freebsd-skq/sys/dev/acpica/acpi_pci_link.c

1058 lines
28 KiB
C
Raw Normal View History

/*-
* Copyright (c) 2002 Mitsuru IWASAKI <iwasaki@jp.freebsd.org>
* All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* 1. Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* 2. Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
*
* THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
* ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
* ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
* FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
* DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
* OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
* HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
* LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
* OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
* SUCH DAMAGE.
*/
#include <sys/cdefs.h>
__FBSDID("$FreeBSD$");
#include "opt_acpi.h"
#include <sys/param.h>
#include <sys/bus.h>
Rework the ACPI PCI link code. - Use a new-bus device driver for the ACPI PCI link devices. The devices are called pci_linkX. The driver includes suspend/resume support so that the ACPI bridge drivers no longer have to poke the links to get them to handle suspend/resume. Also, the code to handle which IRQs a link is routed to and choosing an IRQ when a link is not already routed is all contained in the link driver. The PCI bridge drivers now ask the link driver which IRQ to use once they determine that a _PRT entry does not use a hardwired interrupt number. - The new link driver includes support for multiple IRQ resources per link device as well as preserving any non-IRQ resources when adjusting the IRQ that a link is routed to. - The entire approach to routing when using a link device is now link-centric rather than pci bus/device/pin specific. Thus, when using a tunable to override the default IRQ settings, one now uses a single tunable to route an entire link rather than routing a single device that uses the link (which has great foot-shooting potential if the user tries to route the same link to two different IRQs using two different pci bus/device/pin hints). For example, to adjust the IRQ that \_SB_.LNKA uses, one would set 'hw.pci.link.LNKA.irq=10' from the loader. - As a side effect of having the link driver, unused link devices will now be disabled when they are probed. - The algorithm for choosing an IRQ for a link that doesn't already have an IRQ assigned is now much closer to the one used in $PIR routing. When a link is routed via an ISA IRQ, only known-good IRQs that the BIOS has already used are used for routing instead of using probabilities to guess at which IRQs are probably not used by an ISA device. One change from $PIR is that the SCI is always considered a viable ISA IRQ, so that if the BIOS does not setup any IRQs the kernel will degenerate to routing all interrupts over the SCI. For non ISA IRQs, interrupts are picked from the possible pool using a simplistic weighting algorithm. Tested by: ru, scottl, others on acpi@ Reviewed by: njl
2004-11-23 22:26:44 +00:00
#include <sys/kernel.h>
#include <sys/limits.h>
#include <sys/malloc.h>
#include <sys/module.h>
2005-09-11 18:39:03 +00:00
#include <contrib/dev/acpica/acpi.h>
#include <dev/acpica/acpivar.h>
#include <dev/acpica/acpi_pcibvar.h>
Rework the ACPI PCI link code. - Use a new-bus device driver for the ACPI PCI link devices. The devices are called pci_linkX. The driver includes suspend/resume support so that the ACPI bridge drivers no longer have to poke the links to get them to handle suspend/resume. Also, the code to handle which IRQs a link is routed to and choosing an IRQ when a link is not already routed is all contained in the link driver. The PCI bridge drivers now ask the link driver which IRQ to use once they determine that a _PRT entry does not use a hardwired interrupt number. - The new link driver includes support for multiple IRQ resources per link device as well as preserving any non-IRQ resources when adjusting the IRQ that a link is routed to. - The entire approach to routing when using a link device is now link-centric rather than pci bus/device/pin specific. Thus, when using a tunable to override the default IRQ settings, one now uses a single tunable to route an entire link rather than routing a single device that uses the link (which has great foot-shooting potential if the user tries to route the same link to two different IRQs using two different pci bus/device/pin hints). For example, to adjust the IRQ that \_SB_.LNKA uses, one would set 'hw.pci.link.LNKA.irq=10' from the loader. - As a side effect of having the link driver, unused link devices will now be disabled when they are probed. - The algorithm for choosing an IRQ for a link that doesn't already have an IRQ assigned is now much closer to the one used in $PIR routing. When a link is routed via an ISA IRQ, only known-good IRQs that the BIOS has already used are used for routing instead of using probabilities to guess at which IRQs are probably not used by an ISA device. One change from $PIR is that the SCI is always considered a viable ISA IRQ, so that if the BIOS does not setup any IRQs the kernel will degenerate to routing all interrupts over the SCI. For non ISA IRQs, interrupts are picked from the possible pool using a simplistic weighting algorithm. Tested by: ru, scottl, others on acpi@ Reviewed by: njl
2004-11-23 22:26:44 +00:00
#include <machine/pci_cfgreg.h>
#include <dev/pci/pcireg.h>
#include <dev/pci/pcivar.h>
#include "pcib_if.h"
2004-04-14 03:34:11 +00:00
/* Hooks for the ACPI CA debugging infrastructure. */
#define _COMPONENT ACPI_BUS
ACPI_MODULE_NAME("PCI_LINK")
ACPI_SERIAL_DECL(pci_link, "ACPI PCI link");
Rework the ACPI PCI link code. - Use a new-bus device driver for the ACPI PCI link devices. The devices are called pci_linkX. The driver includes suspend/resume support so that the ACPI bridge drivers no longer have to poke the links to get them to handle suspend/resume. Also, the code to handle which IRQs a link is routed to and choosing an IRQ when a link is not already routed is all contained in the link driver. The PCI bridge drivers now ask the link driver which IRQ to use once they determine that a _PRT entry does not use a hardwired interrupt number. - The new link driver includes support for multiple IRQ resources per link device as well as preserving any non-IRQ resources when adjusting the IRQ that a link is routed to. - The entire approach to routing when using a link device is now link-centric rather than pci bus/device/pin specific. Thus, when using a tunable to override the default IRQ settings, one now uses a single tunable to route an entire link rather than routing a single device that uses the link (which has great foot-shooting potential if the user tries to route the same link to two different IRQs using two different pci bus/device/pin hints). For example, to adjust the IRQ that \_SB_.LNKA uses, one would set 'hw.pci.link.LNKA.irq=10' from the loader. - As a side effect of having the link driver, unused link devices will now be disabled when they are probed. - The algorithm for choosing an IRQ for a link that doesn't already have an IRQ assigned is now much closer to the one used in $PIR routing. When a link is routed via an ISA IRQ, only known-good IRQs that the BIOS has already used are used for routing instead of using probabilities to guess at which IRQs are probably not used by an ISA device. One change from $PIR is that the SCI is always considered a viable ISA IRQ, so that if the BIOS does not setup any IRQs the kernel will degenerate to routing all interrupts over the SCI. For non ISA IRQs, interrupts are picked from the possible pool using a simplistic weighting algorithm. Tested by: ru, scottl, others on acpi@ Reviewed by: njl
2004-11-23 22:26:44 +00:00
#define NUM_ISA_INTERRUPTS 16
#define NUM_ACPI_INTERRUPTS 256
/*
Rework the ACPI PCI link code. - Use a new-bus device driver for the ACPI PCI link devices. The devices are called pci_linkX. The driver includes suspend/resume support so that the ACPI bridge drivers no longer have to poke the links to get them to handle suspend/resume. Also, the code to handle which IRQs a link is routed to and choosing an IRQ when a link is not already routed is all contained in the link driver. The PCI bridge drivers now ask the link driver which IRQ to use once they determine that a _PRT entry does not use a hardwired interrupt number. - The new link driver includes support for multiple IRQ resources per link device as well as preserving any non-IRQ resources when adjusting the IRQ that a link is routed to. - The entire approach to routing when using a link device is now link-centric rather than pci bus/device/pin specific. Thus, when using a tunable to override the default IRQ settings, one now uses a single tunable to route an entire link rather than routing a single device that uses the link (which has great foot-shooting potential if the user tries to route the same link to two different IRQs using two different pci bus/device/pin hints). For example, to adjust the IRQ that \_SB_.LNKA uses, one would set 'hw.pci.link.LNKA.irq=10' from the loader. - As a side effect of having the link driver, unused link devices will now be disabled when they are probed. - The algorithm for choosing an IRQ for a link that doesn't already have an IRQ assigned is now much closer to the one used in $PIR routing. When a link is routed via an ISA IRQ, only known-good IRQs that the BIOS has already used are used for routing instead of using probabilities to guess at which IRQs are probably not used by an ISA device. One change from $PIR is that the SCI is always considered a viable ISA IRQ, so that if the BIOS does not setup any IRQs the kernel will degenerate to routing all interrupts over the SCI. For non ISA IRQs, interrupts are picked from the possible pool using a simplistic weighting algorithm. Tested by: ru, scottl, others on acpi@ Reviewed by: njl
2004-11-23 22:26:44 +00:00
* An ACPI PCI link device may contain multiple links. Each link has its
* own ACPI resource. _PRT entries specify which link is being used via
* the Source Index.
*
* XXX: A note about Source Indices and DPFs: Currently we assume that
* the DPF start and end tags are not counted towards the index that
* Source Index corresponds to. Also, we assume that when DPFs are in use
* they various sets overlap in terms of Indices. Here's an example
* resource list indicating these assumptions:
*
* Resource Index
* -------- -----
* I/O Port 0
* Start DPF -
* IRQ 1
* MemIO 2
* Start DPF -
* IRQ 1
* MemIO 2
* End DPF -
* DMA Channel 3
*
* The XXX is because I'm not sure if this is a valid assumption to make.
*/
/* States during DPF processing. */
#define DPF_OUTSIDE 0
#define DPF_FIRST 1
#define DPF_IGNORE 2
Rework the ACPI PCI link code. - Use a new-bus device driver for the ACPI PCI link devices. The devices are called pci_linkX. The driver includes suspend/resume support so that the ACPI bridge drivers no longer have to poke the links to get them to handle suspend/resume. Also, the code to handle which IRQs a link is routed to and choosing an IRQ when a link is not already routed is all contained in the link driver. The PCI bridge drivers now ask the link driver which IRQ to use once they determine that a _PRT entry does not use a hardwired interrupt number. - The new link driver includes support for multiple IRQ resources per link device as well as preserving any non-IRQ resources when adjusting the IRQ that a link is routed to. - The entire approach to routing when using a link device is now link-centric rather than pci bus/device/pin specific. Thus, when using a tunable to override the default IRQ settings, one now uses a single tunable to route an entire link rather than routing a single device that uses the link (which has great foot-shooting potential if the user tries to route the same link to two different IRQs using two different pci bus/device/pin hints). For example, to adjust the IRQ that \_SB_.LNKA uses, one would set 'hw.pci.link.LNKA.irq=10' from the loader. - As a side effect of having the link driver, unused link devices will now be disabled when they are probed. - The algorithm for choosing an IRQ for a link that doesn't already have an IRQ assigned is now much closer to the one used in $PIR routing. When a link is routed via an ISA IRQ, only known-good IRQs that the BIOS has already used are used for routing instead of using probabilities to guess at which IRQs are probably not used by an ISA device. One change from $PIR is that the SCI is always considered a viable ISA IRQ, so that if the BIOS does not setup any IRQs the kernel will degenerate to routing all interrupts over the SCI. For non ISA IRQs, interrupts are picked from the possible pool using a simplistic weighting algorithm. Tested by: ru, scottl, others on acpi@ Reviewed by: njl
2004-11-23 22:26:44 +00:00
struct link;
Rework the ACPI PCI link code. - Use a new-bus device driver for the ACPI PCI link devices. The devices are called pci_linkX. The driver includes suspend/resume support so that the ACPI bridge drivers no longer have to poke the links to get them to handle suspend/resume. Also, the code to handle which IRQs a link is routed to and choosing an IRQ when a link is not already routed is all contained in the link driver. The PCI bridge drivers now ask the link driver which IRQ to use once they determine that a _PRT entry does not use a hardwired interrupt number. - The new link driver includes support for multiple IRQ resources per link device as well as preserving any non-IRQ resources when adjusting the IRQ that a link is routed to. - The entire approach to routing when using a link device is now link-centric rather than pci bus/device/pin specific. Thus, when using a tunable to override the default IRQ settings, one now uses a single tunable to route an entire link rather than routing a single device that uses the link (which has great foot-shooting potential if the user tries to route the same link to two different IRQs using two different pci bus/device/pin hints). For example, to adjust the IRQ that \_SB_.LNKA uses, one would set 'hw.pci.link.LNKA.irq=10' from the loader. - As a side effect of having the link driver, unused link devices will now be disabled when they are probed. - The algorithm for choosing an IRQ for a link that doesn't already have an IRQ assigned is now much closer to the one used in $PIR routing. When a link is routed via an ISA IRQ, only known-good IRQs that the BIOS has already used are used for routing instead of using probabilities to guess at which IRQs are probably not used by an ISA device. One change from $PIR is that the SCI is always considered a viable ISA IRQ, so that if the BIOS does not setup any IRQs the kernel will degenerate to routing all interrupts over the SCI. For non ISA IRQs, interrupts are picked from the possible pool using a simplistic weighting algorithm. Tested by: ru, scottl, others on acpi@ Reviewed by: njl
2004-11-23 22:26:44 +00:00
struct acpi_pci_link_softc {
int pl_num_links;
int pl_crs_bad;
Rework the ACPI PCI link code. - Use a new-bus device driver for the ACPI PCI link devices. The devices are called pci_linkX. The driver includes suspend/resume support so that the ACPI bridge drivers no longer have to poke the links to get them to handle suspend/resume. Also, the code to handle which IRQs a link is routed to and choosing an IRQ when a link is not already routed is all contained in the link driver. The PCI bridge drivers now ask the link driver which IRQ to use once they determine that a _PRT entry does not use a hardwired interrupt number. - The new link driver includes support for multiple IRQ resources per link device as well as preserving any non-IRQ resources when adjusting the IRQ that a link is routed to. - The entire approach to routing when using a link device is now link-centric rather than pci bus/device/pin specific. Thus, when using a tunable to override the default IRQ settings, one now uses a single tunable to route an entire link rather than routing a single device that uses the link (which has great foot-shooting potential if the user tries to route the same link to two different IRQs using two different pci bus/device/pin hints). For example, to adjust the IRQ that \_SB_.LNKA uses, one would set 'hw.pci.link.LNKA.irq=10' from the loader. - As a side effect of having the link driver, unused link devices will now be disabled when they are probed. - The algorithm for choosing an IRQ for a link that doesn't already have an IRQ assigned is now much closer to the one used in $PIR routing. When a link is routed via an ISA IRQ, only known-good IRQs that the BIOS has already used are used for routing instead of using probabilities to guess at which IRQs are probably not used by an ISA device. One change from $PIR is that the SCI is always considered a viable ISA IRQ, so that if the BIOS does not setup any IRQs the kernel will degenerate to routing all interrupts over the SCI. For non ISA IRQs, interrupts are picked from the possible pool using a simplistic weighting algorithm. Tested by: ru, scottl, others on acpi@ Reviewed by: njl
2004-11-23 22:26:44 +00:00
struct link *pl_links;
device_t pl_dev;
Rework the ACPI PCI link code. - Use a new-bus device driver for the ACPI PCI link devices. The devices are called pci_linkX. The driver includes suspend/resume support so that the ACPI bridge drivers no longer have to poke the links to get them to handle suspend/resume. Also, the code to handle which IRQs a link is routed to and choosing an IRQ when a link is not already routed is all contained in the link driver. The PCI bridge drivers now ask the link driver which IRQ to use once they determine that a _PRT entry does not use a hardwired interrupt number. - The new link driver includes support for multiple IRQ resources per link device as well as preserving any non-IRQ resources when adjusting the IRQ that a link is routed to. - The entire approach to routing when using a link device is now link-centric rather than pci bus/device/pin specific. Thus, when using a tunable to override the default IRQ settings, one now uses a single tunable to route an entire link rather than routing a single device that uses the link (which has great foot-shooting potential if the user tries to route the same link to two different IRQs using two different pci bus/device/pin hints). For example, to adjust the IRQ that \_SB_.LNKA uses, one would set 'hw.pci.link.LNKA.irq=10' from the loader. - As a side effect of having the link driver, unused link devices will now be disabled when they are probed. - The algorithm for choosing an IRQ for a link that doesn't already have an IRQ assigned is now much closer to the one used in $PIR routing. When a link is routed via an ISA IRQ, only known-good IRQs that the BIOS has already used are used for routing instead of using probabilities to guess at which IRQs are probably not used by an ISA device. One change from $PIR is that the SCI is always considered a viable ISA IRQ, so that if the BIOS does not setup any IRQs the kernel will degenerate to routing all interrupts over the SCI. For non ISA IRQs, interrupts are picked from the possible pool using a simplistic weighting algorithm. Tested by: ru, scottl, others on acpi@ Reviewed by: njl
2004-11-23 22:26:44 +00:00
};
Rework the ACPI PCI link code. - Use a new-bus device driver for the ACPI PCI link devices. The devices are called pci_linkX. The driver includes suspend/resume support so that the ACPI bridge drivers no longer have to poke the links to get them to handle suspend/resume. Also, the code to handle which IRQs a link is routed to and choosing an IRQ when a link is not already routed is all contained in the link driver. The PCI bridge drivers now ask the link driver which IRQ to use once they determine that a _PRT entry does not use a hardwired interrupt number. - The new link driver includes support for multiple IRQ resources per link device as well as preserving any non-IRQ resources when adjusting the IRQ that a link is routed to. - The entire approach to routing when using a link device is now link-centric rather than pci bus/device/pin specific. Thus, when using a tunable to override the default IRQ settings, one now uses a single tunable to route an entire link rather than routing a single device that uses the link (which has great foot-shooting potential if the user tries to route the same link to two different IRQs using two different pci bus/device/pin hints). For example, to adjust the IRQ that \_SB_.LNKA uses, one would set 'hw.pci.link.LNKA.irq=10' from the loader. - As a side effect of having the link driver, unused link devices will now be disabled when they are probed. - The algorithm for choosing an IRQ for a link that doesn't already have an IRQ assigned is now much closer to the one used in $PIR routing. When a link is routed via an ISA IRQ, only known-good IRQs that the BIOS has already used are used for routing instead of using probabilities to guess at which IRQs are probably not used by an ISA device. One change from $PIR is that the SCI is always considered a viable ISA IRQ, so that if the BIOS does not setup any IRQs the kernel will degenerate to routing all interrupts over the SCI. For non ISA IRQs, interrupts are picked from the possible pool using a simplistic weighting algorithm. Tested by: ru, scottl, others on acpi@ Reviewed by: njl
2004-11-23 22:26:44 +00:00
struct link {
struct acpi_pci_link_softc *l_sc;
uint8_t l_bios_irq;
uint8_t l_irq;
uint8_t l_initial_irq;
int l_res_index;
int l_num_irqs;
int *l_irqs;
int l_references;
int l_routed:1;
int l_isa_irq:1;
ACPI_RESOURCE l_prs_template;
};
struct link_count_request {
int in_dpf;
int count;
};
Rework the ACPI PCI link code. - Use a new-bus device driver for the ACPI PCI link devices. The devices are called pci_linkX. The driver includes suspend/resume support so that the ACPI bridge drivers no longer have to poke the links to get them to handle suspend/resume. Also, the code to handle which IRQs a link is routed to and choosing an IRQ when a link is not already routed is all contained in the link driver. The PCI bridge drivers now ask the link driver which IRQ to use once they determine that a _PRT entry does not use a hardwired interrupt number. - The new link driver includes support for multiple IRQ resources per link device as well as preserving any non-IRQ resources when adjusting the IRQ that a link is routed to. - The entire approach to routing when using a link device is now link-centric rather than pci bus/device/pin specific. Thus, when using a tunable to override the default IRQ settings, one now uses a single tunable to route an entire link rather than routing a single device that uses the link (which has great foot-shooting potential if the user tries to route the same link to two different IRQs using two different pci bus/device/pin hints). For example, to adjust the IRQ that \_SB_.LNKA uses, one would set 'hw.pci.link.LNKA.irq=10' from the loader. - As a side effect of having the link driver, unused link devices will now be disabled when they are probed. - The algorithm for choosing an IRQ for a link that doesn't already have an IRQ assigned is now much closer to the one used in $PIR routing. When a link is routed via an ISA IRQ, only known-good IRQs that the BIOS has already used are used for routing instead of using probabilities to guess at which IRQs are probably not used by an ISA device. One change from $PIR is that the SCI is always considered a viable ISA IRQ, so that if the BIOS does not setup any IRQs the kernel will degenerate to routing all interrupts over the SCI. For non ISA IRQs, interrupts are picked from the possible pool using a simplistic weighting algorithm. Tested by: ru, scottl, others on acpi@ Reviewed by: njl
2004-11-23 22:26:44 +00:00
struct link_res_request {
struct acpi_pci_link_softc *sc;
int in_dpf;
int res_index;
int link_index;
Rework the ACPI PCI link code. - Use a new-bus device driver for the ACPI PCI link devices. The devices are called pci_linkX. The driver includes suspend/resume support so that the ACPI bridge drivers no longer have to poke the links to get them to handle suspend/resume. Also, the code to handle which IRQs a link is routed to and choosing an IRQ when a link is not already routed is all contained in the link driver. The PCI bridge drivers now ask the link driver which IRQ to use once they determine that a _PRT entry does not use a hardwired interrupt number. - The new link driver includes support for multiple IRQ resources per link device as well as preserving any non-IRQ resources when adjusting the IRQ that a link is routed to. - The entire approach to routing when using a link device is now link-centric rather than pci bus/device/pin specific. Thus, when using a tunable to override the default IRQ settings, one now uses a single tunable to route an entire link rather than routing a single device that uses the link (which has great foot-shooting potential if the user tries to route the same link to two different IRQs using two different pci bus/device/pin hints). For example, to adjust the IRQ that \_SB_.LNKA uses, one would set 'hw.pci.link.LNKA.irq=10' from the loader. - As a side effect of having the link driver, unused link devices will now be disabled when they are probed. - The algorithm for choosing an IRQ for a link that doesn't already have an IRQ assigned is now much closer to the one used in $PIR routing. When a link is routed via an ISA IRQ, only known-good IRQs that the BIOS has already used are used for routing instead of using probabilities to guess at which IRQs are probably not used by an ISA device. One change from $PIR is that the SCI is always considered a viable ISA IRQ, so that if the BIOS does not setup any IRQs the kernel will degenerate to routing all interrupts over the SCI. For non ISA IRQs, interrupts are picked from the possible pool using a simplistic weighting algorithm. Tested by: ru, scottl, others on acpi@ Reviewed by: njl
2004-11-23 22:26:44 +00:00
};
MALLOC_DEFINE(M_PCI_LINK, "pci_link", "ACPI PCI Link structures");
Rework the ACPI PCI link code. - Use a new-bus device driver for the ACPI PCI link devices. The devices are called pci_linkX. The driver includes suspend/resume support so that the ACPI bridge drivers no longer have to poke the links to get them to handle suspend/resume. Also, the code to handle which IRQs a link is routed to and choosing an IRQ when a link is not already routed is all contained in the link driver. The PCI bridge drivers now ask the link driver which IRQ to use once they determine that a _PRT entry does not use a hardwired interrupt number. - The new link driver includes support for multiple IRQ resources per link device as well as preserving any non-IRQ resources when adjusting the IRQ that a link is routed to. - The entire approach to routing when using a link device is now link-centric rather than pci bus/device/pin specific. Thus, when using a tunable to override the default IRQ settings, one now uses a single tunable to route an entire link rather than routing a single device that uses the link (which has great foot-shooting potential if the user tries to route the same link to two different IRQs using two different pci bus/device/pin hints). For example, to adjust the IRQ that \_SB_.LNKA uses, one would set 'hw.pci.link.LNKA.irq=10' from the loader. - As a side effect of having the link driver, unused link devices will now be disabled when they are probed. - The algorithm for choosing an IRQ for a link that doesn't already have an IRQ assigned is now much closer to the one used in $PIR routing. When a link is routed via an ISA IRQ, only known-good IRQs that the BIOS has already used are used for routing instead of using probabilities to guess at which IRQs are probably not used by an ISA device. One change from $PIR is that the SCI is always considered a viable ISA IRQ, so that if the BIOS does not setup any IRQs the kernel will degenerate to routing all interrupts over the SCI. For non ISA IRQs, interrupts are picked from the possible pool using a simplistic weighting algorithm. Tested by: ru, scottl, others on acpi@ Reviewed by: njl
2004-11-23 22:26:44 +00:00
static int pci_link_interrupt_weights[NUM_ACPI_INTERRUPTS];
static int pci_link_bios_isa_irqs;
Rework the ACPI PCI link code. - Use a new-bus device driver for the ACPI PCI link devices. The devices are called pci_linkX. The driver includes suspend/resume support so that the ACPI bridge drivers no longer have to poke the links to get them to handle suspend/resume. Also, the code to handle which IRQs a link is routed to and choosing an IRQ when a link is not already routed is all contained in the link driver. The PCI bridge drivers now ask the link driver which IRQ to use once they determine that a _PRT entry does not use a hardwired interrupt number. - The new link driver includes support for multiple IRQ resources per link device as well as preserving any non-IRQ resources when adjusting the IRQ that a link is routed to. - The entire approach to routing when using a link device is now link-centric rather than pci bus/device/pin specific. Thus, when using a tunable to override the default IRQ settings, one now uses a single tunable to route an entire link rather than routing a single device that uses the link (which has great foot-shooting potential if the user tries to route the same link to two different IRQs using two different pci bus/device/pin hints). For example, to adjust the IRQ that \_SB_.LNKA uses, one would set 'hw.pci.link.LNKA.irq=10' from the loader. - As a side effect of having the link driver, unused link devices will now be disabled when they are probed. - The algorithm for choosing an IRQ for a link that doesn't already have an IRQ assigned is now much closer to the one used in $PIR routing. When a link is routed via an ISA IRQ, only known-good IRQs that the BIOS has already used are used for routing instead of using probabilities to guess at which IRQs are probably not used by an ISA device. One change from $PIR is that the SCI is always considered a viable ISA IRQ, so that if the BIOS does not setup any IRQs the kernel will degenerate to routing all interrupts over the SCI. For non ISA IRQs, interrupts are picked from the possible pool using a simplistic weighting algorithm. Tested by: ru, scottl, others on acpi@ Reviewed by: njl
2004-11-23 22:26:44 +00:00
static char *pci_link_ids[] = { "PNP0C0F", NULL };
Rework the ACPI PCI link code. - Use a new-bus device driver for the ACPI PCI link devices. The devices are called pci_linkX. The driver includes suspend/resume support so that the ACPI bridge drivers no longer have to poke the links to get them to handle suspend/resume. Also, the code to handle which IRQs a link is routed to and choosing an IRQ when a link is not already routed is all contained in the link driver. The PCI bridge drivers now ask the link driver which IRQ to use once they determine that a _PRT entry does not use a hardwired interrupt number. - The new link driver includes support for multiple IRQ resources per link device as well as preserving any non-IRQ resources when adjusting the IRQ that a link is routed to. - The entire approach to routing when using a link device is now link-centric rather than pci bus/device/pin specific. Thus, when using a tunable to override the default IRQ settings, one now uses a single tunable to route an entire link rather than routing a single device that uses the link (which has great foot-shooting potential if the user tries to route the same link to two different IRQs using two different pci bus/device/pin hints). For example, to adjust the IRQ that \_SB_.LNKA uses, one would set 'hw.pci.link.LNKA.irq=10' from the loader. - As a side effect of having the link driver, unused link devices will now be disabled when they are probed. - The algorithm for choosing an IRQ for a link that doesn't already have an IRQ assigned is now much closer to the one used in $PIR routing. When a link is routed via an ISA IRQ, only known-good IRQs that the BIOS has already used are used for routing instead of using probabilities to guess at which IRQs are probably not used by an ISA device. One change from $PIR is that the SCI is always considered a viable ISA IRQ, so that if the BIOS does not setup any IRQs the kernel will degenerate to routing all interrupts over the SCI. For non ISA IRQs, interrupts are picked from the possible pool using a simplistic weighting algorithm. Tested by: ru, scottl, others on acpi@ Reviewed by: njl
2004-11-23 22:26:44 +00:00
/*
* Fetch the short name associated with an ACPI handle and save it in the
* passed in buffer.
*/
static ACPI_STATUS
acpi_short_name(ACPI_HANDLE handle, char *buffer, size_t buflen)
{
ACPI_BUFFER buf;
Rework the ACPI PCI link code. - Use a new-bus device driver for the ACPI PCI link devices. The devices are called pci_linkX. The driver includes suspend/resume support so that the ACPI bridge drivers no longer have to poke the links to get them to handle suspend/resume. Also, the code to handle which IRQs a link is routed to and choosing an IRQ when a link is not already routed is all contained in the link driver. The PCI bridge drivers now ask the link driver which IRQ to use once they determine that a _PRT entry does not use a hardwired interrupt number. - The new link driver includes support for multiple IRQ resources per link device as well as preserving any non-IRQ resources when adjusting the IRQ that a link is routed to. - The entire approach to routing when using a link device is now link-centric rather than pci bus/device/pin specific. Thus, when using a tunable to override the default IRQ settings, one now uses a single tunable to route an entire link rather than routing a single device that uses the link (which has great foot-shooting potential if the user tries to route the same link to two different IRQs using two different pci bus/device/pin hints). For example, to adjust the IRQ that \_SB_.LNKA uses, one would set 'hw.pci.link.LNKA.irq=10' from the loader. - As a side effect of having the link driver, unused link devices will now be disabled when they are probed. - The algorithm for choosing an IRQ for a link that doesn't already have an IRQ assigned is now much closer to the one used in $PIR routing. When a link is routed via an ISA IRQ, only known-good IRQs that the BIOS has already used are used for routing instead of using probabilities to guess at which IRQs are probably not used by an ISA device. One change from $PIR is that the SCI is always considered a viable ISA IRQ, so that if the BIOS does not setup any IRQs the kernel will degenerate to routing all interrupts over the SCI. For non ISA IRQs, interrupts are picked from the possible pool using a simplistic weighting algorithm. Tested by: ru, scottl, others on acpi@ Reviewed by: njl
2004-11-23 22:26:44 +00:00
buf.Length = buflen;
buf.Pointer = buffer;
return (AcpiGetName(handle, ACPI_SINGLE_NAME, &buf));
}
Rework the ACPI PCI link code. - Use a new-bus device driver for the ACPI PCI link devices. The devices are called pci_linkX. The driver includes suspend/resume support so that the ACPI bridge drivers no longer have to poke the links to get them to handle suspend/resume. Also, the code to handle which IRQs a link is routed to and choosing an IRQ when a link is not already routed is all contained in the link driver. The PCI bridge drivers now ask the link driver which IRQ to use once they determine that a _PRT entry does not use a hardwired interrupt number. - The new link driver includes support for multiple IRQ resources per link device as well as preserving any non-IRQ resources when adjusting the IRQ that a link is routed to. - The entire approach to routing when using a link device is now link-centric rather than pci bus/device/pin specific. Thus, when using a tunable to override the default IRQ settings, one now uses a single tunable to route an entire link rather than routing a single device that uses the link (which has great foot-shooting potential if the user tries to route the same link to two different IRQs using two different pci bus/device/pin hints). For example, to adjust the IRQ that \_SB_.LNKA uses, one would set 'hw.pci.link.LNKA.irq=10' from the loader. - As a side effect of having the link driver, unused link devices will now be disabled when they are probed. - The algorithm for choosing an IRQ for a link that doesn't already have an IRQ assigned is now much closer to the one used in $PIR routing. When a link is routed via an ISA IRQ, only known-good IRQs that the BIOS has already used are used for routing instead of using probabilities to guess at which IRQs are probably not used by an ISA device. One change from $PIR is that the SCI is always considered a viable ISA IRQ, so that if the BIOS does not setup any IRQs the kernel will degenerate to routing all interrupts over the SCI. For non ISA IRQs, interrupts are picked from the possible pool using a simplistic weighting algorithm. Tested by: ru, scottl, others on acpi@ Reviewed by: njl
2004-11-23 22:26:44 +00:00
static int
acpi_pci_link_probe(device_t dev)
{
char descr[28], name[12];
Rework the ACPI PCI link code. - Use a new-bus device driver for the ACPI PCI link devices. The devices are called pci_linkX. The driver includes suspend/resume support so that the ACPI bridge drivers no longer have to poke the links to get them to handle suspend/resume. Also, the code to handle which IRQs a link is routed to and choosing an IRQ when a link is not already routed is all contained in the link driver. The PCI bridge drivers now ask the link driver which IRQ to use once they determine that a _PRT entry does not use a hardwired interrupt number. - The new link driver includes support for multiple IRQ resources per link device as well as preserving any non-IRQ resources when adjusting the IRQ that a link is routed to. - The entire approach to routing when using a link device is now link-centric rather than pci bus/device/pin specific. Thus, when using a tunable to override the default IRQ settings, one now uses a single tunable to route an entire link rather than routing a single device that uses the link (which has great foot-shooting potential if the user tries to route the same link to two different IRQs using two different pci bus/device/pin hints). For example, to adjust the IRQ that \_SB_.LNKA uses, one would set 'hw.pci.link.LNKA.irq=10' from the loader. - As a side effect of having the link driver, unused link devices will now be disabled when they are probed. - The algorithm for choosing an IRQ for a link that doesn't already have an IRQ assigned is now much closer to the one used in $PIR routing. When a link is routed via an ISA IRQ, only known-good IRQs that the BIOS has already used are used for routing instead of using probabilities to guess at which IRQs are probably not used by an ISA device. One change from $PIR is that the SCI is always considered a viable ISA IRQ, so that if the BIOS does not setup any IRQs the kernel will degenerate to routing all interrupts over the SCI. For non ISA IRQs, interrupts are picked from the possible pool using a simplistic weighting algorithm. Tested by: ru, scottl, others on acpi@ Reviewed by: njl
2004-11-23 22:26:44 +00:00
/*
* We explicitly do not check _STA since not all systems set it to
* sensible values.
*/
if (acpi_disabled("pci_link") ||
ACPI_ID_PROBE(device_get_parent(dev), dev, pci_link_ids) == NULL)
return (ENXIO);
if (ACPI_SUCCESS(acpi_short_name(acpi_get_handle(dev), name,
sizeof(name)))) {
snprintf(descr, sizeof(descr), "ACPI PCI Link %s", name);
device_set_desc_copy(dev, descr);
} else
device_set_desc(dev, "ACPI PCI Link");
return (0);
}
static ACPI_STATUS
acpi_count_irq_resources(ACPI_RESOURCE *res, void *context)
{
struct link_count_request *req;
req = (struct link_count_request *)context;
2005-11-01 22:44:08 +00:00
switch (res->Type) {
case ACPI_RESOURCE_TYPE_START_DEPENDENT:
switch (req->in_dpf) {
case DPF_OUTSIDE:
/* We've started the first DPF. */
req->in_dpf = DPF_FIRST;
break;
case DPF_FIRST:
/* We've started the second DPF. */
req->in_dpf = DPF_IGNORE;
break;
}
break;
2005-11-01 22:44:08 +00:00
case ACPI_RESOURCE_TYPE_END_DEPENDENT:
/* We are finished with DPF parsing. */
KASSERT(req->in_dpf != DPF_OUTSIDE,
("%s: end dpf when not parsing a dpf", __func__));
req->in_dpf = DPF_OUTSIDE;
break;
2005-11-01 22:44:08 +00:00
case ACPI_RESOURCE_TYPE_IRQ:
case ACPI_RESOURCE_TYPE_EXTENDED_IRQ:
/*
* Don't count resources if we are in a DPF set that we are
* ignoring.
*/
if (req->in_dpf != DPF_IGNORE)
req->count++;
}
Rework the ACPI PCI link code. - Use a new-bus device driver for the ACPI PCI link devices. The devices are called pci_linkX. The driver includes suspend/resume support so that the ACPI bridge drivers no longer have to poke the links to get them to handle suspend/resume. Also, the code to handle which IRQs a link is routed to and choosing an IRQ when a link is not already routed is all contained in the link driver. The PCI bridge drivers now ask the link driver which IRQ to use once they determine that a _PRT entry does not use a hardwired interrupt number. - The new link driver includes support for multiple IRQ resources per link device as well as preserving any non-IRQ resources when adjusting the IRQ that a link is routed to. - The entire approach to routing when using a link device is now link-centric rather than pci bus/device/pin specific. Thus, when using a tunable to override the default IRQ settings, one now uses a single tunable to route an entire link rather than routing a single device that uses the link (which has great foot-shooting potential if the user tries to route the same link to two different IRQs using two different pci bus/device/pin hints). For example, to adjust the IRQ that \_SB_.LNKA uses, one would set 'hw.pci.link.LNKA.irq=10' from the loader. - As a side effect of having the link driver, unused link devices will now be disabled when they are probed. - The algorithm for choosing an IRQ for a link that doesn't already have an IRQ assigned is now much closer to the one used in $PIR routing. When a link is routed via an ISA IRQ, only known-good IRQs that the BIOS has already used are used for routing instead of using probabilities to guess at which IRQs are probably not used by an ISA device. One change from $PIR is that the SCI is always considered a viable ISA IRQ, so that if the BIOS does not setup any IRQs the kernel will degenerate to routing all interrupts over the SCI. For non ISA IRQs, interrupts are picked from the possible pool using a simplistic weighting algorithm. Tested by: ru, scottl, others on acpi@ Reviewed by: njl
2004-11-23 22:26:44 +00:00
return (AE_OK);
}
static ACPI_STATUS
Rework the ACPI PCI link code. - Use a new-bus device driver for the ACPI PCI link devices. The devices are called pci_linkX. The driver includes suspend/resume support so that the ACPI bridge drivers no longer have to poke the links to get them to handle suspend/resume. Also, the code to handle which IRQs a link is routed to and choosing an IRQ when a link is not already routed is all contained in the link driver. The PCI bridge drivers now ask the link driver which IRQ to use once they determine that a _PRT entry does not use a hardwired interrupt number. - The new link driver includes support for multiple IRQ resources per link device as well as preserving any non-IRQ resources when adjusting the IRQ that a link is routed to. - The entire approach to routing when using a link device is now link-centric rather than pci bus/device/pin specific. Thus, when using a tunable to override the default IRQ settings, one now uses a single tunable to route an entire link rather than routing a single device that uses the link (which has great foot-shooting potential if the user tries to route the same link to two different IRQs using two different pci bus/device/pin hints). For example, to adjust the IRQ that \_SB_.LNKA uses, one would set 'hw.pci.link.LNKA.irq=10' from the loader. - As a side effect of having the link driver, unused link devices will now be disabled when they are probed. - The algorithm for choosing an IRQ for a link that doesn't already have an IRQ assigned is now much closer to the one used in $PIR routing. When a link is routed via an ISA IRQ, only known-good IRQs that the BIOS has already used are used for routing instead of using probabilities to guess at which IRQs are probably not used by an ISA device. One change from $PIR is that the SCI is always considered a viable ISA IRQ, so that if the BIOS does not setup any IRQs the kernel will degenerate to routing all interrupts over the SCI. For non ISA IRQs, interrupts are picked from the possible pool using a simplistic weighting algorithm. Tested by: ru, scottl, others on acpi@ Reviewed by: njl
2004-11-23 22:26:44 +00:00
link_add_crs(ACPI_RESOURCE *res, void *context)
{
Rework the ACPI PCI link code. - Use a new-bus device driver for the ACPI PCI link devices. The devices are called pci_linkX. The driver includes suspend/resume support so that the ACPI bridge drivers no longer have to poke the links to get them to handle suspend/resume. Also, the code to handle which IRQs a link is routed to and choosing an IRQ when a link is not already routed is all contained in the link driver. The PCI bridge drivers now ask the link driver which IRQ to use once they determine that a _PRT entry does not use a hardwired interrupt number. - The new link driver includes support for multiple IRQ resources per link device as well as preserving any non-IRQ resources when adjusting the IRQ that a link is routed to. - The entire approach to routing when using a link device is now link-centric rather than pci bus/device/pin specific. Thus, when using a tunable to override the default IRQ settings, one now uses a single tunable to route an entire link rather than routing a single device that uses the link (which has great foot-shooting potential if the user tries to route the same link to two different IRQs using two different pci bus/device/pin hints). For example, to adjust the IRQ that \_SB_.LNKA uses, one would set 'hw.pci.link.LNKA.irq=10' from the loader. - As a side effect of having the link driver, unused link devices will now be disabled when they are probed. - The algorithm for choosing an IRQ for a link that doesn't already have an IRQ assigned is now much closer to the one used in $PIR routing. When a link is routed via an ISA IRQ, only known-good IRQs that the BIOS has already used are used for routing instead of using probabilities to guess at which IRQs are probably not used by an ISA device. One change from $PIR is that the SCI is always considered a viable ISA IRQ, so that if the BIOS does not setup any IRQs the kernel will degenerate to routing all interrupts over the SCI. For non ISA IRQs, interrupts are picked from the possible pool using a simplistic weighting algorithm. Tested by: ru, scottl, others on acpi@ Reviewed by: njl
2004-11-23 22:26:44 +00:00
struct link_res_request *req;
struct link *link;
Rework the ACPI PCI link code. - Use a new-bus device driver for the ACPI PCI link devices. The devices are called pci_linkX. The driver includes suspend/resume support so that the ACPI bridge drivers no longer have to poke the links to get them to handle suspend/resume. Also, the code to handle which IRQs a link is routed to and choosing an IRQ when a link is not already routed is all contained in the link driver. The PCI bridge drivers now ask the link driver which IRQ to use once they determine that a _PRT entry does not use a hardwired interrupt number. - The new link driver includes support for multiple IRQ resources per link device as well as preserving any non-IRQ resources when adjusting the IRQ that a link is routed to. - The entire approach to routing when using a link device is now link-centric rather than pci bus/device/pin specific. Thus, when using a tunable to override the default IRQ settings, one now uses a single tunable to route an entire link rather than routing a single device that uses the link (which has great foot-shooting potential if the user tries to route the same link to two different IRQs using two different pci bus/device/pin hints). For example, to adjust the IRQ that \_SB_.LNKA uses, one would set 'hw.pci.link.LNKA.irq=10' from the loader. - As a side effect of having the link driver, unused link devices will now be disabled when they are probed. - The algorithm for choosing an IRQ for a link that doesn't already have an IRQ assigned is now much closer to the one used in $PIR routing. When a link is routed via an ISA IRQ, only known-good IRQs that the BIOS has already used are used for routing instead of using probabilities to guess at which IRQs are probably not used by an ISA device. One change from $PIR is that the SCI is always considered a viable ISA IRQ, so that if the BIOS does not setup any IRQs the kernel will degenerate to routing all interrupts over the SCI. For non ISA IRQs, interrupts are picked from the possible pool using a simplistic weighting algorithm. Tested by: ru, scottl, others on acpi@ Reviewed by: njl
2004-11-23 22:26:44 +00:00
ACPI_SERIAL_ASSERT(pci_link);
req = (struct link_res_request *)context;
2005-11-01 22:44:08 +00:00
switch (res->Type) {
case ACPI_RESOURCE_TYPE_START_DEPENDENT:
switch (req->in_dpf) {
case DPF_OUTSIDE:
/* We've started the first DPF. */
req->in_dpf = DPF_FIRST;
break;
case DPF_FIRST:
/* We've started the second DPF. */
panic(
"%s: Multiple dependent functions within a current resource",
__func__);
break;
}
break;
2005-11-01 22:44:08 +00:00
case ACPI_RESOURCE_TYPE_END_DEPENDENT:
/* We are finished with DPF parsing. */
KASSERT(req->in_dpf != DPF_OUTSIDE,
("%s: end dpf when not parsing a dpf", __func__));
req->in_dpf = DPF_OUTSIDE;
break;
2005-11-01 22:44:08 +00:00
case ACPI_RESOURCE_TYPE_IRQ:
case ACPI_RESOURCE_TYPE_EXTENDED_IRQ:
KASSERT(req->link_index < req->sc->pl_num_links,
("%s: array boundary violation", __func__));
link = &req->sc->pl_links[req->link_index];
link->l_res_index = req->res_index;
req->link_index++;
req->res_index++;
/*
* Only use the current value if there's one IRQ. Some
* systems return multiple IRQs (which is nonsense for _CRS)
* when the link hasn't been programmed.
*/
2005-11-01 22:44:08 +00:00
if (res->Type == ACPI_RESOURCE_TYPE_IRQ) {
if (res->Data.Irq.InterruptCount == 1)
Rework the ACPI PCI link code. - Use a new-bus device driver for the ACPI PCI link devices. The devices are called pci_linkX. The driver includes suspend/resume support so that the ACPI bridge drivers no longer have to poke the links to get them to handle suspend/resume. Also, the code to handle which IRQs a link is routed to and choosing an IRQ when a link is not already routed is all contained in the link driver. The PCI bridge drivers now ask the link driver which IRQ to use once they determine that a _PRT entry does not use a hardwired interrupt number. - The new link driver includes support for multiple IRQ resources per link device as well as preserving any non-IRQ resources when adjusting the IRQ that a link is routed to. - The entire approach to routing when using a link device is now link-centric rather than pci bus/device/pin specific. Thus, when using a tunable to override the default IRQ settings, one now uses a single tunable to route an entire link rather than routing a single device that uses the link (which has great foot-shooting potential if the user tries to route the same link to two different IRQs using two different pci bus/device/pin hints). For example, to adjust the IRQ that \_SB_.LNKA uses, one would set 'hw.pci.link.LNKA.irq=10' from the loader. - As a side effect of having the link driver, unused link devices will now be disabled when they are probed. - The algorithm for choosing an IRQ for a link that doesn't already have an IRQ assigned is now much closer to the one used in $PIR routing. When a link is routed via an ISA IRQ, only known-good IRQs that the BIOS has already used are used for routing instead of using probabilities to guess at which IRQs are probably not used by an ISA device. One change from $PIR is that the SCI is always considered a viable ISA IRQ, so that if the BIOS does not setup any IRQs the kernel will degenerate to routing all interrupts over the SCI. For non ISA IRQs, interrupts are picked from the possible pool using a simplistic weighting algorithm. Tested by: ru, scottl, others on acpi@ Reviewed by: njl
2004-11-23 22:26:44 +00:00
link->l_irq = res->Data.Irq.Interrupts[0];
2005-11-01 22:44:08 +00:00
} else if (res->Data.ExtendedIrq.InterruptCount == 1)
Rework the ACPI PCI link code. - Use a new-bus device driver for the ACPI PCI link devices. The devices are called pci_linkX. The driver includes suspend/resume support so that the ACPI bridge drivers no longer have to poke the links to get them to handle suspend/resume. Also, the code to handle which IRQs a link is routed to and choosing an IRQ when a link is not already routed is all contained in the link driver. The PCI bridge drivers now ask the link driver which IRQ to use once they determine that a _PRT entry does not use a hardwired interrupt number. - The new link driver includes support for multiple IRQ resources per link device as well as preserving any non-IRQ resources when adjusting the IRQ that a link is routed to. - The entire approach to routing when using a link device is now link-centric rather than pci bus/device/pin specific. Thus, when using a tunable to override the default IRQ settings, one now uses a single tunable to route an entire link rather than routing a single device that uses the link (which has great foot-shooting potential if the user tries to route the same link to two different IRQs using two different pci bus/device/pin hints). For example, to adjust the IRQ that \_SB_.LNKA uses, one would set 'hw.pci.link.LNKA.irq=10' from the loader. - As a side effect of having the link driver, unused link devices will now be disabled when they are probed. - The algorithm for choosing an IRQ for a link that doesn't already have an IRQ assigned is now much closer to the one used in $PIR routing. When a link is routed via an ISA IRQ, only known-good IRQs that the BIOS has already used are used for routing instead of using probabilities to guess at which IRQs are probably not used by an ISA device. One change from $PIR is that the SCI is always considered a viable ISA IRQ, so that if the BIOS does not setup any IRQs the kernel will degenerate to routing all interrupts over the SCI. For non ISA IRQs, interrupts are picked from the possible pool using a simplistic weighting algorithm. Tested by: ru, scottl, others on acpi@ Reviewed by: njl
2004-11-23 22:26:44 +00:00
link->l_irq = res->Data.ExtendedIrq.Interrupts[0];
Rework the ACPI PCI link code. - Use a new-bus device driver for the ACPI PCI link devices. The devices are called pci_linkX. The driver includes suspend/resume support so that the ACPI bridge drivers no longer have to poke the links to get them to handle suspend/resume. Also, the code to handle which IRQs a link is routed to and choosing an IRQ when a link is not already routed is all contained in the link driver. The PCI bridge drivers now ask the link driver which IRQ to use once they determine that a _PRT entry does not use a hardwired interrupt number. - The new link driver includes support for multiple IRQ resources per link device as well as preserving any non-IRQ resources when adjusting the IRQ that a link is routed to. - The entire approach to routing when using a link device is now link-centric rather than pci bus/device/pin specific. Thus, when using a tunable to override the default IRQ settings, one now uses a single tunable to route an entire link rather than routing a single device that uses the link (which has great foot-shooting potential if the user tries to route the same link to two different IRQs using two different pci bus/device/pin hints). For example, to adjust the IRQ that \_SB_.LNKA uses, one would set 'hw.pci.link.LNKA.irq=10' from the loader. - As a side effect of having the link driver, unused link devices will now be disabled when they are probed. - The algorithm for choosing an IRQ for a link that doesn't already have an IRQ assigned is now much closer to the one used in $PIR routing. When a link is routed via an ISA IRQ, only known-good IRQs that the BIOS has already used are used for routing instead of using probabilities to guess at which IRQs are probably not used by an ISA device. One change from $PIR is that the SCI is always considered a viable ISA IRQ, so that if the BIOS does not setup any IRQs the kernel will degenerate to routing all interrupts over the SCI. For non ISA IRQs, interrupts are picked from the possible pool using a simplistic weighting algorithm. Tested by: ru, scottl, others on acpi@ Reviewed by: njl
2004-11-23 22:26:44 +00:00
/*
* An IRQ of zero means that the link isn't routed.
*/
if (link->l_irq == 0)
link->l_irq = PCI_INVALID_IRQ;
break;
default:
req->res_index++;
}
Rework the ACPI PCI link code. - Use a new-bus device driver for the ACPI PCI link devices. The devices are called pci_linkX. The driver includes suspend/resume support so that the ACPI bridge drivers no longer have to poke the links to get them to handle suspend/resume. Also, the code to handle which IRQs a link is routed to and choosing an IRQ when a link is not already routed is all contained in the link driver. The PCI bridge drivers now ask the link driver which IRQ to use once they determine that a _PRT entry does not use a hardwired interrupt number. - The new link driver includes support for multiple IRQ resources per link device as well as preserving any non-IRQ resources when adjusting the IRQ that a link is routed to. - The entire approach to routing when using a link device is now link-centric rather than pci bus/device/pin specific. Thus, when using a tunable to override the default IRQ settings, one now uses a single tunable to route an entire link rather than routing a single device that uses the link (which has great foot-shooting potential if the user tries to route the same link to two different IRQs using two different pci bus/device/pin hints). For example, to adjust the IRQ that \_SB_.LNKA uses, one would set 'hw.pci.link.LNKA.irq=10' from the loader. - As a side effect of having the link driver, unused link devices will now be disabled when they are probed. - The algorithm for choosing an IRQ for a link that doesn't already have an IRQ assigned is now much closer to the one used in $PIR routing. When a link is routed via an ISA IRQ, only known-good IRQs that the BIOS has already used are used for routing instead of using probabilities to guess at which IRQs are probably not used by an ISA device. One change from $PIR is that the SCI is always considered a viable ISA IRQ, so that if the BIOS does not setup any IRQs the kernel will degenerate to routing all interrupts over the SCI. For non ISA IRQs, interrupts are picked from the possible pool using a simplistic weighting algorithm. Tested by: ru, scottl, others on acpi@ Reviewed by: njl
2004-11-23 22:26:44 +00:00
return (AE_OK);
}
Rework the ACPI PCI link code. - Use a new-bus device driver for the ACPI PCI link devices. The devices are called pci_linkX. The driver includes suspend/resume support so that the ACPI bridge drivers no longer have to poke the links to get them to handle suspend/resume. Also, the code to handle which IRQs a link is routed to and choosing an IRQ when a link is not already routed is all contained in the link driver. The PCI bridge drivers now ask the link driver which IRQ to use once they determine that a _PRT entry does not use a hardwired interrupt number. - The new link driver includes support for multiple IRQ resources per link device as well as preserving any non-IRQ resources when adjusting the IRQ that a link is routed to. - The entire approach to routing when using a link device is now link-centric rather than pci bus/device/pin specific. Thus, when using a tunable to override the default IRQ settings, one now uses a single tunable to route an entire link rather than routing a single device that uses the link (which has great foot-shooting potential if the user tries to route the same link to two different IRQs using two different pci bus/device/pin hints). For example, to adjust the IRQ that \_SB_.LNKA uses, one would set 'hw.pci.link.LNKA.irq=10' from the loader. - As a side effect of having the link driver, unused link devices will now be disabled when they are probed. - The algorithm for choosing an IRQ for a link that doesn't already have an IRQ assigned is now much closer to the one used in $PIR routing. When a link is routed via an ISA IRQ, only known-good IRQs that the BIOS has already used are used for routing instead of using probabilities to guess at which IRQs are probably not used by an ISA device. One change from $PIR is that the SCI is always considered a viable ISA IRQ, so that if the BIOS does not setup any IRQs the kernel will degenerate to routing all interrupts over the SCI. For non ISA IRQs, interrupts are picked from the possible pool using a simplistic weighting algorithm. Tested by: ru, scottl, others on acpi@ Reviewed by: njl
2004-11-23 22:26:44 +00:00
/*
* Populate the set of possible IRQs for each device.
*/
static ACPI_STATUS
link_add_prs(ACPI_RESOURCE *res, void *context)
{
struct link_res_request *req;
struct link *link;
2005-11-01 22:44:08 +00:00
UINT8 *irqs = NULL;
UINT32 *ext_irqs = NULL;
int i, is_ext_irq = 1;
Rework the ACPI PCI link code. - Use a new-bus device driver for the ACPI PCI link devices. The devices are called pci_linkX. The driver includes suspend/resume support so that the ACPI bridge drivers no longer have to poke the links to get them to handle suspend/resume. Also, the code to handle which IRQs a link is routed to and choosing an IRQ when a link is not already routed is all contained in the link driver. The PCI bridge drivers now ask the link driver which IRQ to use once they determine that a _PRT entry does not use a hardwired interrupt number. - The new link driver includes support for multiple IRQ resources per link device as well as preserving any non-IRQ resources when adjusting the IRQ that a link is routed to. - The entire approach to routing when using a link device is now link-centric rather than pci bus/device/pin specific. Thus, when using a tunable to override the default IRQ settings, one now uses a single tunable to route an entire link rather than routing a single device that uses the link (which has great foot-shooting potential if the user tries to route the same link to two different IRQs using two different pci bus/device/pin hints). For example, to adjust the IRQ that \_SB_.LNKA uses, one would set 'hw.pci.link.LNKA.irq=10' from the loader. - As a side effect of having the link driver, unused link devices will now be disabled when they are probed. - The algorithm for choosing an IRQ for a link that doesn't already have an IRQ assigned is now much closer to the one used in $PIR routing. When a link is routed via an ISA IRQ, only known-good IRQs that the BIOS has already used are used for routing instead of using probabilities to guess at which IRQs are probably not used by an ISA device. One change from $PIR is that the SCI is always considered a viable ISA IRQ, so that if the BIOS does not setup any IRQs the kernel will degenerate to routing all interrupts over the SCI. For non ISA IRQs, interrupts are picked from the possible pool using a simplistic weighting algorithm. Tested by: ru, scottl, others on acpi@ Reviewed by: njl
2004-11-23 22:26:44 +00:00
ACPI_SERIAL_ASSERT(pci_link);
req = (struct link_res_request *)context;
2005-11-01 22:44:08 +00:00
switch (res->Type) {
case ACPI_RESOURCE_TYPE_START_DEPENDENT:
switch (req->in_dpf) {
case DPF_OUTSIDE:
/* We've started the first DPF. */
req->in_dpf = DPF_FIRST;
break;
case DPF_FIRST:
/* We've started the second DPF. */
req->in_dpf = DPF_IGNORE;
break;
}
break;
2005-11-01 22:44:08 +00:00
case ACPI_RESOURCE_TYPE_END_DEPENDENT:
/* We are finished with DPF parsing. */
KASSERT(req->in_dpf != DPF_OUTSIDE,
("%s: end dpf when not parsing a dpf", __func__));
req->in_dpf = DPF_OUTSIDE;
break;
2005-11-01 22:44:08 +00:00
case ACPI_RESOURCE_TYPE_IRQ:
is_ext_irq = 0;
/* fall through */
case ACPI_RESOURCE_TYPE_EXTENDED_IRQ:
/*
* Don't parse resources if we are in a DPF set that we are
* ignoring.
*/
if (req->in_dpf == DPF_IGNORE)
break;
KASSERT(req->link_index < req->sc->pl_num_links,
("%s: array boundary violation", __func__));
link = &req->sc->pl_links[req->link_index];
if (link->l_res_index == -1) {
KASSERT(req->sc->pl_crs_bad,
("res_index should be set"));
link->l_res_index = req->res_index;
}
req->link_index++;
req->res_index++;
Rework the ACPI PCI link code. - Use a new-bus device driver for the ACPI PCI link devices. The devices are called pci_linkX. The driver includes suspend/resume support so that the ACPI bridge drivers no longer have to poke the links to get them to handle suspend/resume. Also, the code to handle which IRQs a link is routed to and choosing an IRQ when a link is not already routed is all contained in the link driver. The PCI bridge drivers now ask the link driver which IRQ to use once they determine that a _PRT entry does not use a hardwired interrupt number. - The new link driver includes support for multiple IRQ resources per link device as well as preserving any non-IRQ resources when adjusting the IRQ that a link is routed to. - The entire approach to routing when using a link device is now link-centric rather than pci bus/device/pin specific. Thus, when using a tunable to override the default IRQ settings, one now uses a single tunable to route an entire link rather than routing a single device that uses the link (which has great foot-shooting potential if the user tries to route the same link to two different IRQs using two different pci bus/device/pin hints). For example, to adjust the IRQ that \_SB_.LNKA uses, one would set 'hw.pci.link.LNKA.irq=10' from the loader. - As a side effect of having the link driver, unused link devices will now be disabled when they are probed. - The algorithm for choosing an IRQ for a link that doesn't already have an IRQ assigned is now much closer to the one used in $PIR routing. When a link is routed via an ISA IRQ, only known-good IRQs that the BIOS has already used are used for routing instead of using probabilities to guess at which IRQs are probably not used by an ISA device. One change from $PIR is that the SCI is always considered a viable ISA IRQ, so that if the BIOS does not setup any IRQs the kernel will degenerate to routing all interrupts over the SCI. For non ISA IRQs, interrupts are picked from the possible pool using a simplistic weighting algorithm. Tested by: ru, scottl, others on acpi@ Reviewed by: njl
2004-11-23 22:26:44 +00:00
/*
* Stash a copy of the resource for later use when doing
* _SRS.
*/
bcopy(res, &link->l_prs_template, sizeof(ACPI_RESOURCE));
2005-11-01 22:44:08 +00:00
if (is_ext_irq) {
Rework the ACPI PCI link code. - Use a new-bus device driver for the ACPI PCI link devices. The devices are called pci_linkX. The driver includes suspend/resume support so that the ACPI bridge drivers no longer have to poke the links to get them to handle suspend/resume. Also, the code to handle which IRQs a link is routed to and choosing an IRQ when a link is not already routed is all contained in the link driver. The PCI bridge drivers now ask the link driver which IRQ to use once they determine that a _PRT entry does not use a hardwired interrupt number. - The new link driver includes support for multiple IRQ resources per link device as well as preserving any non-IRQ resources when adjusting the IRQ that a link is routed to. - The entire approach to routing when using a link device is now link-centric rather than pci bus/device/pin specific. Thus, when using a tunable to override the default IRQ settings, one now uses a single tunable to route an entire link rather than routing a single device that uses the link (which has great foot-shooting potential if the user tries to route the same link to two different IRQs using two different pci bus/device/pin hints). For example, to adjust the IRQ that \_SB_.LNKA uses, one would set 'hw.pci.link.LNKA.irq=10' from the loader. - As a side effect of having the link driver, unused link devices will now be disabled when they are probed. - The algorithm for choosing an IRQ for a link that doesn't already have an IRQ assigned is now much closer to the one used in $PIR routing. When a link is routed via an ISA IRQ, only known-good IRQs that the BIOS has already used are used for routing instead of using probabilities to guess at which IRQs are probably not used by an ISA device. One change from $PIR is that the SCI is always considered a viable ISA IRQ, so that if the BIOS does not setup any IRQs the kernel will degenerate to routing all interrupts over the SCI. For non ISA IRQs, interrupts are picked from the possible pool using a simplistic weighting algorithm. Tested by: ru, scottl, others on acpi@ Reviewed by: njl
2004-11-23 22:26:44 +00:00
link->l_num_irqs =
2005-11-01 22:44:08 +00:00
res->Data.ExtendedIrq.InterruptCount;
ext_irqs = res->Data.ExtendedIrq.Interrupts;
} else {
link->l_num_irqs = res->Data.Irq.InterruptCount;
irqs = res->Data.Irq.Interrupts;
}
Rework the ACPI PCI link code. - Use a new-bus device driver for the ACPI PCI link devices. The devices are called pci_linkX. The driver includes suspend/resume support so that the ACPI bridge drivers no longer have to poke the links to get them to handle suspend/resume. Also, the code to handle which IRQs a link is routed to and choosing an IRQ when a link is not already routed is all contained in the link driver. The PCI bridge drivers now ask the link driver which IRQ to use once they determine that a _PRT entry does not use a hardwired interrupt number. - The new link driver includes support for multiple IRQ resources per link device as well as preserving any non-IRQ resources when adjusting the IRQ that a link is routed to. - The entire approach to routing when using a link device is now link-centric rather than pci bus/device/pin specific. Thus, when using a tunable to override the default IRQ settings, one now uses a single tunable to route an entire link rather than routing a single device that uses the link (which has great foot-shooting potential if the user tries to route the same link to two different IRQs using two different pci bus/device/pin hints). For example, to adjust the IRQ that \_SB_.LNKA uses, one would set 'hw.pci.link.LNKA.irq=10' from the loader. - As a side effect of having the link driver, unused link devices will now be disabled when they are probed. - The algorithm for choosing an IRQ for a link that doesn't already have an IRQ assigned is now much closer to the one used in $PIR routing. When a link is routed via an ISA IRQ, only known-good IRQs that the BIOS has already used are used for routing instead of using probabilities to guess at which IRQs are probably not used by an ISA device. One change from $PIR is that the SCI is always considered a viable ISA IRQ, so that if the BIOS does not setup any IRQs the kernel will degenerate to routing all interrupts over the SCI. For non ISA IRQs, interrupts are picked from the possible pool using a simplistic weighting algorithm. Tested by: ru, scottl, others on acpi@ Reviewed by: njl
2004-11-23 22:26:44 +00:00
if (link->l_num_irqs == 0)
break;
/*
* Save a list of the valid IRQs. Also, if all of the
* valid IRQs are ISA IRQs, then mark this link as
* routed via an ISA interrupt.
*/
link->l_isa_irq = TRUE;
Rework the ACPI PCI link code. - Use a new-bus device driver for the ACPI PCI link devices. The devices are called pci_linkX. The driver includes suspend/resume support so that the ACPI bridge drivers no longer have to poke the links to get them to handle suspend/resume. Also, the code to handle which IRQs a link is routed to and choosing an IRQ when a link is not already routed is all contained in the link driver. The PCI bridge drivers now ask the link driver which IRQ to use once they determine that a _PRT entry does not use a hardwired interrupt number. - The new link driver includes support for multiple IRQ resources per link device as well as preserving any non-IRQ resources when adjusting the IRQ that a link is routed to. - The entire approach to routing when using a link device is now link-centric rather than pci bus/device/pin specific. Thus, when using a tunable to override the default IRQ settings, one now uses a single tunable to route an entire link rather than routing a single device that uses the link (which has great foot-shooting potential if the user tries to route the same link to two different IRQs using two different pci bus/device/pin hints). For example, to adjust the IRQ that \_SB_.LNKA uses, one would set 'hw.pci.link.LNKA.irq=10' from the loader. - As a side effect of having the link driver, unused link devices will now be disabled when they are probed. - The algorithm for choosing an IRQ for a link that doesn't already have an IRQ assigned is now much closer to the one used in $PIR routing. When a link is routed via an ISA IRQ, only known-good IRQs that the BIOS has already used are used for routing instead of using probabilities to guess at which IRQs are probably not used by an ISA device. One change from $PIR is that the SCI is always considered a viable ISA IRQ, so that if the BIOS does not setup any IRQs the kernel will degenerate to routing all interrupts over the SCI. For non ISA IRQs, interrupts are picked from the possible pool using a simplistic weighting algorithm. Tested by: ru, scottl, others on acpi@ Reviewed by: njl
2004-11-23 22:26:44 +00:00
link->l_irqs = malloc(sizeof(int) * link->l_num_irqs,
M_PCI_LINK, M_WAITOK | M_ZERO);
for (i = 0; i < link->l_num_irqs; i++) {
2005-11-01 22:44:08 +00:00
if (is_ext_irq) {
link->l_irqs[i] = ext_irqs[i];
if (ext_irqs[i] >= NUM_ISA_INTERRUPTS)
link->l_isa_irq = FALSE;
} else {
link->l_irqs[i] = irqs[i];
if (irqs[i] >= NUM_ISA_INTERRUPTS)
link->l_isa_irq = FALSE;
}
}
Rework the ACPI PCI link code. - Use a new-bus device driver for the ACPI PCI link devices. The devices are called pci_linkX. The driver includes suspend/resume support so that the ACPI bridge drivers no longer have to poke the links to get them to handle suspend/resume. Also, the code to handle which IRQs a link is routed to and choosing an IRQ when a link is not already routed is all contained in the link driver. The PCI bridge drivers now ask the link driver which IRQ to use once they determine that a _PRT entry does not use a hardwired interrupt number. - The new link driver includes support for multiple IRQ resources per link device as well as preserving any non-IRQ resources when adjusting the IRQ that a link is routed to. - The entire approach to routing when using a link device is now link-centric rather than pci bus/device/pin specific. Thus, when using a tunable to override the default IRQ settings, one now uses a single tunable to route an entire link rather than routing a single device that uses the link (which has great foot-shooting potential if the user tries to route the same link to two different IRQs using two different pci bus/device/pin hints). For example, to adjust the IRQ that \_SB_.LNKA uses, one would set 'hw.pci.link.LNKA.irq=10' from the loader. - As a side effect of having the link driver, unused link devices will now be disabled when they are probed. - The algorithm for choosing an IRQ for a link that doesn't already have an IRQ assigned is now much closer to the one used in $PIR routing. When a link is routed via an ISA IRQ, only known-good IRQs that the BIOS has already used are used for routing instead of using probabilities to guess at which IRQs are probably not used by an ISA device. One change from $PIR is that the SCI is always considered a viable ISA IRQ, so that if the BIOS does not setup any IRQs the kernel will degenerate to routing all interrupts over the SCI. For non ISA IRQs, interrupts are picked from the possible pool using a simplistic weighting algorithm. Tested by: ru, scottl, others on acpi@ Reviewed by: njl
2004-11-23 22:26:44 +00:00
break;
default:
if (req->in_dpf == DPF_IGNORE)
break;
if (req->sc->pl_crs_bad)
device_printf(req->sc->pl_dev,
"Warning: possible resource %d will be lost during _SRS\n",
req->res_index);
req->res_index++;
}
Rework the ACPI PCI link code. - Use a new-bus device driver for the ACPI PCI link devices. The devices are called pci_linkX. The driver includes suspend/resume support so that the ACPI bridge drivers no longer have to poke the links to get them to handle suspend/resume. Also, the code to handle which IRQs a link is routed to and choosing an IRQ when a link is not already routed is all contained in the link driver. The PCI bridge drivers now ask the link driver which IRQ to use once they determine that a _PRT entry does not use a hardwired interrupt number. - The new link driver includes support for multiple IRQ resources per link device as well as preserving any non-IRQ resources when adjusting the IRQ that a link is routed to. - The entire approach to routing when using a link device is now link-centric rather than pci bus/device/pin specific. Thus, when using a tunable to override the default IRQ settings, one now uses a single tunable to route an entire link rather than routing a single device that uses the link (which has great foot-shooting potential if the user tries to route the same link to two different IRQs using two different pci bus/device/pin hints). For example, to adjust the IRQ that \_SB_.LNKA uses, one would set 'hw.pci.link.LNKA.irq=10' from the loader. - As a side effect of having the link driver, unused link devices will now be disabled when they are probed. - The algorithm for choosing an IRQ for a link that doesn't already have an IRQ assigned is now much closer to the one used in $PIR routing. When a link is routed via an ISA IRQ, only known-good IRQs that the BIOS has already used are used for routing instead of using probabilities to guess at which IRQs are probably not used by an ISA device. One change from $PIR is that the SCI is always considered a viable ISA IRQ, so that if the BIOS does not setup any IRQs the kernel will degenerate to routing all interrupts over the SCI. For non ISA IRQs, interrupts are picked from the possible pool using a simplistic weighting algorithm. Tested by: ru, scottl, others on acpi@ Reviewed by: njl
2004-11-23 22:26:44 +00:00
return (AE_OK);
}
Rework the ACPI PCI link code. - Use a new-bus device driver for the ACPI PCI link devices. The devices are called pci_linkX. The driver includes suspend/resume support so that the ACPI bridge drivers no longer have to poke the links to get them to handle suspend/resume. Also, the code to handle which IRQs a link is routed to and choosing an IRQ when a link is not already routed is all contained in the link driver. The PCI bridge drivers now ask the link driver which IRQ to use once they determine that a _PRT entry does not use a hardwired interrupt number. - The new link driver includes support for multiple IRQ resources per link device as well as preserving any non-IRQ resources when adjusting the IRQ that a link is routed to. - The entire approach to routing when using a link device is now link-centric rather than pci bus/device/pin specific. Thus, when using a tunable to override the default IRQ settings, one now uses a single tunable to route an entire link rather than routing a single device that uses the link (which has great foot-shooting potential if the user tries to route the same link to two different IRQs using two different pci bus/device/pin hints). For example, to adjust the IRQ that \_SB_.LNKA uses, one would set 'hw.pci.link.LNKA.irq=10' from the loader. - As a side effect of having the link driver, unused link devices will now be disabled when they are probed. - The algorithm for choosing an IRQ for a link that doesn't already have an IRQ assigned is now much closer to the one used in $PIR routing. When a link is routed via an ISA IRQ, only known-good IRQs that the BIOS has already used are used for routing instead of using probabilities to guess at which IRQs are probably not used by an ISA device. One change from $PIR is that the SCI is always considered a viable ISA IRQ, so that if the BIOS does not setup any IRQs the kernel will degenerate to routing all interrupts over the SCI. For non ISA IRQs, interrupts are picked from the possible pool using a simplistic weighting algorithm. Tested by: ru, scottl, others on acpi@ Reviewed by: njl
2004-11-23 22:26:44 +00:00
static int
link_valid_irq(struct link *link, int irq)
{
Rework the ACPI PCI link code. - Use a new-bus device driver for the ACPI PCI link devices. The devices are called pci_linkX. The driver includes suspend/resume support so that the ACPI bridge drivers no longer have to poke the links to get them to handle suspend/resume. Also, the code to handle which IRQs a link is routed to and choosing an IRQ when a link is not already routed is all contained in the link driver. The PCI bridge drivers now ask the link driver which IRQ to use once they determine that a _PRT entry does not use a hardwired interrupt number. - The new link driver includes support for multiple IRQ resources per link device as well as preserving any non-IRQ resources when adjusting the IRQ that a link is routed to. - The entire approach to routing when using a link device is now link-centric rather than pci bus/device/pin specific. Thus, when using a tunable to override the default IRQ settings, one now uses a single tunable to route an entire link rather than routing a single device that uses the link (which has great foot-shooting potential if the user tries to route the same link to two different IRQs using two different pci bus/device/pin hints). For example, to adjust the IRQ that \_SB_.LNKA uses, one would set 'hw.pci.link.LNKA.irq=10' from the loader. - As a side effect of having the link driver, unused link devices will now be disabled when they are probed. - The algorithm for choosing an IRQ for a link that doesn't already have an IRQ assigned is now much closer to the one used in $PIR routing. When a link is routed via an ISA IRQ, only known-good IRQs that the BIOS has already used are used for routing instead of using probabilities to guess at which IRQs are probably not used by an ISA device. One change from $PIR is that the SCI is always considered a viable ISA IRQ, so that if the BIOS does not setup any IRQs the kernel will degenerate to routing all interrupts over the SCI. For non ISA IRQs, interrupts are picked from the possible pool using a simplistic weighting algorithm. Tested by: ru, scottl, others on acpi@ Reviewed by: njl
2004-11-23 22:26:44 +00:00
int i;
Rework the ACPI PCI link code. - Use a new-bus device driver for the ACPI PCI link devices. The devices are called pci_linkX. The driver includes suspend/resume support so that the ACPI bridge drivers no longer have to poke the links to get them to handle suspend/resume. Also, the code to handle which IRQs a link is routed to and choosing an IRQ when a link is not already routed is all contained in the link driver. The PCI bridge drivers now ask the link driver which IRQ to use once they determine that a _PRT entry does not use a hardwired interrupt number. - The new link driver includes support for multiple IRQ resources per link device as well as preserving any non-IRQ resources when adjusting the IRQ that a link is routed to. - The entire approach to routing when using a link device is now link-centric rather than pci bus/device/pin specific. Thus, when using a tunable to override the default IRQ settings, one now uses a single tunable to route an entire link rather than routing a single device that uses the link (which has great foot-shooting potential if the user tries to route the same link to two different IRQs using two different pci bus/device/pin hints). For example, to adjust the IRQ that \_SB_.LNKA uses, one would set 'hw.pci.link.LNKA.irq=10' from the loader. - As a side effect of having the link driver, unused link devices will now be disabled when they are probed. - The algorithm for choosing an IRQ for a link that doesn't already have an IRQ assigned is now much closer to the one used in $PIR routing. When a link is routed via an ISA IRQ, only known-good IRQs that the BIOS has already used are used for routing instead of using probabilities to guess at which IRQs are probably not used by an ISA device. One change from $PIR is that the SCI is always considered a viable ISA IRQ, so that if the BIOS does not setup any IRQs the kernel will degenerate to routing all interrupts over the SCI. For non ISA IRQs, interrupts are picked from the possible pool using a simplistic weighting algorithm. Tested by: ru, scottl, others on acpi@ Reviewed by: njl
2004-11-23 22:26:44 +00:00
ACPI_SERIAL_ASSERT(pci_link);
Rework the ACPI PCI link code. - Use a new-bus device driver for the ACPI PCI link devices. The devices are called pci_linkX. The driver includes suspend/resume support so that the ACPI bridge drivers no longer have to poke the links to get them to handle suspend/resume. Also, the code to handle which IRQs a link is routed to and choosing an IRQ when a link is not already routed is all contained in the link driver. The PCI bridge drivers now ask the link driver which IRQ to use once they determine that a _PRT entry does not use a hardwired interrupt number. - The new link driver includes support for multiple IRQ resources per link device as well as preserving any non-IRQ resources when adjusting the IRQ that a link is routed to. - The entire approach to routing when using a link device is now link-centric rather than pci bus/device/pin specific. Thus, when using a tunable to override the default IRQ settings, one now uses a single tunable to route an entire link rather than routing a single device that uses the link (which has great foot-shooting potential if the user tries to route the same link to two different IRQs using two different pci bus/device/pin hints). For example, to adjust the IRQ that \_SB_.LNKA uses, one would set 'hw.pci.link.LNKA.irq=10' from the loader. - As a side effect of having the link driver, unused link devices will now be disabled when they are probed. - The algorithm for choosing an IRQ for a link that doesn't already have an IRQ assigned is now much closer to the one used in $PIR routing. When a link is routed via an ISA IRQ, only known-good IRQs that the BIOS has already used are used for routing instead of using probabilities to guess at which IRQs are probably not used by an ISA device. One change from $PIR is that the SCI is always considered a viable ISA IRQ, so that if the BIOS does not setup any IRQs the kernel will degenerate to routing all interrupts over the SCI. For non ISA IRQs, interrupts are picked from the possible pool using a simplistic weighting algorithm. Tested by: ru, scottl, others on acpi@ Reviewed by: njl
2004-11-23 22:26:44 +00:00
/* Invalid interrupts are never valid. */
if (!PCI_INTERRUPT_VALID(irq))
return (FALSE);
Rework the ACPI PCI link code. - Use a new-bus device driver for the ACPI PCI link devices. The devices are called pci_linkX. The driver includes suspend/resume support so that the ACPI bridge drivers no longer have to poke the links to get them to handle suspend/resume. Also, the code to handle which IRQs a link is routed to and choosing an IRQ when a link is not already routed is all contained in the link driver. The PCI bridge drivers now ask the link driver which IRQ to use once they determine that a _PRT entry does not use a hardwired interrupt number. - The new link driver includes support for multiple IRQ resources per link device as well as preserving any non-IRQ resources when adjusting the IRQ that a link is routed to. - The entire approach to routing when using a link device is now link-centric rather than pci bus/device/pin specific. Thus, when using a tunable to override the default IRQ settings, one now uses a single tunable to route an entire link rather than routing a single device that uses the link (which has great foot-shooting potential if the user tries to route the same link to two different IRQs using two different pci bus/device/pin hints). For example, to adjust the IRQ that \_SB_.LNKA uses, one would set 'hw.pci.link.LNKA.irq=10' from the loader. - As a side effect of having the link driver, unused link devices will now be disabled when they are probed. - The algorithm for choosing an IRQ for a link that doesn't already have an IRQ assigned is now much closer to the one used in $PIR routing. When a link is routed via an ISA IRQ, only known-good IRQs that the BIOS has already used are used for routing instead of using probabilities to guess at which IRQs are probably not used by an ISA device. One change from $PIR is that the SCI is always considered a viable ISA IRQ, so that if the BIOS does not setup any IRQs the kernel will degenerate to routing all interrupts over the SCI. For non ISA IRQs, interrupts are picked from the possible pool using a simplistic weighting algorithm. Tested by: ru, scottl, others on acpi@ Reviewed by: njl
2004-11-23 22:26:44 +00:00
/* Any interrupt in the list of possible interrupts is valid. */
for (i = 0; i < link->l_num_irqs; i++)
if (link->l_irqs[i] == irq)
return (TRUE);
Rework the ACPI PCI link code. - Use a new-bus device driver for the ACPI PCI link devices. The devices are called pci_linkX. The driver includes suspend/resume support so that the ACPI bridge drivers no longer have to poke the links to get them to handle suspend/resume. Also, the code to handle which IRQs a link is routed to and choosing an IRQ when a link is not already routed is all contained in the link driver. The PCI bridge drivers now ask the link driver which IRQ to use once they determine that a _PRT entry does not use a hardwired interrupt number. - The new link driver includes support for multiple IRQ resources per link device as well as preserving any non-IRQ resources when adjusting the IRQ that a link is routed to. - The entire approach to routing when using a link device is now link-centric rather than pci bus/device/pin specific. Thus, when using a tunable to override the default IRQ settings, one now uses a single tunable to route an entire link rather than routing a single device that uses the link (which has great foot-shooting potential if the user tries to route the same link to two different IRQs using two different pci bus/device/pin hints). For example, to adjust the IRQ that \_SB_.LNKA uses, one would set 'hw.pci.link.LNKA.irq=10' from the loader. - As a side effect of having the link driver, unused link devices will now be disabled when they are probed. - The algorithm for choosing an IRQ for a link that doesn't already have an IRQ assigned is now much closer to the one used in $PIR routing. When a link is routed via an ISA IRQ, only known-good IRQs that the BIOS has already used are used for routing instead of using probabilities to guess at which IRQs are probably not used by an ISA device. One change from $PIR is that the SCI is always considered a viable ISA IRQ, so that if the BIOS does not setup any IRQs the kernel will degenerate to routing all interrupts over the SCI. For non ISA IRQs, interrupts are picked from the possible pool using a simplistic weighting algorithm. Tested by: ru, scottl, others on acpi@ Reviewed by: njl
2004-11-23 22:26:44 +00:00
/*
* For links routed via an ISA interrupt, if the SCI is routed via
* an ISA interrupt, the SCI is always treated as a valid IRQ.
*/
if (link->l_isa_irq && AcpiGbl_FADT->SciInt == irq &&
irq < NUM_ISA_INTERRUPTS)
return (TRUE);
Rework the ACPI PCI link code. - Use a new-bus device driver for the ACPI PCI link devices. The devices are called pci_linkX. The driver includes suspend/resume support so that the ACPI bridge drivers no longer have to poke the links to get them to handle suspend/resume. Also, the code to handle which IRQs a link is routed to and choosing an IRQ when a link is not already routed is all contained in the link driver. The PCI bridge drivers now ask the link driver which IRQ to use once they determine that a _PRT entry does not use a hardwired interrupt number. - The new link driver includes support for multiple IRQ resources per link device as well as preserving any non-IRQ resources when adjusting the IRQ that a link is routed to. - The entire approach to routing when using a link device is now link-centric rather than pci bus/device/pin specific. Thus, when using a tunable to override the default IRQ settings, one now uses a single tunable to route an entire link rather than routing a single device that uses the link (which has great foot-shooting potential if the user tries to route the same link to two different IRQs using two different pci bus/device/pin hints). For example, to adjust the IRQ that \_SB_.LNKA uses, one would set 'hw.pci.link.LNKA.irq=10' from the loader. - As a side effect of having the link driver, unused link devices will now be disabled when they are probed. - The algorithm for choosing an IRQ for a link that doesn't already have an IRQ assigned is now much closer to the one used in $PIR routing. When a link is routed via an ISA IRQ, only known-good IRQs that the BIOS has already used are used for routing instead of using probabilities to guess at which IRQs are probably not used by an ISA device. One change from $PIR is that the SCI is always considered a viable ISA IRQ, so that if the BIOS does not setup any IRQs the kernel will degenerate to routing all interrupts over the SCI. For non ISA IRQs, interrupts are picked from the possible pool using a simplistic weighting algorithm. Tested by: ru, scottl, others on acpi@ Reviewed by: njl
2004-11-23 22:26:44 +00:00
/* If the interrupt wasn't found in the list it is not valid. */
return (FALSE);
}
Rework the ACPI PCI link code. - Use a new-bus device driver for the ACPI PCI link devices. The devices are called pci_linkX. The driver includes suspend/resume support so that the ACPI bridge drivers no longer have to poke the links to get them to handle suspend/resume. Also, the code to handle which IRQs a link is routed to and choosing an IRQ when a link is not already routed is all contained in the link driver. The PCI bridge drivers now ask the link driver which IRQ to use once they determine that a _PRT entry does not use a hardwired interrupt number. - The new link driver includes support for multiple IRQ resources per link device as well as preserving any non-IRQ resources when adjusting the IRQ that a link is routed to. - The entire approach to routing when using a link device is now link-centric rather than pci bus/device/pin specific. Thus, when using a tunable to override the default IRQ settings, one now uses a single tunable to route an entire link rather than routing a single device that uses the link (which has great foot-shooting potential if the user tries to route the same link to two different IRQs using two different pci bus/device/pin hints). For example, to adjust the IRQ that \_SB_.LNKA uses, one would set 'hw.pci.link.LNKA.irq=10' from the loader. - As a side effect of having the link driver, unused link devices will now be disabled when they are probed. - The algorithm for choosing an IRQ for a link that doesn't already have an IRQ assigned is now much closer to the one used in $PIR routing. When a link is routed via an ISA IRQ, only known-good IRQs that the BIOS has already used are used for routing instead of using probabilities to guess at which IRQs are probably not used by an ISA device. One change from $PIR is that the SCI is always considered a viable ISA IRQ, so that if the BIOS does not setup any IRQs the kernel will degenerate to routing all interrupts over the SCI. For non ISA IRQs, interrupts are picked from the possible pool using a simplistic weighting algorithm. Tested by: ru, scottl, others on acpi@ Reviewed by: njl
2004-11-23 22:26:44 +00:00
static void
acpi_pci_link_dump(struct acpi_pci_link_softc *sc)
{
Rework the ACPI PCI link code. - Use a new-bus device driver for the ACPI PCI link devices. The devices are called pci_linkX. The driver includes suspend/resume support so that the ACPI bridge drivers no longer have to poke the links to get them to handle suspend/resume. Also, the code to handle which IRQs a link is routed to and choosing an IRQ when a link is not already routed is all contained in the link driver. The PCI bridge drivers now ask the link driver which IRQ to use once they determine that a _PRT entry does not use a hardwired interrupt number. - The new link driver includes support for multiple IRQ resources per link device as well as preserving any non-IRQ resources when adjusting the IRQ that a link is routed to. - The entire approach to routing when using a link device is now link-centric rather than pci bus/device/pin specific. Thus, when using a tunable to override the default IRQ settings, one now uses a single tunable to route an entire link rather than routing a single device that uses the link (which has great foot-shooting potential if the user tries to route the same link to two different IRQs using two different pci bus/device/pin hints). For example, to adjust the IRQ that \_SB_.LNKA uses, one would set 'hw.pci.link.LNKA.irq=10' from the loader. - As a side effect of having the link driver, unused link devices will now be disabled when they are probed. - The algorithm for choosing an IRQ for a link that doesn't already have an IRQ assigned is now much closer to the one used in $PIR routing. When a link is routed via an ISA IRQ, only known-good IRQs that the BIOS has already used are used for routing instead of using probabilities to guess at which IRQs are probably not used by an ISA device. One change from $PIR is that the SCI is always considered a viable ISA IRQ, so that if the BIOS does not setup any IRQs the kernel will degenerate to routing all interrupts over the SCI. For non ISA IRQs, interrupts are picked from the possible pool using a simplistic weighting algorithm. Tested by: ru, scottl, others on acpi@ Reviewed by: njl
2004-11-23 22:26:44 +00:00
struct link *link;
int i, j;
ACPI_SERIAL_ASSERT(pci_link);
Rework the ACPI PCI link code. - Use a new-bus device driver for the ACPI PCI link devices. The devices are called pci_linkX. The driver includes suspend/resume support so that the ACPI bridge drivers no longer have to poke the links to get them to handle suspend/resume. Also, the code to handle which IRQs a link is routed to and choosing an IRQ when a link is not already routed is all contained in the link driver. The PCI bridge drivers now ask the link driver which IRQ to use once they determine that a _PRT entry does not use a hardwired interrupt number. - The new link driver includes support for multiple IRQ resources per link device as well as preserving any non-IRQ resources when adjusting the IRQ that a link is routed to. - The entire approach to routing when using a link device is now link-centric rather than pci bus/device/pin specific. Thus, when using a tunable to override the default IRQ settings, one now uses a single tunable to route an entire link rather than routing a single device that uses the link (which has great foot-shooting potential if the user tries to route the same link to two different IRQs using two different pci bus/device/pin hints). For example, to adjust the IRQ that \_SB_.LNKA uses, one would set 'hw.pci.link.LNKA.irq=10' from the loader. - As a side effect of having the link driver, unused link devices will now be disabled when they are probed. - The algorithm for choosing an IRQ for a link that doesn't already have an IRQ assigned is now much closer to the one used in $PIR routing. When a link is routed via an ISA IRQ, only known-good IRQs that the BIOS has already used are used for routing instead of using probabilities to guess at which IRQs are probably not used by an ISA device. One change from $PIR is that the SCI is always considered a viable ISA IRQ, so that if the BIOS does not setup any IRQs the kernel will degenerate to routing all interrupts over the SCI. For non ISA IRQs, interrupts are picked from the possible pool using a simplistic weighting algorithm. Tested by: ru, scottl, others on acpi@ Reviewed by: njl
2004-11-23 22:26:44 +00:00
printf("Index IRQ Rtd Ref IRQs\n");
for (i = 0; i < sc->pl_num_links; i++) {
link = &sc->pl_links[i];
printf("%5d %3d %c %3d ", i, link->l_irq,
link->l_routed ? 'Y' : 'N', link->l_references);
if (link->l_num_irqs == 0)
printf(" none");
else for (j = 0; j < link->l_num_irqs; j++)
printf(" %d", link->l_irqs[j]);
printf("\n");
}
Rework the ACPI PCI link code. - Use a new-bus device driver for the ACPI PCI link devices. The devices are called pci_linkX. The driver includes suspend/resume support so that the ACPI bridge drivers no longer have to poke the links to get them to handle suspend/resume. Also, the code to handle which IRQs a link is routed to and choosing an IRQ when a link is not already routed is all contained in the link driver. The PCI bridge drivers now ask the link driver which IRQ to use once they determine that a _PRT entry does not use a hardwired interrupt number. - The new link driver includes support for multiple IRQ resources per link device as well as preserving any non-IRQ resources when adjusting the IRQ that a link is routed to. - The entire approach to routing when using a link device is now link-centric rather than pci bus/device/pin specific. Thus, when using a tunable to override the default IRQ settings, one now uses a single tunable to route an entire link rather than routing a single device that uses the link (which has great foot-shooting potential if the user tries to route the same link to two different IRQs using two different pci bus/device/pin hints). For example, to adjust the IRQ that \_SB_.LNKA uses, one would set 'hw.pci.link.LNKA.irq=10' from the loader. - As a side effect of having the link driver, unused link devices will now be disabled when they are probed. - The algorithm for choosing an IRQ for a link that doesn't already have an IRQ assigned is now much closer to the one used in $PIR routing. When a link is routed via an ISA IRQ, only known-good IRQs that the BIOS has already used are used for routing instead of using probabilities to guess at which IRQs are probably not used by an ISA device. One change from $PIR is that the SCI is always considered a viable ISA IRQ, so that if the BIOS does not setup any IRQs the kernel will degenerate to routing all interrupts over the SCI. For non ISA IRQs, interrupts are picked from the possible pool using a simplistic weighting algorithm. Tested by: ru, scottl, others on acpi@ Reviewed by: njl
2004-11-23 22:26:44 +00:00
}
Rework the ACPI PCI link code. - Use a new-bus device driver for the ACPI PCI link devices. The devices are called pci_linkX. The driver includes suspend/resume support so that the ACPI bridge drivers no longer have to poke the links to get them to handle suspend/resume. Also, the code to handle which IRQs a link is routed to and choosing an IRQ when a link is not already routed is all contained in the link driver. The PCI bridge drivers now ask the link driver which IRQ to use once they determine that a _PRT entry does not use a hardwired interrupt number. - The new link driver includes support for multiple IRQ resources per link device as well as preserving any non-IRQ resources when adjusting the IRQ that a link is routed to. - The entire approach to routing when using a link device is now link-centric rather than pci bus/device/pin specific. Thus, when using a tunable to override the default IRQ settings, one now uses a single tunable to route an entire link rather than routing a single device that uses the link (which has great foot-shooting potential if the user tries to route the same link to two different IRQs using two different pci bus/device/pin hints). For example, to adjust the IRQ that \_SB_.LNKA uses, one would set 'hw.pci.link.LNKA.irq=10' from the loader. - As a side effect of having the link driver, unused link devices will now be disabled when they are probed. - The algorithm for choosing an IRQ for a link that doesn't already have an IRQ assigned is now much closer to the one used in $PIR routing. When a link is routed via an ISA IRQ, only known-good IRQs that the BIOS has already used are used for routing instead of using probabilities to guess at which IRQs are probably not used by an ISA device. One change from $PIR is that the SCI is always considered a viable ISA IRQ, so that if the BIOS does not setup any IRQs the kernel will degenerate to routing all interrupts over the SCI. For non ISA IRQs, interrupts are picked from the possible pool using a simplistic weighting algorithm. Tested by: ru, scottl, others on acpi@ Reviewed by: njl
2004-11-23 22:26:44 +00:00
static int
acpi_pci_link_attach(device_t dev)
{
struct acpi_pci_link_softc *sc;
struct link_count_request creq;
struct link_res_request rreq;
Rework the ACPI PCI link code. - Use a new-bus device driver for the ACPI PCI link devices. The devices are called pci_linkX. The driver includes suspend/resume support so that the ACPI bridge drivers no longer have to poke the links to get them to handle suspend/resume. Also, the code to handle which IRQs a link is routed to and choosing an IRQ when a link is not already routed is all contained in the link driver. The PCI bridge drivers now ask the link driver which IRQ to use once they determine that a _PRT entry does not use a hardwired interrupt number. - The new link driver includes support for multiple IRQ resources per link device as well as preserving any non-IRQ resources when adjusting the IRQ that a link is routed to. - The entire approach to routing when using a link device is now link-centric rather than pci bus/device/pin specific. Thus, when using a tunable to override the default IRQ settings, one now uses a single tunable to route an entire link rather than routing a single device that uses the link (which has great foot-shooting potential if the user tries to route the same link to two different IRQs using two different pci bus/device/pin hints). For example, to adjust the IRQ that \_SB_.LNKA uses, one would set 'hw.pci.link.LNKA.irq=10' from the loader. - As a side effect of having the link driver, unused link devices will now be disabled when they are probed. - The algorithm for choosing an IRQ for a link that doesn't already have an IRQ assigned is now much closer to the one used in $PIR routing. When a link is routed via an ISA IRQ, only known-good IRQs that the BIOS has already used are used for routing instead of using probabilities to guess at which IRQs are probably not used by an ISA device. One change from $PIR is that the SCI is always considered a viable ISA IRQ, so that if the BIOS does not setup any IRQs the kernel will degenerate to routing all interrupts over the SCI. For non ISA IRQs, interrupts are picked from the possible pool using a simplistic weighting algorithm. Tested by: ru, scottl, others on acpi@ Reviewed by: njl
2004-11-23 22:26:44 +00:00
ACPI_STATUS status;
int i;
Rework the ACPI PCI link code. - Use a new-bus device driver for the ACPI PCI link devices. The devices are called pci_linkX. The driver includes suspend/resume support so that the ACPI bridge drivers no longer have to poke the links to get them to handle suspend/resume. Also, the code to handle which IRQs a link is routed to and choosing an IRQ when a link is not already routed is all contained in the link driver. The PCI bridge drivers now ask the link driver which IRQ to use once they determine that a _PRT entry does not use a hardwired interrupt number. - The new link driver includes support for multiple IRQ resources per link device as well as preserving any non-IRQ resources when adjusting the IRQ that a link is routed to. - The entire approach to routing when using a link device is now link-centric rather than pci bus/device/pin specific. Thus, when using a tunable to override the default IRQ settings, one now uses a single tunable to route an entire link rather than routing a single device that uses the link (which has great foot-shooting potential if the user tries to route the same link to two different IRQs using two different pci bus/device/pin hints). For example, to adjust the IRQ that \_SB_.LNKA uses, one would set 'hw.pci.link.LNKA.irq=10' from the loader. - As a side effect of having the link driver, unused link devices will now be disabled when they are probed. - The algorithm for choosing an IRQ for a link that doesn't already have an IRQ assigned is now much closer to the one used in $PIR routing. When a link is routed via an ISA IRQ, only known-good IRQs that the BIOS has already used are used for routing instead of using probabilities to guess at which IRQs are probably not used by an ISA device. One change from $PIR is that the SCI is always considered a viable ISA IRQ, so that if the BIOS does not setup any IRQs the kernel will degenerate to routing all interrupts over the SCI. For non ISA IRQs, interrupts are picked from the possible pool using a simplistic weighting algorithm. Tested by: ru, scottl, others on acpi@ Reviewed by: njl
2004-11-23 22:26:44 +00:00
sc = device_get_softc(dev);
sc->pl_dev = dev;
ACPI_SERIAL_BEGIN(pci_link);
/*
Rework the ACPI PCI link code. - Use a new-bus device driver for the ACPI PCI link devices. The devices are called pci_linkX. The driver includes suspend/resume support so that the ACPI bridge drivers no longer have to poke the links to get them to handle suspend/resume. Also, the code to handle which IRQs a link is routed to and choosing an IRQ when a link is not already routed is all contained in the link driver. The PCI bridge drivers now ask the link driver which IRQ to use once they determine that a _PRT entry does not use a hardwired interrupt number. - The new link driver includes support for multiple IRQ resources per link device as well as preserving any non-IRQ resources when adjusting the IRQ that a link is routed to. - The entire approach to routing when using a link device is now link-centric rather than pci bus/device/pin specific. Thus, when using a tunable to override the default IRQ settings, one now uses a single tunable to route an entire link rather than routing a single device that uses the link (which has great foot-shooting potential if the user tries to route the same link to two different IRQs using two different pci bus/device/pin hints). For example, to adjust the IRQ that \_SB_.LNKA uses, one would set 'hw.pci.link.LNKA.irq=10' from the loader. - As a side effect of having the link driver, unused link devices will now be disabled when they are probed. - The algorithm for choosing an IRQ for a link that doesn't already have an IRQ assigned is now much closer to the one used in $PIR routing. When a link is routed via an ISA IRQ, only known-good IRQs that the BIOS has already used are used for routing instead of using probabilities to guess at which IRQs are probably not used by an ISA device. One change from $PIR is that the SCI is always considered a viable ISA IRQ, so that if the BIOS does not setup any IRQs the kernel will degenerate to routing all interrupts over the SCI. For non ISA IRQs, interrupts are picked from the possible pool using a simplistic weighting algorithm. Tested by: ru, scottl, others on acpi@ Reviewed by: njl
2004-11-23 22:26:44 +00:00
* Count the number of current resources so we know how big of
* a link array to allocate. On some systems, _CRS is broken,
* so for those systems try to derive the count from _PRS instead.
*/
creq.in_dpf = DPF_OUTSIDE;
creq.count = 0;
Rework the ACPI PCI link code. - Use a new-bus device driver for the ACPI PCI link devices. The devices are called pci_linkX. The driver includes suspend/resume support so that the ACPI bridge drivers no longer have to poke the links to get them to handle suspend/resume. Also, the code to handle which IRQs a link is routed to and choosing an IRQ when a link is not already routed is all contained in the link driver. The PCI bridge drivers now ask the link driver which IRQ to use once they determine that a _PRT entry does not use a hardwired interrupt number. - The new link driver includes support for multiple IRQ resources per link device as well as preserving any non-IRQ resources when adjusting the IRQ that a link is routed to. - The entire approach to routing when using a link device is now link-centric rather than pci bus/device/pin specific. Thus, when using a tunable to override the default IRQ settings, one now uses a single tunable to route an entire link rather than routing a single device that uses the link (which has great foot-shooting potential if the user tries to route the same link to two different IRQs using two different pci bus/device/pin hints). For example, to adjust the IRQ that \_SB_.LNKA uses, one would set 'hw.pci.link.LNKA.irq=10' from the loader. - As a side effect of having the link driver, unused link devices will now be disabled when they are probed. - The algorithm for choosing an IRQ for a link that doesn't already have an IRQ assigned is now much closer to the one used in $PIR routing. When a link is routed via an ISA IRQ, only known-good IRQs that the BIOS has already used are used for routing instead of using probabilities to guess at which IRQs are probably not used by an ISA device. One change from $PIR is that the SCI is always considered a viable ISA IRQ, so that if the BIOS does not setup any IRQs the kernel will degenerate to routing all interrupts over the SCI. For non ISA IRQs, interrupts are picked from the possible pool using a simplistic weighting algorithm. Tested by: ru, scottl, others on acpi@ Reviewed by: njl
2004-11-23 22:26:44 +00:00
status = AcpiWalkResources(acpi_get_handle(dev), "_CRS",
acpi_count_irq_resources, &creq);
sc->pl_crs_bad = ACPI_FAILURE(status);
if (sc->pl_crs_bad) {
creq.in_dpf = DPF_OUTSIDE;
creq.count = 0;
status = AcpiWalkResources(acpi_get_handle(dev), "_PRS",
acpi_count_irq_resources, &creq);
if (ACPI_FAILURE(status)) {
device_printf(dev,
"Unable to parse _CRS or _PRS: %s\n",
AcpiFormatException(status));
ACPI_SERIAL_END(pci_link);
return (ENXIO);
}
}
sc->pl_num_links = creq.count;
if (creq.count == 0)
Rework the ACPI PCI link code. - Use a new-bus device driver for the ACPI PCI link devices. The devices are called pci_linkX. The driver includes suspend/resume support so that the ACPI bridge drivers no longer have to poke the links to get them to handle suspend/resume. Also, the code to handle which IRQs a link is routed to and choosing an IRQ when a link is not already routed is all contained in the link driver. The PCI bridge drivers now ask the link driver which IRQ to use once they determine that a _PRT entry does not use a hardwired interrupt number. - The new link driver includes support for multiple IRQ resources per link device as well as preserving any non-IRQ resources when adjusting the IRQ that a link is routed to. - The entire approach to routing when using a link device is now link-centric rather than pci bus/device/pin specific. Thus, when using a tunable to override the default IRQ settings, one now uses a single tunable to route an entire link rather than routing a single device that uses the link (which has great foot-shooting potential if the user tries to route the same link to two different IRQs using two different pci bus/device/pin hints). For example, to adjust the IRQ that \_SB_.LNKA uses, one would set 'hw.pci.link.LNKA.irq=10' from the loader. - As a side effect of having the link driver, unused link devices will now be disabled when they are probed. - The algorithm for choosing an IRQ for a link that doesn't already have an IRQ assigned is now much closer to the one used in $PIR routing. When a link is routed via an ISA IRQ, only known-good IRQs that the BIOS has already used are used for routing instead of using probabilities to guess at which IRQs are probably not used by an ISA device. One change from $PIR is that the SCI is always considered a viable ISA IRQ, so that if the BIOS does not setup any IRQs the kernel will degenerate to routing all interrupts over the SCI. For non ISA IRQs, interrupts are picked from the possible pool using a simplistic weighting algorithm. Tested by: ru, scottl, others on acpi@ Reviewed by: njl
2004-11-23 22:26:44 +00:00
return (0);
sc->pl_links = malloc(sizeof(struct link) * sc->pl_num_links,
M_PCI_LINK, M_WAITOK | M_ZERO);
/* Initialize the child links. */
for (i = 0; i < sc->pl_num_links; i++) {
sc->pl_links[i].l_irq = PCI_INVALID_IRQ;
sc->pl_links[i].l_bios_irq = PCI_INVALID_IRQ;
sc->pl_links[i].l_sc = sc;
sc->pl_links[i].l_isa_irq = FALSE;
sc->pl_links[i].l_res_index = -1;
Rework the ACPI PCI link code. - Use a new-bus device driver for the ACPI PCI link devices. The devices are called pci_linkX. The driver includes suspend/resume support so that the ACPI bridge drivers no longer have to poke the links to get them to handle suspend/resume. Also, the code to handle which IRQs a link is routed to and choosing an IRQ when a link is not already routed is all contained in the link driver. The PCI bridge drivers now ask the link driver which IRQ to use once they determine that a _PRT entry does not use a hardwired interrupt number. - The new link driver includes support for multiple IRQ resources per link device as well as preserving any non-IRQ resources when adjusting the IRQ that a link is routed to. - The entire approach to routing when using a link device is now link-centric rather than pci bus/device/pin specific. Thus, when using a tunable to override the default IRQ settings, one now uses a single tunable to route an entire link rather than routing a single device that uses the link (which has great foot-shooting potential if the user tries to route the same link to two different IRQs using two different pci bus/device/pin hints). For example, to adjust the IRQ that \_SB_.LNKA uses, one would set 'hw.pci.link.LNKA.irq=10' from the loader. - As a side effect of having the link driver, unused link devices will now be disabled when they are probed. - The algorithm for choosing an IRQ for a link that doesn't already have an IRQ assigned is now much closer to the one used in $PIR routing. When a link is routed via an ISA IRQ, only known-good IRQs that the BIOS has already used are used for routing instead of using probabilities to guess at which IRQs are probably not used by an ISA device. One change from $PIR is that the SCI is always considered a viable ISA IRQ, so that if the BIOS does not setup any IRQs the kernel will degenerate to routing all interrupts over the SCI. For non ISA IRQs, interrupts are picked from the possible pool using a simplistic weighting algorithm. Tested by: ru, scottl, others on acpi@ Reviewed by: njl
2004-11-23 22:26:44 +00:00
}
/* Try to read the current settings from _CRS if it is valid. */
if (!sc->pl_crs_bad) {
rreq.in_dpf = DPF_OUTSIDE;
rreq.link_index = 0;
rreq.res_index = 0;
rreq.sc = sc;
status = AcpiWalkResources(acpi_get_handle(dev), "_CRS",
link_add_crs, &rreq);
if (ACPI_FAILURE(status)) {
device_printf(dev, "Unable to parse _CRS: %s\n",
AcpiFormatException(status));
goto fail;
}
}
/*
* Try to read the possible settings from _PRS. Note that if the
* _CRS is toast, we depend on having a working _PRS. However, if
* _CRS works, then it is ok for _PRS to be missing.
*/
rreq.in_dpf = DPF_OUTSIDE;
rreq.link_index = 0;
rreq.res_index = 0;
rreq.sc = sc;
Rework the ACPI PCI link code. - Use a new-bus device driver for the ACPI PCI link devices. The devices are called pci_linkX. The driver includes suspend/resume support so that the ACPI bridge drivers no longer have to poke the links to get them to handle suspend/resume. Also, the code to handle which IRQs a link is routed to and choosing an IRQ when a link is not already routed is all contained in the link driver. The PCI bridge drivers now ask the link driver which IRQ to use once they determine that a _PRT entry does not use a hardwired interrupt number. - The new link driver includes support for multiple IRQ resources per link device as well as preserving any non-IRQ resources when adjusting the IRQ that a link is routed to. - The entire approach to routing when using a link device is now link-centric rather than pci bus/device/pin specific. Thus, when using a tunable to override the default IRQ settings, one now uses a single tunable to route an entire link rather than routing a single device that uses the link (which has great foot-shooting potential if the user tries to route the same link to two different IRQs using two different pci bus/device/pin hints). For example, to adjust the IRQ that \_SB_.LNKA uses, one would set 'hw.pci.link.LNKA.irq=10' from the loader. - As a side effect of having the link driver, unused link devices will now be disabled when they are probed. - The algorithm for choosing an IRQ for a link that doesn't already have an IRQ assigned is now much closer to the one used in $PIR routing. When a link is routed via an ISA IRQ, only known-good IRQs that the BIOS has already used are used for routing instead of using probabilities to guess at which IRQs are probably not used by an ISA device. One change from $PIR is that the SCI is always considered a viable ISA IRQ, so that if the BIOS does not setup any IRQs the kernel will degenerate to routing all interrupts over the SCI. For non ISA IRQs, interrupts are picked from the possible pool using a simplistic weighting algorithm. Tested by: ru, scottl, others on acpi@ Reviewed by: njl
2004-11-23 22:26:44 +00:00
status = AcpiWalkResources(acpi_get_handle(dev), "_PRS",
link_add_prs, &rreq);
if (ACPI_FAILURE(status) &&
(status != AE_NOT_FOUND || sc->pl_crs_bad)) {
device_printf(dev, "Unable to parse _PRS: %s\n",
AcpiFormatException(status));
Rework the ACPI PCI link code. - Use a new-bus device driver for the ACPI PCI link devices. The devices are called pci_linkX. The driver includes suspend/resume support so that the ACPI bridge drivers no longer have to poke the links to get them to handle suspend/resume. Also, the code to handle which IRQs a link is routed to and choosing an IRQ when a link is not already routed is all contained in the link driver. The PCI bridge drivers now ask the link driver which IRQ to use once they determine that a _PRT entry does not use a hardwired interrupt number. - The new link driver includes support for multiple IRQ resources per link device as well as preserving any non-IRQ resources when adjusting the IRQ that a link is routed to. - The entire approach to routing when using a link device is now link-centric rather than pci bus/device/pin specific. Thus, when using a tunable to override the default IRQ settings, one now uses a single tunable to route an entire link rather than routing a single device that uses the link (which has great foot-shooting potential if the user tries to route the same link to two different IRQs using two different pci bus/device/pin hints). For example, to adjust the IRQ that \_SB_.LNKA uses, one would set 'hw.pci.link.LNKA.irq=10' from the loader. - As a side effect of having the link driver, unused link devices will now be disabled when they are probed. - The algorithm for choosing an IRQ for a link that doesn't already have an IRQ assigned is now much closer to the one used in $PIR routing. When a link is routed via an ISA IRQ, only known-good IRQs that the BIOS has already used are used for routing instead of using probabilities to guess at which IRQs are probably not used by an ISA device. One change from $PIR is that the SCI is always considered a viable ISA IRQ, so that if the BIOS does not setup any IRQs the kernel will degenerate to routing all interrupts over the SCI. For non ISA IRQs, interrupts are picked from the possible pool using a simplistic weighting algorithm. Tested by: ru, scottl, others on acpi@ Reviewed by: njl
2004-11-23 22:26:44 +00:00
goto fail;
}
Rework the ACPI PCI link code. - Use a new-bus device driver for the ACPI PCI link devices. The devices are called pci_linkX. The driver includes suspend/resume support so that the ACPI bridge drivers no longer have to poke the links to get them to handle suspend/resume. Also, the code to handle which IRQs a link is routed to and choosing an IRQ when a link is not already routed is all contained in the link driver. The PCI bridge drivers now ask the link driver which IRQ to use once they determine that a _PRT entry does not use a hardwired interrupt number. - The new link driver includes support for multiple IRQ resources per link device as well as preserving any non-IRQ resources when adjusting the IRQ that a link is routed to. - The entire approach to routing when using a link device is now link-centric rather than pci bus/device/pin specific. Thus, when using a tunable to override the default IRQ settings, one now uses a single tunable to route an entire link rather than routing a single device that uses the link (which has great foot-shooting potential if the user tries to route the same link to two different IRQs using two different pci bus/device/pin hints). For example, to adjust the IRQ that \_SB_.LNKA uses, one would set 'hw.pci.link.LNKA.irq=10' from the loader. - As a side effect of having the link driver, unused link devices will now be disabled when they are probed. - The algorithm for choosing an IRQ for a link that doesn't already have an IRQ assigned is now much closer to the one used in $PIR routing. When a link is routed via an ISA IRQ, only known-good IRQs that the BIOS has already used are used for routing instead of using probabilities to guess at which IRQs are probably not used by an ISA device. One change from $PIR is that the SCI is always considered a viable ISA IRQ, so that if the BIOS does not setup any IRQs the kernel will degenerate to routing all interrupts over the SCI. For non ISA IRQs, interrupts are picked from the possible pool using a simplistic weighting algorithm. Tested by: ru, scottl, others on acpi@ Reviewed by: njl
2004-11-23 22:26:44 +00:00
if (bootverbose) {
device_printf(dev, "Links after initial probe:\n");
acpi_pci_link_dump(sc);
}
Rework the ACPI PCI link code. - Use a new-bus device driver for the ACPI PCI link devices. The devices are called pci_linkX. The driver includes suspend/resume support so that the ACPI bridge drivers no longer have to poke the links to get them to handle suspend/resume. Also, the code to handle which IRQs a link is routed to and choosing an IRQ when a link is not already routed is all contained in the link driver. The PCI bridge drivers now ask the link driver which IRQ to use once they determine that a _PRT entry does not use a hardwired interrupt number. - The new link driver includes support for multiple IRQ resources per link device as well as preserving any non-IRQ resources when adjusting the IRQ that a link is routed to. - The entire approach to routing when using a link device is now link-centric rather than pci bus/device/pin specific. Thus, when using a tunable to override the default IRQ settings, one now uses a single tunable to route an entire link rather than routing a single device that uses the link (which has great foot-shooting potential if the user tries to route the same link to two different IRQs using two different pci bus/device/pin hints). For example, to adjust the IRQ that \_SB_.LNKA uses, one would set 'hw.pci.link.LNKA.irq=10' from the loader. - As a side effect of having the link driver, unused link devices will now be disabled when they are probed. - The algorithm for choosing an IRQ for a link that doesn't already have an IRQ assigned is now much closer to the one used in $PIR routing. When a link is routed via an ISA IRQ, only known-good IRQs that the BIOS has already used are used for routing instead of using probabilities to guess at which IRQs are probably not used by an ISA device. One change from $PIR is that the SCI is always considered a viable ISA IRQ, so that if the BIOS does not setup any IRQs the kernel will degenerate to routing all interrupts over the SCI. For non ISA IRQs, interrupts are picked from the possible pool using a simplistic weighting algorithm. Tested by: ru, scottl, others on acpi@ Reviewed by: njl
2004-11-23 22:26:44 +00:00
/* Verify initial IRQs if we have _PRS. */
if (status != AE_NOT_FOUND)
for (i = 0; i < sc->pl_num_links; i++)
if (!link_valid_irq(&sc->pl_links[i],
sc->pl_links[i].l_irq))
sc->pl_links[i].l_irq = PCI_INVALID_IRQ;
if (bootverbose) {
device_printf(dev, "Links after initial validation:\n");
acpi_pci_link_dump(sc);
}
Rework the ACPI PCI link code. - Use a new-bus device driver for the ACPI PCI link devices. The devices are called pci_linkX. The driver includes suspend/resume support so that the ACPI bridge drivers no longer have to poke the links to get them to handle suspend/resume. Also, the code to handle which IRQs a link is routed to and choosing an IRQ when a link is not already routed is all contained in the link driver. The PCI bridge drivers now ask the link driver which IRQ to use once they determine that a _PRT entry does not use a hardwired interrupt number. - The new link driver includes support for multiple IRQ resources per link device as well as preserving any non-IRQ resources when adjusting the IRQ that a link is routed to. - The entire approach to routing when using a link device is now link-centric rather than pci bus/device/pin specific. Thus, when using a tunable to override the default IRQ settings, one now uses a single tunable to route an entire link rather than routing a single device that uses the link (which has great foot-shooting potential if the user tries to route the same link to two different IRQs using two different pci bus/device/pin hints). For example, to adjust the IRQ that \_SB_.LNKA uses, one would set 'hw.pci.link.LNKA.irq=10' from the loader. - As a side effect of having the link driver, unused link devices will now be disabled when they are probed. - The algorithm for choosing an IRQ for a link that doesn't already have an IRQ assigned is now much closer to the one used in $PIR routing. When a link is routed via an ISA IRQ, only known-good IRQs that the BIOS has already used are used for routing instead of using probabilities to guess at which IRQs are probably not used by an ISA device. One change from $PIR is that the SCI is always considered a viable ISA IRQ, so that if the BIOS does not setup any IRQs the kernel will degenerate to routing all interrupts over the SCI. For non ISA IRQs, interrupts are picked from the possible pool using a simplistic weighting algorithm. Tested by: ru, scottl, others on acpi@ Reviewed by: njl
2004-11-23 22:26:44 +00:00
/* Save initial IRQs. */
for (i = 0; i < sc->pl_num_links; i++)
sc->pl_links[i].l_initial_irq = sc->pl_links[i].l_irq;
/*
* Try to disable this link. If successful, set the current IRQ to
* zero and flags to indicate this link is not routed. If we can't
* run _DIS (i.e., the method doesn't exist), assume the initial
* IRQ was routed by the BIOS.
*/
Rework the ACPI PCI link code. - Use a new-bus device driver for the ACPI PCI link devices. The devices are called pci_linkX. The driver includes suspend/resume support so that the ACPI bridge drivers no longer have to poke the links to get them to handle suspend/resume. Also, the code to handle which IRQs a link is routed to and choosing an IRQ when a link is not already routed is all contained in the link driver. The PCI bridge drivers now ask the link driver which IRQ to use once they determine that a _PRT entry does not use a hardwired interrupt number. - The new link driver includes support for multiple IRQ resources per link device as well as preserving any non-IRQ resources when adjusting the IRQ that a link is routed to. - The entire approach to routing when using a link device is now link-centric rather than pci bus/device/pin specific. Thus, when using a tunable to override the default IRQ settings, one now uses a single tunable to route an entire link rather than routing a single device that uses the link (which has great foot-shooting potential if the user tries to route the same link to two different IRQs using two different pci bus/device/pin hints). For example, to adjust the IRQ that \_SB_.LNKA uses, one would set 'hw.pci.link.LNKA.irq=10' from the loader. - As a side effect of having the link driver, unused link devices will now be disabled when they are probed. - The algorithm for choosing an IRQ for a link that doesn't already have an IRQ assigned is now much closer to the one used in $PIR routing. When a link is routed via an ISA IRQ, only known-good IRQs that the BIOS has already used are used for routing instead of using probabilities to guess at which IRQs are probably not used by an ISA device. One change from $PIR is that the SCI is always considered a viable ISA IRQ, so that if the BIOS does not setup any IRQs the kernel will degenerate to routing all interrupts over the SCI. For non ISA IRQs, interrupts are picked from the possible pool using a simplistic weighting algorithm. Tested by: ru, scottl, others on acpi@ Reviewed by: njl
2004-11-23 22:26:44 +00:00
if (ACPI_SUCCESS(AcpiEvaluateObject(acpi_get_handle(dev), "_DIS", NULL,
NULL)))
Rework the ACPI PCI link code. - Use a new-bus device driver for the ACPI PCI link devices. The devices are called pci_linkX. The driver includes suspend/resume support so that the ACPI bridge drivers no longer have to poke the links to get them to handle suspend/resume. Also, the code to handle which IRQs a link is routed to and choosing an IRQ when a link is not already routed is all contained in the link driver. The PCI bridge drivers now ask the link driver which IRQ to use once they determine that a _PRT entry does not use a hardwired interrupt number. - The new link driver includes support for multiple IRQ resources per link device as well as preserving any non-IRQ resources when adjusting the IRQ that a link is routed to. - The entire approach to routing when using a link device is now link-centric rather than pci bus/device/pin specific. Thus, when using a tunable to override the default IRQ settings, one now uses a single tunable to route an entire link rather than routing a single device that uses the link (which has great foot-shooting potential if the user tries to route the same link to two different IRQs using two different pci bus/device/pin hints). For example, to adjust the IRQ that \_SB_.LNKA uses, one would set 'hw.pci.link.LNKA.irq=10' from the loader. - As a side effect of having the link driver, unused link devices will now be disabled when they are probed. - The algorithm for choosing an IRQ for a link that doesn't already have an IRQ assigned is now much closer to the one used in $PIR routing. When a link is routed via an ISA IRQ, only known-good IRQs that the BIOS has already used are used for routing instead of using probabilities to guess at which IRQs are probably not used by an ISA device. One change from $PIR is that the SCI is always considered a viable ISA IRQ, so that if the BIOS does not setup any IRQs the kernel will degenerate to routing all interrupts over the SCI. For non ISA IRQs, interrupts are picked from the possible pool using a simplistic weighting algorithm. Tested by: ru, scottl, others on acpi@ Reviewed by: njl
2004-11-23 22:26:44 +00:00
for (i = 0; i < sc->pl_num_links; i++)
sc->pl_links[i].l_irq = PCI_INVALID_IRQ;
else
for (i = 0; i < sc->pl_num_links; i++)
if (PCI_INTERRUPT_VALID(sc->pl_links[i].l_irq))
sc->pl_links[i].l_routed = TRUE;
Rework the ACPI PCI link code. - Use a new-bus device driver for the ACPI PCI link devices. The devices are called pci_linkX. The driver includes suspend/resume support so that the ACPI bridge drivers no longer have to poke the links to get them to handle suspend/resume. Also, the code to handle which IRQs a link is routed to and choosing an IRQ when a link is not already routed is all contained in the link driver. The PCI bridge drivers now ask the link driver which IRQ to use once they determine that a _PRT entry does not use a hardwired interrupt number. - The new link driver includes support for multiple IRQ resources per link device as well as preserving any non-IRQ resources when adjusting the IRQ that a link is routed to. - The entire approach to routing when using a link device is now link-centric rather than pci bus/device/pin specific. Thus, when using a tunable to override the default IRQ settings, one now uses a single tunable to route an entire link rather than routing a single device that uses the link (which has great foot-shooting potential if the user tries to route the same link to two different IRQs using two different pci bus/device/pin hints). For example, to adjust the IRQ that \_SB_.LNKA uses, one would set 'hw.pci.link.LNKA.irq=10' from the loader. - As a side effect of having the link driver, unused link devices will now be disabled when they are probed. - The algorithm for choosing an IRQ for a link that doesn't already have an IRQ assigned is now much closer to the one used in $PIR routing. When a link is routed via an ISA IRQ, only known-good IRQs that the BIOS has already used are used for routing instead of using probabilities to guess at which IRQs are probably not used by an ISA device. One change from $PIR is that the SCI is always considered a viable ISA IRQ, so that if the BIOS does not setup any IRQs the kernel will degenerate to routing all interrupts over the SCI. For non ISA IRQs, interrupts are picked from the possible pool using a simplistic weighting algorithm. Tested by: ru, scottl, others on acpi@ Reviewed by: njl
2004-11-23 22:26:44 +00:00
if (bootverbose) {
device_printf(dev, "Links after disable:\n");
acpi_pci_link_dump(sc);
}
ACPI_SERIAL_END(pci_link);
Rework the ACPI PCI link code. - Use a new-bus device driver for the ACPI PCI link devices. The devices are called pci_linkX. The driver includes suspend/resume support so that the ACPI bridge drivers no longer have to poke the links to get them to handle suspend/resume. Also, the code to handle which IRQs a link is routed to and choosing an IRQ when a link is not already routed is all contained in the link driver. The PCI bridge drivers now ask the link driver which IRQ to use once they determine that a _PRT entry does not use a hardwired interrupt number. - The new link driver includes support for multiple IRQ resources per link device as well as preserving any non-IRQ resources when adjusting the IRQ that a link is routed to. - The entire approach to routing when using a link device is now link-centric rather than pci bus/device/pin specific. Thus, when using a tunable to override the default IRQ settings, one now uses a single tunable to route an entire link rather than routing a single device that uses the link (which has great foot-shooting potential if the user tries to route the same link to two different IRQs using two different pci bus/device/pin hints). For example, to adjust the IRQ that \_SB_.LNKA uses, one would set 'hw.pci.link.LNKA.irq=10' from the loader. - As a side effect of having the link driver, unused link devices will now be disabled when they are probed. - The algorithm for choosing an IRQ for a link that doesn't already have an IRQ assigned is now much closer to the one used in $PIR routing. When a link is routed via an ISA IRQ, only known-good IRQs that the BIOS has already used are used for routing instead of using probabilities to guess at which IRQs are probably not used by an ISA device. One change from $PIR is that the SCI is always considered a viable ISA IRQ, so that if the BIOS does not setup any IRQs the kernel will degenerate to routing all interrupts over the SCI. For non ISA IRQs, interrupts are picked from the possible pool using a simplistic weighting algorithm. Tested by: ru, scottl, others on acpi@ Reviewed by: njl
2004-11-23 22:26:44 +00:00
return (0);
fail:
ACPI_SERIAL_END(pci_link);
Rework the ACPI PCI link code. - Use a new-bus device driver for the ACPI PCI link devices. The devices are called pci_linkX. The driver includes suspend/resume support so that the ACPI bridge drivers no longer have to poke the links to get them to handle suspend/resume. Also, the code to handle which IRQs a link is routed to and choosing an IRQ when a link is not already routed is all contained in the link driver. The PCI bridge drivers now ask the link driver which IRQ to use once they determine that a _PRT entry does not use a hardwired interrupt number. - The new link driver includes support for multiple IRQ resources per link device as well as preserving any non-IRQ resources when adjusting the IRQ that a link is routed to. - The entire approach to routing when using a link device is now link-centric rather than pci bus/device/pin specific. Thus, when using a tunable to override the default IRQ settings, one now uses a single tunable to route an entire link rather than routing a single device that uses the link (which has great foot-shooting potential if the user tries to route the same link to two different IRQs using two different pci bus/device/pin hints). For example, to adjust the IRQ that \_SB_.LNKA uses, one would set 'hw.pci.link.LNKA.irq=10' from the loader. - As a side effect of having the link driver, unused link devices will now be disabled when they are probed. - The algorithm for choosing an IRQ for a link that doesn't already have an IRQ assigned is now much closer to the one used in $PIR routing. When a link is routed via an ISA IRQ, only known-good IRQs that the BIOS has already used are used for routing instead of using probabilities to guess at which IRQs are probably not used by an ISA device. One change from $PIR is that the SCI is always considered a viable ISA IRQ, so that if the BIOS does not setup any IRQs the kernel will degenerate to routing all interrupts over the SCI. For non ISA IRQs, interrupts are picked from the possible pool using a simplistic weighting algorithm. Tested by: ru, scottl, others on acpi@ Reviewed by: njl
2004-11-23 22:26:44 +00:00
for (i = 0; i < sc->pl_num_links; i++)
if (sc->pl_links[i].l_irqs != NULL)
free(sc->pl_links[i].l_irqs, M_PCI_LINK);
free(sc->pl_links, M_PCI_LINK);
return (ENXIO);
}
Rework the ACPI PCI link code. - Use a new-bus device driver for the ACPI PCI link devices. The devices are called pci_linkX. The driver includes suspend/resume support so that the ACPI bridge drivers no longer have to poke the links to get them to handle suspend/resume. Also, the code to handle which IRQs a link is routed to and choosing an IRQ when a link is not already routed is all contained in the link driver. The PCI bridge drivers now ask the link driver which IRQ to use once they determine that a _PRT entry does not use a hardwired interrupt number. - The new link driver includes support for multiple IRQ resources per link device as well as preserving any non-IRQ resources when adjusting the IRQ that a link is routed to. - The entire approach to routing when using a link device is now link-centric rather than pci bus/device/pin specific. Thus, when using a tunable to override the default IRQ settings, one now uses a single tunable to route an entire link rather than routing a single device that uses the link (which has great foot-shooting potential if the user tries to route the same link to two different IRQs using two different pci bus/device/pin hints). For example, to adjust the IRQ that \_SB_.LNKA uses, one would set 'hw.pci.link.LNKA.irq=10' from the loader. - As a side effect of having the link driver, unused link devices will now be disabled when they are probed. - The algorithm for choosing an IRQ for a link that doesn't already have an IRQ assigned is now much closer to the one used in $PIR routing. When a link is routed via an ISA IRQ, only known-good IRQs that the BIOS has already used are used for routing instead of using probabilities to guess at which IRQs are probably not used by an ISA device. One change from $PIR is that the SCI is always considered a viable ISA IRQ, so that if the BIOS does not setup any IRQs the kernel will degenerate to routing all interrupts over the SCI. For non ISA IRQs, interrupts are picked from the possible pool using a simplistic weighting algorithm. Tested by: ru, scottl, others on acpi@ Reviewed by: njl
2004-11-23 22:26:44 +00:00
/* XXX: Note that this is identical to pci_pir_search_irq(). */
static uint8_t
acpi_pci_link_search_irq(int bus, int device, int pin)
{
uint32_t value;
uint8_t func, maxfunc;
/* See if we have a valid device at function 0. */
value = pci_cfgregread(bus, device, 0, PCIR_HDRTYPE, 1);
if ((value & PCIM_HDRTYPE) > PCI_MAXHDRTYPE)
return (PCI_INVALID_IRQ);
if (value & PCIM_MFDEV)
maxfunc = PCI_FUNCMAX;
else
maxfunc = 0;
/* Scan all possible functions at this device. */
for (func = 0; func <= maxfunc; func++) {
value = pci_cfgregread(bus, device, func, PCIR_DEVVENDOR, 4);
if (value == 0xffffffff)
continue;
value = pci_cfgregread(bus, device, func, PCIR_INTPIN, 1);
Rework the ACPI PCI link code. - Use a new-bus device driver for the ACPI PCI link devices. The devices are called pci_linkX. The driver includes suspend/resume support so that the ACPI bridge drivers no longer have to poke the links to get them to handle suspend/resume. Also, the code to handle which IRQs a link is routed to and choosing an IRQ when a link is not already routed is all contained in the link driver. The PCI bridge drivers now ask the link driver which IRQ to use once they determine that a _PRT entry does not use a hardwired interrupt number. - The new link driver includes support for multiple IRQ resources per link device as well as preserving any non-IRQ resources when adjusting the IRQ that a link is routed to. - The entire approach to routing when using a link device is now link-centric rather than pci bus/device/pin specific. Thus, when using a tunable to override the default IRQ settings, one now uses a single tunable to route an entire link rather than routing a single device that uses the link (which has great foot-shooting potential if the user tries to route the same link to two different IRQs using two different pci bus/device/pin hints). For example, to adjust the IRQ that \_SB_.LNKA uses, one would set 'hw.pci.link.LNKA.irq=10' from the loader. - As a side effect of having the link driver, unused link devices will now be disabled when they are probed. - The algorithm for choosing an IRQ for a link that doesn't already have an IRQ assigned is now much closer to the one used in $PIR routing. When a link is routed via an ISA IRQ, only known-good IRQs that the BIOS has already used are used for routing instead of using probabilities to guess at which IRQs are probably not used by an ISA device. One change from $PIR is that the SCI is always considered a viable ISA IRQ, so that if the BIOS does not setup any IRQs the kernel will degenerate to routing all interrupts over the SCI. For non ISA IRQs, interrupts are picked from the possible pool using a simplistic weighting algorithm. Tested by: ru, scottl, others on acpi@ Reviewed by: njl
2004-11-23 22:26:44 +00:00
/*
* See if it uses the pin in question. Note that the passed
* in pin uses 0 for A, .. 3 for D whereas the intpin
* register uses 0 for no interrupt, 1 for A, .. 4 for D.
*/
if (value != pin + 1)
continue;
value = pci_cfgregread(bus, device, func, PCIR_INTLINE, 1);
if (bootverbose)
printf(
"ACPI: Found matching pin for %d.%d.INT%c at func %d: %d\n",
bus, device, pin + 'A', func, value);
if (value != PCI_INVALID_IRQ)
return (value);
}
Rework the ACPI PCI link code. - Use a new-bus device driver for the ACPI PCI link devices. The devices are called pci_linkX. The driver includes suspend/resume support so that the ACPI bridge drivers no longer have to poke the links to get them to handle suspend/resume. Also, the code to handle which IRQs a link is routed to and choosing an IRQ when a link is not already routed is all contained in the link driver. The PCI bridge drivers now ask the link driver which IRQ to use once they determine that a _PRT entry does not use a hardwired interrupt number. - The new link driver includes support for multiple IRQ resources per link device as well as preserving any non-IRQ resources when adjusting the IRQ that a link is routed to. - The entire approach to routing when using a link device is now link-centric rather than pci bus/device/pin specific. Thus, when using a tunable to override the default IRQ settings, one now uses a single tunable to route an entire link rather than routing a single device that uses the link (which has great foot-shooting potential if the user tries to route the same link to two different IRQs using two different pci bus/device/pin hints). For example, to adjust the IRQ that \_SB_.LNKA uses, one would set 'hw.pci.link.LNKA.irq=10' from the loader. - As a side effect of having the link driver, unused link devices will now be disabled when they are probed. - The algorithm for choosing an IRQ for a link that doesn't already have an IRQ assigned is now much closer to the one used in $PIR routing. When a link is routed via an ISA IRQ, only known-good IRQs that the BIOS has already used are used for routing instead of using probabilities to guess at which IRQs are probably not used by an ISA device. One change from $PIR is that the SCI is always considered a viable ISA IRQ, so that if the BIOS does not setup any IRQs the kernel will degenerate to routing all interrupts over the SCI. For non ISA IRQs, interrupts are picked from the possible pool using a simplistic weighting algorithm. Tested by: ru, scottl, others on acpi@ Reviewed by: njl
2004-11-23 22:26:44 +00:00
return (PCI_INVALID_IRQ);
}
/*
* Find the link structure that corresponds to the resource index passed in
* via 'source_index'.
*/
static struct link *
acpi_pci_link_lookup(device_t dev, int source_index)
{
struct acpi_pci_link_softc *sc;
int i;
ACPI_SERIAL_ASSERT(pci_link);
sc = device_get_softc(dev);
for (i = 0; i < sc->pl_num_links; i++)
if (sc->pl_links[i].l_res_index == source_index)
return (&sc->pl_links[i]);
return (NULL);
}
Rework the ACPI PCI link code. - Use a new-bus device driver for the ACPI PCI link devices. The devices are called pci_linkX. The driver includes suspend/resume support so that the ACPI bridge drivers no longer have to poke the links to get them to handle suspend/resume. Also, the code to handle which IRQs a link is routed to and choosing an IRQ when a link is not already routed is all contained in the link driver. The PCI bridge drivers now ask the link driver which IRQ to use once they determine that a _PRT entry does not use a hardwired interrupt number. - The new link driver includes support for multiple IRQ resources per link device as well as preserving any non-IRQ resources when adjusting the IRQ that a link is routed to. - The entire approach to routing when using a link device is now link-centric rather than pci bus/device/pin specific. Thus, when using a tunable to override the default IRQ settings, one now uses a single tunable to route an entire link rather than routing a single device that uses the link (which has great foot-shooting potential if the user tries to route the same link to two different IRQs using two different pci bus/device/pin hints). For example, to adjust the IRQ that \_SB_.LNKA uses, one would set 'hw.pci.link.LNKA.irq=10' from the loader. - As a side effect of having the link driver, unused link devices will now be disabled when they are probed. - The algorithm for choosing an IRQ for a link that doesn't already have an IRQ assigned is now much closer to the one used in $PIR routing. When a link is routed via an ISA IRQ, only known-good IRQs that the BIOS has already used are used for routing instead of using probabilities to guess at which IRQs are probably not used by an ISA device. One change from $PIR is that the SCI is always considered a viable ISA IRQ, so that if the BIOS does not setup any IRQs the kernel will degenerate to routing all interrupts over the SCI. For non ISA IRQs, interrupts are picked from the possible pool using a simplistic weighting algorithm. Tested by: ru, scottl, others on acpi@ Reviewed by: njl
2004-11-23 22:26:44 +00:00
void
acpi_pci_link_add_reference(device_t dev, int index, device_t pcib, int slot,
int pin)
{
Rework the ACPI PCI link code. - Use a new-bus device driver for the ACPI PCI link devices. The devices are called pci_linkX. The driver includes suspend/resume support so that the ACPI bridge drivers no longer have to poke the links to get them to handle suspend/resume. Also, the code to handle which IRQs a link is routed to and choosing an IRQ when a link is not already routed is all contained in the link driver. The PCI bridge drivers now ask the link driver which IRQ to use once they determine that a _PRT entry does not use a hardwired interrupt number. - The new link driver includes support for multiple IRQ resources per link device as well as preserving any non-IRQ resources when adjusting the IRQ that a link is routed to. - The entire approach to routing when using a link device is now link-centric rather than pci bus/device/pin specific. Thus, when using a tunable to override the default IRQ settings, one now uses a single tunable to route an entire link rather than routing a single device that uses the link (which has great foot-shooting potential if the user tries to route the same link to two different IRQs using two different pci bus/device/pin hints). For example, to adjust the IRQ that \_SB_.LNKA uses, one would set 'hw.pci.link.LNKA.irq=10' from the loader. - As a side effect of having the link driver, unused link devices will now be disabled when they are probed. - The algorithm for choosing an IRQ for a link that doesn't already have an IRQ assigned is now much closer to the one used in $PIR routing. When a link is routed via an ISA IRQ, only known-good IRQs that the BIOS has already used are used for routing instead of using probabilities to guess at which IRQs are probably not used by an ISA device. One change from $PIR is that the SCI is always considered a viable ISA IRQ, so that if the BIOS does not setup any IRQs the kernel will degenerate to routing all interrupts over the SCI. For non ISA IRQs, interrupts are picked from the possible pool using a simplistic weighting algorithm. Tested by: ru, scottl, others on acpi@ Reviewed by: njl
2004-11-23 22:26:44 +00:00
struct link *link;
uint8_t bios_irq;
Rework the ACPI PCI link code. - Use a new-bus device driver for the ACPI PCI link devices. The devices are called pci_linkX. The driver includes suspend/resume support so that the ACPI bridge drivers no longer have to poke the links to get them to handle suspend/resume. Also, the code to handle which IRQs a link is routed to and choosing an IRQ when a link is not already routed is all contained in the link driver. The PCI bridge drivers now ask the link driver which IRQ to use once they determine that a _PRT entry does not use a hardwired interrupt number. - The new link driver includes support for multiple IRQ resources per link device as well as preserving any non-IRQ resources when adjusting the IRQ that a link is routed to. - The entire approach to routing when using a link device is now link-centric rather than pci bus/device/pin specific. Thus, when using a tunable to override the default IRQ settings, one now uses a single tunable to route an entire link rather than routing a single device that uses the link (which has great foot-shooting potential if the user tries to route the same link to two different IRQs using two different pci bus/device/pin hints). For example, to adjust the IRQ that \_SB_.LNKA uses, one would set 'hw.pci.link.LNKA.irq=10' from the loader. - As a side effect of having the link driver, unused link devices will now be disabled when they are probed. - The algorithm for choosing an IRQ for a link that doesn't already have an IRQ assigned is now much closer to the one used in $PIR routing. When a link is routed via an ISA IRQ, only known-good IRQs that the BIOS has already used are used for routing instead of using probabilities to guess at which IRQs are probably not used by an ISA device. One change from $PIR is that the SCI is always considered a viable ISA IRQ, so that if the BIOS does not setup any IRQs the kernel will degenerate to routing all interrupts over the SCI. For non ISA IRQs, interrupts are picked from the possible pool using a simplistic weighting algorithm. Tested by: ru, scottl, others on acpi@ Reviewed by: njl
2004-11-23 22:26:44 +00:00
/* Bump the reference count. */
ACPI_SERIAL_BEGIN(pci_link);
link = acpi_pci_link_lookup(dev, index);
if (link == NULL)
panic("%s: apparently invalid index %d", __func__, index);
Rework the ACPI PCI link code. - Use a new-bus device driver for the ACPI PCI link devices. The devices are called pci_linkX. The driver includes suspend/resume support so that the ACPI bridge drivers no longer have to poke the links to get them to handle suspend/resume. Also, the code to handle which IRQs a link is routed to and choosing an IRQ when a link is not already routed is all contained in the link driver. The PCI bridge drivers now ask the link driver which IRQ to use once they determine that a _PRT entry does not use a hardwired interrupt number. - The new link driver includes support for multiple IRQ resources per link device as well as preserving any non-IRQ resources when adjusting the IRQ that a link is routed to. - The entire approach to routing when using a link device is now link-centric rather than pci bus/device/pin specific. Thus, when using a tunable to override the default IRQ settings, one now uses a single tunable to route an entire link rather than routing a single device that uses the link (which has great foot-shooting potential if the user tries to route the same link to two different IRQs using two different pci bus/device/pin hints). For example, to adjust the IRQ that \_SB_.LNKA uses, one would set 'hw.pci.link.LNKA.irq=10' from the loader. - As a side effect of having the link driver, unused link devices will now be disabled when they are probed. - The algorithm for choosing an IRQ for a link that doesn't already have an IRQ assigned is now much closer to the one used in $PIR routing. When a link is routed via an ISA IRQ, only known-good IRQs that the BIOS has already used are used for routing instead of using probabilities to guess at which IRQs are probably not used by an ISA device. One change from $PIR is that the SCI is always considered a viable ISA IRQ, so that if the BIOS does not setup any IRQs the kernel will degenerate to routing all interrupts over the SCI. For non ISA IRQs, interrupts are picked from the possible pool using a simplistic weighting algorithm. Tested by: ru, scottl, others on acpi@ Reviewed by: njl
2004-11-23 22:26:44 +00:00
link->l_references++;
if (link->l_routed)
pci_link_interrupt_weights[link->l_irq]++;
/* Try to find a BIOS IRQ setting from any matching devices. */
bios_irq = acpi_pci_link_search_irq(pcib_get_bus(pcib), slot, pin);
if (!PCI_INTERRUPT_VALID(bios_irq)) {
ACPI_SERIAL_END(pci_link);
return;
}
Rework the ACPI PCI link code. - Use a new-bus device driver for the ACPI PCI link devices. The devices are called pci_linkX. The driver includes suspend/resume support so that the ACPI bridge drivers no longer have to poke the links to get them to handle suspend/resume. Also, the code to handle which IRQs a link is routed to and choosing an IRQ when a link is not already routed is all contained in the link driver. The PCI bridge drivers now ask the link driver which IRQ to use once they determine that a _PRT entry does not use a hardwired interrupt number. - The new link driver includes support for multiple IRQ resources per link device as well as preserving any non-IRQ resources when adjusting the IRQ that a link is routed to. - The entire approach to routing when using a link device is now link-centric rather than pci bus/device/pin specific. Thus, when using a tunable to override the default IRQ settings, one now uses a single tunable to route an entire link rather than routing a single device that uses the link (which has great foot-shooting potential if the user tries to route the same link to two different IRQs using two different pci bus/device/pin hints). For example, to adjust the IRQ that \_SB_.LNKA uses, one would set 'hw.pci.link.LNKA.irq=10' from the loader. - As a side effect of having the link driver, unused link devices will now be disabled when they are probed. - The algorithm for choosing an IRQ for a link that doesn't already have an IRQ assigned is now much closer to the one used in $PIR routing. When a link is routed via an ISA IRQ, only known-good IRQs that the BIOS has already used are used for routing instead of using probabilities to guess at which IRQs are probably not used by an ISA device. One change from $PIR is that the SCI is always considered a viable ISA IRQ, so that if the BIOS does not setup any IRQs the kernel will degenerate to routing all interrupts over the SCI. For non ISA IRQs, interrupts are picked from the possible pool using a simplistic weighting algorithm. Tested by: ru, scottl, others on acpi@ Reviewed by: njl
2004-11-23 22:26:44 +00:00
/* Validate the BIOS IRQ. */
if (!link_valid_irq(link, bios_irq)) {
device_printf(dev, "BIOS IRQ %u for %d.%d.INT%c is invalid\n",
bios_irq, pcib_get_bus(pcib), slot, pin + 'A');
} else if (!PCI_INTERRUPT_VALID(link->l_bios_irq)) {
link->l_bios_irq = bios_irq;
if (bios_irq < NUM_ISA_INTERRUPTS)
pci_link_bios_isa_irqs |= (1 << bios_irq);
if (bios_irq != link->l_initial_irq &&
PCI_INTERRUPT_VALID(link->l_initial_irq))
device_printf(dev,
"BIOS IRQ %u does not match initial IRQ %u\n",
bios_irq, link->l_initial_irq);
} else if (bios_irq != link->l_bios_irq)
device_printf(dev,
"BIOS IRQ %u for %d.%d.INT%c does not match previous BIOS IRQ %u\n",
bios_irq, pcib_get_bus(pcib), slot, pin + 'A',
link->l_bios_irq);
ACPI_SERIAL_END(pci_link);
}
static ACPI_STATUS
acpi_pci_link_srs_from_crs(struct acpi_pci_link_softc *sc, ACPI_BUFFER *srsbuf)
{
Rework the ACPI PCI link code. - Use a new-bus device driver for the ACPI PCI link devices. The devices are called pci_linkX. The driver includes suspend/resume support so that the ACPI bridge drivers no longer have to poke the links to get them to handle suspend/resume. Also, the code to handle which IRQs a link is routed to and choosing an IRQ when a link is not already routed is all contained in the link driver. The PCI bridge drivers now ask the link driver which IRQ to use once they determine that a _PRT entry does not use a hardwired interrupt number. - The new link driver includes support for multiple IRQ resources per link device as well as preserving any non-IRQ resources when adjusting the IRQ that a link is routed to. - The entire approach to routing when using a link device is now link-centric rather than pci bus/device/pin specific. Thus, when using a tunable to override the default IRQ settings, one now uses a single tunable to route an entire link rather than routing a single device that uses the link (which has great foot-shooting potential if the user tries to route the same link to two different IRQs using two different pci bus/device/pin hints). For example, to adjust the IRQ that \_SB_.LNKA uses, one would set 'hw.pci.link.LNKA.irq=10' from the loader. - As a side effect of having the link driver, unused link devices will now be disabled when they are probed. - The algorithm for choosing an IRQ for a link that doesn't already have an IRQ assigned is now much closer to the one used in $PIR routing. When a link is routed via an ISA IRQ, only known-good IRQs that the BIOS has already used are used for routing instead of using probabilities to guess at which IRQs are probably not used by an ISA device. One change from $PIR is that the SCI is always considered a viable ISA IRQ, so that if the BIOS does not setup any IRQs the kernel will degenerate to routing all interrupts over the SCI. For non ISA IRQs, interrupts are picked from the possible pool using a simplistic weighting algorithm. Tested by: ru, scottl, others on acpi@ Reviewed by: njl
2004-11-23 22:26:44 +00:00
ACPI_RESOURCE *resource, *end, newres, *resptr;
ACPI_BUFFER crsbuf;
Rework the ACPI PCI link code. - Use a new-bus device driver for the ACPI PCI link devices. The devices are called pci_linkX. The driver includes suspend/resume support so that the ACPI bridge drivers no longer have to poke the links to get them to handle suspend/resume. Also, the code to handle which IRQs a link is routed to and choosing an IRQ when a link is not already routed is all contained in the link driver. The PCI bridge drivers now ask the link driver which IRQ to use once they determine that a _PRT entry does not use a hardwired interrupt number. - The new link driver includes support for multiple IRQ resources per link device as well as preserving any non-IRQ resources when adjusting the IRQ that a link is routed to. - The entire approach to routing when using a link device is now link-centric rather than pci bus/device/pin specific. Thus, when using a tunable to override the default IRQ settings, one now uses a single tunable to route an entire link rather than routing a single device that uses the link (which has great foot-shooting potential if the user tries to route the same link to two different IRQs using two different pci bus/device/pin hints). For example, to adjust the IRQ that \_SB_.LNKA uses, one would set 'hw.pci.link.LNKA.irq=10' from the loader. - As a side effect of having the link driver, unused link devices will now be disabled when they are probed. - The algorithm for choosing an IRQ for a link that doesn't already have an IRQ assigned is now much closer to the one used in $PIR routing. When a link is routed via an ISA IRQ, only known-good IRQs that the BIOS has already used are used for routing instead of using probabilities to guess at which IRQs are probably not used by an ISA device. One change from $PIR is that the SCI is always considered a viable ISA IRQ, so that if the BIOS does not setup any IRQs the kernel will degenerate to routing all interrupts over the SCI. For non ISA IRQs, interrupts are picked from the possible pool using a simplistic weighting algorithm. Tested by: ru, scottl, others on acpi@ Reviewed by: njl
2004-11-23 22:26:44 +00:00
ACPI_STATUS status;
struct link *link;
int i, in_dpf;
Rework the ACPI PCI link code. - Use a new-bus device driver for the ACPI PCI link devices. The devices are called pci_linkX. The driver includes suspend/resume support so that the ACPI bridge drivers no longer have to poke the links to get them to handle suspend/resume. Also, the code to handle which IRQs a link is routed to and choosing an IRQ when a link is not already routed is all contained in the link driver. The PCI bridge drivers now ask the link driver which IRQ to use once they determine that a _PRT entry does not use a hardwired interrupt number. - The new link driver includes support for multiple IRQ resources per link device as well as preserving any non-IRQ resources when adjusting the IRQ that a link is routed to. - The entire approach to routing when using a link device is now link-centric rather than pci bus/device/pin specific. Thus, when using a tunable to override the default IRQ settings, one now uses a single tunable to route an entire link rather than routing a single device that uses the link (which has great foot-shooting potential if the user tries to route the same link to two different IRQs using two different pci bus/device/pin hints). For example, to adjust the IRQ that \_SB_.LNKA uses, one would set 'hw.pci.link.LNKA.irq=10' from the loader. - As a side effect of having the link driver, unused link devices will now be disabled when they are probed. - The algorithm for choosing an IRQ for a link that doesn't already have an IRQ assigned is now much closer to the one used in $PIR routing. When a link is routed via an ISA IRQ, only known-good IRQs that the BIOS has already used are used for routing instead of using probabilities to guess at which IRQs are probably not used by an ISA device. One change from $PIR is that the SCI is always considered a viable ISA IRQ, so that if the BIOS does not setup any IRQs the kernel will degenerate to routing all interrupts over the SCI. For non ISA IRQs, interrupts are picked from the possible pool using a simplistic weighting algorithm. Tested by: ru, scottl, others on acpi@ Reviewed by: njl
2004-11-23 22:26:44 +00:00
/* Fetch the _CRS. */
ACPI_SERIAL_ASSERT(pci_link);
crsbuf.Pointer = NULL;
Rework the ACPI PCI link code. - Use a new-bus device driver for the ACPI PCI link devices. The devices are called pci_linkX. The driver includes suspend/resume support so that the ACPI bridge drivers no longer have to poke the links to get them to handle suspend/resume. Also, the code to handle which IRQs a link is routed to and choosing an IRQ when a link is not already routed is all contained in the link driver. The PCI bridge drivers now ask the link driver which IRQ to use once they determine that a _PRT entry does not use a hardwired interrupt number. - The new link driver includes support for multiple IRQ resources per link device as well as preserving any non-IRQ resources when adjusting the IRQ that a link is routed to. - The entire approach to routing when using a link device is now link-centric rather than pci bus/device/pin specific. Thus, when using a tunable to override the default IRQ settings, one now uses a single tunable to route an entire link rather than routing a single device that uses the link (which has great foot-shooting potential if the user tries to route the same link to two different IRQs using two different pci bus/device/pin hints). For example, to adjust the IRQ that \_SB_.LNKA uses, one would set 'hw.pci.link.LNKA.irq=10' from the loader. - As a side effect of having the link driver, unused link devices will now be disabled when they are probed. - The algorithm for choosing an IRQ for a link that doesn't already have an IRQ assigned is now much closer to the one used in $PIR routing. When a link is routed via an ISA IRQ, only known-good IRQs that the BIOS has already used are used for routing instead of using probabilities to guess at which IRQs are probably not used by an ISA device. One change from $PIR is that the SCI is always considered a viable ISA IRQ, so that if the BIOS does not setup any IRQs the kernel will degenerate to routing all interrupts over the SCI. For non ISA IRQs, interrupts are picked from the possible pool using a simplistic weighting algorithm. Tested by: ru, scottl, others on acpi@ Reviewed by: njl
2004-11-23 22:26:44 +00:00
crsbuf.Length = ACPI_ALLOCATE_BUFFER;
status = AcpiGetCurrentResources(acpi_get_handle(sc->pl_dev), &crsbuf);
Rework the ACPI PCI link code. - Use a new-bus device driver for the ACPI PCI link devices. The devices are called pci_linkX. The driver includes suspend/resume support so that the ACPI bridge drivers no longer have to poke the links to get them to handle suspend/resume. Also, the code to handle which IRQs a link is routed to and choosing an IRQ when a link is not already routed is all contained in the link driver. The PCI bridge drivers now ask the link driver which IRQ to use once they determine that a _PRT entry does not use a hardwired interrupt number. - The new link driver includes support for multiple IRQ resources per link device as well as preserving any non-IRQ resources when adjusting the IRQ that a link is routed to. - The entire approach to routing when using a link device is now link-centric rather than pci bus/device/pin specific. Thus, when using a tunable to override the default IRQ settings, one now uses a single tunable to route an entire link rather than routing a single device that uses the link (which has great foot-shooting potential if the user tries to route the same link to two different IRQs using two different pci bus/device/pin hints). For example, to adjust the IRQ that \_SB_.LNKA uses, one would set 'hw.pci.link.LNKA.irq=10' from the loader. - As a side effect of having the link driver, unused link devices will now be disabled when they are probed. - The algorithm for choosing an IRQ for a link that doesn't already have an IRQ assigned is now much closer to the one used in $PIR routing. When a link is routed via an ISA IRQ, only known-good IRQs that the BIOS has already used are used for routing instead of using probabilities to guess at which IRQs are probably not used by an ISA device. One change from $PIR is that the SCI is always considered a viable ISA IRQ, so that if the BIOS does not setup any IRQs the kernel will degenerate to routing all interrupts over the SCI. For non ISA IRQs, interrupts are picked from the possible pool using a simplistic weighting algorithm. Tested by: ru, scottl, others on acpi@ Reviewed by: njl
2004-11-23 22:26:44 +00:00
if (ACPI_SUCCESS(status) && crsbuf.Pointer == NULL)
status = AE_NO_MEMORY;
if (ACPI_FAILURE(status)) {
if (bootverbose)
device_printf(sc->pl_dev,
Rework the ACPI PCI link code. - Use a new-bus device driver for the ACPI PCI link devices. The devices are called pci_linkX. The driver includes suspend/resume support so that the ACPI bridge drivers no longer have to poke the links to get them to handle suspend/resume. Also, the code to handle which IRQs a link is routed to and choosing an IRQ when a link is not already routed is all contained in the link driver. The PCI bridge drivers now ask the link driver which IRQ to use once they determine that a _PRT entry does not use a hardwired interrupt number. - The new link driver includes support for multiple IRQ resources per link device as well as preserving any non-IRQ resources when adjusting the IRQ that a link is routed to. - The entire approach to routing when using a link device is now link-centric rather than pci bus/device/pin specific. Thus, when using a tunable to override the default IRQ settings, one now uses a single tunable to route an entire link rather than routing a single device that uses the link (which has great foot-shooting potential if the user tries to route the same link to two different IRQs using two different pci bus/device/pin hints). For example, to adjust the IRQ that \_SB_.LNKA uses, one would set 'hw.pci.link.LNKA.irq=10' from the loader. - As a side effect of having the link driver, unused link devices will now be disabled when they are probed. - The algorithm for choosing an IRQ for a link that doesn't already have an IRQ assigned is now much closer to the one used in $PIR routing. When a link is routed via an ISA IRQ, only known-good IRQs that the BIOS has already used are used for routing instead of using probabilities to guess at which IRQs are probably not used by an ISA device. One change from $PIR is that the SCI is always considered a viable ISA IRQ, so that if the BIOS does not setup any IRQs the kernel will degenerate to routing all interrupts over the SCI. For non ISA IRQs, interrupts are picked from the possible pool using a simplistic weighting algorithm. Tested by: ru, scottl, others on acpi@ Reviewed by: njl
2004-11-23 22:26:44 +00:00
"Unable to fetch current resources: %s\n",
AcpiFormatException(status));
return (status);
}
/* Fill in IRQ resources via link structures. */
srsbuf->Pointer = NULL;
Rework the ACPI PCI link code. - Use a new-bus device driver for the ACPI PCI link devices. The devices are called pci_linkX. The driver includes suspend/resume support so that the ACPI bridge drivers no longer have to poke the links to get them to handle suspend/resume. Also, the code to handle which IRQs a link is routed to and choosing an IRQ when a link is not already routed is all contained in the link driver. The PCI bridge drivers now ask the link driver which IRQ to use once they determine that a _PRT entry does not use a hardwired interrupt number. - The new link driver includes support for multiple IRQ resources per link device as well as preserving any non-IRQ resources when adjusting the IRQ that a link is routed to. - The entire approach to routing when using a link device is now link-centric rather than pci bus/device/pin specific. Thus, when using a tunable to override the default IRQ settings, one now uses a single tunable to route an entire link rather than routing a single device that uses the link (which has great foot-shooting potential if the user tries to route the same link to two different IRQs using two different pci bus/device/pin hints). For example, to adjust the IRQ that \_SB_.LNKA uses, one would set 'hw.pci.link.LNKA.irq=10' from the loader. - As a side effect of having the link driver, unused link devices will now be disabled when they are probed. - The algorithm for choosing an IRQ for a link that doesn't already have an IRQ assigned is now much closer to the one used in $PIR routing. When a link is routed via an ISA IRQ, only known-good IRQs that the BIOS has already used are used for routing instead of using probabilities to guess at which IRQs are probably not used by an ISA device. One change from $PIR is that the SCI is always considered a viable ISA IRQ, so that if the BIOS does not setup any IRQs the kernel will degenerate to routing all interrupts over the SCI. For non ISA IRQs, interrupts are picked from the possible pool using a simplistic weighting algorithm. Tested by: ru, scottl, others on acpi@ Reviewed by: njl
2004-11-23 22:26:44 +00:00
link = sc->pl_links;
i = 0;
in_dpf = DPF_OUTSIDE;
Rework the ACPI PCI link code. - Use a new-bus device driver for the ACPI PCI link devices. The devices are called pci_linkX. The driver includes suspend/resume support so that the ACPI bridge drivers no longer have to poke the links to get them to handle suspend/resume. Also, the code to handle which IRQs a link is routed to and choosing an IRQ when a link is not already routed is all contained in the link driver. The PCI bridge drivers now ask the link driver which IRQ to use once they determine that a _PRT entry does not use a hardwired interrupt number. - The new link driver includes support for multiple IRQ resources per link device as well as preserving any non-IRQ resources when adjusting the IRQ that a link is routed to. - The entire approach to routing when using a link device is now link-centric rather than pci bus/device/pin specific. Thus, when using a tunable to override the default IRQ settings, one now uses a single tunable to route an entire link rather than routing a single device that uses the link (which has great foot-shooting potential if the user tries to route the same link to two different IRQs using two different pci bus/device/pin hints). For example, to adjust the IRQ that \_SB_.LNKA uses, one would set 'hw.pci.link.LNKA.irq=10' from the loader. - As a side effect of having the link driver, unused link devices will now be disabled when they are probed. - The algorithm for choosing an IRQ for a link that doesn't already have an IRQ assigned is now much closer to the one used in $PIR routing. When a link is routed via an ISA IRQ, only known-good IRQs that the BIOS has already used are used for routing instead of using probabilities to guess at which IRQs are probably not used by an ISA device. One change from $PIR is that the SCI is always considered a viable ISA IRQ, so that if the BIOS does not setup any IRQs the kernel will degenerate to routing all interrupts over the SCI. For non ISA IRQs, interrupts are picked from the possible pool using a simplistic weighting algorithm. Tested by: ru, scottl, others on acpi@ Reviewed by: njl
2004-11-23 22:26:44 +00:00
resource = (ACPI_RESOURCE *)crsbuf.Pointer;
end = (ACPI_RESOURCE *)((char *)crsbuf.Pointer + crsbuf.Length);
for (;;) {
2005-11-01 22:44:08 +00:00
switch (resource->Type) {
case ACPI_RESOURCE_TYPE_START_DEPENDENT:
switch (in_dpf) {
case DPF_OUTSIDE:
/* We've started the first DPF. */
in_dpf = DPF_FIRST;
break;
case DPF_FIRST:
/* We've started the second DPF. */
panic(
"%s: Multiple dependent functions within a current resource",
__func__);
break;
}
resptr = NULL;
break;
2005-11-01 22:44:08 +00:00
case ACPI_RESOURCE_TYPE_END_DEPENDENT:
/* We are finished with DPF parsing. */
KASSERT(in_dpf != DPF_OUTSIDE,
("%s: end dpf when not parsing a dpf", __func__));
in_dpf = DPF_OUTSIDE;
resptr = NULL;
break;
2005-11-01 22:44:08 +00:00
case ACPI_RESOURCE_TYPE_IRQ:
Rework the ACPI PCI link code. - Use a new-bus device driver for the ACPI PCI link devices. The devices are called pci_linkX. The driver includes suspend/resume support so that the ACPI bridge drivers no longer have to poke the links to get them to handle suspend/resume. Also, the code to handle which IRQs a link is routed to and choosing an IRQ when a link is not already routed is all contained in the link driver. The PCI bridge drivers now ask the link driver which IRQ to use once they determine that a _PRT entry does not use a hardwired interrupt number. - The new link driver includes support for multiple IRQ resources per link device as well as preserving any non-IRQ resources when adjusting the IRQ that a link is routed to. - The entire approach to routing when using a link device is now link-centric rather than pci bus/device/pin specific. Thus, when using a tunable to override the default IRQ settings, one now uses a single tunable to route an entire link rather than routing a single device that uses the link (which has great foot-shooting potential if the user tries to route the same link to two different IRQs using two different pci bus/device/pin hints). For example, to adjust the IRQ that \_SB_.LNKA uses, one would set 'hw.pci.link.LNKA.irq=10' from the loader. - As a side effect of having the link driver, unused link devices will now be disabled when they are probed. - The algorithm for choosing an IRQ for a link that doesn't already have an IRQ assigned is now much closer to the one used in $PIR routing. When a link is routed via an ISA IRQ, only known-good IRQs that the BIOS has already used are used for routing instead of using probabilities to guess at which IRQs are probably not used by an ISA device. One change from $PIR is that the SCI is always considered a viable ISA IRQ, so that if the BIOS does not setup any IRQs the kernel will degenerate to routing all interrupts over the SCI. For non ISA IRQs, interrupts are picked from the possible pool using a simplistic weighting algorithm. Tested by: ru, scottl, others on acpi@ Reviewed by: njl
2004-11-23 22:26:44 +00:00
MPASS(i < sc->pl_num_links);
2005-11-01 22:44:08 +00:00
MPASS(link->l_prs_template.Type == ACPI_RESOURCE_TYPE_IRQ);
Rework the ACPI PCI link code. - Use a new-bus device driver for the ACPI PCI link devices. The devices are called pci_linkX. The driver includes suspend/resume support so that the ACPI bridge drivers no longer have to poke the links to get them to handle suspend/resume. Also, the code to handle which IRQs a link is routed to and choosing an IRQ when a link is not already routed is all contained in the link driver. The PCI bridge drivers now ask the link driver which IRQ to use once they determine that a _PRT entry does not use a hardwired interrupt number. - The new link driver includes support for multiple IRQ resources per link device as well as preserving any non-IRQ resources when adjusting the IRQ that a link is routed to. - The entire approach to routing when using a link device is now link-centric rather than pci bus/device/pin specific. Thus, when using a tunable to override the default IRQ settings, one now uses a single tunable to route an entire link rather than routing a single device that uses the link (which has great foot-shooting potential if the user tries to route the same link to two different IRQs using two different pci bus/device/pin hints). For example, to adjust the IRQ that \_SB_.LNKA uses, one would set 'hw.pci.link.LNKA.irq=10' from the loader. - As a side effect of having the link driver, unused link devices will now be disabled when they are probed. - The algorithm for choosing an IRQ for a link that doesn't already have an IRQ assigned is now much closer to the one used in $PIR routing. When a link is routed via an ISA IRQ, only known-good IRQs that the BIOS has already used are used for routing instead of using probabilities to guess at which IRQs are probably not used by an ISA device. One change from $PIR is that the SCI is always considered a viable ISA IRQ, so that if the BIOS does not setup any IRQs the kernel will degenerate to routing all interrupts over the SCI. For non ISA IRQs, interrupts are picked from the possible pool using a simplistic weighting algorithm. Tested by: ru, scottl, others on acpi@ Reviewed by: njl
2004-11-23 22:26:44 +00:00
newres = link->l_prs_template;
resptr = &newres;
2005-11-01 22:44:08 +00:00
resptr->Data.Irq.InterruptCount = 1;
if (PCI_INTERRUPT_VALID(link->l_irq)) {
KASSERT(link->l_irq < NUM_ISA_INTERRUPTS,
("%s: can't put non-ISA IRQ %d in legacy IRQ resource type",
__func__, link->l_irq));
Rework the ACPI PCI link code. - Use a new-bus device driver for the ACPI PCI link devices. The devices are called pci_linkX. The driver includes suspend/resume support so that the ACPI bridge drivers no longer have to poke the links to get them to handle suspend/resume. Also, the code to handle which IRQs a link is routed to and choosing an IRQ when a link is not already routed is all contained in the link driver. The PCI bridge drivers now ask the link driver which IRQ to use once they determine that a _PRT entry does not use a hardwired interrupt number. - The new link driver includes support for multiple IRQ resources per link device as well as preserving any non-IRQ resources when adjusting the IRQ that a link is routed to. - The entire approach to routing when using a link device is now link-centric rather than pci bus/device/pin specific. Thus, when using a tunable to override the default IRQ settings, one now uses a single tunable to route an entire link rather than routing a single device that uses the link (which has great foot-shooting potential if the user tries to route the same link to two different IRQs using two different pci bus/device/pin hints). For example, to adjust the IRQ that \_SB_.LNKA uses, one would set 'hw.pci.link.LNKA.irq=10' from the loader. - As a side effect of having the link driver, unused link devices will now be disabled when they are probed. - The algorithm for choosing an IRQ for a link that doesn't already have an IRQ assigned is now much closer to the one used in $PIR routing. When a link is routed via an ISA IRQ, only known-good IRQs that the BIOS has already used are used for routing instead of using probabilities to guess at which IRQs are probably not used by an ISA device. One change from $PIR is that the SCI is always considered a viable ISA IRQ, so that if the BIOS does not setup any IRQs the kernel will degenerate to routing all interrupts over the SCI. For non ISA IRQs, interrupts are picked from the possible pool using a simplistic weighting algorithm. Tested by: ru, scottl, others on acpi@ Reviewed by: njl
2004-11-23 22:26:44 +00:00
resptr->Data.Irq.Interrupts[0] = link->l_irq;
} else
Rework the ACPI PCI link code. - Use a new-bus device driver for the ACPI PCI link devices. The devices are called pci_linkX. The driver includes suspend/resume support so that the ACPI bridge drivers no longer have to poke the links to get them to handle suspend/resume. Also, the code to handle which IRQs a link is routed to and choosing an IRQ when a link is not already routed is all contained in the link driver. The PCI bridge drivers now ask the link driver which IRQ to use once they determine that a _PRT entry does not use a hardwired interrupt number. - The new link driver includes support for multiple IRQ resources per link device as well as preserving any non-IRQ resources when adjusting the IRQ that a link is routed to. - The entire approach to routing when using a link device is now link-centric rather than pci bus/device/pin specific. Thus, when using a tunable to override the default IRQ settings, one now uses a single tunable to route an entire link rather than routing a single device that uses the link (which has great foot-shooting potential if the user tries to route the same link to two different IRQs using two different pci bus/device/pin hints). For example, to adjust the IRQ that \_SB_.LNKA uses, one would set 'hw.pci.link.LNKA.irq=10' from the loader. - As a side effect of having the link driver, unused link devices will now be disabled when they are probed. - The algorithm for choosing an IRQ for a link that doesn't already have an IRQ assigned is now much closer to the one used in $PIR routing. When a link is routed via an ISA IRQ, only known-good IRQs that the BIOS has already used are used for routing instead of using probabilities to guess at which IRQs are probably not used by an ISA device. One change from $PIR is that the SCI is always considered a viable ISA IRQ, so that if the BIOS does not setup any IRQs the kernel will degenerate to routing all interrupts over the SCI. For non ISA IRQs, interrupts are picked from the possible pool using a simplistic weighting algorithm. Tested by: ru, scottl, others on acpi@ Reviewed by: njl
2004-11-23 22:26:44 +00:00
resptr->Data.Irq.Interrupts[0] = 0;
link++;
i++;
Rework the ACPI PCI link code. - Use a new-bus device driver for the ACPI PCI link devices. The devices are called pci_linkX. The driver includes suspend/resume support so that the ACPI bridge drivers no longer have to poke the links to get them to handle suspend/resume. Also, the code to handle which IRQs a link is routed to and choosing an IRQ when a link is not already routed is all contained in the link driver. The PCI bridge drivers now ask the link driver which IRQ to use once they determine that a _PRT entry does not use a hardwired interrupt number. - The new link driver includes support for multiple IRQ resources per link device as well as preserving any non-IRQ resources when adjusting the IRQ that a link is routed to. - The entire approach to routing when using a link device is now link-centric rather than pci bus/device/pin specific. Thus, when using a tunable to override the default IRQ settings, one now uses a single tunable to route an entire link rather than routing a single device that uses the link (which has great foot-shooting potential if the user tries to route the same link to two different IRQs using two different pci bus/device/pin hints). For example, to adjust the IRQ that \_SB_.LNKA uses, one would set 'hw.pci.link.LNKA.irq=10' from the loader. - As a side effect of having the link driver, unused link devices will now be disabled when they are probed. - The algorithm for choosing an IRQ for a link that doesn't already have an IRQ assigned is now much closer to the one used in $PIR routing. When a link is routed via an ISA IRQ, only known-good IRQs that the BIOS has already used are used for routing instead of using probabilities to guess at which IRQs are probably not used by an ISA device. One change from $PIR is that the SCI is always considered a viable ISA IRQ, so that if the BIOS does not setup any IRQs the kernel will degenerate to routing all interrupts over the SCI. For non ISA IRQs, interrupts are picked from the possible pool using a simplistic weighting algorithm. Tested by: ru, scottl, others on acpi@ Reviewed by: njl
2004-11-23 22:26:44 +00:00
break;
2005-11-01 22:44:08 +00:00
case ACPI_RESOURCE_TYPE_EXTENDED_IRQ:
Rework the ACPI PCI link code. - Use a new-bus device driver for the ACPI PCI link devices. The devices are called pci_linkX. The driver includes suspend/resume support so that the ACPI bridge drivers no longer have to poke the links to get them to handle suspend/resume. Also, the code to handle which IRQs a link is routed to and choosing an IRQ when a link is not already routed is all contained in the link driver. The PCI bridge drivers now ask the link driver which IRQ to use once they determine that a _PRT entry does not use a hardwired interrupt number. - The new link driver includes support for multiple IRQ resources per link device as well as preserving any non-IRQ resources when adjusting the IRQ that a link is routed to. - The entire approach to routing when using a link device is now link-centric rather than pci bus/device/pin specific. Thus, when using a tunable to override the default IRQ settings, one now uses a single tunable to route an entire link rather than routing a single device that uses the link (which has great foot-shooting potential if the user tries to route the same link to two different IRQs using two different pci bus/device/pin hints). For example, to adjust the IRQ that \_SB_.LNKA uses, one would set 'hw.pci.link.LNKA.irq=10' from the loader. - As a side effect of having the link driver, unused link devices will now be disabled when they are probed. - The algorithm for choosing an IRQ for a link that doesn't already have an IRQ assigned is now much closer to the one used in $PIR routing. When a link is routed via an ISA IRQ, only known-good IRQs that the BIOS has already used are used for routing instead of using probabilities to guess at which IRQs are probably not used by an ISA device. One change from $PIR is that the SCI is always considered a viable ISA IRQ, so that if the BIOS does not setup any IRQs the kernel will degenerate to routing all interrupts over the SCI. For non ISA IRQs, interrupts are picked from the possible pool using a simplistic weighting algorithm. Tested by: ru, scottl, others on acpi@ Reviewed by: njl
2004-11-23 22:26:44 +00:00
MPASS(i < sc->pl_num_links);
2005-11-01 22:44:08 +00:00
MPASS(link->l_prs_template.Type == ACPI_RESOURCE_TYPE_EXTENDED_IRQ);
Rework the ACPI PCI link code. - Use a new-bus device driver for the ACPI PCI link devices. The devices are called pci_linkX. The driver includes suspend/resume support so that the ACPI bridge drivers no longer have to poke the links to get them to handle suspend/resume. Also, the code to handle which IRQs a link is routed to and choosing an IRQ when a link is not already routed is all contained in the link driver. The PCI bridge drivers now ask the link driver which IRQ to use once they determine that a _PRT entry does not use a hardwired interrupt number. - The new link driver includes support for multiple IRQ resources per link device as well as preserving any non-IRQ resources when adjusting the IRQ that a link is routed to. - The entire approach to routing when using a link device is now link-centric rather than pci bus/device/pin specific. Thus, when using a tunable to override the default IRQ settings, one now uses a single tunable to route an entire link rather than routing a single device that uses the link (which has great foot-shooting potential if the user tries to route the same link to two different IRQs using two different pci bus/device/pin hints). For example, to adjust the IRQ that \_SB_.LNKA uses, one would set 'hw.pci.link.LNKA.irq=10' from the loader. - As a side effect of having the link driver, unused link devices will now be disabled when they are probed. - The algorithm for choosing an IRQ for a link that doesn't already have an IRQ assigned is now much closer to the one used in $PIR routing. When a link is routed via an ISA IRQ, only known-good IRQs that the BIOS has already used are used for routing instead of using probabilities to guess at which IRQs are probably not used by an ISA device. One change from $PIR is that the SCI is always considered a viable ISA IRQ, so that if the BIOS does not setup any IRQs the kernel will degenerate to routing all interrupts over the SCI. For non ISA IRQs, interrupts are picked from the possible pool using a simplistic weighting algorithm. Tested by: ru, scottl, others on acpi@ Reviewed by: njl
2004-11-23 22:26:44 +00:00
newres = link->l_prs_template;
resptr = &newres;
2005-11-01 22:44:08 +00:00
resptr->Data.ExtendedIrq.InterruptCount = 1;
Rework the ACPI PCI link code. - Use a new-bus device driver for the ACPI PCI link devices. The devices are called pci_linkX. The driver includes suspend/resume support so that the ACPI bridge drivers no longer have to poke the links to get them to handle suspend/resume. Also, the code to handle which IRQs a link is routed to and choosing an IRQ when a link is not already routed is all contained in the link driver. The PCI bridge drivers now ask the link driver which IRQ to use once they determine that a _PRT entry does not use a hardwired interrupt number. - The new link driver includes support for multiple IRQ resources per link device as well as preserving any non-IRQ resources when adjusting the IRQ that a link is routed to. - The entire approach to routing when using a link device is now link-centric rather than pci bus/device/pin specific. Thus, when using a tunable to override the default IRQ settings, one now uses a single tunable to route an entire link rather than routing a single device that uses the link (which has great foot-shooting potential if the user tries to route the same link to two different IRQs using two different pci bus/device/pin hints). For example, to adjust the IRQ that \_SB_.LNKA uses, one would set 'hw.pci.link.LNKA.irq=10' from the loader. - As a side effect of having the link driver, unused link devices will now be disabled when they are probed. - The algorithm for choosing an IRQ for a link that doesn't already have an IRQ assigned is now much closer to the one used in $PIR routing. When a link is routed via an ISA IRQ, only known-good IRQs that the BIOS has already used are used for routing instead of using probabilities to guess at which IRQs are probably not used by an ISA device. One change from $PIR is that the SCI is always considered a viable ISA IRQ, so that if the BIOS does not setup any IRQs the kernel will degenerate to routing all interrupts over the SCI. For non ISA IRQs, interrupts are picked from the possible pool using a simplistic weighting algorithm. Tested by: ru, scottl, others on acpi@ Reviewed by: njl
2004-11-23 22:26:44 +00:00
if (PCI_INTERRUPT_VALID(link->l_irq))
resptr->Data.ExtendedIrq.Interrupts[0] =
Rework the ACPI PCI link code. - Use a new-bus device driver for the ACPI PCI link devices. The devices are called pci_linkX. The driver includes suspend/resume support so that the ACPI bridge drivers no longer have to poke the links to get them to handle suspend/resume. Also, the code to handle which IRQs a link is routed to and choosing an IRQ when a link is not already routed is all contained in the link driver. The PCI bridge drivers now ask the link driver which IRQ to use once they determine that a _PRT entry does not use a hardwired interrupt number. - The new link driver includes support for multiple IRQ resources per link device as well as preserving any non-IRQ resources when adjusting the IRQ that a link is routed to. - The entire approach to routing when using a link device is now link-centric rather than pci bus/device/pin specific. Thus, when using a tunable to override the default IRQ settings, one now uses a single tunable to route an entire link rather than routing a single device that uses the link (which has great foot-shooting potential if the user tries to route the same link to two different IRQs using two different pci bus/device/pin hints). For example, to adjust the IRQ that \_SB_.LNKA uses, one would set 'hw.pci.link.LNKA.irq=10' from the loader. - As a side effect of having the link driver, unused link devices will now be disabled when they are probed. - The algorithm for choosing an IRQ for a link that doesn't already have an IRQ assigned is now much closer to the one used in $PIR routing. When a link is routed via an ISA IRQ, only known-good IRQs that the BIOS has already used are used for routing instead of using probabilities to guess at which IRQs are probably not used by an ISA device. One change from $PIR is that the SCI is always considered a viable ISA IRQ, so that if the BIOS does not setup any IRQs the kernel will degenerate to routing all interrupts over the SCI. For non ISA IRQs, interrupts are picked from the possible pool using a simplistic weighting algorithm. Tested by: ru, scottl, others on acpi@ Reviewed by: njl
2004-11-23 22:26:44 +00:00
link->l_irq;
else
resptr->Data.ExtendedIrq.Interrupts[0] = 0;
link++;
i++;
Rework the ACPI PCI link code. - Use a new-bus device driver for the ACPI PCI link devices. The devices are called pci_linkX. The driver includes suspend/resume support so that the ACPI bridge drivers no longer have to poke the links to get them to handle suspend/resume. Also, the code to handle which IRQs a link is routed to and choosing an IRQ when a link is not already routed is all contained in the link driver. The PCI bridge drivers now ask the link driver which IRQ to use once they determine that a _PRT entry does not use a hardwired interrupt number. - The new link driver includes support for multiple IRQ resources per link device as well as preserving any non-IRQ resources when adjusting the IRQ that a link is routed to. - The entire approach to routing when using a link device is now link-centric rather than pci bus/device/pin specific. Thus, when using a tunable to override the default IRQ settings, one now uses a single tunable to route an entire link rather than routing a single device that uses the link (which has great foot-shooting potential if the user tries to route the same link to two different IRQs using two different pci bus/device/pin hints). For example, to adjust the IRQ that \_SB_.LNKA uses, one would set 'hw.pci.link.LNKA.irq=10' from the loader. - As a side effect of having the link driver, unused link devices will now be disabled when they are probed. - The algorithm for choosing an IRQ for a link that doesn't already have an IRQ assigned is now much closer to the one used in $PIR routing. When a link is routed via an ISA IRQ, only known-good IRQs that the BIOS has already used are used for routing instead of using probabilities to guess at which IRQs are probably not used by an ISA device. One change from $PIR is that the SCI is always considered a viable ISA IRQ, so that if the BIOS does not setup any IRQs the kernel will degenerate to routing all interrupts over the SCI. For non ISA IRQs, interrupts are picked from the possible pool using a simplistic weighting algorithm. Tested by: ru, scottl, others on acpi@ Reviewed by: njl
2004-11-23 22:26:44 +00:00
break;
default:
resptr = resource;
}
if (resptr != NULL) {
status = acpi_AppendBufferResource(srsbuf, resptr);
if (ACPI_FAILURE(status)) {
device_printf(sc->pl_dev,
"Unable to build resources: %s\n",
AcpiFormatException(status));
if (srsbuf->Pointer != NULL)
AcpiOsFree(srsbuf->Pointer);
AcpiOsFree(crsbuf.Pointer);
return (status);
}
Rework the ACPI PCI link code. - Use a new-bus device driver for the ACPI PCI link devices. The devices are called pci_linkX. The driver includes suspend/resume support so that the ACPI bridge drivers no longer have to poke the links to get them to handle suspend/resume. Also, the code to handle which IRQs a link is routed to and choosing an IRQ when a link is not already routed is all contained in the link driver. The PCI bridge drivers now ask the link driver which IRQ to use once they determine that a _PRT entry does not use a hardwired interrupt number. - The new link driver includes support for multiple IRQ resources per link device as well as preserving any non-IRQ resources when adjusting the IRQ that a link is routed to. - The entire approach to routing when using a link device is now link-centric rather than pci bus/device/pin specific. Thus, when using a tunable to override the default IRQ settings, one now uses a single tunable to route an entire link rather than routing a single device that uses the link (which has great foot-shooting potential if the user tries to route the same link to two different IRQs using two different pci bus/device/pin hints). For example, to adjust the IRQ that \_SB_.LNKA uses, one would set 'hw.pci.link.LNKA.irq=10' from the loader. - As a side effect of having the link driver, unused link devices will now be disabled when they are probed. - The algorithm for choosing an IRQ for a link that doesn't already have an IRQ assigned is now much closer to the one used in $PIR routing. When a link is routed via an ISA IRQ, only known-good IRQs that the BIOS has already used are used for routing instead of using probabilities to guess at which IRQs are probably not used by an ISA device. One change from $PIR is that the SCI is always considered a viable ISA IRQ, so that if the BIOS does not setup any IRQs the kernel will degenerate to routing all interrupts over the SCI. For non ISA IRQs, interrupts are picked from the possible pool using a simplistic weighting algorithm. Tested by: ru, scottl, others on acpi@ Reviewed by: njl
2004-11-23 22:26:44 +00:00
}
2005-11-01 22:44:08 +00:00
if (resource->Type == ACPI_RESOURCE_TYPE_END_TAG)
Rework the ACPI PCI link code. - Use a new-bus device driver for the ACPI PCI link devices. The devices are called pci_linkX. The driver includes suspend/resume support so that the ACPI bridge drivers no longer have to poke the links to get them to handle suspend/resume. Also, the code to handle which IRQs a link is routed to and choosing an IRQ when a link is not already routed is all contained in the link driver. The PCI bridge drivers now ask the link driver which IRQ to use once they determine that a _PRT entry does not use a hardwired interrupt number. - The new link driver includes support for multiple IRQ resources per link device as well as preserving any non-IRQ resources when adjusting the IRQ that a link is routed to. - The entire approach to routing when using a link device is now link-centric rather than pci bus/device/pin specific. Thus, when using a tunable to override the default IRQ settings, one now uses a single tunable to route an entire link rather than routing a single device that uses the link (which has great foot-shooting potential if the user tries to route the same link to two different IRQs using two different pci bus/device/pin hints). For example, to adjust the IRQ that \_SB_.LNKA uses, one would set 'hw.pci.link.LNKA.irq=10' from the loader. - As a side effect of having the link driver, unused link devices will now be disabled when they are probed. - The algorithm for choosing an IRQ for a link that doesn't already have an IRQ assigned is now much closer to the one used in $PIR routing. When a link is routed via an ISA IRQ, only known-good IRQs that the BIOS has already used are used for routing instead of using probabilities to guess at which IRQs are probably not used by an ISA device. One change from $PIR is that the SCI is always considered a viable ISA IRQ, so that if the BIOS does not setup any IRQs the kernel will degenerate to routing all interrupts over the SCI. For non ISA IRQs, interrupts are picked from the possible pool using a simplistic weighting algorithm. Tested by: ru, scottl, others on acpi@ Reviewed by: njl
2004-11-23 22:26:44 +00:00
break;
resource = ACPI_NEXT_RESOURCE(resource);
if (resource >= end)
break;
}
AcpiOsFree(crsbuf.Pointer);
return (AE_OK);
}
static ACPI_STATUS
acpi_pci_link_srs_from_links(struct acpi_pci_link_softc *sc,
ACPI_BUFFER *srsbuf)
{
ACPI_RESOURCE newres;
ACPI_STATUS status;
struct link *link;
int i;
/* Start off with an empty buffer. */
srsbuf->Pointer = NULL;
link = sc->pl_links;
for (i = 0; i < sc->pl_num_links; i++) {
/* Add a new IRQ resource from each link. */
link = &sc->pl_links[i];
newres = link->l_prs_template;
2005-11-01 22:44:08 +00:00
if (newres.Type == ACPI_RESOURCE_TYPE_IRQ) {
/* Build an IRQ resource. */
2005-11-01 22:44:08 +00:00
newres.Data.Irq.InterruptCount = 1;
if (PCI_INTERRUPT_VALID(link->l_irq)) {
KASSERT(link->l_irq < NUM_ISA_INTERRUPTS,
("%s: can't put non-ISA IRQ %d in legacy IRQ resource type",
__func__, link->l_irq));
newres.Data.Irq.Interrupts[0] = link->l_irq;
} else
newres.Data.Irq.Interrupts[0] = 0;
} else {
/* Build an ExtIRQ resuorce. */
2005-11-01 22:44:08 +00:00
newres.Data.ExtendedIrq.InterruptCount = 1;
if (PCI_INTERRUPT_VALID(link->l_irq))
newres.Data.ExtendedIrq.Interrupts[0] =
link->l_irq;
else
newres.Data.ExtendedIrq.Interrupts[0] = 0;
}
/* Add the new resource to the end of the _SRS buffer. */
status = acpi_AppendBufferResource(srsbuf, &newres);
if (ACPI_FAILURE(status)) {
device_printf(sc->pl_dev,
"Unable to build resources: %s\n",
AcpiFormatException(status));
if (srsbuf->Pointer != NULL)
AcpiOsFree(srsbuf->Pointer);
return (status);
}
}
return (AE_OK);
}
static ACPI_STATUS
acpi_pci_link_route_irqs(device_t dev)
{
struct acpi_pci_link_softc *sc;
ACPI_RESOURCE *resource, *end;
ACPI_BUFFER srsbuf;
ACPI_STATUS status;
struct link *link;
int i;
ACPI_SERIAL_ASSERT(pci_link);
sc = device_get_softc(dev);
if (sc->pl_crs_bad)
status = acpi_pci_link_srs_from_links(sc, &srsbuf);
else
status = acpi_pci_link_srs_from_crs(sc, &srsbuf);
Rework the ACPI PCI link code. - Use a new-bus device driver for the ACPI PCI link devices. The devices are called pci_linkX. The driver includes suspend/resume support so that the ACPI bridge drivers no longer have to poke the links to get them to handle suspend/resume. Also, the code to handle which IRQs a link is routed to and choosing an IRQ when a link is not already routed is all contained in the link driver. The PCI bridge drivers now ask the link driver which IRQ to use once they determine that a _PRT entry does not use a hardwired interrupt number. - The new link driver includes support for multiple IRQ resources per link device as well as preserving any non-IRQ resources when adjusting the IRQ that a link is routed to. - The entire approach to routing when using a link device is now link-centric rather than pci bus/device/pin specific. Thus, when using a tunable to override the default IRQ settings, one now uses a single tunable to route an entire link rather than routing a single device that uses the link (which has great foot-shooting potential if the user tries to route the same link to two different IRQs using two different pci bus/device/pin hints). For example, to adjust the IRQ that \_SB_.LNKA uses, one would set 'hw.pci.link.LNKA.irq=10' from the loader. - As a side effect of having the link driver, unused link devices will now be disabled when they are probed. - The algorithm for choosing an IRQ for a link that doesn't already have an IRQ assigned is now much closer to the one used in $PIR routing. When a link is routed via an ISA IRQ, only known-good IRQs that the BIOS has already used are used for routing instead of using probabilities to guess at which IRQs are probably not used by an ISA device. One change from $PIR is that the SCI is always considered a viable ISA IRQ, so that if the BIOS does not setup any IRQs the kernel will degenerate to routing all interrupts over the SCI. For non ISA IRQs, interrupts are picked from the possible pool using a simplistic weighting algorithm. Tested by: ru, scottl, others on acpi@ Reviewed by: njl
2004-11-23 22:26:44 +00:00
/* Write out new resources via _SRS. */
status = AcpiSetCurrentResources(acpi_get_handle(dev), &srsbuf);
if (ACPI_FAILURE(status)) {
device_printf(dev, "Unable to route IRQs: %s\n",
AcpiFormatException(status));
AcpiOsFree(srsbuf.Pointer);
return (status);
}
/*
Rework the ACPI PCI link code. - Use a new-bus device driver for the ACPI PCI link devices. The devices are called pci_linkX. The driver includes suspend/resume support so that the ACPI bridge drivers no longer have to poke the links to get them to handle suspend/resume. Also, the code to handle which IRQs a link is routed to and choosing an IRQ when a link is not already routed is all contained in the link driver. The PCI bridge drivers now ask the link driver which IRQ to use once they determine that a _PRT entry does not use a hardwired interrupt number. - The new link driver includes support for multiple IRQ resources per link device as well as preserving any non-IRQ resources when adjusting the IRQ that a link is routed to. - The entire approach to routing when using a link device is now link-centric rather than pci bus/device/pin specific. Thus, when using a tunable to override the default IRQ settings, one now uses a single tunable to route an entire link rather than routing a single device that uses the link (which has great foot-shooting potential if the user tries to route the same link to two different IRQs using two different pci bus/device/pin hints). For example, to adjust the IRQ that \_SB_.LNKA uses, one would set 'hw.pci.link.LNKA.irq=10' from the loader. - As a side effect of having the link driver, unused link devices will now be disabled when they are probed. - The algorithm for choosing an IRQ for a link that doesn't already have an IRQ assigned is now much closer to the one used in $PIR routing. When a link is routed via an ISA IRQ, only known-good IRQs that the BIOS has already used are used for routing instead of using probabilities to guess at which IRQs are probably not used by an ISA device. One change from $PIR is that the SCI is always considered a viable ISA IRQ, so that if the BIOS does not setup any IRQs the kernel will degenerate to routing all interrupts over the SCI. For non ISA IRQs, interrupts are picked from the possible pool using a simplistic weighting algorithm. Tested by: ru, scottl, others on acpi@ Reviewed by: njl
2004-11-23 22:26:44 +00:00
* Perform acpi_config_intr() on each IRQ resource if it was just
* routed for the first time.
*/
Rework the ACPI PCI link code. - Use a new-bus device driver for the ACPI PCI link devices. The devices are called pci_linkX. The driver includes suspend/resume support so that the ACPI bridge drivers no longer have to poke the links to get them to handle suspend/resume. Also, the code to handle which IRQs a link is routed to and choosing an IRQ when a link is not already routed is all contained in the link driver. The PCI bridge drivers now ask the link driver which IRQ to use once they determine that a _PRT entry does not use a hardwired interrupt number. - The new link driver includes support for multiple IRQ resources per link device as well as preserving any non-IRQ resources when adjusting the IRQ that a link is routed to. - The entire approach to routing when using a link device is now link-centric rather than pci bus/device/pin specific. Thus, when using a tunable to override the default IRQ settings, one now uses a single tunable to route an entire link rather than routing a single device that uses the link (which has great foot-shooting potential if the user tries to route the same link to two different IRQs using two different pci bus/device/pin hints). For example, to adjust the IRQ that \_SB_.LNKA uses, one would set 'hw.pci.link.LNKA.irq=10' from the loader. - As a side effect of having the link driver, unused link devices will now be disabled when they are probed. - The algorithm for choosing an IRQ for a link that doesn't already have an IRQ assigned is now much closer to the one used in $PIR routing. When a link is routed via an ISA IRQ, only known-good IRQs that the BIOS has already used are used for routing instead of using probabilities to guess at which IRQs are probably not used by an ISA device. One change from $PIR is that the SCI is always considered a viable ISA IRQ, so that if the BIOS does not setup any IRQs the kernel will degenerate to routing all interrupts over the SCI. For non ISA IRQs, interrupts are picked from the possible pool using a simplistic weighting algorithm. Tested by: ru, scottl, others on acpi@ Reviewed by: njl
2004-11-23 22:26:44 +00:00
link = sc->pl_links;
i = 0;
resource = (ACPI_RESOURCE *)srsbuf.Pointer;
end = (ACPI_RESOURCE *)((char *)srsbuf.Pointer + srsbuf.Length);
Rework the ACPI PCI link code. - Use a new-bus device driver for the ACPI PCI link devices. The devices are called pci_linkX. The driver includes suspend/resume support so that the ACPI bridge drivers no longer have to poke the links to get them to handle suspend/resume. Also, the code to handle which IRQs a link is routed to and choosing an IRQ when a link is not already routed is all contained in the link driver. The PCI bridge drivers now ask the link driver which IRQ to use once they determine that a _PRT entry does not use a hardwired interrupt number. - The new link driver includes support for multiple IRQ resources per link device as well as preserving any non-IRQ resources when adjusting the IRQ that a link is routed to. - The entire approach to routing when using a link device is now link-centric rather than pci bus/device/pin specific. Thus, when using a tunable to override the default IRQ settings, one now uses a single tunable to route an entire link rather than routing a single device that uses the link (which has great foot-shooting potential if the user tries to route the same link to two different IRQs using two different pci bus/device/pin hints). For example, to adjust the IRQ that \_SB_.LNKA uses, one would set 'hw.pci.link.LNKA.irq=10' from the loader. - As a side effect of having the link driver, unused link devices will now be disabled when they are probed. - The algorithm for choosing an IRQ for a link that doesn't already have an IRQ assigned is now much closer to the one used in $PIR routing. When a link is routed via an ISA IRQ, only known-good IRQs that the BIOS has already used are used for routing instead of using probabilities to guess at which IRQs are probably not used by an ISA device. One change from $PIR is that the SCI is always considered a viable ISA IRQ, so that if the BIOS does not setup any IRQs the kernel will degenerate to routing all interrupts over the SCI. For non ISA IRQs, interrupts are picked from the possible pool using a simplistic weighting algorithm. Tested by: ru, scottl, others on acpi@ Reviewed by: njl
2004-11-23 22:26:44 +00:00
for (;;) {
2005-11-01 22:44:08 +00:00
if (resource->Type == ACPI_RESOURCE_TYPE_END_TAG)
Rework the ACPI PCI link code. - Use a new-bus device driver for the ACPI PCI link devices. The devices are called pci_linkX. The driver includes suspend/resume support so that the ACPI bridge drivers no longer have to poke the links to get them to handle suspend/resume. Also, the code to handle which IRQs a link is routed to and choosing an IRQ when a link is not already routed is all contained in the link driver. The PCI bridge drivers now ask the link driver which IRQ to use once they determine that a _PRT entry does not use a hardwired interrupt number. - The new link driver includes support for multiple IRQ resources per link device as well as preserving any non-IRQ resources when adjusting the IRQ that a link is routed to. - The entire approach to routing when using a link device is now link-centric rather than pci bus/device/pin specific. Thus, when using a tunable to override the default IRQ settings, one now uses a single tunable to route an entire link rather than routing a single device that uses the link (which has great foot-shooting potential if the user tries to route the same link to two different IRQs using two different pci bus/device/pin hints). For example, to adjust the IRQ that \_SB_.LNKA uses, one would set 'hw.pci.link.LNKA.irq=10' from the loader. - As a side effect of having the link driver, unused link devices will now be disabled when they are probed. - The algorithm for choosing an IRQ for a link that doesn't already have an IRQ assigned is now much closer to the one used in $PIR routing. When a link is routed via an ISA IRQ, only known-good IRQs that the BIOS has already used are used for routing instead of using probabilities to guess at which IRQs are probably not used by an ISA device. One change from $PIR is that the SCI is always considered a viable ISA IRQ, so that if the BIOS does not setup any IRQs the kernel will degenerate to routing all interrupts over the SCI. For non ISA IRQs, interrupts are picked from the possible pool using a simplistic weighting algorithm. Tested by: ru, scottl, others on acpi@ Reviewed by: njl
2004-11-23 22:26:44 +00:00
break;
2005-11-01 22:44:08 +00:00
switch (resource->Type) {
case ACPI_RESOURCE_TYPE_IRQ:
case ACPI_RESOURCE_TYPE_EXTENDED_IRQ:
MPASS(i < sc->pl_num_links);
/*
* Only configure the interrupt and update the
* weights if this link has a valid IRQ and was
* previously unrouted.
*/
if (!link->l_routed &&
PCI_INTERRUPT_VALID(link->l_irq)) {
link->l_routed = TRUE;
acpi_config_intr(dev, resource);
pci_link_interrupt_weights[link->l_irq] +=
link->l_references;
}
link++;
i++;
Rework the ACPI PCI link code. - Use a new-bus device driver for the ACPI PCI link devices. The devices are called pci_linkX. The driver includes suspend/resume support so that the ACPI bridge drivers no longer have to poke the links to get them to handle suspend/resume. Also, the code to handle which IRQs a link is routed to and choosing an IRQ when a link is not already routed is all contained in the link driver. The PCI bridge drivers now ask the link driver which IRQ to use once they determine that a _PRT entry does not use a hardwired interrupt number. - The new link driver includes support for multiple IRQ resources per link device as well as preserving any non-IRQ resources when adjusting the IRQ that a link is routed to. - The entire approach to routing when using a link device is now link-centric rather than pci bus/device/pin specific. Thus, when using a tunable to override the default IRQ settings, one now uses a single tunable to route an entire link rather than routing a single device that uses the link (which has great foot-shooting potential if the user tries to route the same link to two different IRQs using two different pci bus/device/pin hints). For example, to adjust the IRQ that \_SB_.LNKA uses, one would set 'hw.pci.link.LNKA.irq=10' from the loader. - As a side effect of having the link driver, unused link devices will now be disabled when they are probed. - The algorithm for choosing an IRQ for a link that doesn't already have an IRQ assigned is now much closer to the one used in $PIR routing. When a link is routed via an ISA IRQ, only known-good IRQs that the BIOS has already used are used for routing instead of using probabilities to guess at which IRQs are probably not used by an ISA device. One change from $PIR is that the SCI is always considered a viable ISA IRQ, so that if the BIOS does not setup any IRQs the kernel will degenerate to routing all interrupts over the SCI. For non ISA IRQs, interrupts are picked from the possible pool using a simplistic weighting algorithm. Tested by: ru, scottl, others on acpi@ Reviewed by: njl
2004-11-23 22:26:44 +00:00
break;
}
Rework the ACPI PCI link code. - Use a new-bus device driver for the ACPI PCI link devices. The devices are called pci_linkX. The driver includes suspend/resume support so that the ACPI bridge drivers no longer have to poke the links to get them to handle suspend/resume. Also, the code to handle which IRQs a link is routed to and choosing an IRQ when a link is not already routed is all contained in the link driver. The PCI bridge drivers now ask the link driver which IRQ to use once they determine that a _PRT entry does not use a hardwired interrupt number. - The new link driver includes support for multiple IRQ resources per link device as well as preserving any non-IRQ resources when adjusting the IRQ that a link is routed to. - The entire approach to routing when using a link device is now link-centric rather than pci bus/device/pin specific. Thus, when using a tunable to override the default IRQ settings, one now uses a single tunable to route an entire link rather than routing a single device that uses the link (which has great foot-shooting potential if the user tries to route the same link to two different IRQs using two different pci bus/device/pin hints). For example, to adjust the IRQ that \_SB_.LNKA uses, one would set 'hw.pci.link.LNKA.irq=10' from the loader. - As a side effect of having the link driver, unused link devices will now be disabled when they are probed. - The algorithm for choosing an IRQ for a link that doesn't already have an IRQ assigned is now much closer to the one used in $PIR routing. When a link is routed via an ISA IRQ, only known-good IRQs that the BIOS has already used are used for routing instead of using probabilities to guess at which IRQs are probably not used by an ISA device. One change from $PIR is that the SCI is always considered a viable ISA IRQ, so that if the BIOS does not setup any IRQs the kernel will degenerate to routing all interrupts over the SCI. For non ISA IRQs, interrupts are picked from the possible pool using a simplistic weighting algorithm. Tested by: ru, scottl, others on acpi@ Reviewed by: njl
2004-11-23 22:26:44 +00:00
resource = ACPI_NEXT_RESOURCE(resource);
if (resource >= end)
break;
}
Rework the ACPI PCI link code. - Use a new-bus device driver for the ACPI PCI link devices. The devices are called pci_linkX. The driver includes suspend/resume support so that the ACPI bridge drivers no longer have to poke the links to get them to handle suspend/resume. Also, the code to handle which IRQs a link is routed to and choosing an IRQ when a link is not already routed is all contained in the link driver. The PCI bridge drivers now ask the link driver which IRQ to use once they determine that a _PRT entry does not use a hardwired interrupt number. - The new link driver includes support for multiple IRQ resources per link device as well as preserving any non-IRQ resources when adjusting the IRQ that a link is routed to. - The entire approach to routing when using a link device is now link-centric rather than pci bus/device/pin specific. Thus, when using a tunable to override the default IRQ settings, one now uses a single tunable to route an entire link rather than routing a single device that uses the link (which has great foot-shooting potential if the user tries to route the same link to two different IRQs using two different pci bus/device/pin hints). For example, to adjust the IRQ that \_SB_.LNKA uses, one would set 'hw.pci.link.LNKA.irq=10' from the loader. - As a side effect of having the link driver, unused link devices will now be disabled when they are probed. - The algorithm for choosing an IRQ for a link that doesn't already have an IRQ assigned is now much closer to the one used in $PIR routing. When a link is routed via an ISA IRQ, only known-good IRQs that the BIOS has already used are used for routing instead of using probabilities to guess at which IRQs are probably not used by an ISA device. One change from $PIR is that the SCI is always considered a viable ISA IRQ, so that if the BIOS does not setup any IRQs the kernel will degenerate to routing all interrupts over the SCI. For non ISA IRQs, interrupts are picked from the possible pool using a simplistic weighting algorithm. Tested by: ru, scottl, others on acpi@ Reviewed by: njl
2004-11-23 22:26:44 +00:00
AcpiOsFree(srsbuf.Pointer);
return (AE_OK);
}
static int
Rework the ACPI PCI link code. - Use a new-bus device driver for the ACPI PCI link devices. The devices are called pci_linkX. The driver includes suspend/resume support so that the ACPI bridge drivers no longer have to poke the links to get them to handle suspend/resume. Also, the code to handle which IRQs a link is routed to and choosing an IRQ when a link is not already routed is all contained in the link driver. The PCI bridge drivers now ask the link driver which IRQ to use once they determine that a _PRT entry does not use a hardwired interrupt number. - The new link driver includes support for multiple IRQ resources per link device as well as preserving any non-IRQ resources when adjusting the IRQ that a link is routed to. - The entire approach to routing when using a link device is now link-centric rather than pci bus/device/pin specific. Thus, when using a tunable to override the default IRQ settings, one now uses a single tunable to route an entire link rather than routing a single device that uses the link (which has great foot-shooting potential if the user tries to route the same link to two different IRQs using two different pci bus/device/pin hints). For example, to adjust the IRQ that \_SB_.LNKA uses, one would set 'hw.pci.link.LNKA.irq=10' from the loader. - As a side effect of having the link driver, unused link devices will now be disabled when they are probed. - The algorithm for choosing an IRQ for a link that doesn't already have an IRQ assigned is now much closer to the one used in $PIR routing. When a link is routed via an ISA IRQ, only known-good IRQs that the BIOS has already used are used for routing instead of using probabilities to guess at which IRQs are probably not used by an ISA device. One change from $PIR is that the SCI is always considered a viable ISA IRQ, so that if the BIOS does not setup any IRQs the kernel will degenerate to routing all interrupts over the SCI. For non ISA IRQs, interrupts are picked from the possible pool using a simplistic weighting algorithm. Tested by: ru, scottl, others on acpi@ Reviewed by: njl
2004-11-23 22:26:44 +00:00
acpi_pci_link_resume(device_t dev)
{
Rework the ACPI PCI link code. - Use a new-bus device driver for the ACPI PCI link devices. The devices are called pci_linkX. The driver includes suspend/resume support so that the ACPI bridge drivers no longer have to poke the links to get them to handle suspend/resume. Also, the code to handle which IRQs a link is routed to and choosing an IRQ when a link is not already routed is all contained in the link driver. The PCI bridge drivers now ask the link driver which IRQ to use once they determine that a _PRT entry does not use a hardwired interrupt number. - The new link driver includes support for multiple IRQ resources per link device as well as preserving any non-IRQ resources when adjusting the IRQ that a link is routed to. - The entire approach to routing when using a link device is now link-centric rather than pci bus/device/pin specific. Thus, when using a tunable to override the default IRQ settings, one now uses a single tunable to route an entire link rather than routing a single device that uses the link (which has great foot-shooting potential if the user tries to route the same link to two different IRQs using two different pci bus/device/pin hints). For example, to adjust the IRQ that \_SB_.LNKA uses, one would set 'hw.pci.link.LNKA.irq=10' from the loader. - As a side effect of having the link driver, unused link devices will now be disabled when they are probed. - The algorithm for choosing an IRQ for a link that doesn't already have an IRQ assigned is now much closer to the one used in $PIR routing. When a link is routed via an ISA IRQ, only known-good IRQs that the BIOS has already used are used for routing instead of using probabilities to guess at which IRQs are probably not used by an ISA device. One change from $PIR is that the SCI is always considered a viable ISA IRQ, so that if the BIOS does not setup any IRQs the kernel will degenerate to routing all interrupts over the SCI. For non ISA IRQs, interrupts are picked from the possible pool using a simplistic weighting algorithm. Tested by: ru, scottl, others on acpi@ Reviewed by: njl
2004-11-23 22:26:44 +00:00
ACPI_STATUS status;
Rework the ACPI PCI link code. - Use a new-bus device driver for the ACPI PCI link devices. The devices are called pci_linkX. The driver includes suspend/resume support so that the ACPI bridge drivers no longer have to poke the links to get them to handle suspend/resume. Also, the code to handle which IRQs a link is routed to and choosing an IRQ when a link is not already routed is all contained in the link driver. The PCI bridge drivers now ask the link driver which IRQ to use once they determine that a _PRT entry does not use a hardwired interrupt number. - The new link driver includes support for multiple IRQ resources per link device as well as preserving any non-IRQ resources when adjusting the IRQ that a link is routed to. - The entire approach to routing when using a link device is now link-centric rather than pci bus/device/pin specific. Thus, when using a tunable to override the default IRQ settings, one now uses a single tunable to route an entire link rather than routing a single device that uses the link (which has great foot-shooting potential if the user tries to route the same link to two different IRQs using two different pci bus/device/pin hints). For example, to adjust the IRQ that \_SB_.LNKA uses, one would set 'hw.pci.link.LNKA.irq=10' from the loader. - As a side effect of having the link driver, unused link devices will now be disabled when they are probed. - The algorithm for choosing an IRQ for a link that doesn't already have an IRQ assigned is now much closer to the one used in $PIR routing. When a link is routed via an ISA IRQ, only known-good IRQs that the BIOS has already used are used for routing instead of using probabilities to guess at which IRQs are probably not used by an ISA device. One change from $PIR is that the SCI is always considered a viable ISA IRQ, so that if the BIOS does not setup any IRQs the kernel will degenerate to routing all interrupts over the SCI. For non ISA IRQs, interrupts are picked from the possible pool using a simplistic weighting algorithm. Tested by: ru, scottl, others on acpi@ Reviewed by: njl
2004-11-23 22:26:44 +00:00
ACPI_SERIAL_BEGIN(pci_link);
status = acpi_pci_link_route_irqs(dev);
ACPI_SERIAL_END(pci_link);
if (ACPI_FAILURE(status))
return (ENXIO);
else
return (0);
}
Rework the ACPI PCI link code. - Use a new-bus device driver for the ACPI PCI link devices. The devices are called pci_linkX. The driver includes suspend/resume support so that the ACPI bridge drivers no longer have to poke the links to get them to handle suspend/resume. Also, the code to handle which IRQs a link is routed to and choosing an IRQ when a link is not already routed is all contained in the link driver. The PCI bridge drivers now ask the link driver which IRQ to use once they determine that a _PRT entry does not use a hardwired interrupt number. - The new link driver includes support for multiple IRQ resources per link device as well as preserving any non-IRQ resources when adjusting the IRQ that a link is routed to. - The entire approach to routing when using a link device is now link-centric rather than pci bus/device/pin specific. Thus, when using a tunable to override the default IRQ settings, one now uses a single tunable to route an entire link rather than routing a single device that uses the link (which has great foot-shooting potential if the user tries to route the same link to two different IRQs using two different pci bus/device/pin hints). For example, to adjust the IRQ that \_SB_.LNKA uses, one would set 'hw.pci.link.LNKA.irq=10' from the loader. - As a side effect of having the link driver, unused link devices will now be disabled when they are probed. - The algorithm for choosing an IRQ for a link that doesn't already have an IRQ assigned is now much closer to the one used in $PIR routing. When a link is routed via an ISA IRQ, only known-good IRQs that the BIOS has already used are used for routing instead of using probabilities to guess at which IRQs are probably not used by an ISA device. One change from $PIR is that the SCI is always considered a viable ISA IRQ, so that if the BIOS does not setup any IRQs the kernel will degenerate to routing all interrupts over the SCI. For non ISA IRQs, interrupts are picked from the possible pool using a simplistic weighting algorithm. Tested by: ru, scottl, others on acpi@ Reviewed by: njl
2004-11-23 22:26:44 +00:00
/*
* Pick an IRQ to use for this unrouted link.
*/
static uint8_t
acpi_pci_link_choose_irq(device_t dev, struct link *link)
{
Rework the ACPI PCI link code. - Use a new-bus device driver for the ACPI PCI link devices. The devices are called pci_linkX. The driver includes suspend/resume support so that the ACPI bridge drivers no longer have to poke the links to get them to handle suspend/resume. Also, the code to handle which IRQs a link is routed to and choosing an IRQ when a link is not already routed is all contained in the link driver. The PCI bridge drivers now ask the link driver which IRQ to use once they determine that a _PRT entry does not use a hardwired interrupt number. - The new link driver includes support for multiple IRQ resources per link device as well as preserving any non-IRQ resources when adjusting the IRQ that a link is routed to. - The entire approach to routing when using a link device is now link-centric rather than pci bus/device/pin specific. Thus, when using a tunable to override the default IRQ settings, one now uses a single tunable to route an entire link rather than routing a single device that uses the link (which has great foot-shooting potential if the user tries to route the same link to two different IRQs using two different pci bus/device/pin hints). For example, to adjust the IRQ that \_SB_.LNKA uses, one would set 'hw.pci.link.LNKA.irq=10' from the loader. - As a side effect of having the link driver, unused link devices will now be disabled when they are probed. - The algorithm for choosing an IRQ for a link that doesn't already have an IRQ assigned is now much closer to the one used in $PIR routing. When a link is routed via an ISA IRQ, only known-good IRQs that the BIOS has already used are used for routing instead of using probabilities to guess at which IRQs are probably not used by an ISA device. One change from $PIR is that the SCI is always considered a viable ISA IRQ, so that if the BIOS does not setup any IRQs the kernel will degenerate to routing all interrupts over the SCI. For non ISA IRQs, interrupts are picked from the possible pool using a simplistic weighting algorithm. Tested by: ru, scottl, others on acpi@ Reviewed by: njl
2004-11-23 22:26:44 +00:00
char tunable_buffer[64], link_name[5];
u_int8_t best_irq, pos_irq;
int best_weight, pos_weight, i;
KASSERT(!link->l_routed, ("%s: link already routed", __func__));
Rework the ACPI PCI link code. - Use a new-bus device driver for the ACPI PCI link devices. The devices are called pci_linkX. The driver includes suspend/resume support so that the ACPI bridge drivers no longer have to poke the links to get them to handle suspend/resume. Also, the code to handle which IRQs a link is routed to and choosing an IRQ when a link is not already routed is all contained in the link driver. The PCI bridge drivers now ask the link driver which IRQ to use once they determine that a _PRT entry does not use a hardwired interrupt number. - The new link driver includes support for multiple IRQ resources per link device as well as preserving any non-IRQ resources when adjusting the IRQ that a link is routed to. - The entire approach to routing when using a link device is now link-centric rather than pci bus/device/pin specific. Thus, when using a tunable to override the default IRQ settings, one now uses a single tunable to route an entire link rather than routing a single device that uses the link (which has great foot-shooting potential if the user tries to route the same link to two different IRQs using two different pci bus/device/pin hints). For example, to adjust the IRQ that \_SB_.LNKA uses, one would set 'hw.pci.link.LNKA.irq=10' from the loader. - As a side effect of having the link driver, unused link devices will now be disabled when they are probed. - The algorithm for choosing an IRQ for a link that doesn't already have an IRQ assigned is now much closer to the one used in $PIR routing. When a link is routed via an ISA IRQ, only known-good IRQs that the BIOS has already used are used for routing instead of using probabilities to guess at which IRQs are probably not used by an ISA device. One change from $PIR is that the SCI is always considered a viable ISA IRQ, so that if the BIOS does not setup any IRQs the kernel will degenerate to routing all interrupts over the SCI. For non ISA IRQs, interrupts are picked from the possible pool using a simplistic weighting algorithm. Tested by: ru, scottl, others on acpi@ Reviewed by: njl
2004-11-23 22:26:44 +00:00
KASSERT(!PCI_INTERRUPT_VALID(link->l_irq),
("%s: link already has an IRQ", __func__));
/* Check for a tunable override and use it if it is valid. */
if (ACPI_SUCCESS(acpi_short_name(acpi_get_handle(dev), link_name,
sizeof(link_name)))) {
snprintf(tunable_buffer, sizeof(tunable_buffer),
"hw.pci.link.%s.%d.irq", link_name, link->l_res_index);
if (getenv_int(tunable_buffer, &i) &&
PCI_INTERRUPT_VALID(i) && link_valid_irq(link, i))
return (i);
snprintf(tunable_buffer, sizeof(tunable_buffer),
"hw.pci.link.%s.irq", link_name);
if (getenv_int(tunable_buffer, &i) &&
PCI_INTERRUPT_VALID(i) && link_valid_irq(link, i))
return (i);
}
Rework the ACPI PCI link code. - Use a new-bus device driver for the ACPI PCI link devices. The devices are called pci_linkX. The driver includes suspend/resume support so that the ACPI bridge drivers no longer have to poke the links to get them to handle suspend/resume. Also, the code to handle which IRQs a link is routed to and choosing an IRQ when a link is not already routed is all contained in the link driver. The PCI bridge drivers now ask the link driver which IRQ to use once they determine that a _PRT entry does not use a hardwired interrupt number. - The new link driver includes support for multiple IRQ resources per link device as well as preserving any non-IRQ resources when adjusting the IRQ that a link is routed to. - The entire approach to routing when using a link device is now link-centric rather than pci bus/device/pin specific. Thus, when using a tunable to override the default IRQ settings, one now uses a single tunable to route an entire link rather than routing a single device that uses the link (which has great foot-shooting potential if the user tries to route the same link to two different IRQs using two different pci bus/device/pin hints). For example, to adjust the IRQ that \_SB_.LNKA uses, one would set 'hw.pci.link.LNKA.irq=10' from the loader. - As a side effect of having the link driver, unused link devices will now be disabled when they are probed. - The algorithm for choosing an IRQ for a link that doesn't already have an IRQ assigned is now much closer to the one used in $PIR routing. When a link is routed via an ISA IRQ, only known-good IRQs that the BIOS has already used are used for routing instead of using probabilities to guess at which IRQs are probably not used by an ISA device. One change from $PIR is that the SCI is always considered a viable ISA IRQ, so that if the BIOS does not setup any IRQs the kernel will degenerate to routing all interrupts over the SCI. For non ISA IRQs, interrupts are picked from the possible pool using a simplistic weighting algorithm. Tested by: ru, scottl, others on acpi@ Reviewed by: njl
2004-11-23 22:26:44 +00:00
/*
* If we have a valid BIOS IRQ, use that. We trust what the BIOS
* says it routed over what _CRS says the link thinks is routed.
*/
if (PCI_INTERRUPT_VALID(link->l_bios_irq))
return (link->l_bios_irq);
Rework the ACPI PCI link code. - Use a new-bus device driver for the ACPI PCI link devices. The devices are called pci_linkX. The driver includes suspend/resume support so that the ACPI bridge drivers no longer have to poke the links to get them to handle suspend/resume. Also, the code to handle which IRQs a link is routed to and choosing an IRQ when a link is not already routed is all contained in the link driver. The PCI bridge drivers now ask the link driver which IRQ to use once they determine that a _PRT entry does not use a hardwired interrupt number. - The new link driver includes support for multiple IRQ resources per link device as well as preserving any non-IRQ resources when adjusting the IRQ that a link is routed to. - The entire approach to routing when using a link device is now link-centric rather than pci bus/device/pin specific. Thus, when using a tunable to override the default IRQ settings, one now uses a single tunable to route an entire link rather than routing a single device that uses the link (which has great foot-shooting potential if the user tries to route the same link to two different IRQs using two different pci bus/device/pin hints). For example, to adjust the IRQ that \_SB_.LNKA uses, one would set 'hw.pci.link.LNKA.irq=10' from the loader. - As a side effect of having the link driver, unused link devices will now be disabled when they are probed. - The algorithm for choosing an IRQ for a link that doesn't already have an IRQ assigned is now much closer to the one used in $PIR routing. When a link is routed via an ISA IRQ, only known-good IRQs that the BIOS has already used are used for routing instead of using probabilities to guess at which IRQs are probably not used by an ISA device. One change from $PIR is that the SCI is always considered a viable ISA IRQ, so that if the BIOS does not setup any IRQs the kernel will degenerate to routing all interrupts over the SCI. For non ISA IRQs, interrupts are picked from the possible pool using a simplistic weighting algorithm. Tested by: ru, scottl, others on acpi@ Reviewed by: njl
2004-11-23 22:26:44 +00:00
/*
* If we don't have a BIOS IRQ but do have a valid IRQ from _CRS,
* then use that.
*/
if (PCI_INTERRUPT_VALID(link->l_initial_irq))
return (link->l_initial_irq);
Rework the ACPI PCI link code. - Use a new-bus device driver for the ACPI PCI link devices. The devices are called pci_linkX. The driver includes suspend/resume support so that the ACPI bridge drivers no longer have to poke the links to get them to handle suspend/resume. Also, the code to handle which IRQs a link is routed to and choosing an IRQ when a link is not already routed is all contained in the link driver. The PCI bridge drivers now ask the link driver which IRQ to use once they determine that a _PRT entry does not use a hardwired interrupt number. - The new link driver includes support for multiple IRQ resources per link device as well as preserving any non-IRQ resources when adjusting the IRQ that a link is routed to. - The entire approach to routing when using a link device is now link-centric rather than pci bus/device/pin specific. Thus, when using a tunable to override the default IRQ settings, one now uses a single tunable to route an entire link rather than routing a single device that uses the link (which has great foot-shooting potential if the user tries to route the same link to two different IRQs using two different pci bus/device/pin hints). For example, to adjust the IRQ that \_SB_.LNKA uses, one would set 'hw.pci.link.LNKA.irq=10' from the loader. - As a side effect of having the link driver, unused link devices will now be disabled when they are probed. - The algorithm for choosing an IRQ for a link that doesn't already have an IRQ assigned is now much closer to the one used in $PIR routing. When a link is routed via an ISA IRQ, only known-good IRQs that the BIOS has already used are used for routing instead of using probabilities to guess at which IRQs are probably not used by an ISA device. One change from $PIR is that the SCI is always considered a viable ISA IRQ, so that if the BIOS does not setup any IRQs the kernel will degenerate to routing all interrupts over the SCI. For non ISA IRQs, interrupts are picked from the possible pool using a simplistic weighting algorithm. Tested by: ru, scottl, others on acpi@ Reviewed by: njl
2004-11-23 22:26:44 +00:00
/*
* Ok, we have no useful hints, so we have to pick from the
* possible IRQs. For ISA IRQs we only use interrupts that
* have already been used by the BIOS.
*/
best_irq = PCI_INVALID_IRQ;
best_weight = INT_MAX;
for (i = 0; i < link->l_num_irqs; i++) {
pos_irq = link->l_irqs[i];
if (pos_irq < NUM_ISA_INTERRUPTS &&
(pci_link_bios_isa_irqs & 1 << pos_irq) == 0)
continue;
Rework the ACPI PCI link code. - Use a new-bus device driver for the ACPI PCI link devices. The devices are called pci_linkX. The driver includes suspend/resume support so that the ACPI bridge drivers no longer have to poke the links to get them to handle suspend/resume. Also, the code to handle which IRQs a link is routed to and choosing an IRQ when a link is not already routed is all contained in the link driver. The PCI bridge drivers now ask the link driver which IRQ to use once they determine that a _PRT entry does not use a hardwired interrupt number. - The new link driver includes support for multiple IRQ resources per link device as well as preserving any non-IRQ resources when adjusting the IRQ that a link is routed to. - The entire approach to routing when using a link device is now link-centric rather than pci bus/device/pin specific. Thus, when using a tunable to override the default IRQ settings, one now uses a single tunable to route an entire link rather than routing a single device that uses the link (which has great foot-shooting potential if the user tries to route the same link to two different IRQs using two different pci bus/device/pin hints). For example, to adjust the IRQ that \_SB_.LNKA uses, one would set 'hw.pci.link.LNKA.irq=10' from the loader. - As a side effect of having the link driver, unused link devices will now be disabled when they are probed. - The algorithm for choosing an IRQ for a link that doesn't already have an IRQ assigned is now much closer to the one used in $PIR routing. When a link is routed via an ISA IRQ, only known-good IRQs that the BIOS has already used are used for routing instead of using probabilities to guess at which IRQs are probably not used by an ISA device. One change from $PIR is that the SCI is always considered a viable ISA IRQ, so that if the BIOS does not setup any IRQs the kernel will degenerate to routing all interrupts over the SCI. For non ISA IRQs, interrupts are picked from the possible pool using a simplistic weighting algorithm. Tested by: ru, scottl, others on acpi@ Reviewed by: njl
2004-11-23 22:26:44 +00:00
pos_weight = pci_link_interrupt_weights[pos_irq];
if (pos_weight < best_weight) {
best_weight = pos_weight;
best_irq = pos_irq;
}
}
/*
Rework the ACPI PCI link code. - Use a new-bus device driver for the ACPI PCI link devices. The devices are called pci_linkX. The driver includes suspend/resume support so that the ACPI bridge drivers no longer have to poke the links to get them to handle suspend/resume. Also, the code to handle which IRQs a link is routed to and choosing an IRQ when a link is not already routed is all contained in the link driver. The PCI bridge drivers now ask the link driver which IRQ to use once they determine that a _PRT entry does not use a hardwired interrupt number. - The new link driver includes support for multiple IRQ resources per link device as well as preserving any non-IRQ resources when adjusting the IRQ that a link is routed to. - The entire approach to routing when using a link device is now link-centric rather than pci bus/device/pin specific. Thus, when using a tunable to override the default IRQ settings, one now uses a single tunable to route an entire link rather than routing a single device that uses the link (which has great foot-shooting potential if the user tries to route the same link to two different IRQs using two different pci bus/device/pin hints). For example, to adjust the IRQ that \_SB_.LNKA uses, one would set 'hw.pci.link.LNKA.irq=10' from the loader. - As a side effect of having the link driver, unused link devices will now be disabled when they are probed. - The algorithm for choosing an IRQ for a link that doesn't already have an IRQ assigned is now much closer to the one used in $PIR routing. When a link is routed via an ISA IRQ, only known-good IRQs that the BIOS has already used are used for routing instead of using probabilities to guess at which IRQs are probably not used by an ISA device. One change from $PIR is that the SCI is always considered a viable ISA IRQ, so that if the BIOS does not setup any IRQs the kernel will degenerate to routing all interrupts over the SCI. For non ISA IRQs, interrupts are picked from the possible pool using a simplistic weighting algorithm. Tested by: ru, scottl, others on acpi@ Reviewed by: njl
2004-11-23 22:26:44 +00:00
* If this is an ISA IRQ, try using the SCI if it is also an ISA
* interrupt as a fallback.
*/
Rework the ACPI PCI link code. - Use a new-bus device driver for the ACPI PCI link devices. The devices are called pci_linkX. The driver includes suspend/resume support so that the ACPI bridge drivers no longer have to poke the links to get them to handle suspend/resume. Also, the code to handle which IRQs a link is routed to and choosing an IRQ when a link is not already routed is all contained in the link driver. The PCI bridge drivers now ask the link driver which IRQ to use once they determine that a _PRT entry does not use a hardwired interrupt number. - The new link driver includes support for multiple IRQ resources per link device as well as preserving any non-IRQ resources when adjusting the IRQ that a link is routed to. - The entire approach to routing when using a link device is now link-centric rather than pci bus/device/pin specific. Thus, when using a tunable to override the default IRQ settings, one now uses a single tunable to route an entire link rather than routing a single device that uses the link (which has great foot-shooting potential if the user tries to route the same link to two different IRQs using two different pci bus/device/pin hints). For example, to adjust the IRQ that \_SB_.LNKA uses, one would set 'hw.pci.link.LNKA.irq=10' from the loader. - As a side effect of having the link driver, unused link devices will now be disabled when they are probed. - The algorithm for choosing an IRQ for a link that doesn't already have an IRQ assigned is now much closer to the one used in $PIR routing. When a link is routed via an ISA IRQ, only known-good IRQs that the BIOS has already used are used for routing instead of using probabilities to guess at which IRQs are probably not used by an ISA device. One change from $PIR is that the SCI is always considered a viable ISA IRQ, so that if the BIOS does not setup any IRQs the kernel will degenerate to routing all interrupts over the SCI. For non ISA IRQs, interrupts are picked from the possible pool using a simplistic weighting algorithm. Tested by: ru, scottl, others on acpi@ Reviewed by: njl
2004-11-23 22:26:44 +00:00
if (link->l_isa_irq) {
pos_irq = AcpiGbl_FADT->SciInt;
pos_weight = pci_link_interrupt_weights[pos_irq];
if (pos_weight < best_weight) {
best_weight = pos_weight;
best_irq = pos_irq;
}
}
if (PCI_INTERRUPT_VALID(best_irq)) {
if (bootverbose)
Rework the ACPI PCI link code. - Use a new-bus device driver for the ACPI PCI link devices. The devices are called pci_linkX. The driver includes suspend/resume support so that the ACPI bridge drivers no longer have to poke the links to get them to handle suspend/resume. Also, the code to handle which IRQs a link is routed to and choosing an IRQ when a link is not already routed is all contained in the link driver. The PCI bridge drivers now ask the link driver which IRQ to use once they determine that a _PRT entry does not use a hardwired interrupt number. - The new link driver includes support for multiple IRQ resources per link device as well as preserving any non-IRQ resources when adjusting the IRQ that a link is routed to. - The entire approach to routing when using a link device is now link-centric rather than pci bus/device/pin specific. Thus, when using a tunable to override the default IRQ settings, one now uses a single tunable to route an entire link rather than routing a single device that uses the link (which has great foot-shooting potential if the user tries to route the same link to two different IRQs using two different pci bus/device/pin hints). For example, to adjust the IRQ that \_SB_.LNKA uses, one would set 'hw.pci.link.LNKA.irq=10' from the loader. - As a side effect of having the link driver, unused link devices will now be disabled when they are probed. - The algorithm for choosing an IRQ for a link that doesn't already have an IRQ assigned is now much closer to the one used in $PIR routing. When a link is routed via an ISA IRQ, only known-good IRQs that the BIOS has already used are used for routing instead of using probabilities to guess at which IRQs are probably not used by an ISA device. One change from $PIR is that the SCI is always considered a viable ISA IRQ, so that if the BIOS does not setup any IRQs the kernel will degenerate to routing all interrupts over the SCI. For non ISA IRQs, interrupts are picked from the possible pool using a simplistic weighting algorithm. Tested by: ru, scottl, others on acpi@ Reviewed by: njl
2004-11-23 22:26:44 +00:00
device_printf(dev, "Picked IRQ %u with weight %d\n",
best_irq, best_weight);
} else
device_printf(dev, "Unable to choose an IRQ\n");
return (best_irq);
}
int
Rework the ACPI PCI link code. - Use a new-bus device driver for the ACPI PCI link devices. The devices are called pci_linkX. The driver includes suspend/resume support so that the ACPI bridge drivers no longer have to poke the links to get them to handle suspend/resume. Also, the code to handle which IRQs a link is routed to and choosing an IRQ when a link is not already routed is all contained in the link driver. The PCI bridge drivers now ask the link driver which IRQ to use once they determine that a _PRT entry does not use a hardwired interrupt number. - The new link driver includes support for multiple IRQ resources per link device as well as preserving any non-IRQ resources when adjusting the IRQ that a link is routed to. - The entire approach to routing when using a link device is now link-centric rather than pci bus/device/pin specific. Thus, when using a tunable to override the default IRQ settings, one now uses a single tunable to route an entire link rather than routing a single device that uses the link (which has great foot-shooting potential if the user tries to route the same link to two different IRQs using two different pci bus/device/pin hints). For example, to adjust the IRQ that \_SB_.LNKA uses, one would set 'hw.pci.link.LNKA.irq=10' from the loader. - As a side effect of having the link driver, unused link devices will now be disabled when they are probed. - The algorithm for choosing an IRQ for a link that doesn't already have an IRQ assigned is now much closer to the one used in $PIR routing. When a link is routed via an ISA IRQ, only known-good IRQs that the BIOS has already used are used for routing instead of using probabilities to guess at which IRQs are probably not used by an ISA device. One change from $PIR is that the SCI is always considered a viable ISA IRQ, so that if the BIOS does not setup any IRQs the kernel will degenerate to routing all interrupts over the SCI. For non ISA IRQs, interrupts are picked from the possible pool using a simplistic weighting algorithm. Tested by: ru, scottl, others on acpi@ Reviewed by: njl
2004-11-23 22:26:44 +00:00
acpi_pci_link_route_interrupt(device_t dev, int index)
{
Rework the ACPI PCI link code. - Use a new-bus device driver for the ACPI PCI link devices. The devices are called pci_linkX. The driver includes suspend/resume support so that the ACPI bridge drivers no longer have to poke the links to get them to handle suspend/resume. Also, the code to handle which IRQs a link is routed to and choosing an IRQ when a link is not already routed is all contained in the link driver. The PCI bridge drivers now ask the link driver which IRQ to use once they determine that a _PRT entry does not use a hardwired interrupt number. - The new link driver includes support for multiple IRQ resources per link device as well as preserving any non-IRQ resources when adjusting the IRQ that a link is routed to. - The entire approach to routing when using a link device is now link-centric rather than pci bus/device/pin specific. Thus, when using a tunable to override the default IRQ settings, one now uses a single tunable to route an entire link rather than routing a single device that uses the link (which has great foot-shooting potential if the user tries to route the same link to two different IRQs using two different pci bus/device/pin hints). For example, to adjust the IRQ that \_SB_.LNKA uses, one would set 'hw.pci.link.LNKA.irq=10' from the loader. - As a side effect of having the link driver, unused link devices will now be disabled when they are probed. - The algorithm for choosing an IRQ for a link that doesn't already have an IRQ assigned is now much closer to the one used in $PIR routing. When a link is routed via an ISA IRQ, only known-good IRQs that the BIOS has already used are used for routing instead of using probabilities to guess at which IRQs are probably not used by an ISA device. One change from $PIR is that the SCI is always considered a viable ISA IRQ, so that if the BIOS does not setup any IRQs the kernel will degenerate to routing all interrupts over the SCI. For non ISA IRQs, interrupts are picked from the possible pool using a simplistic weighting algorithm. Tested by: ru, scottl, others on acpi@ Reviewed by: njl
2004-11-23 22:26:44 +00:00
struct link *link;
ACPI_SERIAL_BEGIN(pci_link);
link = acpi_pci_link_lookup(dev, index);
if (link == NULL)
panic("%s: apparently invalid index %d", __func__, index);
Rework the ACPI PCI link code. - Use a new-bus device driver for the ACPI PCI link devices. The devices are called pci_linkX. The driver includes suspend/resume support so that the ACPI bridge drivers no longer have to poke the links to get them to handle suspend/resume. Also, the code to handle which IRQs a link is routed to and choosing an IRQ when a link is not already routed is all contained in the link driver. The PCI bridge drivers now ask the link driver which IRQ to use once they determine that a _PRT entry does not use a hardwired interrupt number. - The new link driver includes support for multiple IRQ resources per link device as well as preserving any non-IRQ resources when adjusting the IRQ that a link is routed to. - The entire approach to routing when using a link device is now link-centric rather than pci bus/device/pin specific. Thus, when using a tunable to override the default IRQ settings, one now uses a single tunable to route an entire link rather than routing a single device that uses the link (which has great foot-shooting potential if the user tries to route the same link to two different IRQs using two different pci bus/device/pin hints). For example, to adjust the IRQ that \_SB_.LNKA uses, one would set 'hw.pci.link.LNKA.irq=10' from the loader. - As a side effect of having the link driver, unused link devices will now be disabled when they are probed. - The algorithm for choosing an IRQ for a link that doesn't already have an IRQ assigned is now much closer to the one used in $PIR routing. When a link is routed via an ISA IRQ, only known-good IRQs that the BIOS has already used are used for routing instead of using probabilities to guess at which IRQs are probably not used by an ISA device. One change from $PIR is that the SCI is always considered a viable ISA IRQ, so that if the BIOS does not setup any IRQs the kernel will degenerate to routing all interrupts over the SCI. For non ISA IRQs, interrupts are picked from the possible pool using a simplistic weighting algorithm. Tested by: ru, scottl, others on acpi@ Reviewed by: njl
2004-11-23 22:26:44 +00:00
/*
* If this link device is already routed to an interrupt, just return
* the interrupt it is routed to.
*/
if (link->l_routed) {
KASSERT(PCI_INTERRUPT_VALID(link->l_irq),
("%s: link is routed but has an invalid IRQ", __func__));
ACPI_SERIAL_END(pci_link);
return (link->l_irq);
}
Rework the ACPI PCI link code. - Use a new-bus device driver for the ACPI PCI link devices. The devices are called pci_linkX. The driver includes suspend/resume support so that the ACPI bridge drivers no longer have to poke the links to get them to handle suspend/resume. Also, the code to handle which IRQs a link is routed to and choosing an IRQ when a link is not already routed is all contained in the link driver. The PCI bridge drivers now ask the link driver which IRQ to use once they determine that a _PRT entry does not use a hardwired interrupt number. - The new link driver includes support for multiple IRQ resources per link device as well as preserving any non-IRQ resources when adjusting the IRQ that a link is routed to. - The entire approach to routing when using a link device is now link-centric rather than pci bus/device/pin specific. Thus, when using a tunable to override the default IRQ settings, one now uses a single tunable to route an entire link rather than routing a single device that uses the link (which has great foot-shooting potential if the user tries to route the same link to two different IRQs using two different pci bus/device/pin hints). For example, to adjust the IRQ that \_SB_.LNKA uses, one would set 'hw.pci.link.LNKA.irq=10' from the loader. - As a side effect of having the link driver, unused link devices will now be disabled when they are probed. - The algorithm for choosing an IRQ for a link that doesn't already have an IRQ assigned is now much closer to the one used in $PIR routing. When a link is routed via an ISA IRQ, only known-good IRQs that the BIOS has already used are used for routing instead of using probabilities to guess at which IRQs are probably not used by an ISA device. One change from $PIR is that the SCI is always considered a viable ISA IRQ, so that if the BIOS does not setup any IRQs the kernel will degenerate to routing all interrupts over the SCI. For non ISA IRQs, interrupts are picked from the possible pool using a simplistic weighting algorithm. Tested by: ru, scottl, others on acpi@ Reviewed by: njl
2004-11-23 22:26:44 +00:00
/* Choose an IRQ if we need one. */
if (!PCI_INTERRUPT_VALID(link->l_irq)) {
link->l_irq = acpi_pci_link_choose_irq(dev, link);
/*
Rework the ACPI PCI link code. - Use a new-bus device driver for the ACPI PCI link devices. The devices are called pci_linkX. The driver includes suspend/resume support so that the ACPI bridge drivers no longer have to poke the links to get them to handle suspend/resume. Also, the code to handle which IRQs a link is routed to and choosing an IRQ when a link is not already routed is all contained in the link driver. The PCI bridge drivers now ask the link driver which IRQ to use once they determine that a _PRT entry does not use a hardwired interrupt number. - The new link driver includes support for multiple IRQ resources per link device as well as preserving any non-IRQ resources when adjusting the IRQ that a link is routed to. - The entire approach to routing when using a link device is now link-centric rather than pci bus/device/pin specific. Thus, when using a tunable to override the default IRQ settings, one now uses a single tunable to route an entire link rather than routing a single device that uses the link (which has great foot-shooting potential if the user tries to route the same link to two different IRQs using two different pci bus/device/pin hints). For example, to adjust the IRQ that \_SB_.LNKA uses, one would set 'hw.pci.link.LNKA.irq=10' from the loader. - As a side effect of having the link driver, unused link devices will now be disabled when they are probed. - The algorithm for choosing an IRQ for a link that doesn't already have an IRQ assigned is now much closer to the one used in $PIR routing. When a link is routed via an ISA IRQ, only known-good IRQs that the BIOS has already used are used for routing instead of using probabilities to guess at which IRQs are probably not used by an ISA device. One change from $PIR is that the SCI is always considered a viable ISA IRQ, so that if the BIOS does not setup any IRQs the kernel will degenerate to routing all interrupts over the SCI. For non ISA IRQs, interrupts are picked from the possible pool using a simplistic weighting algorithm. Tested by: ru, scottl, others on acpi@ Reviewed by: njl
2004-11-23 22:26:44 +00:00
* Try to route the interrupt we picked. If it fails, then
* assume the interrupt is not routed.
*/
Rework the ACPI PCI link code. - Use a new-bus device driver for the ACPI PCI link devices. The devices are called pci_linkX. The driver includes suspend/resume support so that the ACPI bridge drivers no longer have to poke the links to get them to handle suspend/resume. Also, the code to handle which IRQs a link is routed to and choosing an IRQ when a link is not already routed is all contained in the link driver. The PCI bridge drivers now ask the link driver which IRQ to use once they determine that a _PRT entry does not use a hardwired interrupt number. - The new link driver includes support for multiple IRQ resources per link device as well as preserving any non-IRQ resources when adjusting the IRQ that a link is routed to. - The entire approach to routing when using a link device is now link-centric rather than pci bus/device/pin specific. Thus, when using a tunable to override the default IRQ settings, one now uses a single tunable to route an entire link rather than routing a single device that uses the link (which has great foot-shooting potential if the user tries to route the same link to two different IRQs using two different pci bus/device/pin hints). For example, to adjust the IRQ that \_SB_.LNKA uses, one would set 'hw.pci.link.LNKA.irq=10' from the loader. - As a side effect of having the link driver, unused link devices will now be disabled when they are probed. - The algorithm for choosing an IRQ for a link that doesn't already have an IRQ assigned is now much closer to the one used in $PIR routing. When a link is routed via an ISA IRQ, only known-good IRQs that the BIOS has already used are used for routing instead of using probabilities to guess at which IRQs are probably not used by an ISA device. One change from $PIR is that the SCI is always considered a viable ISA IRQ, so that if the BIOS does not setup any IRQs the kernel will degenerate to routing all interrupts over the SCI. For non ISA IRQs, interrupts are picked from the possible pool using a simplistic weighting algorithm. Tested by: ru, scottl, others on acpi@ Reviewed by: njl
2004-11-23 22:26:44 +00:00
if (PCI_INTERRUPT_VALID(link->l_irq)) {
acpi_pci_link_route_irqs(dev);
if (!link->l_routed)
link->l_irq = PCI_INVALID_IRQ;
}
}
ACPI_SERIAL_END(pci_link);
Rework the ACPI PCI link code. - Use a new-bus device driver for the ACPI PCI link devices. The devices are called pci_linkX. The driver includes suspend/resume support so that the ACPI bridge drivers no longer have to poke the links to get them to handle suspend/resume. Also, the code to handle which IRQs a link is routed to and choosing an IRQ when a link is not already routed is all contained in the link driver. The PCI bridge drivers now ask the link driver which IRQ to use once they determine that a _PRT entry does not use a hardwired interrupt number. - The new link driver includes support for multiple IRQ resources per link device as well as preserving any non-IRQ resources when adjusting the IRQ that a link is routed to. - The entire approach to routing when using a link device is now link-centric rather than pci bus/device/pin specific. Thus, when using a tunable to override the default IRQ settings, one now uses a single tunable to route an entire link rather than routing a single device that uses the link (which has great foot-shooting potential if the user tries to route the same link to two different IRQs using two different pci bus/device/pin hints). For example, to adjust the IRQ that \_SB_.LNKA uses, one would set 'hw.pci.link.LNKA.irq=10' from the loader. - As a side effect of having the link driver, unused link devices will now be disabled when they are probed. - The algorithm for choosing an IRQ for a link that doesn't already have an IRQ assigned is now much closer to the one used in $PIR routing. When a link is routed via an ISA IRQ, only known-good IRQs that the BIOS has already used are used for routing instead of using probabilities to guess at which IRQs are probably not used by an ISA device. One change from $PIR is that the SCI is always considered a viable ISA IRQ, so that if the BIOS does not setup any IRQs the kernel will degenerate to routing all interrupts over the SCI. For non ISA IRQs, interrupts are picked from the possible pool using a simplistic weighting algorithm. Tested by: ru, scottl, others on acpi@ Reviewed by: njl
2004-11-23 22:26:44 +00:00
return (link->l_irq);
}
/*
Rework the ACPI PCI link code. - Use a new-bus device driver for the ACPI PCI link devices. The devices are called pci_linkX. The driver includes suspend/resume support so that the ACPI bridge drivers no longer have to poke the links to get them to handle suspend/resume. Also, the code to handle which IRQs a link is routed to and choosing an IRQ when a link is not already routed is all contained in the link driver. The PCI bridge drivers now ask the link driver which IRQ to use once they determine that a _PRT entry does not use a hardwired interrupt number. - The new link driver includes support for multiple IRQ resources per link device as well as preserving any non-IRQ resources when adjusting the IRQ that a link is routed to. - The entire approach to routing when using a link device is now link-centric rather than pci bus/device/pin specific. Thus, when using a tunable to override the default IRQ settings, one now uses a single tunable to route an entire link rather than routing a single device that uses the link (which has great foot-shooting potential if the user tries to route the same link to two different IRQs using two different pci bus/device/pin hints). For example, to adjust the IRQ that \_SB_.LNKA uses, one would set 'hw.pci.link.LNKA.irq=10' from the loader. - As a side effect of having the link driver, unused link devices will now be disabled when they are probed. - The algorithm for choosing an IRQ for a link that doesn't already have an IRQ assigned is now much closer to the one used in $PIR routing. When a link is routed via an ISA IRQ, only known-good IRQs that the BIOS has already used are used for routing instead of using probabilities to guess at which IRQs are probably not used by an ISA device. One change from $PIR is that the SCI is always considered a viable ISA IRQ, so that if the BIOS does not setup any IRQs the kernel will degenerate to routing all interrupts over the SCI. For non ISA IRQs, interrupts are picked from the possible pool using a simplistic weighting algorithm. Tested by: ru, scottl, others on acpi@ Reviewed by: njl
2004-11-23 22:26:44 +00:00
* This is gross, but we abuse the identify routine to perform one-time
* SYSINIT() style initialization for the driver.
*/
Rework the ACPI PCI link code. - Use a new-bus device driver for the ACPI PCI link devices. The devices are called pci_linkX. The driver includes suspend/resume support so that the ACPI bridge drivers no longer have to poke the links to get them to handle suspend/resume. Also, the code to handle which IRQs a link is routed to and choosing an IRQ when a link is not already routed is all contained in the link driver. The PCI bridge drivers now ask the link driver which IRQ to use once they determine that a _PRT entry does not use a hardwired interrupt number. - The new link driver includes support for multiple IRQ resources per link device as well as preserving any non-IRQ resources when adjusting the IRQ that a link is routed to. - The entire approach to routing when using a link device is now link-centric rather than pci bus/device/pin specific. Thus, when using a tunable to override the default IRQ settings, one now uses a single tunable to route an entire link rather than routing a single device that uses the link (which has great foot-shooting potential if the user tries to route the same link to two different IRQs using two different pci bus/device/pin hints). For example, to adjust the IRQ that \_SB_.LNKA uses, one would set 'hw.pci.link.LNKA.irq=10' from the loader. - As a side effect of having the link driver, unused link devices will now be disabled when they are probed. - The algorithm for choosing an IRQ for a link that doesn't already have an IRQ assigned is now much closer to the one used in $PIR routing. When a link is routed via an ISA IRQ, only known-good IRQs that the BIOS has already used are used for routing instead of using probabilities to guess at which IRQs are probably not used by an ISA device. One change from $PIR is that the SCI is always considered a viable ISA IRQ, so that if the BIOS does not setup any IRQs the kernel will degenerate to routing all interrupts over the SCI. For non ISA IRQs, interrupts are picked from the possible pool using a simplistic weighting algorithm. Tested by: ru, scottl, others on acpi@ Reviewed by: njl
2004-11-23 22:26:44 +00:00
static void
acpi_pci_link_identify(driver_t *driver, device_t parent)
{
Rework the ACPI PCI link code. - Use a new-bus device driver for the ACPI PCI link devices. The devices are called pci_linkX. The driver includes suspend/resume support so that the ACPI bridge drivers no longer have to poke the links to get them to handle suspend/resume. Also, the code to handle which IRQs a link is routed to and choosing an IRQ when a link is not already routed is all contained in the link driver. The PCI bridge drivers now ask the link driver which IRQ to use once they determine that a _PRT entry does not use a hardwired interrupt number. - The new link driver includes support for multiple IRQ resources per link device as well as preserving any non-IRQ resources when adjusting the IRQ that a link is routed to. - The entire approach to routing when using a link device is now link-centric rather than pci bus/device/pin specific. Thus, when using a tunable to override the default IRQ settings, one now uses a single tunable to route an entire link rather than routing a single device that uses the link (which has great foot-shooting potential if the user tries to route the same link to two different IRQs using two different pci bus/device/pin hints). For example, to adjust the IRQ that \_SB_.LNKA uses, one would set 'hw.pci.link.LNKA.irq=10' from the loader. - As a side effect of having the link driver, unused link devices will now be disabled when they are probed. - The algorithm for choosing an IRQ for a link that doesn't already have an IRQ assigned is now much closer to the one used in $PIR routing. When a link is routed via an ISA IRQ, only known-good IRQs that the BIOS has already used are used for routing instead of using probabilities to guess at which IRQs are probably not used by an ISA device. One change from $PIR is that the SCI is always considered a viable ISA IRQ, so that if the BIOS does not setup any IRQs the kernel will degenerate to routing all interrupts over the SCI. For non ISA IRQs, interrupts are picked from the possible pool using a simplistic weighting algorithm. Tested by: ru, scottl, others on acpi@ Reviewed by: njl
2004-11-23 22:26:44 +00:00
/*
* If the SCI is an ISA IRQ, add it to the bitmask of known good
* ISA IRQs.
*
* XXX: If we are using the APIC, the SCI might have been
* rerouted to an APIC pin in which case this is invalid. However,
* if we are using the APIC, we also shouldn't be having any PCI
* interrupts routed via ISA IRQs, so this is probably ok.
*/
if (AcpiGbl_FADT->SciInt < NUM_ISA_INTERRUPTS)
pci_link_bios_isa_irqs |= (1 << AcpiGbl_FADT->SciInt);
}
Rework the ACPI PCI link code. - Use a new-bus device driver for the ACPI PCI link devices. The devices are called pci_linkX. The driver includes suspend/resume support so that the ACPI bridge drivers no longer have to poke the links to get them to handle suspend/resume. Also, the code to handle which IRQs a link is routed to and choosing an IRQ when a link is not already routed is all contained in the link driver. The PCI bridge drivers now ask the link driver which IRQ to use once they determine that a _PRT entry does not use a hardwired interrupt number. - The new link driver includes support for multiple IRQ resources per link device as well as preserving any non-IRQ resources when adjusting the IRQ that a link is routed to. - The entire approach to routing when using a link device is now link-centric rather than pci bus/device/pin specific. Thus, when using a tunable to override the default IRQ settings, one now uses a single tunable to route an entire link rather than routing a single device that uses the link (which has great foot-shooting potential if the user tries to route the same link to two different IRQs using two different pci bus/device/pin hints). For example, to adjust the IRQ that \_SB_.LNKA uses, one would set 'hw.pci.link.LNKA.irq=10' from the loader. - As a side effect of having the link driver, unused link devices will now be disabled when they are probed. - The algorithm for choosing an IRQ for a link that doesn't already have an IRQ assigned is now much closer to the one used in $PIR routing. When a link is routed via an ISA IRQ, only known-good IRQs that the BIOS has already used are used for routing instead of using probabilities to guess at which IRQs are probably not used by an ISA device. One change from $PIR is that the SCI is always considered a viable ISA IRQ, so that if the BIOS does not setup any IRQs the kernel will degenerate to routing all interrupts over the SCI. For non ISA IRQs, interrupts are picked from the possible pool using a simplistic weighting algorithm. Tested by: ru, scottl, others on acpi@ Reviewed by: njl
2004-11-23 22:26:44 +00:00
static device_method_t acpi_pci_link_methods[] = {
/* Device interface */
DEVMETHOD(device_identify, acpi_pci_link_identify),
DEVMETHOD(device_probe, acpi_pci_link_probe),
DEVMETHOD(device_attach, acpi_pci_link_attach),
DEVMETHOD(device_resume, acpi_pci_link_resume),
Rework the ACPI PCI link code. - Use a new-bus device driver for the ACPI PCI link devices. The devices are called pci_linkX. The driver includes suspend/resume support so that the ACPI bridge drivers no longer have to poke the links to get them to handle suspend/resume. Also, the code to handle which IRQs a link is routed to and choosing an IRQ when a link is not already routed is all contained in the link driver. The PCI bridge drivers now ask the link driver which IRQ to use once they determine that a _PRT entry does not use a hardwired interrupt number. - The new link driver includes support for multiple IRQ resources per link device as well as preserving any non-IRQ resources when adjusting the IRQ that a link is routed to. - The entire approach to routing when using a link device is now link-centric rather than pci bus/device/pin specific. Thus, when using a tunable to override the default IRQ settings, one now uses a single tunable to route an entire link rather than routing a single device that uses the link (which has great foot-shooting potential if the user tries to route the same link to two different IRQs using two different pci bus/device/pin hints). For example, to adjust the IRQ that \_SB_.LNKA uses, one would set 'hw.pci.link.LNKA.irq=10' from the loader. - As a side effect of having the link driver, unused link devices will now be disabled when they are probed. - The algorithm for choosing an IRQ for a link that doesn't already have an IRQ assigned is now much closer to the one used in $PIR routing. When a link is routed via an ISA IRQ, only known-good IRQs that the BIOS has already used are used for routing instead of using probabilities to guess at which IRQs are probably not used by an ISA device. One change from $PIR is that the SCI is always considered a viable ISA IRQ, so that if the BIOS does not setup any IRQs the kernel will degenerate to routing all interrupts over the SCI. For non ISA IRQs, interrupts are picked from the possible pool using a simplistic weighting algorithm. Tested by: ru, scottl, others on acpi@ Reviewed by: njl
2004-11-23 22:26:44 +00:00
{0, 0}
};
Rework the ACPI PCI link code. - Use a new-bus device driver for the ACPI PCI link devices. The devices are called pci_linkX. The driver includes suspend/resume support so that the ACPI bridge drivers no longer have to poke the links to get them to handle suspend/resume. Also, the code to handle which IRQs a link is routed to and choosing an IRQ when a link is not already routed is all contained in the link driver. The PCI bridge drivers now ask the link driver which IRQ to use once they determine that a _PRT entry does not use a hardwired interrupt number. - The new link driver includes support for multiple IRQ resources per link device as well as preserving any non-IRQ resources when adjusting the IRQ that a link is routed to. - The entire approach to routing when using a link device is now link-centric rather than pci bus/device/pin specific. Thus, when using a tunable to override the default IRQ settings, one now uses a single tunable to route an entire link rather than routing a single device that uses the link (which has great foot-shooting potential if the user tries to route the same link to two different IRQs using two different pci bus/device/pin hints). For example, to adjust the IRQ that \_SB_.LNKA uses, one would set 'hw.pci.link.LNKA.irq=10' from the loader. - As a side effect of having the link driver, unused link devices will now be disabled when they are probed. - The algorithm for choosing an IRQ for a link that doesn't already have an IRQ assigned is now much closer to the one used in $PIR routing. When a link is routed via an ISA IRQ, only known-good IRQs that the BIOS has already used are used for routing instead of using probabilities to guess at which IRQs are probably not used by an ISA device. One change from $PIR is that the SCI is always considered a viable ISA IRQ, so that if the BIOS does not setup any IRQs the kernel will degenerate to routing all interrupts over the SCI. For non ISA IRQs, interrupts are picked from the possible pool using a simplistic weighting algorithm. Tested by: ru, scottl, others on acpi@ Reviewed by: njl
2004-11-23 22:26:44 +00:00
static driver_t acpi_pci_link_driver = {
"pci_link",
acpi_pci_link_methods,
sizeof(struct acpi_pci_link_softc),
};
Rework the ACPI PCI link code. - Use a new-bus device driver for the ACPI PCI link devices. The devices are called pci_linkX. The driver includes suspend/resume support so that the ACPI bridge drivers no longer have to poke the links to get them to handle suspend/resume. Also, the code to handle which IRQs a link is routed to and choosing an IRQ when a link is not already routed is all contained in the link driver. The PCI bridge drivers now ask the link driver which IRQ to use once they determine that a _PRT entry does not use a hardwired interrupt number. - The new link driver includes support for multiple IRQ resources per link device as well as preserving any non-IRQ resources when adjusting the IRQ that a link is routed to. - The entire approach to routing when using a link device is now link-centric rather than pci bus/device/pin specific. Thus, when using a tunable to override the default IRQ settings, one now uses a single tunable to route an entire link rather than routing a single device that uses the link (which has great foot-shooting potential if the user tries to route the same link to two different IRQs using two different pci bus/device/pin hints). For example, to adjust the IRQ that \_SB_.LNKA uses, one would set 'hw.pci.link.LNKA.irq=10' from the loader. - As a side effect of having the link driver, unused link devices will now be disabled when they are probed. - The algorithm for choosing an IRQ for a link that doesn't already have an IRQ assigned is now much closer to the one used in $PIR routing. When a link is routed via an ISA IRQ, only known-good IRQs that the BIOS has already used are used for routing instead of using probabilities to guess at which IRQs are probably not used by an ISA device. One change from $PIR is that the SCI is always considered a viable ISA IRQ, so that if the BIOS does not setup any IRQs the kernel will degenerate to routing all interrupts over the SCI. For non ISA IRQs, interrupts are picked from the possible pool using a simplistic weighting algorithm. Tested by: ru, scottl, others on acpi@ Reviewed by: njl
2004-11-23 22:26:44 +00:00
static devclass_t pci_link_devclass;
Rework the ACPI PCI link code. - Use a new-bus device driver for the ACPI PCI link devices. The devices are called pci_linkX. The driver includes suspend/resume support so that the ACPI bridge drivers no longer have to poke the links to get them to handle suspend/resume. Also, the code to handle which IRQs a link is routed to and choosing an IRQ when a link is not already routed is all contained in the link driver. The PCI bridge drivers now ask the link driver which IRQ to use once they determine that a _PRT entry does not use a hardwired interrupt number. - The new link driver includes support for multiple IRQ resources per link device as well as preserving any non-IRQ resources when adjusting the IRQ that a link is routed to. - The entire approach to routing when using a link device is now link-centric rather than pci bus/device/pin specific. Thus, when using a tunable to override the default IRQ settings, one now uses a single tunable to route an entire link rather than routing a single device that uses the link (which has great foot-shooting potential if the user tries to route the same link to two different IRQs using two different pci bus/device/pin hints). For example, to adjust the IRQ that \_SB_.LNKA uses, one would set 'hw.pci.link.LNKA.irq=10' from the loader. - As a side effect of having the link driver, unused link devices will now be disabled when they are probed. - The algorithm for choosing an IRQ for a link that doesn't already have an IRQ assigned is now much closer to the one used in $PIR routing. When a link is routed via an ISA IRQ, only known-good IRQs that the BIOS has already used are used for routing instead of using probabilities to guess at which IRQs are probably not used by an ISA device. One change from $PIR is that the SCI is always considered a viable ISA IRQ, so that if the BIOS does not setup any IRQs the kernel will degenerate to routing all interrupts over the SCI. For non ISA IRQs, interrupts are picked from the possible pool using a simplistic weighting algorithm. Tested by: ru, scottl, others on acpi@ Reviewed by: njl
2004-11-23 22:26:44 +00:00
DRIVER_MODULE(acpi_pci_link, acpi, acpi_pci_link_driver, pci_link_devclass, 0,
0);
MODULE_DEPEND(acpi_pci_link, acpi, 1, 1, 1);