38693e221a87085469be3ad Signed-off-by: Duncan Laurie <dlaurie@chromium.org> Reviewed-on: https://review.coreboot.org/14937 Tested-by: build bot (Jenkins) Reviewed-by: Aaron Durbin <adurbin@chromium.org>
2016-05-10 22:18:17 +02:00
}
acpi: Change device properties to work as a tree There is a second ACPI _DSD document from the UEFI Forum that details how _DSD style tables can be nested, creating a tree of similarly formatted tables. This document is linked from acpi_device.h. In order to support this the device property interface needs to be more flexible and build up a tree of properties to write all entries at once instead of writing each entry as it is generated. In the end this is a more flexible solution that can support drivers that need child tables like the DA7219 codec, while only requiring minor changes to the existing drivers that use the device property interface. This was tested on reef (apollolake) and chell (skylake) boards to ensure that there was no change in the generated SSDT AML. Change-Id: Ia22e3a5fd3982ffa7c324bee1a8d190d49f853dd Signed-off-by: Duncan Laurie <dlaurie@chromium.org> Reviewed-on: https://review.coreboot.org/15537 Tested-by: build bot (Jenkins) Reviewed-by: Aaron Durbin <adurbin@chromium.org>
2016-07-03 04:56:06 +02:00
void acpi_dp_write(struct acpi_dp *table)
acpi_device: Add support for writing ACPI Device Properties The recent ACPI specification extensions have formally defined a method for describing device information with a key=value format that is modeled after the Devicetree/DTS format using a special crafted object named _DSD with a specific UUID for this format. There are three defined Device Property types: Integers, Strings, and References. It is also possible to have arrays of these properties under one key=value pair. Strings and References are both represented as character arrays but result in different generated ACPI OpCodes. Various helpers are provided for writing the Device Property header (to fill in the object name and UUID) and footer (to fill in the property count and device length values) as well as for writing the different Device Property types. A specific helper is provided for writing the defined GPIO binding Device Property that is used to allow GPIOs to be referred to by name rather than resource index. This is all documented in the _DSD Device Properties UUID document: http://uefi.org/sites/default/files/resources/_DSD-device-properties-UUID.pdf This will be used by device drivers to provide device properties that are consumed by the operating system. Devicetree bindings are often described in the linux kernel at Documentation/devicetree/bindings/ A sample driver here has an input GPIO that it needs to describe to the kernel driver: chip.h: struct drivers_generic_sample_config { struct acpi_gpio mode_gpio; }; sample.c: static void acpi_fill_ssdt_generator(struct device *dev) { struct drivers_generic_sample_config *config = dev->chip_info; const char *path = acpi_device_path(dev); ... acpi_device_write_gpio(&config->mode_gpio); ... acpi_dp_write_header(); acpi_dp_write_gpio("mode-gpio", path, 0, 0, 0); acpi_dp_write_footer(); ... } devicetree.cb: device pci 1f.0 on chip drivers/generic/sample register "mode_gpio" = "ACPI_GPIO_INPUT(GPP_B1)" device generic 0 on end end end SSDT.dsl: Name (_CRS, ResourceTemplate () { GpioIo (Exclusive, PullDefault, 0, 0, IoRestrictionInputOnly, "\\_SB.PCI0.GPIO", 0, ResourceConsumer) { 25 } }) Name (_DSD, Package () { ToUUID ("daffd814-6eba-4d8c-8a91-bc9bbf4aa301"), Package () { Package () {"mode-gpio", Package () { \_SB.PCI0.LPCB, 0, 0, 1 }} } }) Change-Id: I93ffd09e59d05c09e38693e221a87085469be3ad Signed-off-by: Duncan Laurie <dlaurie@chromium.org> Reviewed-on: https://review.coreboot.org/14937 Tested-by: build bot (Jenkins) Reviewed-by: Aaron Durbin <adurbin@chromium.org>
2016-05-10 22:18:17 +02:00
{
acpi: Change device properties to work as a tree There is a second ACPI _DSD document from the UEFI Forum that details how _DSD style tables can be nested, creating a tree of similarly formatted tables. This document is linked from acpi_device.h. In order to support this the device property interface needs to be more flexible and build up a tree of properties to write all entries at once instead of writing each entry as it is generated. In the end this is a more flexible solution that can support drivers that need child tables like the DA7219 codec, while only requiring minor changes to the existing drivers that use the device property interface. This was tested on reef (apollolake) and chell (skylake) boards to ensure that there was no change in the generated SSDT AML. Change-Id: Ia22e3a5fd3982ffa7c324bee1a8d190d49f853dd Signed-off-by: Duncan Laurie <dlaurie@chromium.org> Reviewed-on: https://review.coreboot.org/15537 Tested-by: build bot (Jenkins) Reviewed-by: Aaron Durbin <adurbin@chromium.org>
2016-07-03 04:56:06 +02:00
struct acpi_dp *dp, *prop;
acpi: device: avoid empty property list in acpi_dp_write If an acpi_dp table has children but no properties then acpi_dp_write() will write out a properties UUID and package that contains no properties. The existing function will avoid writing out a UUID and empty package when no children exist, but it seems to assume that properties will always be used. With this change properties are handled in a manner akin to children so that a UUID and package are only written if properties exist. BUG=none BRANCH=none TEST=Confirmed that prior to this change a UUID and empty package was present for a device that had children but no properties. Verified that after this change the UUID and empty package are no longer present but the child UUID and package are still present. Change-Id: I6f5597713a1e91ca26b409f36b3ff9eb90a010af Signed-off-by: Matt Delco <delco@chromium.org> Reviewed-on: https://review.coreboot.org/c/31161 Tested-by: build bot (Jenkins) <no-reply@coreboot.org> Reviewed-by: Duncan Laurie <dlaurie@chromium.org> Reviewed-by: Furquan Shaikh <furquan@google.com>
2019-01-30 20:16:08 +01:00
char *dp_count, *prop_count = NULL;
acpi: Change device properties to work as a tree There is a second ACPI _DSD document from the UEFI Forum that details how _DSD style tables can be nested, creating a tree of similarly formatted tables. This document is linked from acpi_device.h. In order to support this the device property interface needs to be more flexible and build up a tree of properties to write all entries at once instead of writing each entry as it is generated. In the end this is a more flexible solution that can support drivers that need child tables like the DA7219 codec, while only requiring minor changes to the existing drivers that use the device property interface. This was tested on reef (apollolake) and chell (skylake) boards to ensure that there was no change in the generated SSDT AML. Change-Id: Ia22e3a5fd3982ffa7c324bee1a8d190d49f853dd Signed-off-by: Duncan Laurie <dlaurie@chromium.org> Reviewed-on: https://review.coreboot.org/15537 Tested-by: build bot (Jenkins) Reviewed-by: Aaron Durbin <adurbin@chromium.org>
2016-07-03 04:56:06 +02:00
int child_count = 0;
if (!table || table->type != ACPI_DP_TYPE_TABLE)
return;
/* Name (name) */
acpigen_write_name(table->name);
/* Device Property list starts with the next entry */
prop = table->next;
acpi: device: avoid empty property list in acpi_dp_write If an acpi_dp table has children but no properties then acpi_dp_write() will write out a properties UUID and package that contains no properties. The existing function will avoid writing out a UUID and empty package when no children exist, but it seems to assume that properties will always be used. With this change properties are handled in a manner akin to children so that a UUID and package are only written if properties exist. BUG=none BRANCH=none TEST=Confirmed that prior to this change a UUID and empty package was present for a device that had children but no properties. Verified that after this change the UUID and empty package are no longer present but the child UUID and package are still present. Change-Id: I6f5597713a1e91ca26b409f36b3ff9eb90a010af Signed-off-by: Matt Delco <delco@chromium.org> Reviewed-on: https://review.coreboot.org/c/31161 Tested-by: build bot (Jenkins) <no-reply@coreboot.org> Reviewed-by: Duncan Laurie <dlaurie@chromium.org> Reviewed-by: Furquan Shaikh <furquan@google.com>
2019-01-30 20:16:08 +01:00
/* Package (DP), default to assuming no properties or children */
dp_count = acpigen_write_package(0);
acpi: Change device properties to work as a tree There is a second ACPI _DSD document from the UEFI Forum that details how _DSD style tables can be nested, creating a tree of similarly formatted tables. This document is linked from acpi_device.h. In order to support this the device property interface needs to be more flexible and build up a tree of properties to write all entries at once instead of writing each entry as it is generated. In the end this is a more flexible solution that can support drivers that need child tables like the DA7219 codec, while only requiring minor changes to the existing drivers that use the device property interface. This was tested on reef (apollolake) and chell (skylake) boards to ensure that there was no change in the generated SSDT AML. Change-Id: Ia22e3a5fd3982ffa7c324bee1a8d190d49f853dd Signed-off-by: Duncan Laurie <dlaurie@chromium.org> Reviewed-on: https://review.coreboot.org/15537 Tested-by: build bot (Jenkins) Reviewed-by: Aaron Durbin <adurbin@chromium.org>
2016-07-03 04:56:06 +02:00
/* Print base properties */
for (dp = prop; dp; dp = dp->next) {
if (dp->type == ACPI_DP_TYPE_CHILD) {
child_count++;
} else {
acpi: device: avoid empty property list in acpi_dp_write If an acpi_dp table has children but no properties then acpi_dp_write() will write out a properties UUID and package that contains no properties. The existing function will avoid writing out a UUID and empty package when no children exist, but it seems to assume that properties will always be used. With this change properties are handled in a manner akin to children so that a UUID and package are only written if properties exist. BUG=none BRANCH=none TEST=Confirmed that prior to this change a UUID and empty package was present for a device that had children but no properties. Verified that after this change the UUID and empty package are no longer present but the child UUID and package are still present. Change-Id: I6f5597713a1e91ca26b409f36b3ff9eb90a010af Signed-off-by: Matt Delco <delco@chromium.org> Reviewed-on: https://review.coreboot.org/c/31161 Tested-by: build bot (Jenkins) <no-reply@coreboot.org> Reviewed-by: Duncan Laurie <dlaurie@chromium.org> Reviewed-by: Furquan Shaikh <furquan@google.com>
2019-01-30 20:16:08 +01:00
/*
* The UUID and package is only added when
* we come across the first property. This
* is to avoid creating a zero-length package
* in situations where there are only children.
*/
if (!prop_count) {
*dp_count += 2;
/* ToUUID (ACPI_DP_UUID) */
acpigen_write_uuid(ACPI_DP_UUID);
/*
* Package (PROP), element count determined as
* it is populated
*/
prop_count = acpigen_write_package(0);
}
acpi: Change device properties to work as a tree There is a second ACPI _DSD document from the UEFI Forum that details how _DSD style tables can be nested, creating a tree of similarly formatted tables. This document is linked from acpi_device.h. In order to support this the device property interface needs to be more flexible and build up a tree of properties to write all entries at once instead of writing each entry as it is generated. In the end this is a more flexible solution that can support drivers that need child tables like the DA7219 codec, while only requiring minor changes to the existing drivers that use the device property interface. This was tested on reef (apollolake) and chell (skylake) boards to ensure that there was no change in the generated SSDT AML. Change-Id: Ia22e3a5fd3982ffa7c324bee1a8d190d49f853dd Signed-off-by: Duncan Laurie <dlaurie@chromium.org> Reviewed-on: https://review.coreboot.org/15537 Tested-by: build bot (Jenkins) Reviewed-by: Aaron Durbin <adurbin@chromium.org>
2016-07-03 04:56:06 +02:00
(*prop_count)++;
acpi_dp_write_property(dp);
}
}
acpi: device: avoid empty property list in acpi_dp_write If an acpi_dp table has children but no properties then acpi_dp_write() will write out a properties UUID and package that contains no properties. The existing function will avoid writing out a UUID and empty package when no children exist, but it seems to assume that properties will always be used. With this change properties are handled in a manner akin to children so that a UUID and package are only written if properties exist. BUG=none BRANCH=none TEST=Confirmed that prior to this change a UUID and empty package was present for a device that had children but no properties. Verified that after this change the UUID and empty package are no longer present but the child UUID and package are still present. Change-Id: I6f5597713a1e91ca26b409f36b3ff9eb90a010af Signed-off-by: Matt Delco <delco@chromium.org> Reviewed-on: https://review.coreboot.org/c/31161 Tested-by: build bot (Jenkins) <no-reply@coreboot.org> Reviewed-by: Duncan Laurie <dlaurie@chromium.org> Reviewed-by: Furquan Shaikh <furquan@google.com>
2019-01-30 20:16:08 +01:00
if (prop_count) {
/* Package (PROP) length, if a package was written */
acpigen_pop_len();
}
acpi: Change device properties to work as a tree There is a second ACPI _DSD document from the UEFI Forum that details how _DSD style tables can be nested, creating a tree of similarly formatted tables. This document is linked from acpi_device.h. In order to support this the device property interface needs to be more flexible and build up a tree of properties to write all entries at once instead of writing each entry as it is generated. In the end this is a more flexible solution that can support drivers that need child tables like the DA7219 codec, while only requiring minor changes to the existing drivers that use the device property interface. This was tested on reef (apollolake) and chell (skylake) boards to ensure that there was no change in the generated SSDT AML. Change-Id: Ia22e3a5fd3982ffa7c324bee1a8d190d49f853dd Signed-off-by: Duncan Laurie <dlaurie@chromium.org> Reviewed-on: https://review.coreboot.org/15537 Tested-by: build bot (Jenkins) Reviewed-by: Aaron Durbin <adurbin@chromium.org>
2016-07-03 04:56:06 +02:00
if (child_count) {
acpi: device: avoid empty property list in acpi_dp_write If an acpi_dp table has children but no properties then acpi_dp_write() will write out a properties UUID and package that contains no properties. The existing function will avoid writing out a UUID and empty package when no children exist, but it seems to assume that properties will always be used. With this change properties are handled in a manner akin to children so that a UUID and package are only written if properties exist. BUG=none BRANCH=none TEST=Confirmed that prior to this change a UUID and empty package was present for a device that had children but no properties. Verified that after this change the UUID and empty package are no longer present but the child UUID and package are still present. Change-Id: I6f5597713a1e91ca26b409f36b3ff9eb90a010af Signed-off-by: Matt Delco <delco@chromium.org> Reviewed-on: https://review.coreboot.org/c/31161 Tested-by: build bot (Jenkins) <no-reply@coreboot.org> Reviewed-by: Duncan Laurie <dlaurie@chromium.org> Reviewed-by: Furquan Shaikh <furquan@google.com>
2019-01-30 20:16:08 +01:00
/* Update DP package count to 2 or 4 */
*dp_count += 2;
acpi: Change device properties to work as a tree There is a second ACPI _DSD document from the UEFI Forum that details how _DSD style tables can be nested, creating a tree of similarly formatted tables. This document is linked from acpi_device.h. In order to support this the device property interface needs to be more flexible and build up a tree of properties to write all entries at once instead of writing each entry as it is generated. In the end this is a more flexible solution that can support drivers that need child tables like the DA7219 codec, while only requiring minor changes to the existing drivers that use the device property interface. This was tested on reef (apollolake) and chell (skylake) boards to ensure that there was no change in the generated SSDT AML. Change-Id: Ia22e3a5fd3982ffa7c324bee1a8d190d49f853dd Signed-off-by: Duncan Laurie <dlaurie@chromium.org> Reviewed-on: https://review.coreboot.org/15537 Tested-by: build bot (Jenkins) Reviewed-by: Aaron Durbin <adurbin@chromium.org>
2016-07-03 04:56:06 +02:00
/* ToUUID (ACPI_DP_CHILD_UUID) */
acpigen_write_uuid(ACPI_DP_CHILD_UUID);
/* Print child pointer properties */
acpigen_write_package(child_count);
for (dp = prop; dp; dp = dp->next)
if (dp->type == ACPI_DP_TYPE_CHILD)
acpi_dp_write_property(dp);
acpi: device: avoid empty property list in acpi_dp_write If an acpi_dp table has children but no properties then acpi_dp_write() will write out a properties UUID and package that contains no properties. The existing function will avoid writing out a UUID and empty package when no children exist, but it seems to assume that properties will always be used. With this change properties are handled in a manner akin to children so that a UUID and package are only written if properties exist. BUG=none BRANCH=none TEST=Confirmed that prior to this change a UUID and empty package was present for a device that had children but no properties. Verified that after this change the UUID and empty package are no longer present but the child UUID and package are still present. Change-Id: I6f5597713a1e91ca26b409f36b3ff9eb90a010af Signed-off-by: Matt Delco <delco@chromium.org> Reviewed-on: https://review.coreboot.org/c/31161 Tested-by: build bot (Jenkins) <no-reply@coreboot.org> Reviewed-by: Duncan Laurie <dlaurie@chromium.org> Reviewed-by: Furquan Shaikh <furquan@google.com>
2019-01-30 20:16:08 +01:00
/* Package (CHILD) length */
acpi: Change device properties to work as a tree There is a second ACPI _DSD document from the UEFI Forum that details how _DSD style tables can be nested, creating a tree of similarly formatted tables. This document is linked from acpi_device.h. In order to support this the device property interface needs to be more flexible and build up a tree of properties to write all entries at once instead of writing each entry as it is generated. In the end this is a more flexible solution that can support drivers that need child tables like the DA7219 codec, while only requiring minor changes to the existing drivers that use the device property interface. This was tested on reef (apollolake) and chell (skylake) boards to ensure that there was no change in the generated SSDT AML. Change-Id: Ia22e3a5fd3982ffa7c324bee1a8d190d49f853dd Signed-off-by: Duncan Laurie <dlaurie@chromium.org> Reviewed-on: https://review.coreboot.org/15537 Tested-by: build bot (Jenkins) Reviewed-by: Aaron Durbin <adurbin@chromium.org>
2016-07-03 04:56:06 +02:00
acpigen_pop_len();
}
/* Package (DP) length */
acpi_device: Add support for writing ACPI Device Properties The recent ACPI specification extensions have formally defined a method for describing device information with a key=value format that is modeled after the Devicetree/DTS format using a special crafted object named _DSD with a specific UUID for this format. There are three defined Device Property types: Integers, Strings, and References. It is also possible to have arrays of these properties under one key=value pair. Strings and References are both represented as character arrays but result in different generated ACPI OpCodes. Various helpers are provided for writing the Device Property header (to fill in the object name and UUID) and footer (to fill in the property count and device length values) as well as for writing the different Device Property types. A specific helper is provided for writing the defined GPIO binding Device Property that is used to allow GPIOs to be referred to by name rather than resource index. This is all documented in the _DSD Device Properties UUID document: http://uefi.org/sites/default/files/resources/_DSD-device-properties-UUID.pdf This will be used by device drivers to provide device properties that are consumed by the operating system. Devicetree bindings are often described in the linux kernel at Documentation/devicetree/bindings/ A sample driver here has an input GPIO that it needs to describe to the kernel driver: chip.h: struct drivers_generic_sample_config { struct acpi_gpio mode_gpio; }; sample.c: static void acpi_fill_ssdt_generator(struct device *dev) { struct drivers_generic_sample_config *config = dev->chip_info; const char *path = acpi_device_path(dev); ... acpi_device_write_gpio(&config->mode_gpio); ... acpi_dp_write_header(); acpi_dp_write_gpio("mode-gpio", path, 0, 0, 0); acpi_dp_write_footer(); ... } devicetree.cb: device pci 1f.0 on chip drivers/generic/sample register "mode_gpio" = "ACPI_GPIO_INPUT(GPP_B1)" device generic 0 on end end end SSDT.dsl: Name (_CRS, ResourceTemplate () { GpioIo (Exclusive, PullDefault, 0, 0, IoRestrictionInputOnly, "\\_SB.PCI0.GPIO", 0, ResourceConsumer) { 25 } }) Name (_DSD, Package () { ToUUID ("daffd814-6eba-4d8c-8a91-bc9bbf4aa301"), Package () { Package () {"mode-gpio", Package () { \_SB.PCI0.LPCB, 0, 0, 1 }} } }) Change-Id: I93ffd09e59d05c09e38693e221a87085469be3ad Signed-off-by: Duncan Laurie <dlaurie@chromium.org> Reviewed-on: https://review.coreboot.org/14937 Tested-by: build bot (Jenkins) Reviewed-by: Aaron Durbin <adurbin@chromium.org>
2016-05-10 22:18:17 +02:00
acpigen_pop_len();
acpi: Change device properties to work as a tree There is a second ACPI _DSD document from the UEFI Forum that details how _DSD style tables can be nested, creating a tree of similarly formatted tables. This document is linked from acpi_device.h. In order to support this the device property interface needs to be more flexible and build up a tree of properties to write all entries at once instead of writing each entry as it is generated. In the end this is a more flexible solution that can support drivers that need child tables like the DA7219 codec, while only requiring minor changes to the existing drivers that use the device property interface. This was tested on reef (apollolake) and chell (skylake) boards to ensure that there was no change in the generated SSDT AML. Change-Id: Ia22e3a5fd3982ffa7c324bee1a8d190d49f853dd Signed-off-by: Duncan Laurie <dlaurie@chromium.org> Reviewed-on: https://review.coreboot.org/15537 Tested-by: build bot (Jenkins) Reviewed-by: Aaron Durbin <adurbin@chromium.org>
2016-07-03 04:56:06 +02:00
/* Recursively parse children into separate tables */
for (dp = prop; dp; dp = dp->next)
if (dp->type == ACPI_DP_TYPE_CHILD)
acpi_dp_write(dp->child);
/* Clean up */
acpi_dp_free(table);
}
static struct acpi_dp *acpi_dp_new(struct acpi_dp *dp, enum acpi_dp_type type,
const char *name)
{
struct acpi_dp *new;
new = malloc(sizeof(struct acpi_dp));
if (!new)
return NULL;
memset(new, 0, sizeof(*new));
new->type = type;
new->name = name;
if (dp) {
/* Add to end of property list */
while (dp->next)
dp = dp->next;
dp->next = new;
}
return new;
}
struct acpi_dp *acpi_dp_new_table(const char *name)
{
return acpi_dp_new(NULL, ACPI_DP_TYPE_TABLE, name);
}
acpi_device: Provide a new function to add a list of properties Provide a new function that will allow adding arbitrary properties to devicetree entries without needing a custom driver for the device. This will allow the 'generic i2c' driver to support kernel drivers that need additional device properties exposed and have those board specific properties set with values from devicetree. BUG=b:63413023 TEST=not used yet, compiles cleanly Change-Id: Id272256639a8525406635e168a3db5ab1ba4df6b Signed-off-by: Duncan Laurie <dlaurie@google.com> Reviewed-on: https://review.coreboot.org/21269 Reviewed-by: Furquan Shaikh <furquan@google.com> Tested-by: build bot (Jenkins) <no-reply@coreboot.org> Reviewed-by: Aaron Durbin <adurbin@chromium.org>
2017-08-29 17:26:50 +02:00
size_t acpi_dp_add_property_list(struct acpi_dp *dp,
const struct acpi_dp *property_list,
size_t property_count)
{
const struct acpi_dp *prop;
size_t i, properties_added = 0;
arch/x86: Do not add properties to null DP packages It doesn't make sense to add a property to a non-existent Device Property package. However, some of these functions will proceed anyway and allocate a new Device Property package, add the property to that, and then immediately leak the new package. This changes all the acpi_dp_add_* functions to ignore a null package. Change-Id: I664dcdbaa6b1b8a3aeb9a0126d622e2ffb736efd Signed-off-by: Jacob Garber <jgarber1@ualberta.ca> Found-by: Coverity CID 135745{6,7}, 138029{2-6} Reviewed-on: https://review.coreboot.org/c/coreboot/+/32971 Tested-by: build bot (Jenkins) <no-reply@coreboot.org> Reviewed-by: Furquan Shaikh <furquan@google.com>
2019-05-23 22:34:58 +02:00
if (!dp || !property_list)
return 0;
acpi_device: Provide a new function to add a list of properties Provide a new function that will allow adding arbitrary properties to devicetree entries without needing a custom driver for the device. This will allow the 'generic i2c' driver to support kernel drivers that need additional device properties exposed and have those board specific properties set with values from devicetree. BUG=b:63413023 TEST=not used yet, compiles cleanly Change-Id: Id272256639a8525406635e168a3db5ab1ba4df6b Signed-off-by: Duncan Laurie <dlaurie@google.com> Reviewed-on: https://review.coreboot.org/21269 Reviewed-by: Furquan Shaikh <furquan@google.com> Tested-by: build bot (Jenkins) <no-reply@coreboot.org> Reviewed-by: Aaron Durbin <adurbin@chromium.org>
2017-08-29 17:26:50 +02:00
for (i = 0; i < property_count; i++) {
prop = &property_list[i];
if (prop->type == ACPI_DP_TYPE_UNKNOWN || !prop->name)
continue;
switch (prop->type) {
case ACPI_DP_TYPE_INTEGER:
acpi_dp_add_integer(dp, prop->name, prop->integer);
break;
case ACPI_DP_TYPE_STRING:
acpi_dp_add_string(dp, prop->name, prop->string);
break;
case ACPI_DP_TYPE_REFERENCE:
acpi_dp_add_reference(dp, prop->name, prop->string);
break;
case ACPI_DP_TYPE_ARRAY:
acpi_dp_add_array(dp, prop->array);
break;
case ACPI_DP_TYPE_CHILD:
acpi_dp_add_child(dp, prop->name, prop->child);
break;
default:
continue;
}
++properties_added;
}
return properties_added;
}
acpi: Change device properties to work as a tree There is a second ACPI _DSD document from the UEFI Forum that details how _DSD style tables can be nested, creating a tree of similarly formatted tables. This document is linked from acpi_device.h. In order to support this the device property interface needs to be more flexible and build up a tree of properties to write all entries at once instead of writing each entry as it is generated. In the end this is a more flexible solution that can support drivers that need child tables like the DA7219 codec, while only requiring minor changes to the existing drivers that use the device property interface. This was tested on reef (apollolake) and chell (skylake) boards to ensure that there was no change in the generated SSDT AML. Change-Id: Ia22e3a5fd3982ffa7c324bee1a8d190d49f853dd Signed-off-by: Duncan Laurie <dlaurie@chromium.org> Reviewed-on: https://review.coreboot.org/15537 Tested-by: build bot (Jenkins) Reviewed-by: Aaron Durbin <adurbin@chromium.org>
2016-07-03 04:56:06 +02:00
struct acpi_dp *acpi_dp_add_integer(struct acpi_dp *dp, const char *name,
uint64_t value)
{
arch/x86: Do not add properties to null DP packages It doesn't make sense to add a property to a non-existent Device Property package. However, some of these functions will proceed anyway and allocate a new Device Property package, add the property to that, and then immediately leak the new package. This changes all the acpi_dp_add_* functions to ignore a null package. Change-Id: I664dcdbaa6b1b8a3aeb9a0126d622e2ffb736efd Signed-off-by: Jacob Garber <jgarber1@ualberta.ca> Found-by: Coverity CID 135745{6,7}, 138029{2-6} Reviewed-on: https://review.coreboot.org/c/coreboot/+/32971 Tested-by: build bot (Jenkins) <no-reply@coreboot.org> Reviewed-by: Furquan Shaikh <furquan@google.com>
2019-05-23 22:34:58 +02:00
if (!dp)
return NULL;
acpi: Change device properties to work as a tree There is a second ACPI _DSD document from the UEFI Forum that details how _DSD style tables can be nested, creating a tree of similarly formatted tables. This document is linked from acpi_device.h. In order to support this the device property interface needs to be more flexible and build up a tree of properties to write all entries at once instead of writing each entry as it is generated. In the end this is a more flexible solution that can support drivers that need child tables like the DA7219 codec, while only requiring minor changes to the existing drivers that use the device property interface. This was tested on reef (apollolake) and chell (skylake) boards to ensure that there was no change in the generated SSDT AML. Change-Id: Ia22e3a5fd3982ffa7c324bee1a8d190d49f853dd Signed-off-by: Duncan Laurie <dlaurie@chromium.org> Reviewed-on: https://review.coreboot.org/15537 Tested-by: build bot (Jenkins) Reviewed-by: Aaron Durbin <adurbin@chromium.org>
2016-07-03 04:56:06 +02:00
struct acpi_dp *new = acpi_dp_new(dp, ACPI_DP_TYPE_INTEGER, name);
if (new)
new->integer = value;
return new;
}
struct acpi_dp *acpi_dp_add_string(struct acpi_dp *dp, const char *name,
const char *string)
{
arch/x86: Do not add properties to null DP packages It doesn't make sense to add a property to a non-existent Device Property package. However, some of these functions will proceed anyway and allocate a new Device Property package, add the property to that, and then immediately leak the new package. This changes all the acpi_dp_add_* functions to ignore a null package. Change-Id: I664dcdbaa6b1b8a3aeb9a0126d622e2ffb736efd Signed-off-by: Jacob Garber <jgarber1@ualberta.ca> Found-by: Coverity CID 135745{6,7}, 138029{2-6} Reviewed-on: https://review.coreboot.org/c/coreboot/+/32971 Tested-by: build bot (Jenkins) <no-reply@coreboot.org> Reviewed-by: Furquan Shaikh <furquan@google.com>
2019-05-23 22:34:58 +02:00
if (!dp)
return NULL;
acpi: Change device properties to work as a tree There is a second ACPI _DSD document from the UEFI Forum that details how _DSD style tables can be nested, creating a tree of similarly formatted tables. This document is linked from acpi_device.h. In order to support this the device property interface needs to be more flexible and build up a tree of properties to write all entries at once instead of writing each entry as it is generated. In the end this is a more flexible solution that can support drivers that need child tables like the DA7219 codec, while only requiring minor changes to the existing drivers that use the device property interface. This was tested on reef (apollolake) and chell (skylake) boards to ensure that there was no change in the generated SSDT AML. Change-Id: Ia22e3a5fd3982ffa7c324bee1a8d190d49f853dd Signed-off-by: Duncan Laurie <dlaurie@chromium.org> Reviewed-on: https://review.coreboot.org/15537 Tested-by: build bot (Jenkins) Reviewed-by: Aaron Durbin <adurbin@chromium.org>
2016-07-03 04:56:06 +02:00
struct acpi_dp *new = acpi_dp_new(dp, ACPI_DP_TYPE_STRING, name);
if (new)
new->string = string;
return new;
}
struct acpi_dp *acpi_dp_add_reference(struct acpi_dp *dp, const char *name,
const char *reference)
{
arch/x86: Do not add properties to null DP packages It doesn't make sense to add a property to a non-existent Device Property package. However, some of these functions will proceed anyway and allocate a new Device Property package, add the property to that, and then immediately leak the new package. This changes all the acpi_dp_add_* functions to ignore a null package. Change-Id: I664dcdbaa6b1b8a3aeb9a0126d622e2ffb736efd Signed-off-by: Jacob Garber <jgarber1@ualberta.ca> Found-by: Coverity CID 135745{6,7}, 138029{2-6} Reviewed-on: https://review.coreboot.org/c/coreboot/+/32971 Tested-by: build bot (Jenkins) <no-reply@coreboot.org> Reviewed-by: Furquan Shaikh <furquan@google.com>
2019-05-23 22:34:58 +02:00
if (!dp)
return NULL;
acpi: Change device properties to work as a tree There is a second ACPI _DSD document from the UEFI Forum that details how _DSD style tables can be nested, creating a tree of similarly formatted tables. This document is linked from acpi_device.h. In order to support this the device property interface needs to be more flexible and build up a tree of properties to write all entries at once instead of writing each entry as it is generated. In the end this is a more flexible solution that can support drivers that need child tables like the DA7219 codec, while only requiring minor changes to the existing drivers that use the device property interface. This was tested on reef (apollolake) and chell (skylake) boards to ensure that there was no change in the generated SSDT AML. Change-Id: Ia22e3a5fd3982ffa7c324bee1a8d190d49f853dd Signed-off-by: Duncan Laurie <dlaurie@chromium.org> Reviewed-on: https://review.coreboot.org/15537 Tested-by: build bot (Jenkins) Reviewed-by: Aaron Durbin <adurbin@chromium.org>
2016-07-03 04:56:06 +02:00
struct acpi_dp *new = acpi_dp_new(dp, ACPI_DP_TYPE_REFERENCE, name);
if (new)
new->string = reference;
return new;
acpi_device: Add support for writing ACPI Device Properties The recent ACPI specification extensions have formally defined a method for describing device information with a key=value format that is modeled after the Devicetree/DTS format using a special crafted object named _DSD with a specific UUID for this format. There are three defined Device Property types: Integers, Strings, and References. It is also possible to have arrays of these properties under one key=value pair. Strings and References are both represented as character arrays but result in different generated ACPI OpCodes. Various helpers are provided for writing the Device Property header (to fill in the object name and UUID) and footer (to fill in the property count and device length values) as well as for writing the different Device Property types. A specific helper is provided for writing the defined GPIO binding Device Property that is used to allow GPIOs to be referred to by name rather than resource index. This is all documented in the _DSD Device Properties UUID document: http://uefi.org/sites/default/files/resources/_DSD-device-properties-UUID.pdf This will be used by device drivers to provide device properties that are consumed by the operating system. Devicetree bindings are often described in the linux kernel at Documentation/devicetree/bindings/ A sample driver here has an input GPIO that it needs to describe to the kernel driver: chip.h: struct drivers_generic_sample_config { struct acpi_gpio mode_gpio; }; sample.c: static void acpi_fill_ssdt_generator(struct device *dev) { struct drivers_generic_sample_config *config = dev->chip_info; const char *path = acpi_device_path(dev); ... acpi_device_write_gpio(&config->mode_gpio); ... acpi_dp_write_header(); acpi_dp_write_gpio("mode-gpio", path, 0, 0, 0); acpi_dp_write_footer(); ... } devicetree.cb: device pci 1f.0 on chip drivers/generic/sample register "mode_gpio" = "ACPI_GPIO_INPUT(GPP_B1)" device generic 0 on end end end SSDT.dsl: Name (_CRS, ResourceTemplate () { GpioIo (Exclusive, PullDefault, 0, 0, IoRestrictionInputOnly, "\\_SB.PCI0.GPIO", 0, ResourceConsumer) { 25 } }) Name (_DSD, Package () { ToUUID ("daffd814-6eba-4d8c-8a91-bc9bbf4aa301"), Package () { Package () {"mode-gpio", Package () { \_SB.PCI0.LPCB, 0, 0, 1 }} } }) Change-Id: I93ffd09e59d05c09e38693e221a87085469be3ad Signed-off-by: Duncan Laurie <dlaurie@chromium.org> Reviewed-on: https://review.coreboot.org/14937 Tested-by: build bot (Jenkins) Reviewed-by: Aaron Durbin <adurbin@chromium.org>
2016-05-10 22:18:17 +02:00
}
acpi: Change device properties to work as a tree There is a second ACPI _DSD document from the UEFI Forum that details how _DSD style tables can be nested, creating a tree of similarly formatted tables. This document is linked from acpi_device.h. In order to support this the device property interface needs to be more flexible and build up a tree of properties to write all entries at once instead of writing each entry as it is generated. In the end this is a more flexible solution that can support drivers that need child tables like the DA7219 codec, while only requiring minor changes to the existing drivers that use the device property interface. This was tested on reef (apollolake) and chell (skylake) boards to ensure that there was no change in the generated SSDT AML. Change-Id: Ia22e3a5fd3982ffa7c324bee1a8d190d49f853dd Signed-off-by: Duncan Laurie <dlaurie@chromium.org> Reviewed-on: https://review.coreboot.org/15537 Tested-by: build bot (Jenkins) Reviewed-by: Aaron Durbin <adurbin@chromium.org>
2016-07-03 04:56:06 +02:00
struct acpi_dp *acpi_dp_add_child(struct acpi_dp *dp, const char *name,
struct acpi_dp *child)
acpi_device: Add support for writing ACPI Device Properties The recent ACPI specification extensions have formally defined a method for describing device information with a key=value format that is modeled after the Devicetree/DTS format using a special crafted object named _DSD with a specific UUID for this format. There are three defined Device Property types: Integers, Strings, and References. It is also possible to have arrays of these properties under one key=value pair. Strings and References are both represented as character arrays but result in different generated ACPI OpCodes. Various helpers are provided for writing the Device Property header (to fill in the object name and UUID) and footer (to fill in the property count and device length values) as well as for writing the different Device Property types. A specific helper is provided for writing the defined GPIO binding Device Property that is used to allow GPIOs to be referred to by name rather than resource index. This is all documented in the _DSD Device Properties UUID document: http://uefi.org/sites/default/files/resources/_DSD-device-properties-UUID.pdf This will be used by device drivers to provide device properties that are consumed by the operating system. Devicetree bindings are often described in the linux kernel at Documentation/devicetree/bindings/ A sample driver here has an input GPIO that it needs to describe to the kernel driver: chip.h: struct drivers_generic_sample_config { struct acpi_gpio mode_gpio; }; sample.c: static void acpi_fill_ssdt_generator(struct device *dev) { struct drivers_generic_sample_config *config = dev->chip_info; const char *path = acpi_device_path(dev); ... acpi_device_write_gpio(&config->mode_gpio); ... acpi_dp_write_header(); acpi_dp_write_gpio("mode-gpio", path, 0, 0, 0); acpi_dp_write_footer(); ... } devicetree.cb: device pci 1f.0 on chip drivers/generic/sample register "mode_gpio" = "ACPI_GPIO_INPUT(GPP_B1)" device generic 0 on end end end SSDT.dsl: Name (_CRS, ResourceTemplate () { GpioIo (Exclusive, PullDefault, 0, 0, IoRestrictionInputOnly, "\\_SB.PCI0.GPIO", 0, ResourceConsumer) { 25 } }) Name (_DSD, Package () { ToUUID ("daffd814-6eba-4d8c-8a91-bc9bbf4aa301"), Package () { Package () {"mode-gpio", Package () { \_SB.PCI0.LPCB, 0, 0, 1 }} } }) Change-Id: I93ffd09e59d05c09e38693e221a87085469be3ad Signed-off-by: Duncan Laurie <dlaurie@chromium.org> Reviewed-on: https://review.coreboot.org/14937 Tested-by: build bot (Jenkins) Reviewed-by: Aaron Durbin <adurbin@chromium.org>
2016-05-10 22:18:17 +02:00
{
acpi: Change device properties to work as a tree There is a second ACPI _DSD document from the UEFI Forum that details how _DSD style tables can be nested, creating a tree of similarly formatted tables. This document is linked from acpi_device.h. In order to support this the device property interface needs to be more flexible and build up a tree of properties to write all entries at once instead of writing each entry as it is generated. In the end this is a more flexible solution that can support drivers that need child tables like the DA7219 codec, while only requiring minor changes to the existing drivers that use the device property interface. This was tested on reef (apollolake) and chell (skylake) boards to ensure that there was no change in the generated SSDT AML. Change-Id: Ia22e3a5fd3982ffa7c324bee1a8d190d49f853dd Signed-off-by: Duncan Laurie <dlaurie@chromium.org> Reviewed-on: https://review.coreboot.org/15537 Tested-by: build bot (Jenkins) Reviewed-by: Aaron Durbin <adurbin@chromium.org>
2016-07-03 04:56:06 +02:00
struct acpi_dp *new;
arch/x86: Do not add properties to null DP packages It doesn't make sense to add a property to a non-existent Device Property package. However, some of these functions will proceed anyway and allocate a new Device Property package, add the property to that, and then immediately leak the new package. This changes all the acpi_dp_add_* functions to ignore a null package. Change-Id: I664dcdbaa6b1b8a3aeb9a0126d622e2ffb736efd Signed-off-by: Jacob Garber <jgarber1@ualberta.ca> Found-by: Coverity CID 135745{6,7}, 138029{2-6} Reviewed-on: https://review.coreboot.org/c/coreboot/+/32971 Tested-by: build bot (Jenkins) <no-reply@coreboot.org> Reviewed-by: Furquan Shaikh <furquan@google.com>
2019-05-23 22:34:58 +02:00
if (!dp || !child || child->type != ACPI_DP_TYPE_TABLE)
acpi: Change device properties to work as a tree There is a second ACPI _DSD document from the UEFI Forum that details how _DSD style tables can be nested, creating a tree of similarly formatted tables. This document is linked from acpi_device.h. In order to support this the device property interface needs to be more flexible and build up a tree of properties to write all entries at once instead of writing each entry as it is generated. In the end this is a more flexible solution that can support drivers that need child tables like the DA7219 codec, while only requiring minor changes to the existing drivers that use the device property interface. This was tested on reef (apollolake) and chell (skylake) boards to ensure that there was no change in the generated SSDT AML. Change-Id: Ia22e3a5fd3982ffa7c324bee1a8d190d49f853dd Signed-off-by: Duncan Laurie <dlaurie@chromium.org> Reviewed-on: https://review.coreboot.org/15537 Tested-by: build bot (Jenkins) Reviewed-by: Aaron Durbin <adurbin@chromium.org>
2016-07-03 04:56:06 +02:00
return NULL;
new = acpi_dp_new(dp, ACPI_DP_TYPE_CHILD, name);
if (new) {
new->child = child;
new->string = child->name;
}
return new;
}
struct acpi_dp *acpi_dp_add_array(struct acpi_dp *dp, struct acpi_dp *array)
{
struct acpi_dp *new;
arch/x86: Do not add properties to null DP packages It doesn't make sense to add a property to a non-existent Device Property package. However, some of these functions will proceed anyway and allocate a new Device Property package, add the property to that, and then immediately leak the new package. This changes all the acpi_dp_add_* functions to ignore a null package. Change-Id: I664dcdbaa6b1b8a3aeb9a0126d622e2ffb736efd Signed-off-by: Jacob Garber <jgarber1@ualberta.ca> Found-by: Coverity CID 135745{6,7}, 138029{2-6} Reviewed-on: https://review.coreboot.org/c/coreboot/+/32971 Tested-by: build bot (Jenkins) <no-reply@coreboot.org> Reviewed-by: Furquan Shaikh <furquan@google.com>
2019-05-23 22:34:58 +02:00
if (!dp || !array || array->type != ACPI_DP_TYPE_TABLE)
acpi: Change device properties to work as a tree There is a second ACPI _DSD document from the UEFI Forum that details how _DSD style tables can be nested, creating a tree of similarly formatted tables. This document is linked from acpi_device.h. In order to support this the device property interface needs to be more flexible and build up a tree of properties to write all entries at once instead of writing each entry as it is generated. In the end this is a more flexible solution that can support drivers that need child tables like the DA7219 codec, while only requiring minor changes to the existing drivers that use the device property interface. This was tested on reef (apollolake) and chell (skylake) boards to ensure that there was no change in the generated SSDT AML. Change-Id: Ia22e3a5fd3982ffa7c324bee1a8d190d49f853dd Signed-off-by: Duncan Laurie <dlaurie@chromium.org> Reviewed-on: https://review.coreboot.org/15537 Tested-by: build bot (Jenkins) Reviewed-by: Aaron Durbin <adurbin@chromium.org>
2016-07-03 04:56:06 +02:00
return NULL;
new = acpi_dp_new(dp, ACPI_DP_TYPE_ARRAY, array->name);
if (new)
new->array = array;
return new;
}
struct acpi_dp *acpi_dp_add_integer_array(struct acpi_dp *dp, const char *name,
uint64_t *array, int len)
{
struct acpi_dp *dp_array;
acpi_device: Add support for writing ACPI Device Properties The recent ACPI specification extensions have formally defined a method for describing device information with a key=value format that is modeled after the Devicetree/DTS format using a special crafted object named _DSD with a specific UUID for this format. There are three defined Device Property types: Integers, Strings, and References. It is also possible to have arrays of these properties under one key=value pair. Strings and References are both represented as character arrays but result in different generated ACPI OpCodes. Various helpers are provided for writing the Device Property header (to fill in the object name and UUID) and footer (to fill in the property count and device length values) as well as for writing the different Device Property types. A specific helper is provided for writing the defined GPIO binding Device Property that is used to allow GPIOs to be referred to by name rather than resource index. This is all documented in the _DSD Device Properties UUID document: http://uefi.org/sites/default/files/resources/_DSD-device-properties-UUID.pdf This will be used by device drivers to provide device properties that are consumed by the operating system. Devicetree bindings are often described in the linux kernel at Documentation/devicetree/bindings/ A sample driver here has an input GPIO that it needs to describe to the kernel driver: chip.h: struct drivers_generic_sample_config { struct acpi_gpio mode_gpio; }; sample.c: static void acpi_fill_ssdt_generator(struct device *dev) { struct drivers_generic_sample_config *config = dev->chip_info; const char *path = acpi_device_path(dev); ... acpi_device_write_gpio(&config->mode_gpio); ... acpi_dp_write_header(); acpi_dp_write_gpio("mode-gpio", path, 0, 0, 0); acpi_dp_write_footer(); ... } devicetree.cb: device pci 1f.0 on chip drivers/generic/sample register "mode_gpio" = "ACPI_GPIO_INPUT(GPP_B1)" device generic 0 on end end end SSDT.dsl: Name (_CRS, ResourceTemplate () { GpioIo (Exclusive, PullDefault, 0, 0, IoRestrictionInputOnly, "\\_SB.PCI0.GPIO", 0, ResourceConsumer) { 25 } }) Name (_DSD, Package () { ToUUID ("daffd814-6eba-4d8c-8a91-bc9bbf4aa301"), Package () { Package () {"mode-gpio", Package () { \_SB.PCI0.LPCB, 0, 0, 1 }} } }) Change-Id: I93ffd09e59d05c09e38693e221a87085469be3ad Signed-off-by: Duncan Laurie <dlaurie@chromium.org> Reviewed-on: https://review.coreboot.org/14937 Tested-by: build bot (Jenkins) Reviewed-by: Aaron Durbin <adurbin@chromium.org>
2016-05-10 22:18:17 +02:00
int i;
acpi: Change device properties to work as a tree There is a second ACPI _DSD document from the UEFI Forum that details how _DSD style tables can be nested, creating a tree of similarly formatted tables. This document is linked from acpi_device.h. In order to support this the device property interface needs to be more flexible and build up a tree of properties to write all entries at once instead of writing each entry as it is generated. In the end this is a more flexible solution that can support drivers that need child tables like the DA7219 codec, while only requiring minor changes to the existing drivers that use the device property interface. This was tested on reef (apollolake) and chell (skylake) boards to ensure that there was no change in the generated SSDT AML. Change-Id: Ia22e3a5fd3982ffa7c324bee1a8d190d49f853dd Signed-off-by: Duncan Laurie <dlaurie@chromium.org> Reviewed-on: https://review.coreboot.org/15537 Tested-by: build bot (Jenkins) Reviewed-by: Aaron Durbin <adurbin@chromium.org>
2016-07-03 04:56:06 +02:00
arch/x86: Do not add properties to null DP packages It doesn't make sense to add a property to a non-existent Device Property package. However, some of these functions will proceed anyway and allocate a new Device Property package, add the property to that, and then immediately leak the new package. This changes all the acpi_dp_add_* functions to ignore a null package. Change-Id: I664dcdbaa6b1b8a3aeb9a0126d622e2ffb736efd Signed-off-by: Jacob Garber <jgarber1@ualberta.ca> Found-by: Coverity CID 135745{6,7}, 138029{2-6} Reviewed-on: https://review.coreboot.org/c/coreboot/+/32971 Tested-by: build bot (Jenkins) <no-reply@coreboot.org> Reviewed-by: Furquan Shaikh <furquan@google.com>
2019-05-23 22:34:58 +02:00
if (!dp || len <= 0)
acpi: Change device properties to work as a tree There is a second ACPI _DSD document from the UEFI Forum that details how _DSD style tables can be nested, creating a tree of similarly formatted tables. This document is linked from acpi_device.h. In order to support this the device property interface needs to be more flexible and build up a tree of properties to write all entries at once instead of writing each entry as it is generated. In the end this is a more flexible solution that can support drivers that need child tables like the DA7219 codec, while only requiring minor changes to the existing drivers that use the device property interface. This was tested on reef (apollolake) and chell (skylake) boards to ensure that there was no change in the generated SSDT AML. Change-Id: Ia22e3a5fd3982ffa7c324bee1a8d190d49f853dd Signed-off-by: Duncan Laurie <dlaurie@chromium.org> Reviewed-on: https://review.coreboot.org/15537 Tested-by: build bot (Jenkins) Reviewed-by: Aaron Durbin <adurbin@chromium.org>
2016-07-03 04:56:06 +02:00
return NULL;
dp_array = acpi_dp_new_table(name);
if (!dp_array)
return NULL;
acpi_device: Add support for writing ACPI Device Properties The recent ACPI specification extensions have formally defined a method for describing device information with a key=value format that is modeled after the Devicetree/DTS format using a special crafted object named _DSD with a specific UUID for this format. There are three defined Device Property types: Integers, Strings, and References. It is also possible to have arrays of these properties under one key=value pair. Strings and References are both represented as character arrays but result in different generated ACPI OpCodes. Various helpers are provided for writing the Device Property header (to fill in the object name and UUID) and footer (to fill in the property count and device length values) as well as for writing the different Device Property types. A specific helper is provided for writing the defined GPIO binding Device Property that is used to allow GPIOs to be referred to by name rather than resource index. This is all documented in the _DSD Device Properties UUID document: http://uefi.org/sites/default/files/resources/_DSD-device-properties-UUID.pdf This will be used by device drivers to provide device properties that are consumed by the operating system. Devicetree bindings are often described in the linux kernel at Documentation/devicetree/bindings/ A sample driver here has an input GPIO that it needs to describe to the kernel driver: chip.h: struct drivers_generic_sample_config { struct acpi_gpio mode_gpio; }; sample.c: static void acpi_fill_ssdt_generator(struct device *dev) { struct drivers_generic_sample_config *config = dev->chip_info; const char *path = acpi_device_path(dev); ... acpi_device_write_gpio(&config->mode_gpio); ... acpi_dp_write_header(); acpi_dp_write_gpio("mode-gpio", path, 0, 0, 0); acpi_dp_write_footer(); ... } devicetree.cb: device pci 1f.0 on chip drivers/generic/sample register "mode_gpio" = "ACPI_GPIO_INPUT(GPP_B1)" device generic 0 on end end end SSDT.dsl: Name (_CRS, ResourceTemplate () { GpioIo (Exclusive, PullDefault, 0, 0, IoRestrictionInputOnly, "\\_SB.PCI0.GPIO", 0, ResourceConsumer) { 25 } }) Name (_DSD, Package () { ToUUID ("daffd814-6eba-4d8c-8a91-bc9bbf4aa301"), Package () { Package () {"mode-gpio", Package () { \_SB.PCI0.LPCB, 0, 0, 1 }} } }) Change-Id: I93ffd09e59d05c09e38693e221a87085469be3ad Signed-off-by: Duncan Laurie <dlaurie@chromium.org> Reviewed-on: https://review.coreboot.org/14937 Tested-by: build bot (Jenkins) Reviewed-by: Aaron Durbin <adurbin@chromium.org>
2016-05-10 22:18:17 +02:00
for (i = 0; i < len; i++)
acpi: Change device properties to work as a tree There is a second ACPI _DSD document from the UEFI Forum that details how _DSD style tables can be nested, creating a tree of similarly formatted tables. This document is linked from acpi_device.h. In order to support this the device property interface needs to be more flexible and build up a tree of properties to write all entries at once instead of writing each entry as it is generated. In the end this is a more flexible solution that can support drivers that need child tables like the DA7219 codec, while only requiring minor changes to the existing drivers that use the device property interface. This was tested on reef (apollolake) and chell (skylake) boards to ensure that there was no change in the generated SSDT AML. Change-Id: Ia22e3a5fd3982ffa7c324bee1a8d190d49f853dd Signed-off-by: Duncan Laurie <dlaurie@chromium.org> Reviewed-on: https://review.coreboot.org/15537 Tested-by: build bot (Jenkins) Reviewed-by: Aaron Durbin <adurbin@chromium.org>
2016-07-03 04:56:06 +02:00
if (!acpi_dp_add_integer(dp_array, NULL, array[i]))
break;
acpi_dp_add_array(dp, dp_array);
return dp_array;
acpi_device: Add support for writing ACPI Device Properties The recent ACPI specification extensions have formally defined a method for describing device information with a key=value format that is modeled after the Devicetree/DTS format using a special crafted object named _DSD with a specific UUID for this format. There are three defined Device Property types: Integers, Strings, and References. It is also possible to have arrays of these properties under one key=value pair. Strings and References are both represented as character arrays but result in different generated ACPI OpCodes. Various helpers are provided for writing the Device Property header (to fill in the object name and UUID) and footer (to fill in the property count and device length values) as well as for writing the different Device Property types. A specific helper is provided for writing the defined GPIO binding Device Property that is used to allow GPIOs to be referred to by name rather than resource index. This is all documented in the _DSD Device Properties UUID document: http://uefi.org/sites/default/files/resources/_DSD-device-properties-UUID.pdf This will be used by device drivers to provide device properties that are consumed by the operating system. Devicetree bindings are often described in the linux kernel at Documentation/devicetree/bindings/ A sample driver here has an input GPIO that it needs to describe to the kernel driver: chip.h: struct drivers_generic_sample_config { struct acpi_gpio mode_gpio; }; sample.c: static void acpi_fill_ssdt_generator(struct device *dev) { struct drivers_generic_sample_config *config = dev->chip_info; const char *path = acpi_device_path(dev); ... acpi_device_write_gpio(&config->mode_gpio); ... acpi_dp_write_header(); acpi_dp_write_gpio("mode-gpio", path, 0, 0, 0); acpi_dp_write_footer(); ... } devicetree.cb: device pci 1f.0 on chip drivers/generic/sample register "mode_gpio" = "ACPI_GPIO_INPUT(GPP_B1)" device generic 0 on end end end SSDT.dsl: Name (_CRS, ResourceTemplate () { GpioIo (Exclusive, PullDefault, 0, 0, IoRestrictionInputOnly, "\\_SB.PCI0.GPIO", 0, ResourceConsumer) { 25 } }) Name (_DSD, Package () { ToUUID ("daffd814-6eba-4d8c-8a91-bc9bbf4aa301"), Package () { Package () {"mode-gpio", Package () { \_SB.PCI0.LPCB, 0, 0, 1 }} } }) Change-Id: I93ffd09e59d05c09e38693e221a87085469be3ad Signed-off-by: Duncan Laurie <dlaurie@chromium.org> Reviewed-on: https://review.coreboot.org/14937 Tested-by: build bot (Jenkins) Reviewed-by: Aaron Durbin <adurbin@chromium.org>
2016-05-10 22:18:17 +02:00
}
acpi: Change device properties to work as a tree There is a second ACPI _DSD document from the UEFI Forum that details how _DSD style tables can be nested, creating a tree of similarly formatted tables. This document is linked from acpi_device.h. In order to support this the device property interface needs to be more flexible and build up a tree of properties to write all entries at once instead of writing each entry as it is generated. In the end this is a more flexible solution that can support drivers that need child tables like the DA7219 codec, while only requiring minor changes to the existing drivers that use the device property interface. This was tested on reef (apollolake) and chell (skylake) boards to ensure that there was no change in the generated SSDT AML. Change-Id: Ia22e3a5fd3982ffa7c324bee1a8d190d49f853dd Signed-off-by: Duncan Laurie <dlaurie@chromium.org> Reviewed-on: https://review.coreboot.org/15537 Tested-by: build bot (Jenkins) Reviewed-by: Aaron Durbin <adurbin@chromium.org>
2016-07-03 04:56:06 +02:00
struct acpi_dp *acpi_dp_add_gpio(struct acpi_dp *dp, const char *name,
const char *ref, int index, int pin,
int active_low)
acpi_device: Add support for writing ACPI Device Properties The recent ACPI specification extensions have formally defined a method for describing device information with a key=value format that is modeled after the Devicetree/DTS format using a special crafted object named _DSD with a specific UUID for this format. There are three defined Device Property types: Integers, Strings, and References. It is also possible to have arrays of these properties under one key=value pair. Strings and References are both represented as character arrays but result in different generated ACPI OpCodes. Various helpers are provided for writing the Device Property header (to fill in the object name and UUID) and footer (to fill in the property count and device length values) as well as for writing the different Device Property types. A specific helper is provided for writing the defined GPIO binding Device Property that is used to allow GPIOs to be referred to by name rather than resource index. This is all documented in the _DSD Device Properties UUID document: http://uefi.org/sites/default/files/resources/_DSD-device-properties-UUID.pdf This will be used by device drivers to provide device properties that are consumed by the operating system. Devicetree bindings are often described in the linux kernel at Documentation/devicetree/bindings/ A sample driver here has an input GPIO that it needs to describe to the kernel driver: chip.h: struct drivers_generic_sample_config { struct acpi_gpio mode_gpio; }; sample.c: static void acpi_fill_ssdt_generator(struct device *dev) { struct drivers_generic_sample_config *config = dev->chip_info; const char *path = acpi_device_path(dev); ... acpi_device_write_gpio(&config->mode_gpio); ... acpi_dp_write_header(); acpi_dp_write_gpio("mode-gpio", path, 0, 0, 0); acpi_dp_write_footer(); ... } devicetree.cb: device pci 1f.0 on chip drivers/generic/sample register "mode_gpio" = "ACPI_GPIO_INPUT(GPP_B1)" device generic 0 on end end end SSDT.dsl: Name (_CRS, ResourceTemplate () { GpioIo (Exclusive, PullDefault, 0, 0, IoRestrictionInputOnly, "\\_SB.PCI0.GPIO", 0, ResourceConsumer) { 25 } }) Name (_DSD, Package () { ToUUID ("daffd814-6eba-4d8c-8a91-bc9bbf4aa301"), Package () { Package () {"mode-gpio", Package () { \_SB.PCI0.LPCB, 0, 0, 1 }} } }) Change-Id: I93ffd09e59d05c09e38693e221a87085469be3ad Signed-off-by: Duncan Laurie <dlaurie@chromium.org> Reviewed-on: https://review.coreboot.org/14937 Tested-by: build bot (Jenkins) Reviewed-by: Aaron Durbin <adurbin@chromium.org>
2016-05-10 22:18:17 +02:00
{
arch/x86: Do not add properties to null DP packages It doesn't make sense to add a property to a non-existent Device Property package. However, some of these functions will proceed anyway and allocate a new Device Property package, add the property to that, and then immediately leak the new package. This changes all the acpi_dp_add_* functions to ignore a null package. Change-Id: I664dcdbaa6b1b8a3aeb9a0126d622e2ffb736efd Signed-off-by: Jacob Garber <jgarber1@ualberta.ca> Found-by: Coverity CID 135745{6,7}, 138029{2-6} Reviewed-on: https://review.coreboot.org/c/coreboot/+/32971 Tested-by: build bot (Jenkins) <no-reply@coreboot.org> Reviewed-by: Furquan Shaikh <furquan@google.com>
2019-05-23 22:34:58 +02:00
if (!dp)
return NULL;
acpi: Change device properties to work as a tree There is a second ACPI _DSD document from the UEFI Forum that details how _DSD style tables can be nested, creating a tree of similarly formatted tables. This document is linked from acpi_device.h. In order to support this the device property interface needs to be more flexible and build up a tree of properties to write all entries at once instead of writing each entry as it is generated. In the end this is a more flexible solution that can support drivers that need child tables like the DA7219 codec, while only requiring minor changes to the existing drivers that use the device property interface. This was tested on reef (apollolake) and chell (skylake) boards to ensure that there was no change in the generated SSDT AML. Change-Id: Ia22e3a5fd3982ffa7c324bee1a8d190d49f853dd Signed-off-by: Duncan Laurie <dlaurie@chromium.org> Reviewed-on: https://review.coreboot.org/15537 Tested-by: build bot (Jenkins) Reviewed-by: Aaron Durbin <adurbin@chromium.org>
2016-07-03 04:56:06 +02:00
struct acpi_dp *gpio = acpi_dp_new_table(name);
if (!gpio)
return NULL;
/* The device that has _CRS containing GpioIO()/GpioInt() */
acpi_dp_add_reference(gpio, NULL, ref);
/* Index of the GPIO resource in _CRS starting from zero */
acpi_dp_add_integer(gpio, NULL, index);
/* Pin in the GPIO resource, typically zero */
acpi_dp_add_integer(gpio, NULL, pin);
/* Set if pin is active low */
acpi_dp_add_integer(gpio, NULL, active_low);
acpi_dp_add_array(dp, gpio);
return gpio;
acpi_device: Add support for writing ACPI Device Properties The recent ACPI specification extensions have formally defined a method for describing device information with a key=value format that is modeled after the Devicetree/DTS format using a special crafted object named _DSD with a specific UUID for this format. There are three defined Device Property types: Integers, Strings, and References. It is also possible to have arrays of these properties under one key=value pair. Strings and References are both represented as character arrays but result in different generated ACPI OpCodes. Various helpers are provided for writing the Device Property header (to fill in the object name and UUID) and footer (to fill in the property count and device length values) as well as for writing the different Device Property types. A specific helper is provided for writing the defined GPIO binding Device Property that is used to allow GPIOs to be referred to by name rather than resource index. This is all documented in the _DSD Device Properties UUID document: http://uefi.org/sites/default/files/resources/_DSD-device-properties-UUID.pdf This will be used by device drivers to provide device properties that are consumed by the operating system. Devicetree bindings are often described in the linux kernel at Documentation/devicetree/bindings/ A sample driver here has an input GPIO that it needs to describe to the kernel driver: chip.h: struct drivers_generic_sample_config { struct acpi_gpio mode_gpio; }; sample.c: static void acpi_fill_ssdt_generator(struct device *dev) { struct drivers_generic_sample_config *config = dev->chip_info; const char *path = acpi_device_path(dev); ... acpi_device_write_gpio(&config->mode_gpio); ... acpi_dp_write_header(); acpi_dp_write_gpio("mode-gpio", path, 0, 0, 0); acpi_dp_write_footer(); ... } devicetree.cb: device pci 1f.0 on chip drivers/generic/sample register "mode_gpio" = "ACPI_GPIO_INPUT(GPP_B1)" device generic 0 on end end end SSDT.dsl: Name (_CRS, ResourceTemplate () { GpioIo (Exclusive, PullDefault, 0, 0, IoRestrictionInputOnly, "\\_SB.PCI0.GPIO", 0, ResourceConsumer) { 25 } }) Name (_DSD, Package () { ToUUID ("daffd814-6eba-4d8c-8a91-bc9bbf4aa301"), Package () { Package () {"mode-gpio", Package () { \_SB.PCI0.LPCB, 0, 0, 1 }} } }) Change-Id: I93ffd09e59d05c09e38693e221a87085469be3ad Signed-off-by: Duncan Laurie <dlaurie@chromium.org> Reviewed-on: https://review.coreboot.org/14937 Tested-by: build bot (Jenkins) Reviewed-by: Aaron Durbin <adurbin@chromium.org>
2016-05-10 22:18:17 +02:00
}
arch/x86/acpi_device: Add a helper function to write PCI device This change adds a helper function to write a PCI device with _ADR and _STA defined for it. BUG=b:153858769 Signed-off-by: Furquan Shaikh <furquan@google.com> Change-Id: I932af917d91198876fe8e90af9bb7a2531bd8960 Reviewed-on: https://review.coreboot.org/c/coreboot/+/40674 Reviewed-by: Aaron Durbin <adurbin@chromium.org> Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
2020-04-23 21:51:42 +02:00
/*
* This function writes a PCI device with _ADR object:
* Example:
* Scope (\_SB.PCI0)
* {
* Device (IGFX)
* {
* Name (_ADR, 0x0000000000000000)
* Method (_STA, 0, NotSerialized) { Return (status) }
* }
* }
*/
device: Constify struct device * parameter to acpi_fill_ssdt() .acpi_fill_ssdt() does not need to modify the device structure. This change makes the struct device * parameter to acpi_fill_ssdt() as const. Change-Id: I110f4c67c3b6671c9ac0a82e02609902a8ee5d5c Signed-off-by: Furquan Shaikh <furquan@google.com> Reviewed-on: https://review.coreboot.org/c/coreboot/+/40710 Reviewed-by: HAOUAS Elyes <ehaouas@noos.fr> Reviewed-by: Aaron Durbin <adurbin@chromium.org> Reviewed-by: Paul Menzel <paulepanter@users.sourceforge.net> Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
2020-04-25 06:59:21 +02:00
void acpi_device_write_pci_dev(const struct device *dev)
arch/x86/acpi_device: Add a helper function to write PCI device This change adds a helper function to write a PCI device with _ADR and _STA defined for it. BUG=b:153858769 Signed-off-by: Furquan Shaikh <furquan@google.com> Change-Id: I932af917d91198876fe8e90af9bb7a2531bd8960 Reviewed-on: https://review.coreboot.org/c/coreboot/+/40674 Reviewed-by: Aaron Durbin <adurbin@chromium.org> Tested-by: build bot (Jenkins) <no-reply@coreboot.org>
2020-04-23 21:51:42 +02:00
{
const char *scope = acpi_device_scope(dev);
const char *name = acpi_device_name(dev);
assert(dev->path.type == DEVICE_PATH_PCI);
assert(name);
assert(scope);
acpigen_write_scope(scope);
acpigen_write_device(name);
acpigen_write_ADR_pci_device(dev);
acpigen_write_STA(acpi_device_status(dev));
acpigen_pop_len(); /* Device */
acpigen_pop_len(); /* Scope */
}
Reference in a new issue Copy permalink