freebsd-skq/sys/sparc64/sbus/sbus.c

939 lines
26 KiB
C
Raw Normal View History

/*-
* Copyright (c) 1999-2002 Eduardo Horvath
* 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.
* 3. The name of the author may not be used to endorse or promote products
* derived from this software without specific prior written permission.
*
* THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``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 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.
*
* from: NetBSD: sbus.c,v 1.50 2002/06/20 18:26:24 eeh Exp
*/
/*-
* Copyright (c) 2002 by Thomas Moestl <tmm@FreeBSD.org>.
Rototill the sparc64 nexus(4) (actually this brings in the code the sun4v nexus(4) in turn is based on): o Change nexus(4) to manage the resources of its children so the respective device drivers don't need to figure them out of OFW themselves. o Change nexus(4) to provide the ofw_bus KOBJ interface instead of using IVARs for supplying the OFW node and the subset of standard properties of its children. Together with the previous change this also allows to fully take advantage of newbus in that drivers like fhc(4), which attach on multiple parent busses, no longer require different bus front-ends as obtaining the OFW node and properties as well as resource allocation works the same for all supported busses. As such this change also is part 4/4 of allowing creator(4) to work in USIII-based machines as it allows this driver to attach on both nexus(4) and upa(4). On the other hand removing these IVARs breaks API compatibility with the powerpc nexus(4) but which isn't that bad as a) sparc64 currently doesn't share any device driver hanging off of nexus(4) with powerpc and b) they were no longer compatible regarding OFW-related extensions at the pci(4) level since quite some time. o Provide bus_get_dma_tag methods in nexus(4) and its children in order to handle DMA tags in a hierarchical way and get rid of the sparc64_root_dma_tag kludge. Together with the previous two items this changes also allows to completely get rid of the nexus(4) IVAR interface. It also includes: - pushing the constraints previously specified by the nexus_dmatag down into the DMA tags of psycho(4) and sbus(4) as it's their IOMMUs which induce these restrictions (and nothing at the nexus(4) or anything that would warrant specifying them there), - fixing some obviously wrong constraints of the psycho(4) and sbus(4) DMA tags, which happened to not actually be used with the sparc64_root_dma_tag kludge in place and therefore didn't cause problems so far, - replacing magic constants for constraints with macros as far as it is obvious as to where they come from. This doesn't include taking advantage of the newbus way to get the parent DMA tags implemented by this change in order to divorce the IOTSBs of the PCI and SBus IOMMUs or for implementing the workaround for the DMA sync bug in Sabre (and Tomatillo) bridges, yet, though. o Get rid of the notion that nexus(4) (mostly) reflects an UPA bus by replacing ofw_upa.h and with ofw_nexus.h (which was repo-copied from ofw_upa.h) and renaming its content, which actually applies to all of Fireplane/Safari, JBus and UPA (in the host bus case), as appropriate. o Just use M_DEVBUF instead of a separate M_NEXUS malloc type for allocating the device info for the children of nexus(4). This is done in order to not need to export M_NEXUS when deriving drivers for subordinate busses from the nexus(4) class. o Use the DEFINE_CLASS_0() macro to declare the nexus(4) driver so we can derive subclasses from it. o Const'ify the nexus_excl_name and nexus_excl_type arrays as well as add 'associations' and 'rsc', which are pseudo-devices without resources and therefore of no real interest for nexus(4), to the former. o Let the nexus(4) device memory rman manage the entire 64-bit address space instead of just the UPA_MEMSTART to UPA_MEMEND subregion as Fireplane/Safari- and JBus-based machines use multiple ranges, which can't be as easily divided as in the case of UPA (limiting the address space only served for sanity checking anyway). o Use M_WAITOK instead of M_NOWAIT when allocating the device info for children of nexus(4) in order to give one less opportunity for adding devices to nexus(4) to fail. o While adapting the drivers affected by the above nexus(4) changes, change them to take advantage of rman_get_rid() instead of caching the RIDs assigned to allocated resources, now that the RIDs of resources are correctly set. o In iommu(4) and nexus(4) replace hard-coded functions names, which actually became outdated in several places, in panic strings and status massages with __func__. [1] o Use driver_filter_t in prototypes where appropriate. o Add my copyright to creator(4), fhc(4), nexus(4), psycho(4) and sbus(4) as I changed considerable amounts of these drivers as well as added a bunch of new features, workarounds for silicon bugs etc. o Fix some white space nits. Due to lack of access to Exx00 hardware, these changes, i.e. central(4) and fhc(4), couldn't be runtime tested on such a machine. Exx00 are currently reported to panic before trying to attach nexus(4) anyway though. PR: 76052 [1] Approved by: re (kensmith)
2007-03-07 21:13:51 +00:00
* Copyright (c) 2005 Marius Strobl <marius@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.
*
* THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``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 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("$FreeBSD$");
/*
* SBus support.
*/
#include <sys/param.h>
#include <sys/systm.h>
#include <sys/bus.h>
#include <sys/kernel.h>
#include <sys/malloc.h>
2004-06-04 11:52:25 +00:00
#include <sys/module.h>
#include <sys/pcpu.h>
- Divorce the IOTSBs, which so far where handled via a global list instead of per IOMMU, so we no longer need to program all of them identically in systems having multiple IOMMUs. This continues the rototilling of the nexus(4) done about 5 months ago, which amongst others changed nexus(4) and the drivers for host-to-foo bridges to provide bus_get_dma_tag methods, allowing to handle DMA tags in a hierarchical way and to link them with devices. This still doesn't move the silicon bug workarounds for Sabre (and in the uncommitted schizo(4) for Tomatillo) bridges into special bus_dma_tag_create() and bus_dmamap_sync() methods though, as w/o fully newbus'ified bus_dma_tag_create() and bus_dma_tag_destroy() this still requires too much hackery, i.e. per-child parent DMA tags in the parent driver. - Let the host-to-foo drivers supply the maximum physical address of the IOMMU accompanying the bridges. Previously iommu(4) hard- coded an upper limit of 16GB, which actually only applies to the IOMMUs of the Hummingbird and Sabre bridges. The Psycho variants as well as the U2S in fact can can translate to up to 2TB, i.e. translate to 41-bit physical addresses. According to the recently available Tomatillo documentation these bridges even translate to 43-bit physical addresses and hints at the Schizo bridges doing 43 bits as well. This fixes the issue the FreeBSD 6.0 todo list item "Max RAM on sparc64" was refering to and pretty much obsoletes the lack of support for bounce buffers on sparc64. Thanks to Nathan Whitehorn for pointing me at the Tomatillo manual. Approved by: re (kensmith)
2007-08-05 11:56:44 +00:00
#include <sys/queue.h>
#include <sys/reboot.h>
- Divorce the IOTSBs, which so far where handled via a global list instead of per IOMMU, so we no longer need to program all of them identically in systems having multiple IOMMUs. This continues the rototilling of the nexus(4) done about 5 months ago, which amongst others changed nexus(4) and the drivers for host-to-foo bridges to provide bus_get_dma_tag methods, allowing to handle DMA tags in a hierarchical way and to link them with devices. This still doesn't move the silicon bug workarounds for Sabre (and in the uncommitted schizo(4) for Tomatillo) bridges into special bus_dma_tag_create() and bus_dmamap_sync() methods though, as w/o fully newbus'ified bus_dma_tag_create() and bus_dma_tag_destroy() this still requires too much hackery, i.e. per-child parent DMA tags in the parent driver. - Let the host-to-foo drivers supply the maximum physical address of the IOMMU accompanying the bridges. Previously iommu(4) hard- coded an upper limit of 16GB, which actually only applies to the IOMMUs of the Hummingbird and Sabre bridges. The Psycho variants as well as the U2S in fact can can translate to up to 2TB, i.e. translate to 41-bit physical addresses. According to the recently available Tomatillo documentation these bridges even translate to 43-bit physical addresses and hints at the Schizo bridges doing 43 bits as well. This fixes the issue the FreeBSD 6.0 todo list item "Max RAM on sparc64" was refering to and pretty much obsoletes the lack of support for bounce buffers on sparc64. Thanks to Nathan Whitehorn for pointing me at the Tomatillo manual. Approved by: re (kensmith)
2007-08-05 11:56:44 +00:00
#include <sys/rman.h>
- Introduce an ofw_bus kobj-interface for retrieving the OFW node and a subset ("compatible", "device_type", "model" and "name") of the standard properties in drivers for devices on Open Firmware supported busses. The standard properties "reg", "interrupts" und "address" are not covered by this interface because they are only of interest in the respective bridge code. There's a remaining standard property "status" which is unclear how to support properly but which also isn't used in FreeBSD at present. This ofw_bus kobj-interface allows to replace the various (ebus_get_node(), ofw_pci_get_node(), etc.) and partially inconsistent (central_get_type() vs. sbus_get_device_type(), etc.) existing IVAR ones with a common one. This in turn allows to simplify and remove code-duplication in drivers for devices that can hang off of more than one OFW supported bus. - Convert the sparc64 Central, EBus, FHC, PCI and SBus bus drivers and the drivers for their children to use the ofw_bus kobj-interface. The IVAR- interfaces of the Central, EBus and FHC are entirely replaced by this. The PCI bus driver used its own kobj-interface and now also uses the ofw_bus one. The IVARs special to the SBus, e.g. for retrieving the burst size, remain. Beware: this causes an ABI-breakage for modules of drivers which used the IVAR-interfaces, i.e. esp(4), hme(4), isp(4) and uart(4), which need to be recompiled. The style-inconsistencies introduced in some of the bus drivers will be fixed by tmm@ in a generic clean-up of the respective drivers later (he requested to add the changes in the "new" style). - Convert the powerpc MacIO bus driver and the drivers for its children to use the ofw_bus kobj-interface. This invloves removing the IVARs related to the "reg" property which were unused and a leftover from the NetBSD origini of the code. There's no ABI-breakage caused by this because none of these driver are currently built as modules. There are other powerpc bus drivers which can be converted to the ofw_bus kobj-interface, e.g. the PCI bus driver, which should be done together with converting powerpc to use the OFW PCI code from sparc64. - Make the SBus and FHC front-end of zs(4) and the sparc64 eeprom(4) take advantage of the ofw_bus kobj-interface and simplify them a bit. Reviewed by: grehan, tmm Approved by: re (scottl) Discussed with: tmm Tested with: Sun AX1105, AXe, Ultra 2, Ultra 60; PPC cross-build on i386
2004-08-12 17:41:33 +00:00
#include <dev/ofw/ofw_bus.h>
#include <dev/ofw/ofw_bus_subr.h>
#include <dev/ofw/openfirm.h>
#include <machine/bus.h>
- Divorce the IOTSBs, which so far where handled via a global list instead of per IOMMU, so we no longer need to program all of them identically in systems having multiple IOMMUs. This continues the rototilling of the nexus(4) done about 5 months ago, which amongst others changed nexus(4) and the drivers for host-to-foo bridges to provide bus_get_dma_tag methods, allowing to handle DMA tags in a hierarchical way and to link them with devices. This still doesn't move the silicon bug workarounds for Sabre (and in the uncommitted schizo(4) for Tomatillo) bridges into special bus_dma_tag_create() and bus_dmamap_sync() methods though, as w/o fully newbus'ified bus_dma_tag_create() and bus_dma_tag_destroy() this still requires too much hackery, i.e. per-child parent DMA tags in the parent driver. - Let the host-to-foo drivers supply the maximum physical address of the IOMMU accompanying the bridges. Previously iommu(4) hard- coded an upper limit of 16GB, which actually only applies to the IOMMUs of the Hummingbird and Sabre bridges. The Psycho variants as well as the U2S in fact can can translate to up to 2TB, i.e. translate to 41-bit physical addresses. According to the recently available Tomatillo documentation these bridges even translate to 43-bit physical addresses and hints at the Schizo bridges doing 43 bits as well. This fixes the issue the FreeBSD 6.0 todo list item "Max RAM on sparc64" was refering to and pretty much obsoletes the lack of support for bounce buffers on sparc64. Thanks to Nathan Whitehorn for pointing me at the Tomatillo manual. Approved by: re (kensmith)
2007-08-05 11:56:44 +00:00
#include <machine/bus_common.h>
#include <machine/bus_private.h>
#include <machine/iommureg.h>
#include <machine/iommuvar.h>
- Divorce the IOTSBs, which so far where handled via a global list instead of per IOMMU, so we no longer need to program all of them identically in systems having multiple IOMMUs. This continues the rototilling of the nexus(4) done about 5 months ago, which amongst others changed nexus(4) and the drivers for host-to-foo bridges to provide bus_get_dma_tag methods, allowing to handle DMA tags in a hierarchical way and to link them with devices. This still doesn't move the silicon bug workarounds for Sabre (and in the uncommitted schizo(4) for Tomatillo) bridges into special bus_dma_tag_create() and bus_dmamap_sync() methods though, as w/o fully newbus'ified bus_dma_tag_create() and bus_dma_tag_destroy() this still requires too much hackery, i.e. per-child parent DMA tags in the parent driver. - Let the host-to-foo drivers supply the maximum physical address of the IOMMU accompanying the bridges. Previously iommu(4) hard- coded an upper limit of 16GB, which actually only applies to the IOMMUs of the Hummingbird and Sabre bridges. The Psycho variants as well as the U2S in fact can can translate to up to 2TB, i.e. translate to 41-bit physical addresses. According to the recently available Tomatillo documentation these bridges even translate to 43-bit physical addresses and hints at the Schizo bridges doing 43 bits as well. This fixes the issue the FreeBSD 6.0 todo list item "Max RAM on sparc64" was refering to and pretty much obsoletes the lack of support for bounce buffers on sparc64. Thanks to Nathan Whitehorn for pointing me at the Tomatillo manual. Approved by: re (kensmith)
2007-08-05 11:56:44 +00:00
#include <machine/resource.h>
#include <sparc64/sbus/ofw_sbus.h>
#include <sparc64/sbus/sbusreg.h>
#include <sparc64/sbus/sbusvar.h>
struct sbus_devinfo {
int sdi_burstsz;
int sdi_clockfreq;
int sdi_slot;
struct ofw_bus_devinfo sdi_obdinfo;
struct resource_list sdi_rl;
};
/* Range descriptor, allocated for each sc_range. */
struct sbus_rd {
bus_addr_t rd_poffset;
bus_addr_t rd_pend;
int rd_slot;
bus_addr_t rd_coffset;
bus_addr_t rd_cend;
struct rman rd_rman;
bus_space_handle_t rd_bushandle;
struct resource *rd_res;
};
struct sbus_softc {
o Revamp the sparc64 interrupt code in order to be able to interface with the INTR_FILTER-enabled MI code. Basically this consists of registering an interrupt controller (of which there can be multiple and optionally different ones either per host-to-foo bridge or shared amongst host-to-foo bridges in any one machine) along with an interrupt vector as specific argument for all the interrupt vectors used by a given host-to-foo bridge (roughly similar to registering interrupt sources on amd64 and i386), providing functions to enable, clear and disable the interrupts of the children beneath the bridge. This also includes: - No longer entering a critical section in tl0_intr() and tl1_intr() for executing interrupt handlers but rather let the handlers enter it themselves so in the case of intr_event_handle() we don't enter a nested critical section. - Adding infrastructure for binding delivery of interrupt vectors to specific CPUs which later on can be interfaced with the code from amd64/i386 for binding interrupts to specific CPUs. - Getting rid of the wrapper hack introduced along the lines of the API changes for INTR_FILTER which as a side-effect caused interrupts associated with ithread handlers only to get the elevated priority of those associated with filters ("fast handlers") (this removes the hack also in the non-INTR_FILTER case). - Disabling (by not clearing) an interrupt in the interrupt controller until all associated handlers have been executed, which is crucial for the typical locking strategy of NIC drivers in order to work correctly in case of shared interrupts. This was a more or less theoretical problem on sparc64 though, as shared interrupts are rather uncommon there except for the on-board SCCs and UARTs. Note that due to the behavior of at least of some of the interrupt controllers used on sparc64 an enable+EOI instead of a disable+EOI approach (as implied by the INTR_FILTER MI code and implemented on other architectures) is used as the latter can cause lost interrupts or in the worst case interrupt starvation. o Correct a typo in sbus_alloc_resource() which caused (pass-through) allocations to only work down to the grandchildren of the bus, which wasn't a real problem so far as we don't support any devices which are great-grandchildren or greater of a U2S bridge, yet. o In fhc(4) use bus_{read,write}_4() instead of bus_space_{read,write}_4() in order to get rid of sc_bh and sc_bt in the fhc_softc. Also get rid of some other unneeded members in fhc_softc. Reviewed by: marcel (earlier version) Approved by: re (kensmith)
2007-09-06 19:16:30 +00:00
device_t sc_dev;
bus_dma_tag_t sc_cdmatag;
int sc_clockfreq; /* clock frequency (in Hz) */
int sc_nrange;
struct sbus_rd *sc_rd;
int sc_burst; /* burst transfer sizes supp. */
struct resource *sc_sysio_res;
int sc_ign; /* IGN for this sysio */
struct iommu_state sc_is; /* IOMMU state (iommuvar.h) */
struct resource *sc_ot_ires;
void *sc_ot_ihand;
struct resource *sc_pf_ires;
void *sc_pf_ihand;
};
#define SYSIO_READ8(sc, off) \
bus_read_8((sc)->sc_sysio_res, (off))
#define SYSIO_WRITE8(sc, off, v) \
bus_write_8((sc)->sc_sysio_res, (off), (v))
static device_probe_t sbus_probe;
static device_attach_t sbus_attach;
static bus_print_child_t sbus_print_child;
static bus_probe_nomatch_t sbus_probe_nomatch;
static bus_read_ivar_t sbus_read_ivar;
static bus_get_resource_list_t sbus_get_resource_list;
static bus_setup_intr_t sbus_setup_intr;
static bus_alloc_resource_t sbus_alloc_resource;
static bus_activate_resource_t sbus_activate_resource;
Make sparc64 compatible with NEW_PCIB and enable it: - Implement bus_adjust_resource() methods as far as necessary and in non-PCI bridge drivers as far as feasible without rototilling them. - As NEW_PCIB does a layering violation by activating resources at layers above pci(4) without previously bubbling up their allocation there, move the assignment of bus tags and handles from the bus_alloc_resource() to the bus_activate_resource() methods like at least the other NEW_PCIB enabled architectures do. This is somewhat unfortunate as previously sparc64 (ab)used resource activation to indicate whether SYS_RES_MEMORY resources should be mapped into KVA, which is only necessary if their going to be accessed via the pointer returned from rman_get_virtual() but not for bus_space(9) as the later always uses physical access on sparc64. Besides wasting KVA if we always map in SYS_RES_MEMORY resources, a driver also may deliberately not map them in if the firmware already has done so, possibly in a special way. So in order to still allow a driver to decide whether a SYS_RES_MEMORY resource should be mapped into KVA we let it indicate that by calling bus_space_map(9) with BUS_SPACE_MAP_LINEAR as actually documented in the bus_space(9) page. This is implemented by allocating a separate bus tag per SYS_RES_MEMORY resource and passing the resource via the previously unused bus tag cookie so we later on can call rman_set_virtual() in sparc64_bus_mem_map(). As a side effect this now also allows to actually indicate that a SYS_RES_MEMORY resource should be mapped in as cacheable and/or read-only via BUS_SPACE_MAP_CACHEABLE and BUS_SPACE_MAP_READONLY respectively. - Do some minor cleanup like taking advantage of rman_init_from_resource(), factor out the common part of bus tag allocation into a newly added sparc64_alloc_bus_tag(), hook up some missing newbus methods and replace some homegrown versions with the generic counterparts etc. - While at it, let apb_attach() (which can't use the generic NEW_PCIB code as APB bridges just don't have the base and limit registers implemented) regarding the config space registers cached in pcib_softc and the SYSCTL reporting nodes set up.
2011-10-02 23:22:38 +00:00
static bus_adjust_resource_t sbus_adjust_resource;
static bus_release_resource_t sbus_release_resource;
Rototill the sparc64 nexus(4) (actually this brings in the code the sun4v nexus(4) in turn is based on): o Change nexus(4) to manage the resources of its children so the respective device drivers don't need to figure them out of OFW themselves. o Change nexus(4) to provide the ofw_bus KOBJ interface instead of using IVARs for supplying the OFW node and the subset of standard properties of its children. Together with the previous change this also allows to fully take advantage of newbus in that drivers like fhc(4), which attach on multiple parent busses, no longer require different bus front-ends as obtaining the OFW node and properties as well as resource allocation works the same for all supported busses. As such this change also is part 4/4 of allowing creator(4) to work in USIII-based machines as it allows this driver to attach on both nexus(4) and upa(4). On the other hand removing these IVARs breaks API compatibility with the powerpc nexus(4) but which isn't that bad as a) sparc64 currently doesn't share any device driver hanging off of nexus(4) with powerpc and b) they were no longer compatible regarding OFW-related extensions at the pci(4) level since quite some time. o Provide bus_get_dma_tag methods in nexus(4) and its children in order to handle DMA tags in a hierarchical way and get rid of the sparc64_root_dma_tag kludge. Together with the previous two items this changes also allows to completely get rid of the nexus(4) IVAR interface. It also includes: - pushing the constraints previously specified by the nexus_dmatag down into the DMA tags of psycho(4) and sbus(4) as it's their IOMMUs which induce these restrictions (and nothing at the nexus(4) or anything that would warrant specifying them there), - fixing some obviously wrong constraints of the psycho(4) and sbus(4) DMA tags, which happened to not actually be used with the sparc64_root_dma_tag kludge in place and therefore didn't cause problems so far, - replacing magic constants for constraints with macros as far as it is obvious as to where they come from. This doesn't include taking advantage of the newbus way to get the parent DMA tags implemented by this change in order to divorce the IOTSBs of the PCI and SBus IOMMUs or for implementing the workaround for the DMA sync bug in Sabre (and Tomatillo) bridges, yet, though. o Get rid of the notion that nexus(4) (mostly) reflects an UPA bus by replacing ofw_upa.h and with ofw_nexus.h (which was repo-copied from ofw_upa.h) and renaming its content, which actually applies to all of Fireplane/Safari, JBus and UPA (in the host bus case), as appropriate. o Just use M_DEVBUF instead of a separate M_NEXUS malloc type for allocating the device info for the children of nexus(4). This is done in order to not need to export M_NEXUS when deriving drivers for subordinate busses from the nexus(4) class. o Use the DEFINE_CLASS_0() macro to declare the nexus(4) driver so we can derive subclasses from it. o Const'ify the nexus_excl_name and nexus_excl_type arrays as well as add 'associations' and 'rsc', which are pseudo-devices without resources and therefore of no real interest for nexus(4), to the former. o Let the nexus(4) device memory rman manage the entire 64-bit address space instead of just the UPA_MEMSTART to UPA_MEMEND subregion as Fireplane/Safari- and JBus-based machines use multiple ranges, which can't be as easily divided as in the case of UPA (limiting the address space only served for sanity checking anyway). o Use M_WAITOK instead of M_NOWAIT when allocating the device info for children of nexus(4) in order to give one less opportunity for adding devices to nexus(4) to fail. o While adapting the drivers affected by the above nexus(4) changes, change them to take advantage of rman_get_rid() instead of caching the RIDs assigned to allocated resources, now that the RIDs of resources are correctly set. o In iommu(4) and nexus(4) replace hard-coded functions names, which actually became outdated in several places, in panic strings and status massages with __func__. [1] o Use driver_filter_t in prototypes where appropriate. o Add my copyright to creator(4), fhc(4), nexus(4), psycho(4) and sbus(4) as I changed considerable amounts of these drivers as well as added a bunch of new features, workarounds for silicon bugs etc. o Fix some white space nits. Due to lack of access to Exx00 hardware, these changes, i.e. central(4) and fhc(4), couldn't be runtime tested on such a machine. Exx00 are currently reported to panic before trying to attach nexus(4) anyway though. PR: 76052 [1] Approved by: re (kensmith)
2007-03-07 21:13:51 +00:00
static bus_get_dma_tag_t sbus_get_dma_tag;
static ofw_bus_get_devinfo_t sbus_get_devinfo;
static int sbus_inlist(const char *, const char *const *);
static struct sbus_devinfo * sbus_setup_dinfo(device_t, struct sbus_softc *,
phandle_t);
static void sbus_destroy_dinfo(struct sbus_devinfo *);
o Revamp the sparc64 interrupt code in order to be able to interface with the INTR_FILTER-enabled MI code. Basically this consists of registering an interrupt controller (of which there can be multiple and optionally different ones either per host-to-foo bridge or shared amongst host-to-foo bridges in any one machine) along with an interrupt vector as specific argument for all the interrupt vectors used by a given host-to-foo bridge (roughly similar to registering interrupt sources on amd64 and i386), providing functions to enable, clear and disable the interrupts of the children beneath the bridge. This also includes: - No longer entering a critical section in tl0_intr() and tl1_intr() for executing interrupt handlers but rather let the handlers enter it themselves so in the case of intr_event_handle() we don't enter a nested critical section. - Adding infrastructure for binding delivery of interrupt vectors to specific CPUs which later on can be interfaced with the code from amd64/i386 for binding interrupts to specific CPUs. - Getting rid of the wrapper hack introduced along the lines of the API changes for INTR_FILTER which as a side-effect caused interrupts associated with ithread handlers only to get the elevated priority of those associated with filters ("fast handlers") (this removes the hack also in the non-INTR_FILTER case). - Disabling (by not clearing) an interrupt in the interrupt controller until all associated handlers have been executed, which is crucial for the typical locking strategy of NIC drivers in order to work correctly in case of shared interrupts. This was a more or less theoretical problem on sparc64 though, as shared interrupts are rather uncommon there except for the on-board SCCs and UARTs. Note that due to the behavior of at least of some of the interrupt controllers used on sparc64 an enable+EOI instead of a disable+EOI approach (as implied by the INTR_FILTER MI code and implemented on other architectures) is used as the latter can cause lost interrupts or in the worst case interrupt starvation. o Correct a typo in sbus_alloc_resource() which caused (pass-through) allocations to only work down to the grandchildren of the bus, which wasn't a real problem so far as we don't support any devices which are great-grandchildren or greater of a U2S bridge, yet. o In fhc(4) use bus_{read,write}_4() instead of bus_space_{read,write}_4() in order to get rid of sc_bh and sc_bt in the fhc_softc. Also get rid of some other unneeded members in fhc_softc. Reviewed by: marcel (earlier version) Approved by: re (kensmith)
2007-09-06 19:16:30 +00:00
static void sbus_intr_enable(void *);
static void sbus_intr_disable(void *);
static void sbus_intr_assign(void *);
static void sbus_intr_clear(void *);
o Revamp the sparc64 interrupt code in order to be able to interface with the INTR_FILTER-enabled MI code. Basically this consists of registering an interrupt controller (of which there can be multiple and optionally different ones either per host-to-foo bridge or shared amongst host-to-foo bridges in any one machine) along with an interrupt vector as specific argument for all the interrupt vectors used by a given host-to-foo bridge (roughly similar to registering interrupt sources on amd64 and i386), providing functions to enable, clear and disable the interrupts of the children beneath the bridge. This also includes: - No longer entering a critical section in tl0_intr() and tl1_intr() for executing interrupt handlers but rather let the handlers enter it themselves so in the case of intr_event_handle() we don't enter a nested critical section. - Adding infrastructure for binding delivery of interrupt vectors to specific CPUs which later on can be interfaced with the code from amd64/i386 for binding interrupts to specific CPUs. - Getting rid of the wrapper hack introduced along the lines of the API changes for INTR_FILTER which as a side-effect caused interrupts associated with ithread handlers only to get the elevated priority of those associated with filters ("fast handlers") (this removes the hack also in the non-INTR_FILTER case). - Disabling (by not clearing) an interrupt in the interrupt controller until all associated handlers have been executed, which is crucial for the typical locking strategy of NIC drivers in order to work correctly in case of shared interrupts. This was a more or less theoretical problem on sparc64 though, as shared interrupts are rather uncommon there except for the on-board SCCs and UARTs. Note that due to the behavior of at least of some of the interrupt controllers used on sparc64 an enable+EOI instead of a disable+EOI approach (as implied by the INTR_FILTER MI code and implemented on other architectures) is used as the latter can cause lost interrupts or in the worst case interrupt starvation. o Correct a typo in sbus_alloc_resource() which caused (pass-through) allocations to only work down to the grandchildren of the bus, which wasn't a real problem so far as we don't support any devices which are great-grandchildren or greater of a U2S bridge, yet. o In fhc(4) use bus_{read,write}_4() instead of bus_space_{read,write}_4() in order to get rid of sc_bh and sc_bt in the fhc_softc. Also get rid of some other unneeded members in fhc_softc. Reviewed by: marcel (earlier version) Approved by: re (kensmith)
2007-09-06 19:16:30 +00:00
static int sbus_find_intrmap(struct sbus_softc *, u_int, bus_addr_t *,
bus_addr_t *);
static driver_intr_t sbus_overtemp;
static driver_intr_t sbus_pwrfail;
static int sbus_print_res(struct sbus_devinfo *);
static device_method_t sbus_methods[] = {
/* Device interface */
DEVMETHOD(device_probe, sbus_probe),
DEVMETHOD(device_attach, sbus_attach),
DEVMETHOD(device_shutdown, bus_generic_shutdown),
DEVMETHOD(device_suspend, bus_generic_suspend),
DEVMETHOD(device_resume, bus_generic_resume),
/* Bus interface */
DEVMETHOD(bus_print_child, sbus_print_child),
DEVMETHOD(bus_probe_nomatch, sbus_probe_nomatch),
DEVMETHOD(bus_read_ivar, sbus_read_ivar),
DEVMETHOD(bus_alloc_resource, sbus_alloc_resource),
Make sparc64 compatible with NEW_PCIB and enable it: - Implement bus_adjust_resource() methods as far as necessary and in non-PCI bridge drivers as far as feasible without rototilling them. - As NEW_PCIB does a layering violation by activating resources at layers above pci(4) without previously bubbling up their allocation there, move the assignment of bus tags and handles from the bus_alloc_resource() to the bus_activate_resource() methods like at least the other NEW_PCIB enabled architectures do. This is somewhat unfortunate as previously sparc64 (ab)used resource activation to indicate whether SYS_RES_MEMORY resources should be mapped into KVA, which is only necessary if their going to be accessed via the pointer returned from rman_get_virtual() but not for bus_space(9) as the later always uses physical access on sparc64. Besides wasting KVA if we always map in SYS_RES_MEMORY resources, a driver also may deliberately not map them in if the firmware already has done so, possibly in a special way. So in order to still allow a driver to decide whether a SYS_RES_MEMORY resource should be mapped into KVA we let it indicate that by calling bus_space_map(9) with BUS_SPACE_MAP_LINEAR as actually documented in the bus_space(9) page. This is implemented by allocating a separate bus tag per SYS_RES_MEMORY resource and passing the resource via the previously unused bus tag cookie so we later on can call rman_set_virtual() in sparc64_bus_mem_map(). As a side effect this now also allows to actually indicate that a SYS_RES_MEMORY resource should be mapped in as cacheable and/or read-only via BUS_SPACE_MAP_CACHEABLE and BUS_SPACE_MAP_READONLY respectively. - Do some minor cleanup like taking advantage of rman_init_from_resource(), factor out the common part of bus tag allocation into a newly added sparc64_alloc_bus_tag(), hook up some missing newbus methods and replace some homegrown versions with the generic counterparts etc. - While at it, let apb_attach() (which can't use the generic NEW_PCIB code as APB bridges just don't have the base and limit registers implemented) regarding the config space registers cached in pcib_softc and the SYSCTL reporting nodes set up.
2011-10-02 23:22:38 +00:00
DEVMETHOD(bus_activate_resource, sbus_activate_resource),
DEVMETHOD(bus_deactivate_resource, bus_generic_deactivate_resource),
DEVMETHOD(bus_adjust_resource, sbus_adjust_resource),
DEVMETHOD(bus_release_resource, sbus_release_resource),
Make sparc64 compatible with NEW_PCIB and enable it: - Implement bus_adjust_resource() methods as far as necessary and in non-PCI bridge drivers as far as feasible without rototilling them. - As NEW_PCIB does a layering violation by activating resources at layers above pci(4) without previously bubbling up their allocation there, move the assignment of bus tags and handles from the bus_alloc_resource() to the bus_activate_resource() methods like at least the other NEW_PCIB enabled architectures do. This is somewhat unfortunate as previously sparc64 (ab)used resource activation to indicate whether SYS_RES_MEMORY resources should be mapped into KVA, which is only necessary if their going to be accessed via the pointer returned from rman_get_virtual() but not for bus_space(9) as the later always uses physical access on sparc64. Besides wasting KVA if we always map in SYS_RES_MEMORY resources, a driver also may deliberately not map them in if the firmware already has done so, possibly in a special way. So in order to still allow a driver to decide whether a SYS_RES_MEMORY resource should be mapped into KVA we let it indicate that by calling bus_space_map(9) with BUS_SPACE_MAP_LINEAR as actually documented in the bus_space(9) page. This is implemented by allocating a separate bus tag per SYS_RES_MEMORY resource and passing the resource via the previously unused bus tag cookie so we later on can call rman_set_virtual() in sparc64_bus_mem_map(). As a side effect this now also allows to actually indicate that a SYS_RES_MEMORY resource should be mapped in as cacheable and/or read-only via BUS_SPACE_MAP_CACHEABLE and BUS_SPACE_MAP_READONLY respectively. - Do some minor cleanup like taking advantage of rman_init_from_resource(), factor out the common part of bus tag allocation into a newly added sparc64_alloc_bus_tag(), hook up some missing newbus methods and replace some homegrown versions with the generic counterparts etc. - While at it, let apb_attach() (which can't use the generic NEW_PCIB code as APB bridges just don't have the base and limit registers implemented) regarding the config space registers cached in pcib_softc and the SYSCTL reporting nodes set up.
2011-10-02 23:22:38 +00:00
DEVMETHOD(bus_setup_intr, sbus_setup_intr),
DEVMETHOD(bus_teardown_intr, bus_generic_teardown_intr),
DEVMETHOD(bus_get_resource, bus_generic_rl_get_resource),
DEVMETHOD(bus_get_resource_list, sbus_get_resource_list),
DEVMETHOD(bus_child_pnpinfo_str, ofw_bus_gen_child_pnpinfo_str),
Rototill the sparc64 nexus(4) (actually this brings in the code the sun4v nexus(4) in turn is based on): o Change nexus(4) to manage the resources of its children so the respective device drivers don't need to figure them out of OFW themselves. o Change nexus(4) to provide the ofw_bus KOBJ interface instead of using IVARs for supplying the OFW node and the subset of standard properties of its children. Together with the previous change this also allows to fully take advantage of newbus in that drivers like fhc(4), which attach on multiple parent busses, no longer require different bus front-ends as obtaining the OFW node and properties as well as resource allocation works the same for all supported busses. As such this change also is part 4/4 of allowing creator(4) to work in USIII-based machines as it allows this driver to attach on both nexus(4) and upa(4). On the other hand removing these IVARs breaks API compatibility with the powerpc nexus(4) but which isn't that bad as a) sparc64 currently doesn't share any device driver hanging off of nexus(4) with powerpc and b) they were no longer compatible regarding OFW-related extensions at the pci(4) level since quite some time. o Provide bus_get_dma_tag methods in nexus(4) and its children in order to handle DMA tags in a hierarchical way and get rid of the sparc64_root_dma_tag kludge. Together with the previous two items this changes also allows to completely get rid of the nexus(4) IVAR interface. It also includes: - pushing the constraints previously specified by the nexus_dmatag down into the DMA tags of psycho(4) and sbus(4) as it's their IOMMUs which induce these restrictions (and nothing at the nexus(4) or anything that would warrant specifying them there), - fixing some obviously wrong constraints of the psycho(4) and sbus(4) DMA tags, which happened to not actually be used with the sparc64_root_dma_tag kludge in place and therefore didn't cause problems so far, - replacing magic constants for constraints with macros as far as it is obvious as to where they come from. This doesn't include taking advantage of the newbus way to get the parent DMA tags implemented by this change in order to divorce the IOTSBs of the PCI and SBus IOMMUs or for implementing the workaround for the DMA sync bug in Sabre (and Tomatillo) bridges, yet, though. o Get rid of the notion that nexus(4) (mostly) reflects an UPA bus by replacing ofw_upa.h and with ofw_nexus.h (which was repo-copied from ofw_upa.h) and renaming its content, which actually applies to all of Fireplane/Safari, JBus and UPA (in the host bus case), as appropriate. o Just use M_DEVBUF instead of a separate M_NEXUS malloc type for allocating the device info for the children of nexus(4). This is done in order to not need to export M_NEXUS when deriving drivers for subordinate busses from the nexus(4) class. o Use the DEFINE_CLASS_0() macro to declare the nexus(4) driver so we can derive subclasses from it. o Const'ify the nexus_excl_name and nexus_excl_type arrays as well as add 'associations' and 'rsc', which are pseudo-devices without resources and therefore of no real interest for nexus(4), to the former. o Let the nexus(4) device memory rman manage the entire 64-bit address space instead of just the UPA_MEMSTART to UPA_MEMEND subregion as Fireplane/Safari- and JBus-based machines use multiple ranges, which can't be as easily divided as in the case of UPA (limiting the address space only served for sanity checking anyway). o Use M_WAITOK instead of M_NOWAIT when allocating the device info for children of nexus(4) in order to give one less opportunity for adding devices to nexus(4) to fail. o While adapting the drivers affected by the above nexus(4) changes, change them to take advantage of rman_get_rid() instead of caching the RIDs assigned to allocated resources, now that the RIDs of resources are correctly set. o In iommu(4) and nexus(4) replace hard-coded functions names, which actually became outdated in several places, in panic strings and status massages with __func__. [1] o Use driver_filter_t in prototypes where appropriate. o Add my copyright to creator(4), fhc(4), nexus(4), psycho(4) and sbus(4) as I changed considerable amounts of these drivers as well as added a bunch of new features, workarounds for silicon bugs etc. o Fix some white space nits. Due to lack of access to Exx00 hardware, these changes, i.e. central(4) and fhc(4), couldn't be runtime tested on such a machine. Exx00 are currently reported to panic before trying to attach nexus(4) anyway though. PR: 76052 [1] Approved by: re (kensmith)
2007-03-07 21:13:51 +00:00
DEVMETHOD(bus_get_dma_tag, sbus_get_dma_tag),
- Introduce an ofw_bus kobj-interface for retrieving the OFW node and a subset ("compatible", "device_type", "model" and "name") of the standard properties in drivers for devices on Open Firmware supported busses. The standard properties "reg", "interrupts" und "address" are not covered by this interface because they are only of interest in the respective bridge code. There's a remaining standard property "status" which is unclear how to support properly but which also isn't used in FreeBSD at present. This ofw_bus kobj-interface allows to replace the various (ebus_get_node(), ofw_pci_get_node(), etc.) and partially inconsistent (central_get_type() vs. sbus_get_device_type(), etc.) existing IVAR ones with a common one. This in turn allows to simplify and remove code-duplication in drivers for devices that can hang off of more than one OFW supported bus. - Convert the sparc64 Central, EBus, FHC, PCI and SBus bus drivers and the drivers for their children to use the ofw_bus kobj-interface. The IVAR- interfaces of the Central, EBus and FHC are entirely replaced by this. The PCI bus driver used its own kobj-interface and now also uses the ofw_bus one. The IVARs special to the SBus, e.g. for retrieving the burst size, remain. Beware: this causes an ABI-breakage for modules of drivers which used the IVAR-interfaces, i.e. esp(4), hme(4), isp(4) and uart(4), which need to be recompiled. The style-inconsistencies introduced in some of the bus drivers will be fixed by tmm@ in a generic clean-up of the respective drivers later (he requested to add the changes in the "new" style). - Convert the powerpc MacIO bus driver and the drivers for its children to use the ofw_bus kobj-interface. This invloves removing the IVARs related to the "reg" property which were unused and a leftover from the NetBSD origini of the code. There's no ABI-breakage caused by this because none of these driver are currently built as modules. There are other powerpc bus drivers which can be converted to the ofw_bus kobj-interface, e.g. the PCI bus driver, which should be done together with converting powerpc to use the OFW PCI code from sparc64. - Make the SBus and FHC front-end of zs(4) and the sparc64 eeprom(4) take advantage of the ofw_bus kobj-interface and simplify them a bit. Reviewed by: grehan, tmm Approved by: re (scottl) Discussed with: tmm Tested with: Sun AX1105, AXe, Ultra 2, Ultra 60; PPC cross-build on i386
2004-08-12 17:41:33 +00:00
/* ofw_bus interface */
DEVMETHOD(ofw_bus_get_devinfo, sbus_get_devinfo),
DEVMETHOD(ofw_bus_get_compat, ofw_bus_gen_get_compat),
DEVMETHOD(ofw_bus_get_model, ofw_bus_gen_get_model),
DEVMETHOD(ofw_bus_get_name, ofw_bus_gen_get_name),
DEVMETHOD(ofw_bus_get_node, ofw_bus_gen_get_node),
DEVMETHOD(ofw_bus_get_type, ofw_bus_gen_get_type),
- Introduce an ofw_bus kobj-interface for retrieving the OFW node and a subset ("compatible", "device_type", "model" and "name") of the standard properties in drivers for devices on Open Firmware supported busses. The standard properties "reg", "interrupts" und "address" are not covered by this interface because they are only of interest in the respective bridge code. There's a remaining standard property "status" which is unclear how to support properly but which also isn't used in FreeBSD at present. This ofw_bus kobj-interface allows to replace the various (ebus_get_node(), ofw_pci_get_node(), etc.) and partially inconsistent (central_get_type() vs. sbus_get_device_type(), etc.) existing IVAR ones with a common one. This in turn allows to simplify and remove code-duplication in drivers for devices that can hang off of more than one OFW supported bus. - Convert the sparc64 Central, EBus, FHC, PCI and SBus bus drivers and the drivers for their children to use the ofw_bus kobj-interface. The IVAR- interfaces of the Central, EBus and FHC are entirely replaced by this. The PCI bus driver used its own kobj-interface and now also uses the ofw_bus one. The IVARs special to the SBus, e.g. for retrieving the burst size, remain. Beware: this causes an ABI-breakage for modules of drivers which used the IVAR-interfaces, i.e. esp(4), hme(4), isp(4) and uart(4), which need to be recompiled. The style-inconsistencies introduced in some of the bus drivers will be fixed by tmm@ in a generic clean-up of the respective drivers later (he requested to add the changes in the "new" style). - Convert the powerpc MacIO bus driver and the drivers for its children to use the ofw_bus kobj-interface. This invloves removing the IVARs related to the "reg" property which were unused and a leftover from the NetBSD origini of the code. There's no ABI-breakage caused by this because none of these driver are currently built as modules. There are other powerpc bus drivers which can be converted to the ofw_bus kobj-interface, e.g. the PCI bus driver, which should be done together with converting powerpc to use the OFW PCI code from sparc64. - Make the SBus and FHC front-end of zs(4) and the sparc64 eeprom(4) take advantage of the ofw_bus kobj-interface and simplify them a bit. Reviewed by: grehan, tmm Approved by: re (scottl) Discussed with: tmm Tested with: Sun AX1105, AXe, Ultra 2, Ultra 60; PPC cross-build on i386
2004-08-12 17:41:33 +00:00
KOBJMETHOD_END
};
static driver_t sbus_driver = {
"sbus",
sbus_methods,
sizeof(struct sbus_softc),
};
static devclass_t sbus_devclass;
EARLY_DRIVER_MODULE(sbus, nexus, sbus_driver, sbus_devclass, 0, 0,
BUS_PASS_BUS);
MODULE_DEPEND(sbus, nexus, 1, 1, 1);
MODULE_VERSION(sbus, 1);
#define OFW_SBUS_TYPE "sbus"
#define OFW_SBUS_NAME "sbus"
o Revamp the sparc64 interrupt code in order to be able to interface with the INTR_FILTER-enabled MI code. Basically this consists of registering an interrupt controller (of which there can be multiple and optionally different ones either per host-to-foo bridge or shared amongst host-to-foo bridges in any one machine) along with an interrupt vector as specific argument for all the interrupt vectors used by a given host-to-foo bridge (roughly similar to registering interrupt sources on amd64 and i386), providing functions to enable, clear and disable the interrupts of the children beneath the bridge. This also includes: - No longer entering a critical section in tl0_intr() and tl1_intr() for executing interrupt handlers but rather let the handlers enter it themselves so in the case of intr_event_handle() we don't enter a nested critical section. - Adding infrastructure for binding delivery of interrupt vectors to specific CPUs which later on can be interfaced with the code from amd64/i386 for binding interrupts to specific CPUs. - Getting rid of the wrapper hack introduced along the lines of the API changes for INTR_FILTER which as a side-effect caused interrupts associated with ithread handlers only to get the elevated priority of those associated with filters ("fast handlers") (this removes the hack also in the non-INTR_FILTER case). - Disabling (by not clearing) an interrupt in the interrupt controller until all associated handlers have been executed, which is crucial for the typical locking strategy of NIC drivers in order to work correctly in case of shared interrupts. This was a more or less theoretical problem on sparc64 though, as shared interrupts are rather uncommon there except for the on-board SCCs and UARTs. Note that due to the behavior of at least of some of the interrupt controllers used on sparc64 an enable+EOI instead of a disable+EOI approach (as implied by the INTR_FILTER MI code and implemented on other architectures) is used as the latter can cause lost interrupts or in the worst case interrupt starvation. o Correct a typo in sbus_alloc_resource() which caused (pass-through) allocations to only work down to the grandchildren of the bus, which wasn't a real problem so far as we don't support any devices which are great-grandchildren or greater of a U2S bridge, yet. o In fhc(4) use bus_{read,write}_4() instead of bus_space_{read,write}_4() in order to get rid of sc_bh and sc_bt in the fhc_softc. Also get rid of some other unneeded members in fhc_softc. Reviewed by: marcel (earlier version) Approved by: re (kensmith)
2007-09-06 19:16:30 +00:00
static const struct intr_controller sbus_ic = {
sbus_intr_enable,
sbus_intr_disable,
sbus_intr_assign,
sbus_intr_clear
o Revamp the sparc64 interrupt code in order to be able to interface with the INTR_FILTER-enabled MI code. Basically this consists of registering an interrupt controller (of which there can be multiple and optionally different ones either per host-to-foo bridge or shared amongst host-to-foo bridges in any one machine) along with an interrupt vector as specific argument for all the interrupt vectors used by a given host-to-foo bridge (roughly similar to registering interrupt sources on amd64 and i386), providing functions to enable, clear and disable the interrupts of the children beneath the bridge. This also includes: - No longer entering a critical section in tl0_intr() and tl1_intr() for executing interrupt handlers but rather let the handlers enter it themselves so in the case of intr_event_handle() we don't enter a nested critical section. - Adding infrastructure for binding delivery of interrupt vectors to specific CPUs which later on can be interfaced with the code from amd64/i386 for binding interrupts to specific CPUs. - Getting rid of the wrapper hack introduced along the lines of the API changes for INTR_FILTER which as a side-effect caused interrupts associated with ithread handlers only to get the elevated priority of those associated with filters ("fast handlers") (this removes the hack also in the non-INTR_FILTER case). - Disabling (by not clearing) an interrupt in the interrupt controller until all associated handlers have been executed, which is crucial for the typical locking strategy of NIC drivers in order to work correctly in case of shared interrupts. This was a more or less theoretical problem on sparc64 though, as shared interrupts are rather uncommon there except for the on-board SCCs and UARTs. Note that due to the behavior of at least of some of the interrupt controllers used on sparc64 an enable+EOI instead of a disable+EOI approach (as implied by the INTR_FILTER MI code and implemented on other architectures) is used as the latter can cause lost interrupts or in the worst case interrupt starvation. o Correct a typo in sbus_alloc_resource() which caused (pass-through) allocations to only work down to the grandchildren of the bus, which wasn't a real problem so far as we don't support any devices which are great-grandchildren or greater of a U2S bridge, yet. o In fhc(4) use bus_{read,write}_4() instead of bus_space_{read,write}_4() in order to get rid of sc_bh and sc_bt in the fhc_softc. Also get rid of some other unneeded members in fhc_softc. Reviewed by: marcel (earlier version) Approved by: re (kensmith)
2007-09-06 19:16:30 +00:00
};
struct sbus_icarg {
struct sbus_softc *sica_sc;
bus_addr_t sica_map;
bus_addr_t sica_clr;
};
static const char *const sbus_order_first[] = {
- Add an IVAR for retrieving the interrupt group number of the parent Sbus device and which also applies to the children. This is very usefull for drivers for the various subordinate busses so they don't need to fiddle with the OFW node of their parent themselves. As SBus busses hang of the nexus and we don't use the ofw_bus interface for nexus devices, yet, this would also require special knowledge about this in the drivers for the SBus children which these shouldn't need to have. This includes switching to use an unshifted IGN in the sc_ign member of the sbus(4) softc internally. - For SBus child devices where there are variants that are actually split split into two SBus devices (as opposed to the first half of the device being a SBus device and the second half hanging off of the first one) like 'auxio' and 'SUNW,fdtwo' or 'dma' and 'esp' probe the SBus device which is a prerequisite to the driver attaching to the second one with a lower order. This saves us from dealing with different probe orders in the respective device drivers which generally is more hackish. - Remove a stale comment about the 'specials' array above the attaching of the child devices. This is a remnant of the NetBSD/sparc origin of this code. There the 'specials' array is also used to probe certain devices which are prerequisites to others first. Why NetBSD soley relies on the devices having the expected order in the OFW tree on sparc64 isn't clear to me, as far as I can tell OFW doesn't guaranteed such things.
2005-05-19 14:47:31 +00:00
"auxio",
"dma",
NULL
};
static int
sbus_inlist(const char *name, const char *const *list)
- Add an IVAR for retrieving the interrupt group number of the parent Sbus device and which also applies to the children. This is very usefull for drivers for the various subordinate busses so they don't need to fiddle with the OFW node of their parent themselves. As SBus busses hang of the nexus and we don't use the ofw_bus interface for nexus devices, yet, this would also require special knowledge about this in the drivers for the SBus children which these shouldn't need to have. This includes switching to use an unshifted IGN in the sc_ign member of the sbus(4) softc internally. - For SBus child devices where there are variants that are actually split split into two SBus devices (as opposed to the first half of the device being a SBus device and the second half hanging off of the first one) like 'auxio' and 'SUNW,fdtwo' or 'dma' and 'esp' probe the SBus device which is a prerequisite to the driver attaching to the second one with a lower order. This saves us from dealing with different probe orders in the respective device drivers which generally is more hackish. - Remove a stale comment about the 'specials' array above the attaching of the child devices. This is a remnant of the NetBSD/sparc origin of this code. There the 'specials' array is also used to probe certain devices which are prerequisites to others first. Why NetBSD soley relies on the devices having the expected order in the OFW tree on sparc64 isn't clear to me, as far as I can tell OFW doesn't guaranteed such things.
2005-05-19 14:47:31 +00:00
{
int i;
if (name == NULL)
return (0);
for (i = 0; list[i] != NULL; i++) {
if (strcmp(name, list[i]) == 0)
return (1);
}
return (0);
}
static int
sbus_probe(device_t dev)
{
Rototill the sparc64 nexus(4) (actually this brings in the code the sun4v nexus(4) in turn is based on): o Change nexus(4) to manage the resources of its children so the respective device drivers don't need to figure them out of OFW themselves. o Change nexus(4) to provide the ofw_bus KOBJ interface instead of using IVARs for supplying the OFW node and the subset of standard properties of its children. Together with the previous change this also allows to fully take advantage of newbus in that drivers like fhc(4), which attach on multiple parent busses, no longer require different bus front-ends as obtaining the OFW node and properties as well as resource allocation works the same for all supported busses. As such this change also is part 4/4 of allowing creator(4) to work in USIII-based machines as it allows this driver to attach on both nexus(4) and upa(4). On the other hand removing these IVARs breaks API compatibility with the powerpc nexus(4) but which isn't that bad as a) sparc64 currently doesn't share any device driver hanging off of nexus(4) with powerpc and b) they were no longer compatible regarding OFW-related extensions at the pci(4) level since quite some time. o Provide bus_get_dma_tag methods in nexus(4) and its children in order to handle DMA tags in a hierarchical way and get rid of the sparc64_root_dma_tag kludge. Together with the previous two items this changes also allows to completely get rid of the nexus(4) IVAR interface. It also includes: - pushing the constraints previously specified by the nexus_dmatag down into the DMA tags of psycho(4) and sbus(4) as it's their IOMMUs which induce these restrictions (and nothing at the nexus(4) or anything that would warrant specifying them there), - fixing some obviously wrong constraints of the psycho(4) and sbus(4) DMA tags, which happened to not actually be used with the sparc64_root_dma_tag kludge in place and therefore didn't cause problems so far, - replacing magic constants for constraints with macros as far as it is obvious as to where they come from. This doesn't include taking advantage of the newbus way to get the parent DMA tags implemented by this change in order to divorce the IOTSBs of the PCI and SBus IOMMUs or for implementing the workaround for the DMA sync bug in Sabre (and Tomatillo) bridges, yet, though. o Get rid of the notion that nexus(4) (mostly) reflects an UPA bus by replacing ofw_upa.h and with ofw_nexus.h (which was repo-copied from ofw_upa.h) and renaming its content, which actually applies to all of Fireplane/Safari, JBus and UPA (in the host bus case), as appropriate. o Just use M_DEVBUF instead of a separate M_NEXUS malloc type for allocating the device info for the children of nexus(4). This is done in order to not need to export M_NEXUS when deriving drivers for subordinate busses from the nexus(4) class. o Use the DEFINE_CLASS_0() macro to declare the nexus(4) driver so we can derive subclasses from it. o Const'ify the nexus_excl_name and nexus_excl_type arrays as well as add 'associations' and 'rsc', which are pseudo-devices without resources and therefore of no real interest for nexus(4), to the former. o Let the nexus(4) device memory rman manage the entire 64-bit address space instead of just the UPA_MEMSTART to UPA_MEMEND subregion as Fireplane/Safari- and JBus-based machines use multiple ranges, which can't be as easily divided as in the case of UPA (limiting the address space only served for sanity checking anyway). o Use M_WAITOK instead of M_NOWAIT when allocating the device info for children of nexus(4) in order to give one less opportunity for adding devices to nexus(4) to fail. o While adapting the drivers affected by the above nexus(4) changes, change them to take advantage of rman_get_rid() instead of caching the RIDs assigned to allocated resources, now that the RIDs of resources are correctly set. o In iommu(4) and nexus(4) replace hard-coded functions names, which actually became outdated in several places, in panic strings and status massages with __func__. [1] o Use driver_filter_t in prototypes where appropriate. o Add my copyright to creator(4), fhc(4), nexus(4), psycho(4) and sbus(4) as I changed considerable amounts of these drivers as well as added a bunch of new features, workarounds for silicon bugs etc. o Fix some white space nits. Due to lack of access to Exx00 hardware, these changes, i.e. central(4) and fhc(4), couldn't be runtime tested on such a machine. Exx00 are currently reported to panic before trying to attach nexus(4) anyway though. PR: 76052 [1] Approved by: re (kensmith)
2007-03-07 21:13:51 +00:00
const char *t;
Rototill the sparc64 nexus(4) (actually this brings in the code the sun4v nexus(4) in turn is based on): o Change nexus(4) to manage the resources of its children so the respective device drivers don't need to figure them out of OFW themselves. o Change nexus(4) to provide the ofw_bus KOBJ interface instead of using IVARs for supplying the OFW node and the subset of standard properties of its children. Together with the previous change this also allows to fully take advantage of newbus in that drivers like fhc(4), which attach on multiple parent busses, no longer require different bus front-ends as obtaining the OFW node and properties as well as resource allocation works the same for all supported busses. As such this change also is part 4/4 of allowing creator(4) to work in USIII-based machines as it allows this driver to attach on both nexus(4) and upa(4). On the other hand removing these IVARs breaks API compatibility with the powerpc nexus(4) but which isn't that bad as a) sparc64 currently doesn't share any device driver hanging off of nexus(4) with powerpc and b) they were no longer compatible regarding OFW-related extensions at the pci(4) level since quite some time. o Provide bus_get_dma_tag methods in nexus(4) and its children in order to handle DMA tags in a hierarchical way and get rid of the sparc64_root_dma_tag kludge. Together with the previous two items this changes also allows to completely get rid of the nexus(4) IVAR interface. It also includes: - pushing the constraints previously specified by the nexus_dmatag down into the DMA tags of psycho(4) and sbus(4) as it's their IOMMUs which induce these restrictions (and nothing at the nexus(4) or anything that would warrant specifying them there), - fixing some obviously wrong constraints of the psycho(4) and sbus(4) DMA tags, which happened to not actually be used with the sparc64_root_dma_tag kludge in place and therefore didn't cause problems so far, - replacing magic constants for constraints with macros as far as it is obvious as to where they come from. This doesn't include taking advantage of the newbus way to get the parent DMA tags implemented by this change in order to divorce the IOTSBs of the PCI and SBus IOMMUs or for implementing the workaround for the DMA sync bug in Sabre (and Tomatillo) bridges, yet, though. o Get rid of the notion that nexus(4) (mostly) reflects an UPA bus by replacing ofw_upa.h and with ofw_nexus.h (which was repo-copied from ofw_upa.h) and renaming its content, which actually applies to all of Fireplane/Safari, JBus and UPA (in the host bus case), as appropriate. o Just use M_DEVBUF instead of a separate M_NEXUS malloc type for allocating the device info for the children of nexus(4). This is done in order to not need to export M_NEXUS when deriving drivers for subordinate busses from the nexus(4) class. o Use the DEFINE_CLASS_0() macro to declare the nexus(4) driver so we can derive subclasses from it. o Const'ify the nexus_excl_name and nexus_excl_type arrays as well as add 'associations' and 'rsc', which are pseudo-devices without resources and therefore of no real interest for nexus(4), to the former. o Let the nexus(4) device memory rman manage the entire 64-bit address space instead of just the UPA_MEMSTART to UPA_MEMEND subregion as Fireplane/Safari- and JBus-based machines use multiple ranges, which can't be as easily divided as in the case of UPA (limiting the address space only served for sanity checking anyway). o Use M_WAITOK instead of M_NOWAIT when allocating the device info for children of nexus(4) in order to give one less opportunity for adding devices to nexus(4) to fail. o While adapting the drivers affected by the above nexus(4) changes, change them to take advantage of rman_get_rid() instead of caching the RIDs assigned to allocated resources, now that the RIDs of resources are correctly set. o In iommu(4) and nexus(4) replace hard-coded functions names, which actually became outdated in several places, in panic strings and status massages with __func__. [1] o Use driver_filter_t in prototypes where appropriate. o Add my copyright to creator(4), fhc(4), nexus(4), psycho(4) and sbus(4) as I changed considerable amounts of these drivers as well as added a bunch of new features, workarounds for silicon bugs etc. o Fix some white space nits. Due to lack of access to Exx00 hardware, these changes, i.e. central(4) and fhc(4), couldn't be runtime tested on such a machine. Exx00 are currently reported to panic before trying to attach nexus(4) anyway though. PR: 76052 [1] Approved by: re (kensmith)
2007-03-07 21:13:51 +00:00
t = ofw_bus_get_type(dev);
if (((t == NULL || strcmp(t, OFW_SBUS_TYPE) != 0)) &&
Rototill the sparc64 nexus(4) (actually this brings in the code the sun4v nexus(4) in turn is based on): o Change nexus(4) to manage the resources of its children so the respective device drivers don't need to figure them out of OFW themselves. o Change nexus(4) to provide the ofw_bus KOBJ interface instead of using IVARs for supplying the OFW node and the subset of standard properties of its children. Together with the previous change this also allows to fully take advantage of newbus in that drivers like fhc(4), which attach on multiple parent busses, no longer require different bus front-ends as obtaining the OFW node and properties as well as resource allocation works the same for all supported busses. As such this change also is part 4/4 of allowing creator(4) to work in USIII-based machines as it allows this driver to attach on both nexus(4) and upa(4). On the other hand removing these IVARs breaks API compatibility with the powerpc nexus(4) but which isn't that bad as a) sparc64 currently doesn't share any device driver hanging off of nexus(4) with powerpc and b) they were no longer compatible regarding OFW-related extensions at the pci(4) level since quite some time. o Provide bus_get_dma_tag methods in nexus(4) and its children in order to handle DMA tags in a hierarchical way and get rid of the sparc64_root_dma_tag kludge. Together with the previous two items this changes also allows to completely get rid of the nexus(4) IVAR interface. It also includes: - pushing the constraints previously specified by the nexus_dmatag down into the DMA tags of psycho(4) and sbus(4) as it's their IOMMUs which induce these restrictions (and nothing at the nexus(4) or anything that would warrant specifying them there), - fixing some obviously wrong constraints of the psycho(4) and sbus(4) DMA tags, which happened to not actually be used with the sparc64_root_dma_tag kludge in place and therefore didn't cause problems so far, - replacing magic constants for constraints with macros as far as it is obvious as to where they come from. This doesn't include taking advantage of the newbus way to get the parent DMA tags implemented by this change in order to divorce the IOTSBs of the PCI and SBus IOMMUs or for implementing the workaround for the DMA sync bug in Sabre (and Tomatillo) bridges, yet, though. o Get rid of the notion that nexus(4) (mostly) reflects an UPA bus by replacing ofw_upa.h and with ofw_nexus.h (which was repo-copied from ofw_upa.h) and renaming its content, which actually applies to all of Fireplane/Safari, JBus and UPA (in the host bus case), as appropriate. o Just use M_DEVBUF instead of a separate M_NEXUS malloc type for allocating the device info for the children of nexus(4). This is done in order to not need to export M_NEXUS when deriving drivers for subordinate busses from the nexus(4) class. o Use the DEFINE_CLASS_0() macro to declare the nexus(4) driver so we can derive subclasses from it. o Const'ify the nexus_excl_name and nexus_excl_type arrays as well as add 'associations' and 'rsc', which are pseudo-devices without resources and therefore of no real interest for nexus(4), to the former. o Let the nexus(4) device memory rman manage the entire 64-bit address space instead of just the UPA_MEMSTART to UPA_MEMEND subregion as Fireplane/Safari- and JBus-based machines use multiple ranges, which can't be as easily divided as in the case of UPA (limiting the address space only served for sanity checking anyway). o Use M_WAITOK instead of M_NOWAIT when allocating the device info for children of nexus(4) in order to give one less opportunity for adding devices to nexus(4) to fail. o While adapting the drivers affected by the above nexus(4) changes, change them to take advantage of rman_get_rid() instead of caching the RIDs assigned to allocated resources, now that the RIDs of resources are correctly set. o In iommu(4) and nexus(4) replace hard-coded functions names, which actually became outdated in several places, in panic strings and status massages with __func__. [1] o Use driver_filter_t in prototypes where appropriate. o Add my copyright to creator(4), fhc(4), nexus(4), psycho(4) and sbus(4) as I changed considerable amounts of these drivers as well as added a bunch of new features, workarounds for silicon bugs etc. o Fix some white space nits. Due to lack of access to Exx00 hardware, these changes, i.e. central(4) and fhc(4), couldn't be runtime tested on such a machine. Exx00 are currently reported to panic before trying to attach nexus(4) anyway though. PR: 76052 [1] Approved by: re (kensmith)
2007-03-07 21:13:51 +00:00
strcmp(ofw_bus_get_name(dev), OFW_SBUS_NAME) != 0)
return (ENXIO);
device_set_desc(dev, "U2S UPA-SBus bridge");
return (0);
}
static int
sbus_attach(device_t dev)
{
struct sbus_softc *sc;
struct sbus_devinfo *sdi;
o Revamp the sparc64 interrupt code in order to be able to interface with the INTR_FILTER-enabled MI code. Basically this consists of registering an interrupt controller (of which there can be multiple and optionally different ones either per host-to-foo bridge or shared amongst host-to-foo bridges in any one machine) along with an interrupt vector as specific argument for all the interrupt vectors used by a given host-to-foo bridge (roughly similar to registering interrupt sources on amd64 and i386), providing functions to enable, clear and disable the interrupts of the children beneath the bridge. This also includes: - No longer entering a critical section in tl0_intr() and tl1_intr() for executing interrupt handlers but rather let the handlers enter it themselves so in the case of intr_event_handle() we don't enter a nested critical section. - Adding infrastructure for binding delivery of interrupt vectors to specific CPUs which later on can be interfaced with the code from amd64/i386 for binding interrupts to specific CPUs. - Getting rid of the wrapper hack introduced along the lines of the API changes for INTR_FILTER which as a side-effect caused interrupts associated with ithread handlers only to get the elevated priority of those associated with filters ("fast handlers") (this removes the hack also in the non-INTR_FILTER case). - Disabling (by not clearing) an interrupt in the interrupt controller until all associated handlers have been executed, which is crucial for the typical locking strategy of NIC drivers in order to work correctly in case of shared interrupts. This was a more or less theoretical problem on sparc64 though, as shared interrupts are rather uncommon there except for the on-board SCCs and UARTs. Note that due to the behavior of at least of some of the interrupt controllers used on sparc64 an enable+EOI instead of a disable+EOI approach (as implied by the INTR_FILTER MI code and implemented on other architectures) is used as the latter can cause lost interrupts or in the worst case interrupt starvation. o Correct a typo in sbus_alloc_resource() which caused (pass-through) allocations to only work down to the grandchildren of the bus, which wasn't a real problem so far as we don't support any devices which are great-grandchildren or greater of a U2S bridge, yet. o In fhc(4) use bus_{read,write}_4() instead of bus_space_{read,write}_4() in order to get rid of sc_bh and sc_bt in the fhc_softc. Also get rid of some other unneeded members in fhc_softc. Reviewed by: marcel (earlier version) Approved by: re (kensmith)
2007-09-06 19:16:30 +00:00
struct sbus_icarg *sica;
struct sbus_ranges *range;
struct resource *res;
Rototill the sparc64 nexus(4) (actually this brings in the code the sun4v nexus(4) in turn is based on): o Change nexus(4) to manage the resources of its children so the respective device drivers don't need to figure them out of OFW themselves. o Change nexus(4) to provide the ofw_bus KOBJ interface instead of using IVARs for supplying the OFW node and the subset of standard properties of its children. Together with the previous change this also allows to fully take advantage of newbus in that drivers like fhc(4), which attach on multiple parent busses, no longer require different bus front-ends as obtaining the OFW node and properties as well as resource allocation works the same for all supported busses. As such this change also is part 4/4 of allowing creator(4) to work in USIII-based machines as it allows this driver to attach on both nexus(4) and upa(4). On the other hand removing these IVARs breaks API compatibility with the powerpc nexus(4) but which isn't that bad as a) sparc64 currently doesn't share any device driver hanging off of nexus(4) with powerpc and b) they were no longer compatible regarding OFW-related extensions at the pci(4) level since quite some time. o Provide bus_get_dma_tag methods in nexus(4) and its children in order to handle DMA tags in a hierarchical way and get rid of the sparc64_root_dma_tag kludge. Together with the previous two items this changes also allows to completely get rid of the nexus(4) IVAR interface. It also includes: - pushing the constraints previously specified by the nexus_dmatag down into the DMA tags of psycho(4) and sbus(4) as it's their IOMMUs which induce these restrictions (and nothing at the nexus(4) or anything that would warrant specifying them there), - fixing some obviously wrong constraints of the psycho(4) and sbus(4) DMA tags, which happened to not actually be used with the sparc64_root_dma_tag kludge in place and therefore didn't cause problems so far, - replacing magic constants for constraints with macros as far as it is obvious as to where they come from. This doesn't include taking advantage of the newbus way to get the parent DMA tags implemented by this change in order to divorce the IOTSBs of the PCI and SBus IOMMUs or for implementing the workaround for the DMA sync bug in Sabre (and Tomatillo) bridges, yet, though. o Get rid of the notion that nexus(4) (mostly) reflects an UPA bus by replacing ofw_upa.h and with ofw_nexus.h (which was repo-copied from ofw_upa.h) and renaming its content, which actually applies to all of Fireplane/Safari, JBus and UPA (in the host bus case), as appropriate. o Just use M_DEVBUF instead of a separate M_NEXUS malloc type for allocating the device info for the children of nexus(4). This is done in order to not need to export M_NEXUS when deriving drivers for subordinate busses from the nexus(4) class. o Use the DEFINE_CLASS_0() macro to declare the nexus(4) driver so we can derive subclasses from it. o Const'ify the nexus_excl_name and nexus_excl_type arrays as well as add 'associations' and 'rsc', which are pseudo-devices without resources and therefore of no real interest for nexus(4), to the former. o Let the nexus(4) device memory rman manage the entire 64-bit address space instead of just the UPA_MEMSTART to UPA_MEMEND subregion as Fireplane/Safari- and JBus-based machines use multiple ranges, which can't be as easily divided as in the case of UPA (limiting the address space only served for sanity checking anyway). o Use M_WAITOK instead of M_NOWAIT when allocating the device info for children of nexus(4) in order to give one less opportunity for adding devices to nexus(4) to fail. o While adapting the drivers affected by the above nexus(4) changes, change them to take advantage of rman_get_rid() instead of caching the RIDs assigned to allocated resources, now that the RIDs of resources are correctly set. o In iommu(4) and nexus(4) replace hard-coded functions names, which actually became outdated in several places, in panic strings and status massages with __func__. [1] o Use driver_filter_t in prototypes where appropriate. o Add my copyright to creator(4), fhc(4), nexus(4), psycho(4) and sbus(4) as I changed considerable amounts of these drivers as well as added a bunch of new features, workarounds for silicon bugs etc. o Fix some white space nits. Due to lack of access to Exx00 hardware, these changes, i.e. central(4) and fhc(4), couldn't be runtime tested on such a machine. Exx00 are currently reported to panic before trying to attach nexus(4) anyway though. PR: 76052 [1] Approved by: re (kensmith)
2007-03-07 21:13:51 +00:00
struct resource_list *rl;
device_t cdev;
o Revamp the sparc64 interrupt code in order to be able to interface with the INTR_FILTER-enabled MI code. Basically this consists of registering an interrupt controller (of which there can be multiple and optionally different ones either per host-to-foo bridge or shared amongst host-to-foo bridges in any one machine) along with an interrupt vector as specific argument for all the interrupt vectors used by a given host-to-foo bridge (roughly similar to registering interrupt sources on amd64 and i386), providing functions to enable, clear and disable the interrupts of the children beneath the bridge. This also includes: - No longer entering a critical section in tl0_intr() and tl1_intr() for executing interrupt handlers but rather let the handlers enter it themselves so in the case of intr_event_handle() we don't enter a nested critical section. - Adding infrastructure for binding delivery of interrupt vectors to specific CPUs which later on can be interfaced with the code from amd64/i386 for binding interrupts to specific CPUs. - Getting rid of the wrapper hack introduced along the lines of the API changes for INTR_FILTER which as a side-effect caused interrupts associated with ithread handlers only to get the elevated priority of those associated with filters ("fast handlers") (this removes the hack also in the non-INTR_FILTER case). - Disabling (by not clearing) an interrupt in the interrupt controller until all associated handlers have been executed, which is crucial for the typical locking strategy of NIC drivers in order to work correctly in case of shared interrupts. This was a more or less theoretical problem on sparc64 though, as shared interrupts are rather uncommon there except for the on-board SCCs and UARTs. Note that due to the behavior of at least of some of the interrupt controllers used on sparc64 an enable+EOI instead of a disable+EOI approach (as implied by the INTR_FILTER MI code and implemented on other architectures) is used as the latter can cause lost interrupts or in the worst case interrupt starvation. o Correct a typo in sbus_alloc_resource() which caused (pass-through) allocations to only work down to the grandchildren of the bus, which wasn't a real problem so far as we don't support any devices which are great-grandchildren or greater of a U2S bridge, yet. o In fhc(4) use bus_{read,write}_4() instead of bus_space_{read,write}_4() in order to get rid of sc_bh and sc_bt in the fhc_softc. Also get rid of some other unneeded members in fhc_softc. Reviewed by: marcel (earlier version) Approved by: re (kensmith)
2007-09-06 19:16:30 +00:00
bus_addr_t intrclr, intrmap, phys;
bus_size_t size;
o Revamp the sparc64 interrupt code in order to be able to interface with the INTR_FILTER-enabled MI code. Basically this consists of registering an interrupt controller (of which there can be multiple and optionally different ones either per host-to-foo bridge or shared amongst host-to-foo bridges in any one machine) along with an interrupt vector as specific argument for all the interrupt vectors used by a given host-to-foo bridge (roughly similar to registering interrupt sources on amd64 and i386), providing functions to enable, clear and disable the interrupts of the children beneath the bridge. This also includes: - No longer entering a critical section in tl0_intr() and tl1_intr() for executing interrupt handlers but rather let the handlers enter it themselves so in the case of intr_event_handle() we don't enter a nested critical section. - Adding infrastructure for binding delivery of interrupt vectors to specific CPUs which later on can be interfaced with the code from amd64/i386 for binding interrupts to specific CPUs. - Getting rid of the wrapper hack introduced along the lines of the API changes for INTR_FILTER which as a side-effect caused interrupts associated with ithread handlers only to get the elevated priority of those associated with filters ("fast handlers") (this removes the hack also in the non-INTR_FILTER case). - Disabling (by not clearing) an interrupt in the interrupt controller until all associated handlers have been executed, which is crucial for the typical locking strategy of NIC drivers in order to work correctly in case of shared interrupts. This was a more or less theoretical problem on sparc64 though, as shared interrupts are rather uncommon there except for the on-board SCCs and UARTs. Note that due to the behavior of at least of some of the interrupt controllers used on sparc64 an enable+EOI instead of a disable+EOI approach (as implied by the INTR_FILTER MI code and implemented on other architectures) is used as the latter can cause lost interrupts or in the worst case interrupt starvation. o Correct a typo in sbus_alloc_resource() which caused (pass-through) allocations to only work down to the grandchildren of the bus, which wasn't a real problem so far as we don't support any devices which are great-grandchildren or greater of a U2S bridge, yet. o In fhc(4) use bus_{read,write}_4() instead of bus_space_{read,write}_4() in order to get rid of sc_bh and sc_bt in the fhc_softc. Also get rid of some other unneeded members in fhc_softc. Reviewed by: marcel (earlier version) Approved by: re (kensmith)
2007-09-06 19:16:30 +00:00
u_long vec;
phandle_t child, node;
uint32_t prop;
int i, j;
sc = device_get_softc(dev);
o Revamp the sparc64 interrupt code in order to be able to interface with the INTR_FILTER-enabled MI code. Basically this consists of registering an interrupt controller (of which there can be multiple and optionally different ones either per host-to-foo bridge or shared amongst host-to-foo bridges in any one machine) along with an interrupt vector as specific argument for all the interrupt vectors used by a given host-to-foo bridge (roughly similar to registering interrupt sources on amd64 and i386), providing functions to enable, clear and disable the interrupts of the children beneath the bridge. This also includes: - No longer entering a critical section in tl0_intr() and tl1_intr() for executing interrupt handlers but rather let the handlers enter it themselves so in the case of intr_event_handle() we don't enter a nested critical section. - Adding infrastructure for binding delivery of interrupt vectors to specific CPUs which later on can be interfaced with the code from amd64/i386 for binding interrupts to specific CPUs. - Getting rid of the wrapper hack introduced along the lines of the API changes for INTR_FILTER which as a side-effect caused interrupts associated with ithread handlers only to get the elevated priority of those associated with filters ("fast handlers") (this removes the hack also in the non-INTR_FILTER case). - Disabling (by not clearing) an interrupt in the interrupt controller until all associated handlers have been executed, which is crucial for the typical locking strategy of NIC drivers in order to work correctly in case of shared interrupts. This was a more or less theoretical problem on sparc64 though, as shared interrupts are rather uncommon there except for the on-board SCCs and UARTs. Note that due to the behavior of at least of some of the interrupt controllers used on sparc64 an enable+EOI instead of a disable+EOI approach (as implied by the INTR_FILTER MI code and implemented on other architectures) is used as the latter can cause lost interrupts or in the worst case interrupt starvation. o Correct a typo in sbus_alloc_resource() which caused (pass-through) allocations to only work down to the grandchildren of the bus, which wasn't a real problem so far as we don't support any devices which are great-grandchildren or greater of a U2S bridge, yet. o In fhc(4) use bus_{read,write}_4() instead of bus_space_{read,write}_4() in order to get rid of sc_bh and sc_bt in the fhc_softc. Also get rid of some other unneeded members in fhc_softc. Reviewed by: marcel (earlier version) Approved by: re (kensmith)
2007-09-06 19:16:30 +00:00
sc->sc_dev = dev;
Rototill the sparc64 nexus(4) (actually this brings in the code the sun4v nexus(4) in turn is based on): o Change nexus(4) to manage the resources of its children so the respective device drivers don't need to figure them out of OFW themselves. o Change nexus(4) to provide the ofw_bus KOBJ interface instead of using IVARs for supplying the OFW node and the subset of standard properties of its children. Together with the previous change this also allows to fully take advantage of newbus in that drivers like fhc(4), which attach on multiple parent busses, no longer require different bus front-ends as obtaining the OFW node and properties as well as resource allocation works the same for all supported busses. As such this change also is part 4/4 of allowing creator(4) to work in USIII-based machines as it allows this driver to attach on both nexus(4) and upa(4). On the other hand removing these IVARs breaks API compatibility with the powerpc nexus(4) but which isn't that bad as a) sparc64 currently doesn't share any device driver hanging off of nexus(4) with powerpc and b) they were no longer compatible regarding OFW-related extensions at the pci(4) level since quite some time. o Provide bus_get_dma_tag methods in nexus(4) and its children in order to handle DMA tags in a hierarchical way and get rid of the sparc64_root_dma_tag kludge. Together with the previous two items this changes also allows to completely get rid of the nexus(4) IVAR interface. It also includes: - pushing the constraints previously specified by the nexus_dmatag down into the DMA tags of psycho(4) and sbus(4) as it's their IOMMUs which induce these restrictions (and nothing at the nexus(4) or anything that would warrant specifying them there), - fixing some obviously wrong constraints of the psycho(4) and sbus(4) DMA tags, which happened to not actually be used with the sparc64_root_dma_tag kludge in place and therefore didn't cause problems so far, - replacing magic constants for constraints with macros as far as it is obvious as to where they come from. This doesn't include taking advantage of the newbus way to get the parent DMA tags implemented by this change in order to divorce the IOTSBs of the PCI and SBus IOMMUs or for implementing the workaround for the DMA sync bug in Sabre (and Tomatillo) bridges, yet, though. o Get rid of the notion that nexus(4) (mostly) reflects an UPA bus by replacing ofw_upa.h and with ofw_nexus.h (which was repo-copied from ofw_upa.h) and renaming its content, which actually applies to all of Fireplane/Safari, JBus and UPA (in the host bus case), as appropriate. o Just use M_DEVBUF instead of a separate M_NEXUS malloc type for allocating the device info for the children of nexus(4). This is done in order to not need to export M_NEXUS when deriving drivers for subordinate busses from the nexus(4) class. o Use the DEFINE_CLASS_0() macro to declare the nexus(4) driver so we can derive subclasses from it. o Const'ify the nexus_excl_name and nexus_excl_type arrays as well as add 'associations' and 'rsc', which are pseudo-devices without resources and therefore of no real interest for nexus(4), to the former. o Let the nexus(4) device memory rman manage the entire 64-bit address space instead of just the UPA_MEMSTART to UPA_MEMEND subregion as Fireplane/Safari- and JBus-based machines use multiple ranges, which can't be as easily divided as in the case of UPA (limiting the address space only served for sanity checking anyway). o Use M_WAITOK instead of M_NOWAIT when allocating the device info for children of nexus(4) in order to give one less opportunity for adding devices to nexus(4) to fail. o While adapting the drivers affected by the above nexus(4) changes, change them to take advantage of rman_get_rid() instead of caching the RIDs assigned to allocated resources, now that the RIDs of resources are correctly set. o In iommu(4) and nexus(4) replace hard-coded functions names, which actually became outdated in several places, in panic strings and status massages with __func__. [1] o Use driver_filter_t in prototypes where appropriate. o Add my copyright to creator(4), fhc(4), nexus(4), psycho(4) and sbus(4) as I changed considerable amounts of these drivers as well as added a bunch of new features, workarounds for silicon bugs etc. o Fix some white space nits. Due to lack of access to Exx00 hardware, these changes, i.e. central(4) and fhc(4), couldn't be runtime tested on such a machine. Exx00 are currently reported to panic before trying to attach nexus(4) anyway though. PR: 76052 [1] Approved by: re (kensmith)
2007-03-07 21:13:51 +00:00
node = ofw_bus_get_node(dev);
i = 0;
sc->sc_sysio_res = bus_alloc_resource_any(dev, SYS_RES_MEMORY, &i,
Rototill the sparc64 nexus(4) (actually this brings in the code the sun4v nexus(4) in turn is based on): o Change nexus(4) to manage the resources of its children so the respective device drivers don't need to figure them out of OFW themselves. o Change nexus(4) to provide the ofw_bus KOBJ interface instead of using IVARs for supplying the OFW node and the subset of standard properties of its children. Together with the previous change this also allows to fully take advantage of newbus in that drivers like fhc(4), which attach on multiple parent busses, no longer require different bus front-ends as obtaining the OFW node and properties as well as resource allocation works the same for all supported busses. As such this change also is part 4/4 of allowing creator(4) to work in USIII-based machines as it allows this driver to attach on both nexus(4) and upa(4). On the other hand removing these IVARs breaks API compatibility with the powerpc nexus(4) but which isn't that bad as a) sparc64 currently doesn't share any device driver hanging off of nexus(4) with powerpc and b) they were no longer compatible regarding OFW-related extensions at the pci(4) level since quite some time. o Provide bus_get_dma_tag methods in nexus(4) and its children in order to handle DMA tags in a hierarchical way and get rid of the sparc64_root_dma_tag kludge. Together with the previous two items this changes also allows to completely get rid of the nexus(4) IVAR interface. It also includes: - pushing the constraints previously specified by the nexus_dmatag down into the DMA tags of psycho(4) and sbus(4) as it's their IOMMUs which induce these restrictions (and nothing at the nexus(4) or anything that would warrant specifying them there), - fixing some obviously wrong constraints of the psycho(4) and sbus(4) DMA tags, which happened to not actually be used with the sparc64_root_dma_tag kludge in place and therefore didn't cause problems so far, - replacing magic constants for constraints with macros as far as it is obvious as to where they come from. This doesn't include taking advantage of the newbus way to get the parent DMA tags implemented by this change in order to divorce the IOTSBs of the PCI and SBus IOMMUs or for implementing the workaround for the DMA sync bug in Sabre (and Tomatillo) bridges, yet, though. o Get rid of the notion that nexus(4) (mostly) reflects an UPA bus by replacing ofw_upa.h and with ofw_nexus.h (which was repo-copied from ofw_upa.h) and renaming its content, which actually applies to all of Fireplane/Safari, JBus and UPA (in the host bus case), as appropriate. o Just use M_DEVBUF instead of a separate M_NEXUS malloc type for allocating the device info for the children of nexus(4). This is done in order to not need to export M_NEXUS when deriving drivers for subordinate busses from the nexus(4) class. o Use the DEFINE_CLASS_0() macro to declare the nexus(4) driver so we can derive subclasses from it. o Const'ify the nexus_excl_name and nexus_excl_type arrays as well as add 'associations' and 'rsc', which are pseudo-devices without resources and therefore of no real interest for nexus(4), to the former. o Let the nexus(4) device memory rman manage the entire 64-bit address space instead of just the UPA_MEMSTART to UPA_MEMEND subregion as Fireplane/Safari- and JBus-based machines use multiple ranges, which can't be as easily divided as in the case of UPA (limiting the address space only served for sanity checking anyway). o Use M_WAITOK instead of M_NOWAIT when allocating the device info for children of nexus(4) in order to give one less opportunity for adding devices to nexus(4) to fail. o While adapting the drivers affected by the above nexus(4) changes, change them to take advantage of rman_get_rid() instead of caching the RIDs assigned to allocated resources, now that the RIDs of resources are correctly set. o In iommu(4) and nexus(4) replace hard-coded functions names, which actually became outdated in several places, in panic strings and status massages with __func__. [1] o Use driver_filter_t in prototypes where appropriate. o Add my copyright to creator(4), fhc(4), nexus(4), psycho(4) and sbus(4) as I changed considerable amounts of these drivers as well as added a bunch of new features, workarounds for silicon bugs etc. o Fix some white space nits. Due to lack of access to Exx00 hardware, these changes, i.e. central(4) and fhc(4), couldn't be runtime tested on such a machine. Exx00 are currently reported to panic before trying to attach nexus(4) anyway though. PR: 76052 [1] Approved by: re (kensmith)
2007-03-07 21:13:51 +00:00
RF_ACTIVE);
if (sc->sc_sysio_res == NULL)
panic("%s: cannot allocate device memory", __func__);
if (OF_getprop(node, "interrupts", &prop, sizeof(prop)) == -1)
panic("%s: cannot get IGN", __func__);
sc->sc_ign = INTIGN(prop);
/*
* Record clock frequency for synchronous SCSI.
* IS THIS THE CORRECT DEFAULT??
*/
if (OF_getprop(node, "clock-frequency", &prop, sizeof(prop)) == -1)
prop = 25000000;
sc->sc_clockfreq = prop;
prop /= 1000;
device_printf(dev, "clock %d.%03d MHz\n", prop / 1000, prop % 1000);
/*
* Collect address translations from the OBP.
*/
if ((sc->sc_nrange = OF_getprop_alloc(node, "ranges",
sizeof(*range), (void **)&range)) == -1) {
panic("%s: error getting ranges property", __func__);
}
sc->sc_rd = malloc(sizeof(*sc->sc_rd) * sc->sc_nrange, M_DEVBUF,
M_NOWAIT | M_ZERO);
if (sc->sc_rd == NULL)
panic("%s: cannot allocate rmans", __func__);
/*
* Preallocate all space that the SBus bridge decodes, so that nothing
* else gets in the way; set up rmans etc.
*/
Rototill the sparc64 nexus(4) (actually this brings in the code the sun4v nexus(4) in turn is based on): o Change nexus(4) to manage the resources of its children so the respective device drivers don't need to figure them out of OFW themselves. o Change nexus(4) to provide the ofw_bus KOBJ interface instead of using IVARs for supplying the OFW node and the subset of standard properties of its children. Together with the previous change this also allows to fully take advantage of newbus in that drivers like fhc(4), which attach on multiple parent busses, no longer require different bus front-ends as obtaining the OFW node and properties as well as resource allocation works the same for all supported busses. As such this change also is part 4/4 of allowing creator(4) to work in USIII-based machines as it allows this driver to attach on both nexus(4) and upa(4). On the other hand removing these IVARs breaks API compatibility with the powerpc nexus(4) but which isn't that bad as a) sparc64 currently doesn't share any device driver hanging off of nexus(4) with powerpc and b) they were no longer compatible regarding OFW-related extensions at the pci(4) level since quite some time. o Provide bus_get_dma_tag methods in nexus(4) and its children in order to handle DMA tags in a hierarchical way and get rid of the sparc64_root_dma_tag kludge. Together with the previous two items this changes also allows to completely get rid of the nexus(4) IVAR interface. It also includes: - pushing the constraints previously specified by the nexus_dmatag down into the DMA tags of psycho(4) and sbus(4) as it's their IOMMUs which induce these restrictions (and nothing at the nexus(4) or anything that would warrant specifying them there), - fixing some obviously wrong constraints of the psycho(4) and sbus(4) DMA tags, which happened to not actually be used with the sparc64_root_dma_tag kludge in place and therefore didn't cause problems so far, - replacing magic constants for constraints with macros as far as it is obvious as to where they come from. This doesn't include taking advantage of the newbus way to get the parent DMA tags implemented by this change in order to divorce the IOTSBs of the PCI and SBus IOMMUs or for implementing the workaround for the DMA sync bug in Sabre (and Tomatillo) bridges, yet, though. o Get rid of the notion that nexus(4) (mostly) reflects an UPA bus by replacing ofw_upa.h and with ofw_nexus.h (which was repo-copied from ofw_upa.h) and renaming its content, which actually applies to all of Fireplane/Safari, JBus and UPA (in the host bus case), as appropriate. o Just use M_DEVBUF instead of a separate M_NEXUS malloc type for allocating the device info for the children of nexus(4). This is done in order to not need to export M_NEXUS when deriving drivers for subordinate busses from the nexus(4) class. o Use the DEFINE_CLASS_0() macro to declare the nexus(4) driver so we can derive subclasses from it. o Const'ify the nexus_excl_name and nexus_excl_type arrays as well as add 'associations' and 'rsc', which are pseudo-devices without resources and therefore of no real interest for nexus(4), to the former. o Let the nexus(4) device memory rman manage the entire 64-bit address space instead of just the UPA_MEMSTART to UPA_MEMEND subregion as Fireplane/Safari- and JBus-based machines use multiple ranges, which can't be as easily divided as in the case of UPA (limiting the address space only served for sanity checking anyway). o Use M_WAITOK instead of M_NOWAIT when allocating the device info for children of nexus(4) in order to give one less opportunity for adding devices to nexus(4) to fail. o While adapting the drivers affected by the above nexus(4) changes, change them to take advantage of rman_get_rid() instead of caching the RIDs assigned to allocated resources, now that the RIDs of resources are correctly set. o In iommu(4) and nexus(4) replace hard-coded functions names, which actually became outdated in several places, in panic strings and status massages with __func__. [1] o Use driver_filter_t in prototypes where appropriate. o Add my copyright to creator(4), fhc(4), nexus(4), psycho(4) and sbus(4) as I changed considerable amounts of these drivers as well as added a bunch of new features, workarounds for silicon bugs etc. o Fix some white space nits. Due to lack of access to Exx00 hardware, these changes, i.e. central(4) and fhc(4), couldn't be runtime tested on such a machine. Exx00 are currently reported to panic before trying to attach nexus(4) anyway though. PR: 76052 [1] Approved by: re (kensmith)
2007-03-07 21:13:51 +00:00
rl = BUS_GET_RESOURCE_LIST(device_get_parent(dev), dev);
for (i = 0; i < sc->sc_nrange; i++) {
phys = range[i].poffset | ((bus_addr_t)range[i].pspace << 32);
size = range[i].size;
sc->sc_rd[i].rd_slot = range[i].cspace;
sc->sc_rd[i].rd_coffset = range[i].coffset;
sc->sc_rd[i].rd_cend = sc->sc_rd[i].rd_coffset + size;
j = resource_list_add_next(rl, SYS_RES_MEMORY, phys,
Rototill the sparc64 nexus(4) (actually this brings in the code the sun4v nexus(4) in turn is based on): o Change nexus(4) to manage the resources of its children so the respective device drivers don't need to figure them out of OFW themselves. o Change nexus(4) to provide the ofw_bus KOBJ interface instead of using IVARs for supplying the OFW node and the subset of standard properties of its children. Together with the previous change this also allows to fully take advantage of newbus in that drivers like fhc(4), which attach on multiple parent busses, no longer require different bus front-ends as obtaining the OFW node and properties as well as resource allocation works the same for all supported busses. As such this change also is part 4/4 of allowing creator(4) to work in USIII-based machines as it allows this driver to attach on both nexus(4) and upa(4). On the other hand removing these IVARs breaks API compatibility with the powerpc nexus(4) but which isn't that bad as a) sparc64 currently doesn't share any device driver hanging off of nexus(4) with powerpc and b) they were no longer compatible regarding OFW-related extensions at the pci(4) level since quite some time. o Provide bus_get_dma_tag methods in nexus(4) and its children in order to handle DMA tags in a hierarchical way and get rid of the sparc64_root_dma_tag kludge. Together with the previous two items this changes also allows to completely get rid of the nexus(4) IVAR interface. It also includes: - pushing the constraints previously specified by the nexus_dmatag down into the DMA tags of psycho(4) and sbus(4) as it's their IOMMUs which induce these restrictions (and nothing at the nexus(4) or anything that would warrant specifying them there), - fixing some obviously wrong constraints of the psycho(4) and sbus(4) DMA tags, which happened to not actually be used with the sparc64_root_dma_tag kludge in place and therefore didn't cause problems so far, - replacing magic constants for constraints with macros as far as it is obvious as to where they come from. This doesn't include taking advantage of the newbus way to get the parent DMA tags implemented by this change in order to divorce the IOTSBs of the PCI and SBus IOMMUs or for implementing the workaround for the DMA sync bug in Sabre (and Tomatillo) bridges, yet, though. o Get rid of the notion that nexus(4) (mostly) reflects an UPA bus by replacing ofw_upa.h and with ofw_nexus.h (which was repo-copied from ofw_upa.h) and renaming its content, which actually applies to all of Fireplane/Safari, JBus and UPA (in the host bus case), as appropriate. o Just use M_DEVBUF instead of a separate M_NEXUS malloc type for allocating the device info for the children of nexus(4). This is done in order to not need to export M_NEXUS when deriving drivers for subordinate busses from the nexus(4) class. o Use the DEFINE_CLASS_0() macro to declare the nexus(4) driver so we can derive subclasses from it. o Const'ify the nexus_excl_name and nexus_excl_type arrays as well as add 'associations' and 'rsc', which are pseudo-devices without resources and therefore of no real interest for nexus(4), to the former. o Let the nexus(4) device memory rman manage the entire 64-bit address space instead of just the UPA_MEMSTART to UPA_MEMEND subregion as Fireplane/Safari- and JBus-based machines use multiple ranges, which can't be as easily divided as in the case of UPA (limiting the address space only served for sanity checking anyway). o Use M_WAITOK instead of M_NOWAIT when allocating the device info for children of nexus(4) in order to give one less opportunity for adding devices to nexus(4) to fail. o While adapting the drivers affected by the above nexus(4) changes, change them to take advantage of rman_get_rid() instead of caching the RIDs assigned to allocated resources, now that the RIDs of resources are correctly set. o In iommu(4) and nexus(4) replace hard-coded functions names, which actually became outdated in several places, in panic strings and status massages with __func__. [1] o Use driver_filter_t in prototypes where appropriate. o Add my copyright to creator(4), fhc(4), nexus(4), psycho(4) and sbus(4) as I changed considerable amounts of these drivers as well as added a bunch of new features, workarounds for silicon bugs etc. o Fix some white space nits. Due to lack of access to Exx00 hardware, these changes, i.e. central(4) and fhc(4), couldn't be runtime tested on such a machine. Exx00 are currently reported to panic before trying to attach nexus(4) anyway though. PR: 76052 [1] Approved by: re (kensmith)
2007-03-07 21:13:51 +00:00
phys + size - 1, size);
if ((res = bus_alloc_resource_any(dev, SYS_RES_MEMORY, &j,
Rototill the sparc64 nexus(4) (actually this brings in the code the sun4v nexus(4) in turn is based on): o Change nexus(4) to manage the resources of its children so the respective device drivers don't need to figure them out of OFW themselves. o Change nexus(4) to provide the ofw_bus KOBJ interface instead of using IVARs for supplying the OFW node and the subset of standard properties of its children. Together with the previous change this also allows to fully take advantage of newbus in that drivers like fhc(4), which attach on multiple parent busses, no longer require different bus front-ends as obtaining the OFW node and properties as well as resource allocation works the same for all supported busses. As such this change also is part 4/4 of allowing creator(4) to work in USIII-based machines as it allows this driver to attach on both nexus(4) and upa(4). On the other hand removing these IVARs breaks API compatibility with the powerpc nexus(4) but which isn't that bad as a) sparc64 currently doesn't share any device driver hanging off of nexus(4) with powerpc and b) they were no longer compatible regarding OFW-related extensions at the pci(4) level since quite some time. o Provide bus_get_dma_tag methods in nexus(4) and its children in order to handle DMA tags in a hierarchical way and get rid of the sparc64_root_dma_tag kludge. Together with the previous two items this changes also allows to completely get rid of the nexus(4) IVAR interface. It also includes: - pushing the constraints previously specified by the nexus_dmatag down into the DMA tags of psycho(4) and sbus(4) as it's their IOMMUs which induce these restrictions (and nothing at the nexus(4) or anything that would warrant specifying them there), - fixing some obviously wrong constraints of the psycho(4) and sbus(4) DMA tags, which happened to not actually be used with the sparc64_root_dma_tag kludge in place and therefore didn't cause problems so far, - replacing magic constants for constraints with macros as far as it is obvious as to where they come from. This doesn't include taking advantage of the newbus way to get the parent DMA tags implemented by this change in order to divorce the IOTSBs of the PCI and SBus IOMMUs or for implementing the workaround for the DMA sync bug in Sabre (and Tomatillo) bridges, yet, though. o Get rid of the notion that nexus(4) (mostly) reflects an UPA bus by replacing ofw_upa.h and with ofw_nexus.h (which was repo-copied from ofw_upa.h) and renaming its content, which actually applies to all of Fireplane/Safari, JBus and UPA (in the host bus case), as appropriate. o Just use M_DEVBUF instead of a separate M_NEXUS malloc type for allocating the device info for the children of nexus(4). This is done in order to not need to export M_NEXUS when deriving drivers for subordinate busses from the nexus(4) class. o Use the DEFINE_CLASS_0() macro to declare the nexus(4) driver so we can derive subclasses from it. o Const'ify the nexus_excl_name and nexus_excl_type arrays as well as add 'associations' and 'rsc', which are pseudo-devices without resources and therefore of no real interest for nexus(4), to the former. o Let the nexus(4) device memory rman manage the entire 64-bit address space instead of just the UPA_MEMSTART to UPA_MEMEND subregion as Fireplane/Safari- and JBus-based machines use multiple ranges, which can't be as easily divided as in the case of UPA (limiting the address space only served for sanity checking anyway). o Use M_WAITOK instead of M_NOWAIT when allocating the device info for children of nexus(4) in order to give one less opportunity for adding devices to nexus(4) to fail. o While adapting the drivers affected by the above nexus(4) changes, change them to take advantage of rman_get_rid() instead of caching the RIDs assigned to allocated resources, now that the RIDs of resources are correctly set. o In iommu(4) and nexus(4) replace hard-coded functions names, which actually became outdated in several places, in panic strings and status massages with __func__. [1] o Use driver_filter_t in prototypes where appropriate. o Add my copyright to creator(4), fhc(4), nexus(4), psycho(4) and sbus(4) as I changed considerable amounts of these drivers as well as added a bunch of new features, workarounds for silicon bugs etc. o Fix some white space nits. Due to lack of access to Exx00 hardware, these changes, i.e. central(4) and fhc(4), couldn't be runtime tested on such a machine. Exx00 are currently reported to panic before trying to attach nexus(4) anyway though. PR: 76052 [1] Approved by: re (kensmith)
2007-03-07 21:13:51 +00:00
RF_ACTIVE)) == NULL)
panic("%s: cannot allocate decoded range", __func__);
sc->sc_rd[i].rd_bushandle = rman_get_bushandle(res);
sc->sc_rd[i].rd_rman.rm_type = RMAN_ARRAY;
sc->sc_rd[i].rd_rman.rm_descr = "SBus Device Memory";
if (rman_init(&sc->sc_rd[i].rd_rman) != 0 ||
rman_manage_region(&sc->sc_rd[i].rd_rman, 0, size) != 0)
panic("%s: failed to set up memory rman", __func__);
sc->sc_rd[i].rd_poffset = phys;
sc->sc_rd[i].rd_pend = phys + size;
sc->sc_rd[i].rd_res = res;
}
free(range, M_OFWPROP);
/*
* Get the SBus burst transfer size if burst transfers are supported.
*/
if (OF_getprop(node, "up-burst-sizes", &sc->sc_burst,
sizeof(sc->sc_burst)) == -1 || sc->sc_burst == 0)
sc->sc_burst =
(SBUS_BURST64_DEF << SBUS_BURST64_SHIFT) | SBUS_BURST_DEF;
/* initalise the IOMMU */
/* punch in our copies */
- Divorce the IOTSBs, which so far where handled via a global list instead of per IOMMU, so we no longer need to program all of them identically in systems having multiple IOMMUs. This continues the rototilling of the nexus(4) done about 5 months ago, which amongst others changed nexus(4) and the drivers for host-to-foo bridges to provide bus_get_dma_tag methods, allowing to handle DMA tags in a hierarchical way and to link them with devices. This still doesn't move the silicon bug workarounds for Sabre (and in the uncommitted schizo(4) for Tomatillo) bridges into special bus_dma_tag_create() and bus_dmamap_sync() methods though, as w/o fully newbus'ified bus_dma_tag_create() and bus_dma_tag_destroy() this still requires too much hackery, i.e. per-child parent DMA tags in the parent driver. - Let the host-to-foo drivers supply the maximum physical address of the IOMMU accompanying the bridges. Previously iommu(4) hard- coded an upper limit of 16GB, which actually only applies to the IOMMUs of the Hummingbird and Sabre bridges. The Psycho variants as well as the U2S in fact can can translate to up to 2TB, i.e. translate to 41-bit physical addresses. According to the recently available Tomatillo documentation these bridges even translate to 43-bit physical addresses and hints at the Schizo bridges doing 43 bits as well. This fixes the issue the FreeBSD 6.0 todo list item "Max RAM on sparc64" was refering to and pretty much obsoletes the lack of support for bounce buffers on sparc64. Thanks to Nathan Whitehorn for pointing me at the Tomatillo manual. Approved by: re (kensmith)
2007-08-05 11:56:44 +00:00
sc->sc_is.is_pmaxaddr = IOMMU_MAXADDR(SBUS_IOMMU_BITS);
sc->sc_is.is_bustag = rman_get_bustag(sc->sc_sysio_res);
sc->sc_is.is_bushandle = rman_get_bushandle(sc->sc_sysio_res);
sc->sc_is.is_iommu = SBR_IOMMU;
sc->sc_is.is_dtag = SBR_IOMMU_TLB_TAG_DIAG;
sc->sc_is.is_ddram = SBR_IOMMU_TLB_DATA_DIAG;
sc->sc_is.is_dqueue = SBR_IOMMU_QUEUE_DIAG;
sc->sc_is.is_dva = SBR_IOMMU_SVADIAG;
sc->sc_is.is_dtcmp = 0;
sc->sc_is.is_sb[0] = SBR_STRBUF;
2003-12-26 14:22:26 +00:00
sc->sc_is.is_sb[1] = 0;
/*
* Note: the SBus IOMMU ignores the high bits of an address, so a NULL
* DMA pointer will be translated by the first page of the IOTSB.
* To detect bugs we'll allocate and ignore the first entry.
*/
iommu_init(device_get_nameunit(dev), &sc->sc_is, 3, -1, 1);
/* Create the DMA tag. */
- Divorce the IOTSBs, which so far where handled via a global list instead of per IOMMU, so we no longer need to program all of them identically in systems having multiple IOMMUs. This continues the rototilling of the nexus(4) done about 5 months ago, which amongst others changed nexus(4) and the drivers for host-to-foo bridges to provide bus_get_dma_tag methods, allowing to handle DMA tags in a hierarchical way and to link them with devices. This still doesn't move the silicon bug workarounds for Sabre (and in the uncommitted schizo(4) for Tomatillo) bridges into special bus_dma_tag_create() and bus_dmamap_sync() methods though, as w/o fully newbus'ified bus_dma_tag_create() and bus_dma_tag_destroy() this still requires too much hackery, i.e. per-child parent DMA tags in the parent driver. - Let the host-to-foo drivers supply the maximum physical address of the IOMMU accompanying the bridges. Previously iommu(4) hard- coded an upper limit of 16GB, which actually only applies to the IOMMUs of the Hummingbird and Sabre bridges. The Psycho variants as well as the U2S in fact can can translate to up to 2TB, i.e. translate to 41-bit physical addresses. According to the recently available Tomatillo documentation these bridges even translate to 43-bit physical addresses and hints at the Schizo bridges doing 43 bits as well. This fixes the issue the FreeBSD 6.0 todo list item "Max RAM on sparc64" was refering to and pretty much obsoletes the lack of support for bounce buffers on sparc64. Thanks to Nathan Whitehorn for pointing me at the Tomatillo manual. Approved by: re (kensmith)
2007-08-05 11:56:44 +00:00
if (bus_dma_tag_create(bus_get_dma_tag(dev), 8, 0,
sc->sc_is.is_pmaxaddr, ~0, NULL, NULL, sc->sc_is.is_pmaxaddr,
0xff, 0xffffffff, 0, NULL, NULL, &sc->sc_cdmatag) != 0)
panic("%s: bus_dma_tag_create failed", __func__);
/* Customize the tag. */
sc->sc_cdmatag->dt_cookie = &sc->sc_is;
sc->sc_cdmatag->dt_mt = &iommu_dma_methods;
Make sparc64 compatible with NEW_PCIB and enable it: - Implement bus_adjust_resource() methods as far as necessary and in non-PCI bridge drivers as far as feasible without rototilling them. - As NEW_PCIB does a layering violation by activating resources at layers above pci(4) without previously bubbling up their allocation there, move the assignment of bus tags and handles from the bus_alloc_resource() to the bus_activate_resource() methods like at least the other NEW_PCIB enabled architectures do. This is somewhat unfortunate as previously sparc64 (ab)used resource activation to indicate whether SYS_RES_MEMORY resources should be mapped into KVA, which is only necessary if their going to be accessed via the pointer returned from rman_get_virtual() but not for bus_space(9) as the later always uses physical access on sparc64. Besides wasting KVA if we always map in SYS_RES_MEMORY resources, a driver also may deliberately not map them in if the firmware already has done so, possibly in a special way. So in order to still allow a driver to decide whether a SYS_RES_MEMORY resource should be mapped into KVA we let it indicate that by calling bus_space_map(9) with BUS_SPACE_MAP_LINEAR as actually documented in the bus_space(9) page. This is implemented by allocating a separate bus tag per SYS_RES_MEMORY resource and passing the resource via the previously unused bus tag cookie so we later on can call rman_set_virtual() in sparc64_bus_mem_map(). As a side effect this now also allows to actually indicate that a SYS_RES_MEMORY resource should be mapped in as cacheable and/or read-only via BUS_SPACE_MAP_CACHEABLE and BUS_SPACE_MAP_READONLY respectively. - Do some minor cleanup like taking advantage of rman_init_from_resource(), factor out the common part of bus tag allocation into a newly added sparc64_alloc_bus_tag(), hook up some missing newbus methods and replace some homegrown versions with the generic counterparts etc. - While at it, let apb_attach() (which can't use the generic NEW_PCIB code as APB bridges just don't have the base and limit registers implemented) regarding the config space registers cached in pcib_softc and the SYSCTL reporting nodes set up.
2011-10-02 23:22:38 +00:00
/*
o Revamp the sparc64 interrupt code in order to be able to interface with the INTR_FILTER-enabled MI code. Basically this consists of registering an interrupt controller (of which there can be multiple and optionally different ones either per host-to-foo bridge or shared amongst host-to-foo bridges in any one machine) along with an interrupt vector as specific argument for all the interrupt vectors used by a given host-to-foo bridge (roughly similar to registering interrupt sources on amd64 and i386), providing functions to enable, clear and disable the interrupts of the children beneath the bridge. This also includes: - No longer entering a critical section in tl0_intr() and tl1_intr() for executing interrupt handlers but rather let the handlers enter it themselves so in the case of intr_event_handle() we don't enter a nested critical section. - Adding infrastructure for binding delivery of interrupt vectors to specific CPUs which later on can be interfaced with the code from amd64/i386 for binding interrupts to specific CPUs. - Getting rid of the wrapper hack introduced along the lines of the API changes for INTR_FILTER which as a side-effect caused interrupts associated with ithread handlers only to get the elevated priority of those associated with filters ("fast handlers") (this removes the hack also in the non-INTR_FILTER case). - Disabling (by not clearing) an interrupt in the interrupt controller until all associated handlers have been executed, which is crucial for the typical locking strategy of NIC drivers in order to work correctly in case of shared interrupts. This was a more or less theoretical problem on sparc64 though, as shared interrupts are rather uncommon there except for the on-board SCCs and UARTs. Note that due to the behavior of at least of some of the interrupt controllers used on sparc64 an enable+EOI instead of a disable+EOI approach (as implied by the INTR_FILTER MI code and implemented on other architectures) is used as the latter can cause lost interrupts or in the worst case interrupt starvation. o Correct a typo in sbus_alloc_resource() which caused (pass-through) allocations to only work down to the grandchildren of the bus, which wasn't a real problem so far as we don't support any devices which are great-grandchildren or greater of a U2S bridge, yet. o In fhc(4) use bus_{read,write}_4() instead of bus_space_{read,write}_4() in order to get rid of sc_bh and sc_bt in the fhc_softc. Also get rid of some other unneeded members in fhc_softc. Reviewed by: marcel (earlier version) Approved by: re (kensmith)
2007-09-06 19:16:30 +00:00
* Hunt through all the interrupt mapping regs and register our
* interrupt controller for the corresponding interrupt vectors.
* We do this early in order to be able to catch stray interrupts.
o Revamp the sparc64 interrupt code in order to be able to interface with the INTR_FILTER-enabled MI code. Basically this consists of registering an interrupt controller (of which there can be multiple and optionally different ones either per host-to-foo bridge or shared amongst host-to-foo bridges in any one machine) along with an interrupt vector as specific argument for all the interrupt vectors used by a given host-to-foo bridge (roughly similar to registering interrupt sources on amd64 and i386), providing functions to enable, clear and disable the interrupts of the children beneath the bridge. This also includes: - No longer entering a critical section in tl0_intr() and tl1_intr() for executing interrupt handlers but rather let the handlers enter it themselves so in the case of intr_event_handle() we don't enter a nested critical section. - Adding infrastructure for binding delivery of interrupt vectors to specific CPUs which later on can be interfaced with the code from amd64/i386 for binding interrupts to specific CPUs. - Getting rid of the wrapper hack introduced along the lines of the API changes for INTR_FILTER which as a side-effect caused interrupts associated with ithread handlers only to get the elevated priority of those associated with filters ("fast handlers") (this removes the hack also in the non-INTR_FILTER case). - Disabling (by not clearing) an interrupt in the interrupt controller until all associated handlers have been executed, which is crucial for the typical locking strategy of NIC drivers in order to work correctly in case of shared interrupts. This was a more or less theoretical problem on sparc64 though, as shared interrupts are rather uncommon there except for the on-board SCCs and UARTs. Note that due to the behavior of at least of some of the interrupt controllers used on sparc64 an enable+EOI instead of a disable+EOI approach (as implied by the INTR_FILTER MI code and implemented on other architectures) is used as the latter can cause lost interrupts or in the worst case interrupt starvation. o Correct a typo in sbus_alloc_resource() which caused (pass-through) allocations to only work down to the grandchildren of the bus, which wasn't a real problem so far as we don't support any devices which are great-grandchildren or greater of a U2S bridge, yet. o In fhc(4) use bus_{read,write}_4() instead of bus_space_{read,write}_4() in order to get rid of sc_bh and sc_bt in the fhc_softc. Also get rid of some other unneeded members in fhc_softc. Reviewed by: marcel (earlier version) Approved by: re (kensmith)
2007-09-06 19:16:30 +00:00
*/
for (i = 0; i <= SBUS_MAX_INO; i++) {
if (sbus_find_intrmap(sc, i, &intrmap, &intrclr) == 0)
continue;
sica = malloc(sizeof(*sica), M_DEVBUF, M_NOWAIT);
if (sica == NULL)
panic("%s: could not allocate interrupt controller "
"argument", __func__);
sica->sica_sc = sc;
sica->sica_map = intrmap;
sica->sica_clr = intrclr;
#ifdef SBUS_DEBUG
device_printf(dev,
"intr map (INO %d, %s) %#lx: %#lx, clr: %#lx\n",
i, (i & INTMAP_OBIO_MASK) == 0 ? "SBus slot" : "OBIO",
(u_long)intrmap, (u_long)SYSIO_READ8(sc, intrmap),
(u_long)intrclr);
#endif
j = intr_controller_register(INTMAP_VEC(sc->sc_ign, i),
&sbus_ic, sica);
if (j != 0)
device_printf(dev, "could not register interrupt "
"controller for INO %d (%d)\n", i, j);
o Revamp the sparc64 interrupt code in order to be able to interface with the INTR_FILTER-enabled MI code. Basically this consists of registering an interrupt controller (of which there can be multiple and optionally different ones either per host-to-foo bridge or shared amongst host-to-foo bridges in any one machine) along with an interrupt vector as specific argument for all the interrupt vectors used by a given host-to-foo bridge (roughly similar to registering interrupt sources on amd64 and i386), providing functions to enable, clear and disable the interrupts of the children beneath the bridge. This also includes: - No longer entering a critical section in tl0_intr() and tl1_intr() for executing interrupt handlers but rather let the handlers enter it themselves so in the case of intr_event_handle() we don't enter a nested critical section. - Adding infrastructure for binding delivery of interrupt vectors to specific CPUs which later on can be interfaced with the code from amd64/i386 for binding interrupts to specific CPUs. - Getting rid of the wrapper hack introduced along the lines of the API changes for INTR_FILTER which as a side-effect caused interrupts associated with ithread handlers only to get the elevated priority of those associated with filters ("fast handlers") (this removes the hack also in the non-INTR_FILTER case). - Disabling (by not clearing) an interrupt in the interrupt controller until all associated handlers have been executed, which is crucial for the typical locking strategy of NIC drivers in order to work correctly in case of shared interrupts. This was a more or less theoretical problem on sparc64 though, as shared interrupts are rather uncommon there except for the on-board SCCs and UARTs. Note that due to the behavior of at least of some of the interrupt controllers used on sparc64 an enable+EOI instead of a disable+EOI approach (as implied by the INTR_FILTER MI code and implemented on other architectures) is used as the latter can cause lost interrupts or in the worst case interrupt starvation. o Correct a typo in sbus_alloc_resource() which caused (pass-through) allocations to only work down to the grandchildren of the bus, which wasn't a real problem so far as we don't support any devices which are great-grandchildren or greater of a U2S bridge, yet. o In fhc(4) use bus_{read,write}_4() instead of bus_space_{read,write}_4() in order to get rid of sc_bh and sc_bt in the fhc_softc. Also get rid of some other unneeded members in fhc_softc. Reviewed by: marcel (earlier version) Approved by: re (kensmith)
2007-09-06 19:16:30 +00:00
}
/* Enable the over-temperature and power-fail interrupts. */
i = 4;
sc->sc_ot_ires = bus_alloc_resource_any(dev, SYS_RES_IRQ, &i,
Rototill the sparc64 nexus(4) (actually this brings in the code the sun4v nexus(4) in turn is based on): o Change nexus(4) to manage the resources of its children so the respective device drivers don't need to figure them out of OFW themselves. o Change nexus(4) to provide the ofw_bus KOBJ interface instead of using IVARs for supplying the OFW node and the subset of standard properties of its children. Together with the previous change this also allows to fully take advantage of newbus in that drivers like fhc(4), which attach on multiple parent busses, no longer require different bus front-ends as obtaining the OFW node and properties as well as resource allocation works the same for all supported busses. As such this change also is part 4/4 of allowing creator(4) to work in USIII-based machines as it allows this driver to attach on both nexus(4) and upa(4). On the other hand removing these IVARs breaks API compatibility with the powerpc nexus(4) but which isn't that bad as a) sparc64 currently doesn't share any device driver hanging off of nexus(4) with powerpc and b) they were no longer compatible regarding OFW-related extensions at the pci(4) level since quite some time. o Provide bus_get_dma_tag methods in nexus(4) and its children in order to handle DMA tags in a hierarchical way and get rid of the sparc64_root_dma_tag kludge. Together with the previous two items this changes also allows to completely get rid of the nexus(4) IVAR interface. It also includes: - pushing the constraints previously specified by the nexus_dmatag down into the DMA tags of psycho(4) and sbus(4) as it's their IOMMUs which induce these restrictions (and nothing at the nexus(4) or anything that would warrant specifying them there), - fixing some obviously wrong constraints of the psycho(4) and sbus(4) DMA tags, which happened to not actually be used with the sparc64_root_dma_tag kludge in place and therefore didn't cause problems so far, - replacing magic constants for constraints with macros as far as it is obvious as to where they come from. This doesn't include taking advantage of the newbus way to get the parent DMA tags implemented by this change in order to divorce the IOTSBs of the PCI and SBus IOMMUs or for implementing the workaround for the DMA sync bug in Sabre (and Tomatillo) bridges, yet, though. o Get rid of the notion that nexus(4) (mostly) reflects an UPA bus by replacing ofw_upa.h and with ofw_nexus.h (which was repo-copied from ofw_upa.h) and renaming its content, which actually applies to all of Fireplane/Safari, JBus and UPA (in the host bus case), as appropriate. o Just use M_DEVBUF instead of a separate M_NEXUS malloc type for allocating the device info for the children of nexus(4). This is done in order to not need to export M_NEXUS when deriving drivers for subordinate busses from the nexus(4) class. o Use the DEFINE_CLASS_0() macro to declare the nexus(4) driver so we can derive subclasses from it. o Const'ify the nexus_excl_name and nexus_excl_type arrays as well as add 'associations' and 'rsc', which are pseudo-devices without resources and therefore of no real interest for nexus(4), to the former. o Let the nexus(4) device memory rman manage the entire 64-bit address space instead of just the UPA_MEMSTART to UPA_MEMEND subregion as Fireplane/Safari- and JBus-based machines use multiple ranges, which can't be as easily divided as in the case of UPA (limiting the address space only served for sanity checking anyway). o Use M_WAITOK instead of M_NOWAIT when allocating the device info for children of nexus(4) in order to give one less opportunity for adding devices to nexus(4) to fail. o While adapting the drivers affected by the above nexus(4) changes, change them to take advantage of rman_get_rid() instead of caching the RIDs assigned to allocated resources, now that the RIDs of resources are correctly set. o In iommu(4) and nexus(4) replace hard-coded functions names, which actually became outdated in several places, in panic strings and status massages with __func__. [1] o Use driver_filter_t in prototypes where appropriate. o Add my copyright to creator(4), fhc(4), nexus(4), psycho(4) and sbus(4) as I changed considerable amounts of these drivers as well as added a bunch of new features, workarounds for silicon bugs etc. o Fix some white space nits. Due to lack of access to Exx00 hardware, these changes, i.e. central(4) and fhc(4), couldn't be runtime tested on such a machine. Exx00 are currently reported to panic before trying to attach nexus(4) anyway though. PR: 76052 [1] Approved by: re (kensmith)
2007-03-07 21:13:51 +00:00
RF_ACTIVE);
if (sc->sc_ot_ires == NULL ||
o Revamp the sparc64 interrupt code in order to be able to interface with the INTR_FILTER-enabled MI code. Basically this consists of registering an interrupt controller (of which there can be multiple and optionally different ones either per host-to-foo bridge or shared amongst host-to-foo bridges in any one machine) along with an interrupt vector as specific argument for all the interrupt vectors used by a given host-to-foo bridge (roughly similar to registering interrupt sources on amd64 and i386), providing functions to enable, clear and disable the interrupts of the children beneath the bridge. This also includes: - No longer entering a critical section in tl0_intr() and tl1_intr() for executing interrupt handlers but rather let the handlers enter it themselves so in the case of intr_event_handle() we don't enter a nested critical section. - Adding infrastructure for binding delivery of interrupt vectors to specific CPUs which later on can be interfaced with the code from amd64/i386 for binding interrupts to specific CPUs. - Getting rid of the wrapper hack introduced along the lines of the API changes for INTR_FILTER which as a side-effect caused interrupts associated with ithread handlers only to get the elevated priority of those associated with filters ("fast handlers") (this removes the hack also in the non-INTR_FILTER case). - Disabling (by not clearing) an interrupt in the interrupt controller until all associated handlers have been executed, which is crucial for the typical locking strategy of NIC drivers in order to work correctly in case of shared interrupts. This was a more or less theoretical problem on sparc64 though, as shared interrupts are rather uncommon there except for the on-board SCCs and UARTs. Note that due to the behavior of at least of some of the interrupt controllers used on sparc64 an enable+EOI instead of a disable+EOI approach (as implied by the INTR_FILTER MI code and implemented on other architectures) is used as the latter can cause lost interrupts or in the worst case interrupt starvation. o Correct a typo in sbus_alloc_resource() which caused (pass-through) allocations to only work down to the grandchildren of the bus, which wasn't a real problem so far as we don't support any devices which are great-grandchildren or greater of a U2S bridge, yet. o In fhc(4) use bus_{read,write}_4() instead of bus_space_{read,write}_4() in order to get rid of sc_bh and sc_bt in the fhc_softc. Also get rid of some other unneeded members in fhc_softc. Reviewed by: marcel (earlier version) Approved by: re (kensmith)
2007-09-06 19:16:30 +00:00
INTIGN(vec = rman_get_start(sc->sc_ot_ires)) != sc->sc_ign ||
INTVEC(SYSIO_READ8(sc, SBR_THERM_INT_MAP)) != vec ||
intr_vectors[vec].iv_ic != &sbus_ic ||
bus_setup_intr(dev, sc->sc_ot_ires, INTR_TYPE_MISC | INTR_BRIDGE,
NULL, sbus_overtemp, sc, &sc->sc_ot_ihand) != 0)
panic("%s: failed to set up temperature interrupt", __func__);
i = 3;
sc->sc_pf_ires = bus_alloc_resource_any(dev, SYS_RES_IRQ, &i,
Rototill the sparc64 nexus(4) (actually this brings in the code the sun4v nexus(4) in turn is based on): o Change nexus(4) to manage the resources of its children so the respective device drivers don't need to figure them out of OFW themselves. o Change nexus(4) to provide the ofw_bus KOBJ interface instead of using IVARs for supplying the OFW node and the subset of standard properties of its children. Together with the previous change this also allows to fully take advantage of newbus in that drivers like fhc(4), which attach on multiple parent busses, no longer require different bus front-ends as obtaining the OFW node and properties as well as resource allocation works the same for all supported busses. As such this change also is part 4/4 of allowing creator(4) to work in USIII-based machines as it allows this driver to attach on both nexus(4) and upa(4). On the other hand removing these IVARs breaks API compatibility with the powerpc nexus(4) but which isn't that bad as a) sparc64 currently doesn't share any device driver hanging off of nexus(4) with powerpc and b) they were no longer compatible regarding OFW-related extensions at the pci(4) level since quite some time. o Provide bus_get_dma_tag methods in nexus(4) and its children in order to handle DMA tags in a hierarchical way and get rid of the sparc64_root_dma_tag kludge. Together with the previous two items this changes also allows to completely get rid of the nexus(4) IVAR interface. It also includes: - pushing the constraints previously specified by the nexus_dmatag down into the DMA tags of psycho(4) and sbus(4) as it's their IOMMUs which induce these restrictions (and nothing at the nexus(4) or anything that would warrant specifying them there), - fixing some obviously wrong constraints of the psycho(4) and sbus(4) DMA tags, which happened to not actually be used with the sparc64_root_dma_tag kludge in place and therefore didn't cause problems so far, - replacing magic constants for constraints with macros as far as it is obvious as to where they come from. This doesn't include taking advantage of the newbus way to get the parent DMA tags implemented by this change in order to divorce the IOTSBs of the PCI and SBus IOMMUs or for implementing the workaround for the DMA sync bug in Sabre (and Tomatillo) bridges, yet, though. o Get rid of the notion that nexus(4) (mostly) reflects an UPA bus by replacing ofw_upa.h and with ofw_nexus.h (which was repo-copied from ofw_upa.h) and renaming its content, which actually applies to all of Fireplane/Safari, JBus and UPA (in the host bus case), as appropriate. o Just use M_DEVBUF instead of a separate M_NEXUS malloc type for allocating the device info for the children of nexus(4). This is done in order to not need to export M_NEXUS when deriving drivers for subordinate busses from the nexus(4) class. o Use the DEFINE_CLASS_0() macro to declare the nexus(4) driver so we can derive subclasses from it. o Const'ify the nexus_excl_name and nexus_excl_type arrays as well as add 'associations' and 'rsc', which are pseudo-devices without resources and therefore of no real interest for nexus(4), to the former. o Let the nexus(4) device memory rman manage the entire 64-bit address space instead of just the UPA_MEMSTART to UPA_MEMEND subregion as Fireplane/Safari- and JBus-based machines use multiple ranges, which can't be as easily divided as in the case of UPA (limiting the address space only served for sanity checking anyway). o Use M_WAITOK instead of M_NOWAIT when allocating the device info for children of nexus(4) in order to give one less opportunity for adding devices to nexus(4) to fail. o While adapting the drivers affected by the above nexus(4) changes, change them to take advantage of rman_get_rid() instead of caching the RIDs assigned to allocated resources, now that the RIDs of resources are correctly set. o In iommu(4) and nexus(4) replace hard-coded functions names, which actually became outdated in several places, in panic strings and status massages with __func__. [1] o Use driver_filter_t in prototypes where appropriate. o Add my copyright to creator(4), fhc(4), nexus(4), psycho(4) and sbus(4) as I changed considerable amounts of these drivers as well as added a bunch of new features, workarounds for silicon bugs etc. o Fix some white space nits. Due to lack of access to Exx00 hardware, these changes, i.e. central(4) and fhc(4), couldn't be runtime tested on such a machine. Exx00 are currently reported to panic before trying to attach nexus(4) anyway though. PR: 76052 [1] Approved by: re (kensmith)
2007-03-07 21:13:51 +00:00
RF_ACTIVE);
if (sc->sc_pf_ires == NULL ||
o Revamp the sparc64 interrupt code in order to be able to interface with the INTR_FILTER-enabled MI code. Basically this consists of registering an interrupt controller (of which there can be multiple and optionally different ones either per host-to-foo bridge or shared amongst host-to-foo bridges in any one machine) along with an interrupt vector as specific argument for all the interrupt vectors used by a given host-to-foo bridge (roughly similar to registering interrupt sources on amd64 and i386), providing functions to enable, clear and disable the interrupts of the children beneath the bridge. This also includes: - No longer entering a critical section in tl0_intr() and tl1_intr() for executing interrupt handlers but rather let the handlers enter it themselves so in the case of intr_event_handle() we don't enter a nested critical section. - Adding infrastructure for binding delivery of interrupt vectors to specific CPUs which later on can be interfaced with the code from amd64/i386 for binding interrupts to specific CPUs. - Getting rid of the wrapper hack introduced along the lines of the API changes for INTR_FILTER which as a side-effect caused interrupts associated with ithread handlers only to get the elevated priority of those associated with filters ("fast handlers") (this removes the hack also in the non-INTR_FILTER case). - Disabling (by not clearing) an interrupt in the interrupt controller until all associated handlers have been executed, which is crucial for the typical locking strategy of NIC drivers in order to work correctly in case of shared interrupts. This was a more or less theoretical problem on sparc64 though, as shared interrupts are rather uncommon there except for the on-board SCCs and UARTs. Note that due to the behavior of at least of some of the interrupt controllers used on sparc64 an enable+EOI instead of a disable+EOI approach (as implied by the INTR_FILTER MI code and implemented on other architectures) is used as the latter can cause lost interrupts or in the worst case interrupt starvation. o Correct a typo in sbus_alloc_resource() which caused (pass-through) allocations to only work down to the grandchildren of the bus, which wasn't a real problem so far as we don't support any devices which are great-grandchildren or greater of a U2S bridge, yet. o In fhc(4) use bus_{read,write}_4() instead of bus_space_{read,write}_4() in order to get rid of sc_bh and sc_bt in the fhc_softc. Also get rid of some other unneeded members in fhc_softc. Reviewed by: marcel (earlier version) Approved by: re (kensmith)
2007-09-06 19:16:30 +00:00
INTIGN(vec = rman_get_start(sc->sc_pf_ires)) != sc->sc_ign ||
INTVEC(SYSIO_READ8(sc, SBR_POWER_INT_MAP)) != vec ||
intr_vectors[vec].iv_ic != &sbus_ic ||
bus_setup_intr(dev, sc->sc_pf_ires, INTR_TYPE_MISC | INTR_BRIDGE,
NULL, sbus_pwrfail, sc, &sc->sc_pf_ihand) != 0)
panic("%s: failed to set up power fail interrupt", __func__);
/* Initialize the counter-timer. */
sparc64_counter_init(device_get_nameunit(dev),
rman_get_bustag(sc->sc_sysio_res),
rman_get_bushandle(sc->sc_sysio_res), SBR_TC0);
/*
* Loop through ROM children, fixing any relative addresses
* and then configuring each device.
*/
for (child = OF_child(node); child != 0; child = OF_peer(child)) {
if ((sdi = sbus_setup_dinfo(dev, sc, child)) == NULL)
continue;
- Add an IVAR for retrieving the interrupt group number of the parent Sbus device and which also applies to the children. This is very usefull for drivers for the various subordinate busses so they don't need to fiddle with the OFW node of their parent themselves. As SBus busses hang of the nexus and we don't use the ofw_bus interface for nexus devices, yet, this would also require special knowledge about this in the drivers for the SBus children which these shouldn't need to have. This includes switching to use an unshifted IGN in the sc_ign member of the sbus(4) softc internally. - For SBus child devices where there are variants that are actually split split into two SBus devices (as opposed to the first half of the device being a SBus device and the second half hanging off of the first one) like 'auxio' and 'SUNW,fdtwo' or 'dma' and 'esp' probe the SBus device which is a prerequisite to the driver attaching to the second one with a lower order. This saves us from dealing with different probe orders in the respective device drivers which generally is more hackish. - Remove a stale comment about the 'specials' array above the attaching of the child devices. This is a remnant of the NetBSD/sparc origin of this code. There the 'specials' array is also used to probe certain devices which are prerequisites to others first. Why NetBSD soley relies on the devices having the expected order in the OFW tree on sparc64 isn't clear to me, as far as I can tell OFW doesn't guaranteed such things.
2005-05-19 14:47:31 +00:00
/*
* For devices where there are variants that are actually
* split into two SBus devices (as opposed to the first
* half of the device being a SBus device and the second
* half hanging off of the first one) like 'auxio' and
* 'SUNW,fdtwo' or 'dma' and 'esp' probe the SBus device
* which is a prerequisite to the driver attaching to the
* second one with a lower order. Saves us from dealing
* with different probe orders in the respective device
* drivers which generally is more hackish.
*/
cdev = device_add_child_ordered(dev, (OF_child(child) == 0 &&
sbus_inlist(sdi->sdi_obdinfo.obd_name, sbus_order_first)) ?
SBUS_ORDER_FIRST : SBUS_ORDER_NORMAL, NULL, -1);
if (cdev == NULL) {
device_printf(dev,
"<%s>: device_add_child_ordered failed\n",
sdi->sdi_obdinfo.obd_name);
sbus_destroy_dinfo(sdi);
continue;
}
device_set_ivars(cdev, sdi);
}
return (bus_generic_attach(dev));
}
static struct sbus_devinfo *
sbus_setup_dinfo(device_t dev, struct sbus_softc *sc, phandle_t node)
{
struct sbus_devinfo *sdi;
struct sbus_regs *reg;
u_int32_t base, iv, *intr;
int i, nreg, nintr, slot, rslot;
sdi = malloc(sizeof(*sdi), M_DEVBUF, M_ZERO | M_WAITOK);
if (ofw_bus_gen_setup_devinfo(&sdi->sdi_obdinfo, node) != 0) {
free(sdi, M_DEVBUF);
return (NULL);
}
resource_list_init(&sdi->sdi_rl);
slot = -1;
nreg = OF_getprop_alloc(node, "reg", sizeof(*reg), (void **)&reg);
if (nreg == -1) {
if (sdi->sdi_obdinfo.obd_type == NULL ||
strcmp(sdi->sdi_obdinfo.obd_type, "hierarchical") != 0) {
device_printf(dev, "<%s>: incomplete\n",
sdi->sdi_obdinfo.obd_name);
goto fail;
}
} else {
for (i = 0; i < nreg; i++) {
base = reg[i].sbr_offset;
if (SBUS_ABS(base)) {
rslot = SBUS_ABS_TO_SLOT(base);
base = SBUS_ABS_TO_OFFSET(base);
} else
rslot = reg[i].sbr_slot;
if (slot != -1 && slot != rslot) {
device_printf(dev, "<%s>: multiple slots\n",
sdi->sdi_obdinfo.obd_name);
free(reg, M_OFWPROP);
goto fail;
}
slot = rslot;
resource_list_add(&sdi->sdi_rl, SYS_RES_MEMORY, i,
base, base + reg[i].sbr_size, reg[i].sbr_size);
}
free(reg, M_OFWPROP);
}
sdi->sdi_slot = slot;
/*
* The `interrupts' property contains the SBus interrupt level.
*/
nintr = OF_getprop_alloc(node, "interrupts", sizeof(*intr),
(void **)&intr);
if (nintr != -1) {
for (i = 0; i < nintr; i++) {
iv = intr[i];
/*
* SBus card devices need the slot number encoded into
* the vector as this is generally not done.
*/
if ((iv & INTMAP_OBIO_MASK) == 0)
iv |= slot << 3;
o Revamp the sparc64 interrupt code in order to be able to interface with the INTR_FILTER-enabled MI code. Basically this consists of registering an interrupt controller (of which there can be multiple and optionally different ones either per host-to-foo bridge or shared amongst host-to-foo bridges in any one machine) along with an interrupt vector as specific argument for all the interrupt vectors used by a given host-to-foo bridge (roughly similar to registering interrupt sources on amd64 and i386), providing functions to enable, clear and disable the interrupts of the children beneath the bridge. This also includes: - No longer entering a critical section in tl0_intr() and tl1_intr() for executing interrupt handlers but rather let the handlers enter it themselves so in the case of intr_event_handle() we don't enter a nested critical section. - Adding infrastructure for binding delivery of interrupt vectors to specific CPUs which later on can be interfaced with the code from amd64/i386 for binding interrupts to specific CPUs. - Getting rid of the wrapper hack introduced along the lines of the API changes for INTR_FILTER which as a side-effect caused interrupts associated with ithread handlers only to get the elevated priority of those associated with filters ("fast handlers") (this removes the hack also in the non-INTR_FILTER case). - Disabling (by not clearing) an interrupt in the interrupt controller until all associated handlers have been executed, which is crucial for the typical locking strategy of NIC drivers in order to work correctly in case of shared interrupts. This was a more or less theoretical problem on sparc64 though, as shared interrupts are rather uncommon there except for the on-board SCCs and UARTs. Note that due to the behavior of at least of some of the interrupt controllers used on sparc64 an enable+EOI instead of a disable+EOI approach (as implied by the INTR_FILTER MI code and implemented on other architectures) is used as the latter can cause lost interrupts or in the worst case interrupt starvation. o Correct a typo in sbus_alloc_resource() which caused (pass-through) allocations to only work down to the grandchildren of the bus, which wasn't a real problem so far as we don't support any devices which are great-grandchildren or greater of a U2S bridge, yet. o In fhc(4) use bus_{read,write}_4() instead of bus_space_{read,write}_4() in order to get rid of sc_bh and sc_bt in the fhc_softc. Also get rid of some other unneeded members in fhc_softc. Reviewed by: marcel (earlier version) Approved by: re (kensmith)
2007-09-06 19:16:30 +00:00
iv = INTMAP_VEC(sc->sc_ign, iv);
resource_list_add(&sdi->sdi_rl, SYS_RES_IRQ, i,
iv, iv, 1);
}
free(intr, M_OFWPROP);
}
if (OF_getprop(node, "burst-sizes", &sdi->sdi_burstsz,
sizeof(sdi->sdi_burstsz)) == -1)
sdi->sdi_burstsz = sc->sc_burst;
else
sdi->sdi_burstsz &= sc->sc_burst;
if (OF_getprop(node, "clock-frequency", &sdi->sdi_clockfreq,
sizeof(sdi->sdi_clockfreq)) == -1)
sdi->sdi_clockfreq = sc->sc_clockfreq;
return (sdi);
fail:
sbus_destroy_dinfo(sdi);
return (NULL);
}
static void
sbus_destroy_dinfo(struct sbus_devinfo *dinfo)
{
resource_list_free(&dinfo->sdi_rl);
ofw_bus_gen_destroy_devinfo(&dinfo->sdi_obdinfo);
free(dinfo, M_DEVBUF);
}
static int
sbus_print_child(device_t dev, device_t child)
{
int rv;
rv = bus_print_child_header(dev, child);
rv += sbus_print_res(device_get_ivars(child));
rv += bus_print_child_footer(dev, child);
return (rv);
}
static void
sbus_probe_nomatch(device_t dev, device_t child)
{
const char *type;
device_printf(dev, "<%s>", ofw_bus_get_name(child));
sbus_print_res(device_get_ivars(child));
type = ofw_bus_get_type(child);
printf(" type %s (no driver attached)\n",
type != NULL ? type : "unknown");
}
static int
sbus_read_ivar(device_t dev, device_t child, int which, uintptr_t *result)
{
- Add an IVAR for retrieving the interrupt group number of the parent Sbus device and which also applies to the children. This is very usefull for drivers for the various subordinate busses so they don't need to fiddle with the OFW node of their parent themselves. As SBus busses hang of the nexus and we don't use the ofw_bus interface for nexus devices, yet, this would also require special knowledge about this in the drivers for the SBus children which these shouldn't need to have. This includes switching to use an unshifted IGN in the sc_ign member of the sbus(4) softc internally. - For SBus child devices where there are variants that are actually split split into two SBus devices (as opposed to the first half of the device being a SBus device and the second half hanging off of the first one) like 'auxio' and 'SUNW,fdtwo' or 'dma' and 'esp' probe the SBus device which is a prerequisite to the driver attaching to the second one with a lower order. This saves us from dealing with different probe orders in the respective device drivers which generally is more hackish. - Remove a stale comment about the 'specials' array above the attaching of the child devices. This is a remnant of the NetBSD/sparc origin of this code. There the 'specials' array is also used to probe certain devices which are prerequisites to others first. Why NetBSD soley relies on the devices having the expected order in the OFW tree on sparc64 isn't clear to me, as far as I can tell OFW doesn't guaranteed such things.
2005-05-19 14:47:31 +00:00
struct sbus_softc *sc;
struct sbus_devinfo *dinfo;
- Add an IVAR for retrieving the interrupt group number of the parent Sbus device and which also applies to the children. This is very usefull for drivers for the various subordinate busses so they don't need to fiddle with the OFW node of their parent themselves. As SBus busses hang of the nexus and we don't use the ofw_bus interface for nexus devices, yet, this would also require special knowledge about this in the drivers for the SBus children which these shouldn't need to have. This includes switching to use an unshifted IGN in the sc_ign member of the sbus(4) softc internally. - For SBus child devices where there are variants that are actually split split into two SBus devices (as opposed to the first half of the device being a SBus device and the second half hanging off of the first one) like 'auxio' and 'SUNW,fdtwo' or 'dma' and 'esp' probe the SBus device which is a prerequisite to the driver attaching to the second one with a lower order. This saves us from dealing with different probe orders in the respective device drivers which generally is more hackish. - Remove a stale comment about the 'specials' array above the attaching of the child devices. This is a remnant of the NetBSD/sparc origin of this code. There the 'specials' array is also used to probe certain devices which are prerequisites to others first. Why NetBSD soley relies on the devices having the expected order in the OFW tree on sparc64 isn't clear to me, as far as I can tell OFW doesn't guaranteed such things.
2005-05-19 14:47:31 +00:00
sc = device_get_softc(dev);
if ((dinfo = device_get_ivars(child)) == NULL)
return (ENOENT);
switch (which) {
case SBUS_IVAR_BURSTSZ:
*result = dinfo->sdi_burstsz;
break;
case SBUS_IVAR_CLOCKFREQ:
*result = dinfo->sdi_clockfreq;
break;
- Add an IVAR for retrieving the interrupt group number of the parent Sbus device and which also applies to the children. This is very usefull for drivers for the various subordinate busses so they don't need to fiddle with the OFW node of their parent themselves. As SBus busses hang of the nexus and we don't use the ofw_bus interface for nexus devices, yet, this would also require special knowledge about this in the drivers for the SBus children which these shouldn't need to have. This includes switching to use an unshifted IGN in the sc_ign member of the sbus(4) softc internally. - For SBus child devices where there are variants that are actually split split into two SBus devices (as opposed to the first half of the device being a SBus device and the second half hanging off of the first one) like 'auxio' and 'SUNW,fdtwo' or 'dma' and 'esp' probe the SBus device which is a prerequisite to the driver attaching to the second one with a lower order. This saves us from dealing with different probe orders in the respective device drivers which generally is more hackish. - Remove a stale comment about the 'specials' array above the attaching of the child devices. This is a remnant of the NetBSD/sparc origin of this code. There the 'specials' array is also used to probe certain devices which are prerequisites to others first. Why NetBSD soley relies on the devices having the expected order in the OFW tree on sparc64 isn't clear to me, as far as I can tell OFW doesn't guaranteed such things.
2005-05-19 14:47:31 +00:00
case SBUS_IVAR_IGN:
*result = sc->sc_ign;
break;
case SBUS_IVAR_SLOT:
*result = dinfo->sdi_slot;
break;
default:
return (ENOENT);
}
- Add an IVAR for retrieving the interrupt group number of the parent Sbus device and which also applies to the children. This is very usefull for drivers for the various subordinate busses so they don't need to fiddle with the OFW node of their parent themselves. As SBus busses hang of the nexus and we don't use the ofw_bus interface for nexus devices, yet, this would also require special knowledge about this in the drivers for the SBus children which these shouldn't need to have. This includes switching to use an unshifted IGN in the sc_ign member of the sbus(4) softc internally. - For SBus child devices where there are variants that are actually split split into two SBus devices (as opposed to the first half of the device being a SBus device and the second half hanging off of the first one) like 'auxio' and 'SUNW,fdtwo' or 'dma' and 'esp' probe the SBus device which is a prerequisite to the driver attaching to the second one with a lower order. This saves us from dealing with different probe orders in the respective device drivers which generally is more hackish. - Remove a stale comment about the 'specials' array above the attaching of the child devices. This is a remnant of the NetBSD/sparc origin of this code. There the 'specials' array is also used to probe certain devices which are prerequisites to others first. Why NetBSD soley relies on the devices having the expected order in the OFW tree on sparc64 isn't clear to me, as far as I can tell OFW doesn't guaranteed such things.
2005-05-19 14:47:31 +00:00
return (0);
}
static struct resource_list *
sbus_get_resource_list(device_t dev, device_t child)
{
struct sbus_devinfo *sdi;
sdi = device_get_ivars(child);
return (&sdi->sdi_rl);
}
o Revamp the sparc64 interrupt code in order to be able to interface with the INTR_FILTER-enabled MI code. Basically this consists of registering an interrupt controller (of which there can be multiple and optionally different ones either per host-to-foo bridge or shared amongst host-to-foo bridges in any one machine) along with an interrupt vector as specific argument for all the interrupt vectors used by a given host-to-foo bridge (roughly similar to registering interrupt sources on amd64 and i386), providing functions to enable, clear and disable the interrupts of the children beneath the bridge. This also includes: - No longer entering a critical section in tl0_intr() and tl1_intr() for executing interrupt handlers but rather let the handlers enter it themselves so in the case of intr_event_handle() we don't enter a nested critical section. - Adding infrastructure for binding delivery of interrupt vectors to specific CPUs which later on can be interfaced with the code from amd64/i386 for binding interrupts to specific CPUs. - Getting rid of the wrapper hack introduced along the lines of the API changes for INTR_FILTER which as a side-effect caused interrupts associated with ithread handlers only to get the elevated priority of those associated with filters ("fast handlers") (this removes the hack also in the non-INTR_FILTER case). - Disabling (by not clearing) an interrupt in the interrupt controller until all associated handlers have been executed, which is crucial for the typical locking strategy of NIC drivers in order to work correctly in case of shared interrupts. This was a more or less theoretical problem on sparc64 though, as shared interrupts are rather uncommon there except for the on-board SCCs and UARTs. Note that due to the behavior of at least of some of the interrupt controllers used on sparc64 an enable+EOI instead of a disable+EOI approach (as implied by the INTR_FILTER MI code and implemented on other architectures) is used as the latter can cause lost interrupts or in the worst case interrupt starvation. o Correct a typo in sbus_alloc_resource() which caused (pass-through) allocations to only work down to the grandchildren of the bus, which wasn't a real problem so far as we don't support any devices which are great-grandchildren or greater of a U2S bridge, yet. o In fhc(4) use bus_{read,write}_4() instead of bus_space_{read,write}_4() in order to get rid of sc_bh and sc_bt in the fhc_softc. Also get rid of some other unneeded members in fhc_softc. Reviewed by: marcel (earlier version) Approved by: re (kensmith)
2007-09-06 19:16:30 +00:00
static void
sbus_intr_enable(void *arg)
{
struct intr_vector *iv = arg;
struct sbus_icarg *sica = iv->iv_icarg;
SYSIO_WRITE8(sica->sica_sc, sica->sica_map,
INTMAP_ENABLE(iv->iv_vec, iv->iv_mid));
}
o Revamp the sparc64 interrupt code in order to be able to interface with the INTR_FILTER-enabled MI code. Basically this consists of registering an interrupt controller (of which there can be multiple and optionally different ones either per host-to-foo bridge or shared amongst host-to-foo bridges in any one machine) along with an interrupt vector as specific argument for all the interrupt vectors used by a given host-to-foo bridge (roughly similar to registering interrupt sources on amd64 and i386), providing functions to enable, clear and disable the interrupts of the children beneath the bridge. This also includes: - No longer entering a critical section in tl0_intr() and tl1_intr() for executing interrupt handlers but rather let the handlers enter it themselves so in the case of intr_event_handle() we don't enter a nested critical section. - Adding infrastructure for binding delivery of interrupt vectors to specific CPUs which later on can be interfaced with the code from amd64/i386 for binding interrupts to specific CPUs. - Getting rid of the wrapper hack introduced along the lines of the API changes for INTR_FILTER which as a side-effect caused interrupts associated with ithread handlers only to get the elevated priority of those associated with filters ("fast handlers") (this removes the hack also in the non-INTR_FILTER case). - Disabling (by not clearing) an interrupt in the interrupt controller until all associated handlers have been executed, which is crucial for the typical locking strategy of NIC drivers in order to work correctly in case of shared interrupts. This was a more or less theoretical problem on sparc64 though, as shared interrupts are rather uncommon there except for the on-board SCCs and UARTs. Note that due to the behavior of at least of some of the interrupt controllers used on sparc64 an enable+EOI instead of a disable+EOI approach (as implied by the INTR_FILTER MI code and implemented on other architectures) is used as the latter can cause lost interrupts or in the worst case interrupt starvation. o Correct a typo in sbus_alloc_resource() which caused (pass-through) allocations to only work down to the grandchildren of the bus, which wasn't a real problem so far as we don't support any devices which are great-grandchildren or greater of a U2S bridge, yet. o In fhc(4) use bus_{read,write}_4() instead of bus_space_{read,write}_4() in order to get rid of sc_bh and sc_bt in the fhc_softc. Also get rid of some other unneeded members in fhc_softc. Reviewed by: marcel (earlier version) Approved by: re (kensmith)
2007-09-06 19:16:30 +00:00
static void
sbus_intr_disable(void *arg)
{
o Revamp the sparc64 interrupt code in order to be able to interface with the INTR_FILTER-enabled MI code. Basically this consists of registering an interrupt controller (of which there can be multiple and optionally different ones either per host-to-foo bridge or shared amongst host-to-foo bridges in any one machine) along with an interrupt vector as specific argument for all the interrupt vectors used by a given host-to-foo bridge (roughly similar to registering interrupt sources on amd64 and i386), providing functions to enable, clear and disable the interrupts of the children beneath the bridge. This also includes: - No longer entering a critical section in tl0_intr() and tl1_intr() for executing interrupt handlers but rather let the handlers enter it themselves so in the case of intr_event_handle() we don't enter a nested critical section. - Adding infrastructure for binding delivery of interrupt vectors to specific CPUs which later on can be interfaced with the code from amd64/i386 for binding interrupts to specific CPUs. - Getting rid of the wrapper hack introduced along the lines of the API changes for INTR_FILTER which as a side-effect caused interrupts associated with ithread handlers only to get the elevated priority of those associated with filters ("fast handlers") (this removes the hack also in the non-INTR_FILTER case). - Disabling (by not clearing) an interrupt in the interrupt controller until all associated handlers have been executed, which is crucial for the typical locking strategy of NIC drivers in order to work correctly in case of shared interrupts. This was a more or less theoretical problem on sparc64 though, as shared interrupts are rather uncommon there except for the on-board SCCs and UARTs. Note that due to the behavior of at least of some of the interrupt controllers used on sparc64 an enable+EOI instead of a disable+EOI approach (as implied by the INTR_FILTER MI code and implemented on other architectures) is used as the latter can cause lost interrupts or in the worst case interrupt starvation. o Correct a typo in sbus_alloc_resource() which caused (pass-through) allocations to only work down to the grandchildren of the bus, which wasn't a real problem so far as we don't support any devices which are great-grandchildren or greater of a U2S bridge, yet. o In fhc(4) use bus_{read,write}_4() instead of bus_space_{read,write}_4() in order to get rid of sc_bh and sc_bt in the fhc_softc. Also get rid of some other unneeded members in fhc_softc. Reviewed by: marcel (earlier version) Approved by: re (kensmith)
2007-09-06 19:16:30 +00:00
struct intr_vector *iv = arg;
struct sbus_icarg *sica = iv->iv_icarg;
SYSIO_WRITE8(sica->sica_sc, sica->sica_map, iv->iv_vec);
}
static void
sbus_intr_assign(void *arg)
{
struct intr_vector *iv = arg;
struct sbus_icarg *sica = iv->iv_icarg;
SYSIO_WRITE8(sica->sica_sc, sica->sica_map, INTMAP_TID(
SYSIO_READ8(sica->sica_sc, sica->sica_map), iv->iv_mid));
}
static void
sbus_intr_clear(void *arg)
{
o Revamp the sparc64 interrupt code in order to be able to interface with the INTR_FILTER-enabled MI code. Basically this consists of registering an interrupt controller (of which there can be multiple and optionally different ones either per host-to-foo bridge or shared amongst host-to-foo bridges in any one machine) along with an interrupt vector as specific argument for all the interrupt vectors used by a given host-to-foo bridge (roughly similar to registering interrupt sources on amd64 and i386), providing functions to enable, clear and disable the interrupts of the children beneath the bridge. This also includes: - No longer entering a critical section in tl0_intr() and tl1_intr() for executing interrupt handlers but rather let the handlers enter it themselves so in the case of intr_event_handle() we don't enter a nested critical section. - Adding infrastructure for binding delivery of interrupt vectors to specific CPUs which later on can be interfaced with the code from amd64/i386 for binding interrupts to specific CPUs. - Getting rid of the wrapper hack introduced along the lines of the API changes for INTR_FILTER which as a side-effect caused interrupts associated with ithread handlers only to get the elevated priority of those associated with filters ("fast handlers") (this removes the hack also in the non-INTR_FILTER case). - Disabling (by not clearing) an interrupt in the interrupt controller until all associated handlers have been executed, which is crucial for the typical locking strategy of NIC drivers in order to work correctly in case of shared interrupts. This was a more or less theoretical problem on sparc64 though, as shared interrupts are rather uncommon there except for the on-board SCCs and UARTs. Note that due to the behavior of at least of some of the interrupt controllers used on sparc64 an enable+EOI instead of a disable+EOI approach (as implied by the INTR_FILTER MI code and implemented on other architectures) is used as the latter can cause lost interrupts or in the worst case interrupt starvation. o Correct a typo in sbus_alloc_resource() which caused (pass-through) allocations to only work down to the grandchildren of the bus, which wasn't a real problem so far as we don't support any devices which are great-grandchildren or greater of a U2S bridge, yet. o In fhc(4) use bus_{read,write}_4() instead of bus_space_{read,write}_4() in order to get rid of sc_bh and sc_bt in the fhc_softc. Also get rid of some other unneeded members in fhc_softc. Reviewed by: marcel (earlier version) Approved by: re (kensmith)
2007-09-06 19:16:30 +00:00
struct intr_vector *iv = arg;
struct sbus_icarg *sica = iv->iv_icarg;
SYSIO_WRITE8(sica->sica_sc, sica->sica_clr, INTCLR_IDLE);
}
static int
o Revamp the sparc64 interrupt code in order to be able to interface with the INTR_FILTER-enabled MI code. Basically this consists of registering an interrupt controller (of which there can be multiple and optionally different ones either per host-to-foo bridge or shared amongst host-to-foo bridges in any one machine) along with an interrupt vector as specific argument for all the interrupt vectors used by a given host-to-foo bridge (roughly similar to registering interrupt sources on amd64 and i386), providing functions to enable, clear and disable the interrupts of the children beneath the bridge. This also includes: - No longer entering a critical section in tl0_intr() and tl1_intr() for executing interrupt handlers but rather let the handlers enter it themselves so in the case of intr_event_handle() we don't enter a nested critical section. - Adding infrastructure for binding delivery of interrupt vectors to specific CPUs which later on can be interfaced with the code from amd64/i386 for binding interrupts to specific CPUs. - Getting rid of the wrapper hack introduced along the lines of the API changes for INTR_FILTER which as a side-effect caused interrupts associated with ithread handlers only to get the elevated priority of those associated with filters ("fast handlers") (this removes the hack also in the non-INTR_FILTER case). - Disabling (by not clearing) an interrupt in the interrupt controller until all associated handlers have been executed, which is crucial for the typical locking strategy of NIC drivers in order to work correctly in case of shared interrupts. This was a more or less theoretical problem on sparc64 though, as shared interrupts are rather uncommon there except for the on-board SCCs and UARTs. Note that due to the behavior of at least of some of the interrupt controllers used on sparc64 an enable+EOI instead of a disable+EOI approach (as implied by the INTR_FILTER MI code and implemented on other architectures) is used as the latter can cause lost interrupts or in the worst case interrupt starvation. o Correct a typo in sbus_alloc_resource() which caused (pass-through) allocations to only work down to the grandchildren of the bus, which wasn't a real problem so far as we don't support any devices which are great-grandchildren or greater of a U2S bridge, yet. o In fhc(4) use bus_{read,write}_4() instead of bus_space_{read,write}_4() in order to get rid of sc_bh and sc_bt in the fhc_softc. Also get rid of some other unneeded members in fhc_softc. Reviewed by: marcel (earlier version) Approved by: re (kensmith)
2007-09-06 19:16:30 +00:00
sbus_find_intrmap(struct sbus_softc *sc, u_int ino, bus_addr_t *intrmapptr,
bus_addr_t *intrclrptr)
{
o Revamp the sparc64 interrupt code in order to be able to interface with the INTR_FILTER-enabled MI code. Basically this consists of registering an interrupt controller (of which there can be multiple and optionally different ones either per host-to-foo bridge or shared amongst host-to-foo bridges in any one machine) along with an interrupt vector as specific argument for all the interrupt vectors used by a given host-to-foo bridge (roughly similar to registering interrupt sources on amd64 and i386), providing functions to enable, clear and disable the interrupts of the children beneath the bridge. This also includes: - No longer entering a critical section in tl0_intr() and tl1_intr() for executing interrupt handlers but rather let the handlers enter it themselves so in the case of intr_event_handle() we don't enter a nested critical section. - Adding infrastructure for binding delivery of interrupt vectors to specific CPUs which later on can be interfaced with the code from amd64/i386 for binding interrupts to specific CPUs. - Getting rid of the wrapper hack introduced along the lines of the API changes for INTR_FILTER which as a side-effect caused interrupts associated with ithread handlers only to get the elevated priority of those associated with filters ("fast handlers") (this removes the hack also in the non-INTR_FILTER case). - Disabling (by not clearing) an interrupt in the interrupt controller until all associated handlers have been executed, which is crucial for the typical locking strategy of NIC drivers in order to work correctly in case of shared interrupts. This was a more or less theoretical problem on sparc64 though, as shared interrupts are rather uncommon there except for the on-board SCCs and UARTs. Note that due to the behavior of at least of some of the interrupt controllers used on sparc64 an enable+EOI instead of a disable+EOI approach (as implied by the INTR_FILTER MI code and implemented on other architectures) is used as the latter can cause lost interrupts or in the worst case interrupt starvation. o Correct a typo in sbus_alloc_resource() which caused (pass-through) allocations to only work down to the grandchildren of the bus, which wasn't a real problem so far as we don't support any devices which are great-grandchildren or greater of a U2S bridge, yet. o In fhc(4) use bus_{read,write}_4() instead of bus_space_{read,write}_4() in order to get rid of sc_bh and sc_bt in the fhc_softc. Also get rid of some other unneeded members in fhc_softc. Reviewed by: marcel (earlier version) Approved by: re (kensmith)
2007-09-06 19:16:30 +00:00
bus_addr_t intrclr, intrmap;
int i;
o Revamp the sparc64 interrupt code in order to be able to interface with the INTR_FILTER-enabled MI code. Basically this consists of registering an interrupt controller (of which there can be multiple and optionally different ones either per host-to-foo bridge or shared amongst host-to-foo bridges in any one machine) along with an interrupt vector as specific argument for all the interrupt vectors used by a given host-to-foo bridge (roughly similar to registering interrupt sources on amd64 and i386), providing functions to enable, clear and disable the interrupts of the children beneath the bridge. This also includes: - No longer entering a critical section in tl0_intr() and tl1_intr() for executing interrupt handlers but rather let the handlers enter it themselves so in the case of intr_event_handle() we don't enter a nested critical section. - Adding infrastructure for binding delivery of interrupt vectors to specific CPUs which later on can be interfaced with the code from amd64/i386 for binding interrupts to specific CPUs. - Getting rid of the wrapper hack introduced along the lines of the API changes for INTR_FILTER which as a side-effect caused interrupts associated with ithread handlers only to get the elevated priority of those associated with filters ("fast handlers") (this removes the hack also in the non-INTR_FILTER case). - Disabling (by not clearing) an interrupt in the interrupt controller until all associated handlers have been executed, which is crucial for the typical locking strategy of NIC drivers in order to work correctly in case of shared interrupts. This was a more or less theoretical problem on sparc64 though, as shared interrupts are rather uncommon there except for the on-board SCCs and UARTs. Note that due to the behavior of at least of some of the interrupt controllers used on sparc64 an enable+EOI instead of a disable+EOI approach (as implied by the INTR_FILTER MI code and implemented on other architectures) is used as the latter can cause lost interrupts or in the worst case interrupt starvation. o Correct a typo in sbus_alloc_resource() which caused (pass-through) allocations to only work down to the grandchildren of the bus, which wasn't a real problem so far as we don't support any devices which are great-grandchildren or greater of a U2S bridge, yet. o In fhc(4) use bus_{read,write}_4() instead of bus_space_{read,write}_4() in order to get rid of sc_bh and sc_bt in the fhc_softc. Also get rid of some other unneeded members in fhc_softc. Reviewed by: marcel (earlier version) Approved by: re (kensmith)
2007-09-06 19:16:30 +00:00
if (ino > SBUS_MAX_INO) {
device_printf(sc->sc_dev, "out of range INO %d requested\n",
ino);
return (0);
}
o Revamp the sparc64 interrupt code in order to be able to interface with the INTR_FILTER-enabled MI code. Basically this consists of registering an interrupt controller (of which there can be multiple and optionally different ones either per host-to-foo bridge or shared amongst host-to-foo bridges in any one machine) along with an interrupt vector as specific argument for all the interrupt vectors used by a given host-to-foo bridge (roughly similar to registering interrupt sources on amd64 and i386), providing functions to enable, clear and disable the interrupts of the children beneath the bridge. This also includes: - No longer entering a critical section in tl0_intr() and tl1_intr() for executing interrupt handlers but rather let the handlers enter it themselves so in the case of intr_event_handle() we don't enter a nested critical section. - Adding infrastructure for binding delivery of interrupt vectors to specific CPUs which later on can be interfaced with the code from amd64/i386 for binding interrupts to specific CPUs. - Getting rid of the wrapper hack introduced along the lines of the API changes for INTR_FILTER which as a side-effect caused interrupts associated with ithread handlers only to get the elevated priority of those associated with filters ("fast handlers") (this removes the hack also in the non-INTR_FILTER case). - Disabling (by not clearing) an interrupt in the interrupt controller until all associated handlers have been executed, which is crucial for the typical locking strategy of NIC drivers in order to work correctly in case of shared interrupts. This was a more or less theoretical problem on sparc64 though, as shared interrupts are rather uncommon there except for the on-board SCCs and UARTs. Note that due to the behavior of at least of some of the interrupt controllers used on sparc64 an enable+EOI instead of a disable+EOI approach (as implied by the INTR_FILTER MI code and implemented on other architectures) is used as the latter can cause lost interrupts or in the worst case interrupt starvation. o Correct a typo in sbus_alloc_resource() which caused (pass-through) allocations to only work down to the grandchildren of the bus, which wasn't a real problem so far as we don't support any devices which are great-grandchildren or greater of a U2S bridge, yet. o In fhc(4) use bus_{read,write}_4() instead of bus_space_{read,write}_4() in order to get rid of sc_bh and sc_bt in the fhc_softc. Also get rid of some other unneeded members in fhc_softc. Reviewed by: marcel (earlier version) Approved by: re (kensmith)
2007-09-06 19:16:30 +00:00
if ((ino & INTMAP_OBIO_MASK) == 0) {
intrmap = SBR_SLOT0_INT_MAP + INTSLOT(ino) * 8;
intrclr = SBR_SLOT0_INT_CLR +
(INTSLOT(ino) * 8 * 8) + (INTPRI(ino) * 8);
} else {
intrclr = 0;
for (i = 0, intrmap = SBR_SCSI_INT_MAP;
intrmap <= SBR_RESERVED_INT_MAP; intrmap += 8, i++) {
if (INTVEC(SYSIO_READ8(sc, intrmap)) ==
INTMAP_VEC(sc->sc_ign, ino)) {
intrclr = SBR_SCSI_INT_CLR + i * 8;
break;
}
}
if (intrclr == 0)
return (0);
}
if (intrmapptr != NULL)
*intrmapptr = intrmap;
if (intrclrptr != NULL)
*intrclrptr = intrclr;
return (1);
}
static int
o Revamp the sparc64 interrupt code in order to be able to interface with the INTR_FILTER-enabled MI code. Basically this consists of registering an interrupt controller (of which there can be multiple and optionally different ones either per host-to-foo bridge or shared amongst host-to-foo bridges in any one machine) along with an interrupt vector as specific argument for all the interrupt vectors used by a given host-to-foo bridge (roughly similar to registering interrupt sources on amd64 and i386), providing functions to enable, clear and disable the interrupts of the children beneath the bridge. This also includes: - No longer entering a critical section in tl0_intr() and tl1_intr() for executing interrupt handlers but rather let the handlers enter it themselves so in the case of intr_event_handle() we don't enter a nested critical section. - Adding infrastructure for binding delivery of interrupt vectors to specific CPUs which later on can be interfaced with the code from amd64/i386 for binding interrupts to specific CPUs. - Getting rid of the wrapper hack introduced along the lines of the API changes for INTR_FILTER which as a side-effect caused interrupts associated with ithread handlers only to get the elevated priority of those associated with filters ("fast handlers") (this removes the hack also in the non-INTR_FILTER case). - Disabling (by not clearing) an interrupt in the interrupt controller until all associated handlers have been executed, which is crucial for the typical locking strategy of NIC drivers in order to work correctly in case of shared interrupts. This was a more or less theoretical problem on sparc64 though, as shared interrupts are rather uncommon there except for the on-board SCCs and UARTs. Note that due to the behavior of at least of some of the interrupt controllers used on sparc64 an enable+EOI instead of a disable+EOI approach (as implied by the INTR_FILTER MI code and implemented on other architectures) is used as the latter can cause lost interrupts or in the worst case interrupt starvation. o Correct a typo in sbus_alloc_resource() which caused (pass-through) allocations to only work down to the grandchildren of the bus, which wasn't a real problem so far as we don't support any devices which are great-grandchildren or greater of a U2S bridge, yet. o In fhc(4) use bus_{read,write}_4() instead of bus_space_{read,write}_4() in order to get rid of sc_bh and sc_bt in the fhc_softc. Also get rid of some other unneeded members in fhc_softc. Reviewed by: marcel (earlier version) Approved by: re (kensmith)
2007-09-06 19:16:30 +00:00
sbus_setup_intr(device_t dev, device_t child, struct resource *ires, int flags,
driver_filter_t *filt, driver_intr_t *intr, void *arg, void **cookiep)
{
o Revamp the sparc64 interrupt code in order to be able to interface with the INTR_FILTER-enabled MI code. Basically this consists of registering an interrupt controller (of which there can be multiple and optionally different ones either per host-to-foo bridge or shared amongst host-to-foo bridges in any one machine) along with an interrupt vector as specific argument for all the interrupt vectors used by a given host-to-foo bridge (roughly similar to registering interrupt sources on amd64 and i386), providing functions to enable, clear and disable the interrupts of the children beneath the bridge. This also includes: - No longer entering a critical section in tl0_intr() and tl1_intr() for executing interrupt handlers but rather let the handlers enter it themselves so in the case of intr_event_handle() we don't enter a nested critical section. - Adding infrastructure for binding delivery of interrupt vectors to specific CPUs which later on can be interfaced with the code from amd64/i386 for binding interrupts to specific CPUs. - Getting rid of the wrapper hack introduced along the lines of the API changes for INTR_FILTER which as a side-effect caused interrupts associated with ithread handlers only to get the elevated priority of those associated with filters ("fast handlers") (this removes the hack also in the non-INTR_FILTER case). - Disabling (by not clearing) an interrupt in the interrupt controller until all associated handlers have been executed, which is crucial for the typical locking strategy of NIC drivers in order to work correctly in case of shared interrupts. This was a more or less theoretical problem on sparc64 though, as shared interrupts are rather uncommon there except for the on-board SCCs and UARTs. Note that due to the behavior of at least of some of the interrupt controllers used on sparc64 an enable+EOI instead of a disable+EOI approach (as implied by the INTR_FILTER MI code and implemented on other architectures) is used as the latter can cause lost interrupts or in the worst case interrupt starvation. o Correct a typo in sbus_alloc_resource() which caused (pass-through) allocations to only work down to the grandchildren of the bus, which wasn't a real problem so far as we don't support any devices which are great-grandchildren or greater of a U2S bridge, yet. o In fhc(4) use bus_{read,write}_4() instead of bus_space_{read,write}_4() in order to get rid of sc_bh and sc_bt in the fhc_softc. Also get rid of some other unneeded members in fhc_softc. Reviewed by: marcel (earlier version) Approved by: re (kensmith)
2007-09-06 19:16:30 +00:00
struct sbus_softc *sc;
u_long vec;
o Revamp the sparc64 interrupt code in order to be able to interface with the INTR_FILTER-enabled MI code. Basically this consists of registering an interrupt controller (of which there can be multiple and optionally different ones either per host-to-foo bridge or shared amongst host-to-foo bridges in any one machine) along with an interrupt vector as specific argument for all the interrupt vectors used by a given host-to-foo bridge (roughly similar to registering interrupt sources on amd64 and i386), providing functions to enable, clear and disable the interrupts of the children beneath the bridge. This also includes: - No longer entering a critical section in tl0_intr() and tl1_intr() for executing interrupt handlers but rather let the handlers enter it themselves so in the case of intr_event_handle() we don't enter a nested critical section. - Adding infrastructure for binding delivery of interrupt vectors to specific CPUs which later on can be interfaced with the code from amd64/i386 for binding interrupts to specific CPUs. - Getting rid of the wrapper hack introduced along the lines of the API changes for INTR_FILTER which as a side-effect caused interrupts associated with ithread handlers only to get the elevated priority of those associated with filters ("fast handlers") (this removes the hack also in the non-INTR_FILTER case). - Disabling (by not clearing) an interrupt in the interrupt controller until all associated handlers have been executed, which is crucial for the typical locking strategy of NIC drivers in order to work correctly in case of shared interrupts. This was a more or less theoretical problem on sparc64 though, as shared interrupts are rather uncommon there except for the on-board SCCs and UARTs. Note that due to the behavior of at least of some of the interrupt controllers used on sparc64 an enable+EOI instead of a disable+EOI approach (as implied by the INTR_FILTER MI code and implemented on other architectures) is used as the latter can cause lost interrupts or in the worst case interrupt starvation. o Correct a typo in sbus_alloc_resource() which caused (pass-through) allocations to only work down to the grandchildren of the bus, which wasn't a real problem so far as we don't support any devices which are great-grandchildren or greater of a U2S bridge, yet. o In fhc(4) use bus_{read,write}_4() instead of bus_space_{read,write}_4() in order to get rid of sc_bh and sc_bt in the fhc_softc. Also get rid of some other unneeded members in fhc_softc. Reviewed by: marcel (earlier version) Approved by: re (kensmith)
2007-09-06 19:16:30 +00:00
sc = device_get_softc(dev);
/*
o Revamp the sparc64 interrupt code in order to be able to interface with the INTR_FILTER-enabled MI code. Basically this consists of registering an interrupt controller (of which there can be multiple and optionally different ones either per host-to-foo bridge or shared amongst host-to-foo bridges in any one machine) along with an interrupt vector as specific argument for all the interrupt vectors used by a given host-to-foo bridge (roughly similar to registering interrupt sources on amd64 and i386), providing functions to enable, clear and disable the interrupts of the children beneath the bridge. This also includes: - No longer entering a critical section in tl0_intr() and tl1_intr() for executing interrupt handlers but rather let the handlers enter it themselves so in the case of intr_event_handle() we don't enter a nested critical section. - Adding infrastructure for binding delivery of interrupt vectors to specific CPUs which later on can be interfaced with the code from amd64/i386 for binding interrupts to specific CPUs. - Getting rid of the wrapper hack introduced along the lines of the API changes for INTR_FILTER which as a side-effect caused interrupts associated with ithread handlers only to get the elevated priority of those associated with filters ("fast handlers") (this removes the hack also in the non-INTR_FILTER case). - Disabling (by not clearing) an interrupt in the interrupt controller until all associated handlers have been executed, which is crucial for the typical locking strategy of NIC drivers in order to work correctly in case of shared interrupts. This was a more or less theoretical problem on sparc64 though, as shared interrupts are rather uncommon there except for the on-board SCCs and UARTs. Note that due to the behavior of at least of some of the interrupt controllers used on sparc64 an enable+EOI instead of a disable+EOI approach (as implied by the INTR_FILTER MI code and implemented on other architectures) is used as the latter can cause lost interrupts or in the worst case interrupt starvation. o Correct a typo in sbus_alloc_resource() which caused (pass-through) allocations to only work down to the grandchildren of the bus, which wasn't a real problem so far as we don't support any devices which are great-grandchildren or greater of a U2S bridge, yet. o In fhc(4) use bus_{read,write}_4() instead of bus_space_{read,write}_4() in order to get rid of sc_bh and sc_bt in the fhc_softc. Also get rid of some other unneeded members in fhc_softc. Reviewed by: marcel (earlier version) Approved by: re (kensmith)
2007-09-06 19:16:30 +00:00
* Make sure the vector is fully specified and we registered
* our interrupt controller for it.
Make sparc64 compatible with NEW_PCIB and enable it: - Implement bus_adjust_resource() methods as far as necessary and in non-PCI bridge drivers as far as feasible without rototilling them. - As NEW_PCIB does a layering violation by activating resources at layers above pci(4) without previously bubbling up their allocation there, move the assignment of bus tags and handles from the bus_alloc_resource() to the bus_activate_resource() methods like at least the other NEW_PCIB enabled architectures do. This is somewhat unfortunate as previously sparc64 (ab)used resource activation to indicate whether SYS_RES_MEMORY resources should be mapped into KVA, which is only necessary if their going to be accessed via the pointer returned from rman_get_virtual() but not for bus_space(9) as the later always uses physical access on sparc64. Besides wasting KVA if we always map in SYS_RES_MEMORY resources, a driver also may deliberately not map them in if the firmware already has done so, possibly in a special way. So in order to still allow a driver to decide whether a SYS_RES_MEMORY resource should be mapped into KVA we let it indicate that by calling bus_space_map(9) with BUS_SPACE_MAP_LINEAR as actually documented in the bus_space(9) page. This is implemented by allocating a separate bus tag per SYS_RES_MEMORY resource and passing the resource via the previously unused bus tag cookie so we later on can call rman_set_virtual() in sparc64_bus_mem_map(). As a side effect this now also allows to actually indicate that a SYS_RES_MEMORY resource should be mapped in as cacheable and/or read-only via BUS_SPACE_MAP_CACHEABLE and BUS_SPACE_MAP_READONLY respectively. - Do some minor cleanup like taking advantage of rman_init_from_resource(), factor out the common part of bus tag allocation into a newly added sparc64_alloc_bus_tag(), hook up some missing newbus methods and replace some homegrown versions with the generic counterparts etc. - While at it, let apb_attach() (which can't use the generic NEW_PCIB code as APB bridges just don't have the base and limit registers implemented) regarding the config space registers cached in pcib_softc and the SYSCTL reporting nodes set up.
2011-10-02 23:22:38 +00:00
*/
o Revamp the sparc64 interrupt code in order to be able to interface with the INTR_FILTER-enabled MI code. Basically this consists of registering an interrupt controller (of which there can be multiple and optionally different ones either per host-to-foo bridge or shared amongst host-to-foo bridges in any one machine) along with an interrupt vector as specific argument for all the interrupt vectors used by a given host-to-foo bridge (roughly similar to registering interrupt sources on amd64 and i386), providing functions to enable, clear and disable the interrupts of the children beneath the bridge. This also includes: - No longer entering a critical section in tl0_intr() and tl1_intr() for executing interrupt handlers but rather let the handlers enter it themselves so in the case of intr_event_handle() we don't enter a nested critical section. - Adding infrastructure for binding delivery of interrupt vectors to specific CPUs which later on can be interfaced with the code from amd64/i386 for binding interrupts to specific CPUs. - Getting rid of the wrapper hack introduced along the lines of the API changes for INTR_FILTER which as a side-effect caused interrupts associated with ithread handlers only to get the elevated priority of those associated with filters ("fast handlers") (this removes the hack also in the non-INTR_FILTER case). - Disabling (by not clearing) an interrupt in the interrupt controller until all associated handlers have been executed, which is crucial for the typical locking strategy of NIC drivers in order to work correctly in case of shared interrupts. This was a more or less theoretical problem on sparc64 though, as shared interrupts are rather uncommon there except for the on-board SCCs and UARTs. Note that due to the behavior of at least of some of the interrupt controllers used on sparc64 an enable+EOI instead of a disable+EOI approach (as implied by the INTR_FILTER MI code and implemented on other architectures) is used as the latter can cause lost interrupts or in the worst case interrupt starvation. o Correct a typo in sbus_alloc_resource() which caused (pass-through) allocations to only work down to the grandchildren of the bus, which wasn't a real problem so far as we don't support any devices which are great-grandchildren or greater of a U2S bridge, yet. o In fhc(4) use bus_{read,write}_4() instead of bus_space_{read,write}_4() in order to get rid of sc_bh and sc_bt in the fhc_softc. Also get rid of some other unneeded members in fhc_softc. Reviewed by: marcel (earlier version) Approved by: re (kensmith)
2007-09-06 19:16:30 +00:00
vec = rman_get_start(ires);
if (INTIGN(vec) != sc->sc_ign || intr_vectors[vec].iv_ic != &sbus_ic) {
device_printf(dev, "invalid interrupt vector 0x%lx\n", vec);
Make sparc64 compatible with NEW_PCIB and enable it: - Implement bus_adjust_resource() methods as far as necessary and in non-PCI bridge drivers as far as feasible without rototilling them. - As NEW_PCIB does a layering violation by activating resources at layers above pci(4) without previously bubbling up their allocation there, move the assignment of bus tags and handles from the bus_alloc_resource() to the bus_activate_resource() methods like at least the other NEW_PCIB enabled architectures do. This is somewhat unfortunate as previously sparc64 (ab)used resource activation to indicate whether SYS_RES_MEMORY resources should be mapped into KVA, which is only necessary if their going to be accessed via the pointer returned from rman_get_virtual() but not for bus_space(9) as the later always uses physical access on sparc64. Besides wasting KVA if we always map in SYS_RES_MEMORY resources, a driver also may deliberately not map them in if the firmware already has done so, possibly in a special way. So in order to still allow a driver to decide whether a SYS_RES_MEMORY resource should be mapped into KVA we let it indicate that by calling bus_space_map(9) with BUS_SPACE_MAP_LINEAR as actually documented in the bus_space(9) page. This is implemented by allocating a separate bus tag per SYS_RES_MEMORY resource and passing the resource via the previously unused bus tag cookie so we later on can call rman_set_virtual() in sparc64_bus_mem_map(). As a side effect this now also allows to actually indicate that a SYS_RES_MEMORY resource should be mapped in as cacheable and/or read-only via BUS_SPACE_MAP_CACHEABLE and BUS_SPACE_MAP_READONLY respectively. - Do some minor cleanup like taking advantage of rman_init_from_resource(), factor out the common part of bus tag allocation into a newly added sparc64_alloc_bus_tag(), hook up some missing newbus methods and replace some homegrown versions with the generic counterparts etc. - While at it, let apb_attach() (which can't use the generic NEW_PCIB code as APB bridges just don't have the base and limit registers implemented) regarding the config space registers cached in pcib_softc and the SYSCTL reporting nodes set up.
2011-10-02 23:22:38 +00:00
return (EINVAL);
}
o Revamp the sparc64 interrupt code in order to be able to interface with the INTR_FILTER-enabled MI code. Basically this consists of registering an interrupt controller (of which there can be multiple and optionally different ones either per host-to-foo bridge or shared amongst host-to-foo bridges in any one machine) along with an interrupt vector as specific argument for all the interrupt vectors used by a given host-to-foo bridge (roughly similar to registering interrupt sources on amd64 and i386), providing functions to enable, clear and disable the interrupts of the children beneath the bridge. This also includes: - No longer entering a critical section in tl0_intr() and tl1_intr() for executing interrupt handlers but rather let the handlers enter it themselves so in the case of intr_event_handle() we don't enter a nested critical section. - Adding infrastructure for binding delivery of interrupt vectors to specific CPUs which later on can be interfaced with the code from amd64/i386 for binding interrupts to specific CPUs. - Getting rid of the wrapper hack introduced along the lines of the API changes for INTR_FILTER which as a side-effect caused interrupts associated with ithread handlers only to get the elevated priority of those associated with filters ("fast handlers") (this removes the hack also in the non-INTR_FILTER case). - Disabling (by not clearing) an interrupt in the interrupt controller until all associated handlers have been executed, which is crucial for the typical locking strategy of NIC drivers in order to work correctly in case of shared interrupts. This was a more or less theoretical problem on sparc64 though, as shared interrupts are rather uncommon there except for the on-board SCCs and UARTs. Note that due to the behavior of at least of some of the interrupt controllers used on sparc64 an enable+EOI instead of a disable+EOI approach (as implied by the INTR_FILTER MI code and implemented on other architectures) is used as the latter can cause lost interrupts or in the worst case interrupt starvation. o Correct a typo in sbus_alloc_resource() which caused (pass-through) allocations to only work down to the grandchildren of the bus, which wasn't a real problem so far as we don't support any devices which are great-grandchildren or greater of a U2S bridge, yet. o In fhc(4) use bus_{read,write}_4() instead of bus_space_{read,write}_4() in order to get rid of sc_bh and sc_bt in the fhc_softc. Also get rid of some other unneeded members in fhc_softc. Reviewed by: marcel (earlier version) Approved by: re (kensmith)
2007-09-06 19:16:30 +00:00
return (bus_generic_setup_intr(dev, child, ires, flags, filt, intr,
arg, cookiep));
}
static struct resource *
sbus_alloc_resource(device_t bus, device_t child, int type, int *rid,
u_long start, u_long end, u_long count, u_int flags)
{
struct sbus_softc *sc;
struct rman *rm;
struct resource *rv;
struct resource_list *rl;
struct resource_list_entry *rle;
device_t schild;
bus_addr_t toffs;
bus_size_t tend;
int i, slot;
Make sparc64 compatible with NEW_PCIB and enable it: - Implement bus_adjust_resource() methods as far as necessary and in non-PCI bridge drivers as far as feasible without rototilling them. - As NEW_PCIB does a layering violation by activating resources at layers above pci(4) without previously bubbling up their allocation there, move the assignment of bus tags and handles from the bus_alloc_resource() to the bus_activate_resource() methods like at least the other NEW_PCIB enabled architectures do. This is somewhat unfortunate as previously sparc64 (ab)used resource activation to indicate whether SYS_RES_MEMORY resources should be mapped into KVA, which is only necessary if their going to be accessed via the pointer returned from rman_get_virtual() but not for bus_space(9) as the later always uses physical access on sparc64. Besides wasting KVA if we always map in SYS_RES_MEMORY resources, a driver also may deliberately not map them in if the firmware already has done so, possibly in a special way. So in order to still allow a driver to decide whether a SYS_RES_MEMORY resource should be mapped into KVA we let it indicate that by calling bus_space_map(9) with BUS_SPACE_MAP_LINEAR as actually documented in the bus_space(9) page. This is implemented by allocating a separate bus tag per SYS_RES_MEMORY resource and passing the resource via the previously unused bus tag cookie so we later on can call rman_set_virtual() in sparc64_bus_mem_map(). As a side effect this now also allows to actually indicate that a SYS_RES_MEMORY resource should be mapped in as cacheable and/or read-only via BUS_SPACE_MAP_CACHEABLE and BUS_SPACE_MAP_READONLY respectively. - Do some minor cleanup like taking advantage of rman_init_from_resource(), factor out the common part of bus tag allocation into a newly added sparc64_alloc_bus_tag(), hook up some missing newbus methods and replace some homegrown versions with the generic counterparts etc. - While at it, let apb_attach() (which can't use the generic NEW_PCIB code as APB bridges just don't have the base and limit registers implemented) regarding the config space registers cached in pcib_softc and the SYSCTL reporting nodes set up.
2011-10-02 23:22:38 +00:00
int isdefault, passthrough;
isdefault = (start == 0UL && end == ~0UL);
passthrough = (device_get_parent(child) != bus);
rle = NULL;
sc = device_get_softc(bus);
rl = BUS_GET_RESOURCE_LIST(bus, child);
switch (type) {
case SYS_RES_IRQ:
return (resource_list_alloc(rl, bus, child, type, rid, start,
end, count, flags));
case SYS_RES_MEMORY:
if (!passthrough) {
rle = resource_list_find(rl, type, *rid);
if (rle == NULL)
return (NULL);
if (rle->res != NULL)
panic("%s: resource entry is busy", __func__);
if (isdefault) {
start = rle->start;
count = ulmax(count, rle->count);
end = ulmax(rle->end, start + count - 1);
}
}
rm = NULL;
schild = child;
while (device_get_parent(schild) != bus)
o Revamp the sparc64 interrupt code in order to be able to interface with the INTR_FILTER-enabled MI code. Basically this consists of registering an interrupt controller (of which there can be multiple and optionally different ones either per host-to-foo bridge or shared amongst host-to-foo bridges in any one machine) along with an interrupt vector as specific argument for all the interrupt vectors used by a given host-to-foo bridge (roughly similar to registering interrupt sources on amd64 and i386), providing functions to enable, clear and disable the interrupts of the children beneath the bridge. This also includes: - No longer entering a critical section in tl0_intr() and tl1_intr() for executing interrupt handlers but rather let the handlers enter it themselves so in the case of intr_event_handle() we don't enter a nested critical section. - Adding infrastructure for binding delivery of interrupt vectors to specific CPUs which later on can be interfaced with the code from amd64/i386 for binding interrupts to specific CPUs. - Getting rid of the wrapper hack introduced along the lines of the API changes for INTR_FILTER which as a side-effect caused interrupts associated with ithread handlers only to get the elevated priority of those associated with filters ("fast handlers") (this removes the hack also in the non-INTR_FILTER case). - Disabling (by not clearing) an interrupt in the interrupt controller until all associated handlers have been executed, which is crucial for the typical locking strategy of NIC drivers in order to work correctly in case of shared interrupts. This was a more or less theoretical problem on sparc64 though, as shared interrupts are rather uncommon there except for the on-board SCCs and UARTs. Note that due to the behavior of at least of some of the interrupt controllers used on sparc64 an enable+EOI instead of a disable+EOI approach (as implied by the INTR_FILTER MI code and implemented on other architectures) is used as the latter can cause lost interrupts or in the worst case interrupt starvation. o Correct a typo in sbus_alloc_resource() which caused (pass-through) allocations to only work down to the grandchildren of the bus, which wasn't a real problem so far as we don't support any devices which are great-grandchildren or greater of a U2S bridge, yet. o In fhc(4) use bus_{read,write}_4() instead of bus_space_{read,write}_4() in order to get rid of sc_bh and sc_bt in the fhc_softc. Also get rid of some other unneeded members in fhc_softc. Reviewed by: marcel (earlier version) Approved by: re (kensmith)
2007-09-06 19:16:30 +00:00
schild = device_get_parent(schild);
slot = sbus_get_slot(schild);
for (i = 0; i < sc->sc_nrange; i++) {
if (sc->sc_rd[i].rd_slot != slot ||
start < sc->sc_rd[i].rd_coffset ||
start > sc->sc_rd[i].rd_cend)
continue;
/* Disallow cross-range allocations. */
if (end > sc->sc_rd[i].rd_cend)
return (NULL);
/* We've found the connection to the parent bus */
toffs = start - sc->sc_rd[i].rd_coffset;
tend = end - sc->sc_rd[i].rd_coffset;
rm = &sc->sc_rd[i].rd_rman;
break;
}
if (rm == NULL)
return (NULL);
Make sparc64 compatible with NEW_PCIB and enable it: - Implement bus_adjust_resource() methods as far as necessary and in non-PCI bridge drivers as far as feasible without rototilling them. - As NEW_PCIB does a layering violation by activating resources at layers above pci(4) without previously bubbling up their allocation there, move the assignment of bus tags and handles from the bus_alloc_resource() to the bus_activate_resource() methods like at least the other NEW_PCIB enabled architectures do. This is somewhat unfortunate as previously sparc64 (ab)used resource activation to indicate whether SYS_RES_MEMORY resources should be mapped into KVA, which is only necessary if their going to be accessed via the pointer returned from rman_get_virtual() but not for bus_space(9) as the later always uses physical access on sparc64. Besides wasting KVA if we always map in SYS_RES_MEMORY resources, a driver also may deliberately not map them in if the firmware already has done so, possibly in a special way. So in order to still allow a driver to decide whether a SYS_RES_MEMORY resource should be mapped into KVA we let it indicate that by calling bus_space_map(9) with BUS_SPACE_MAP_LINEAR as actually documented in the bus_space(9) page. This is implemented by allocating a separate bus tag per SYS_RES_MEMORY resource and passing the resource via the previously unused bus tag cookie so we later on can call rman_set_virtual() in sparc64_bus_mem_map(). As a side effect this now also allows to actually indicate that a SYS_RES_MEMORY resource should be mapped in as cacheable and/or read-only via BUS_SPACE_MAP_CACHEABLE and BUS_SPACE_MAP_READONLY respectively. - Do some minor cleanup like taking advantage of rman_init_from_resource(), factor out the common part of bus tag allocation into a newly added sparc64_alloc_bus_tag(), hook up some missing newbus methods and replace some homegrown versions with the generic counterparts etc. - While at it, let apb_attach() (which can't use the generic NEW_PCIB code as APB bridges just don't have the base and limit registers implemented) regarding the config space registers cached in pcib_softc and the SYSCTL reporting nodes set up.
2011-10-02 23:22:38 +00:00
rv = rman_reserve_resource(rm, toffs, tend, count, flags &
~RF_ACTIVE, child);
if (rv == NULL)
return (NULL);
rman_set_rid(rv, *rid);
Make sparc64 compatible with NEW_PCIB and enable it: - Implement bus_adjust_resource() methods as far as necessary and in non-PCI bridge drivers as far as feasible without rototilling them. - As NEW_PCIB does a layering violation by activating resources at layers above pci(4) without previously bubbling up their allocation there, move the assignment of bus tags and handles from the bus_alloc_resource() to the bus_activate_resource() methods like at least the other NEW_PCIB enabled architectures do. This is somewhat unfortunate as previously sparc64 (ab)used resource activation to indicate whether SYS_RES_MEMORY resources should be mapped into KVA, which is only necessary if their going to be accessed via the pointer returned from rman_get_virtual() but not for bus_space(9) as the later always uses physical access on sparc64. Besides wasting KVA if we always map in SYS_RES_MEMORY resources, a driver also may deliberately not map them in if the firmware already has done so, possibly in a special way. So in order to still allow a driver to decide whether a SYS_RES_MEMORY resource should be mapped into KVA we let it indicate that by calling bus_space_map(9) with BUS_SPACE_MAP_LINEAR as actually documented in the bus_space(9) page. This is implemented by allocating a separate bus tag per SYS_RES_MEMORY resource and passing the resource via the previously unused bus tag cookie so we later on can call rman_set_virtual() in sparc64_bus_mem_map(). As a side effect this now also allows to actually indicate that a SYS_RES_MEMORY resource should be mapped in as cacheable and/or read-only via BUS_SPACE_MAP_CACHEABLE and BUS_SPACE_MAP_READONLY respectively. - Do some minor cleanup like taking advantage of rman_init_from_resource(), factor out the common part of bus tag allocation into a newly added sparc64_alloc_bus_tag(), hook up some missing newbus methods and replace some homegrown versions with the generic counterparts etc. - While at it, let apb_attach() (which can't use the generic NEW_PCIB code as APB bridges just don't have the base and limit registers implemented) regarding the config space registers cached in pcib_softc and the SYSCTL reporting nodes set up.
2011-10-02 23:22:38 +00:00
if ((flags & RF_ACTIVE) != 0 && bus_activate_resource(child,
type, *rid, rv)) {
rman_release_resource(rv);
return (NULL);
}
if (!passthrough)
rle->res = rv;
return (rv);
default:
return (NULL);
}
}
static int
sbus_activate_resource(device_t bus, device_t child, int type, int rid,
struct resource *r)
{
Make sparc64 compatible with NEW_PCIB and enable it: - Implement bus_adjust_resource() methods as far as necessary and in non-PCI bridge drivers as far as feasible without rototilling them. - As NEW_PCIB does a layering violation by activating resources at layers above pci(4) without previously bubbling up their allocation there, move the assignment of bus tags and handles from the bus_alloc_resource() to the bus_activate_resource() methods like at least the other NEW_PCIB enabled architectures do. This is somewhat unfortunate as previously sparc64 (ab)used resource activation to indicate whether SYS_RES_MEMORY resources should be mapped into KVA, which is only necessary if their going to be accessed via the pointer returned from rman_get_virtual() but not for bus_space(9) as the later always uses physical access on sparc64. Besides wasting KVA if we always map in SYS_RES_MEMORY resources, a driver also may deliberately not map them in if the firmware already has done so, possibly in a special way. So in order to still allow a driver to decide whether a SYS_RES_MEMORY resource should be mapped into KVA we let it indicate that by calling bus_space_map(9) with BUS_SPACE_MAP_LINEAR as actually documented in the bus_space(9) page. This is implemented by allocating a separate bus tag per SYS_RES_MEMORY resource and passing the resource via the previously unused bus tag cookie so we later on can call rman_set_virtual() in sparc64_bus_mem_map(). As a side effect this now also allows to actually indicate that a SYS_RES_MEMORY resource should be mapped in as cacheable and/or read-only via BUS_SPACE_MAP_CACHEABLE and BUS_SPACE_MAP_READONLY respectively. - Do some minor cleanup like taking advantage of rman_init_from_resource(), factor out the common part of bus tag allocation into a newly added sparc64_alloc_bus_tag(), hook up some missing newbus methods and replace some homegrown versions with the generic counterparts etc. - While at it, let apb_attach() (which can't use the generic NEW_PCIB code as APB bridges just don't have the base and limit registers implemented) regarding the config space registers cached in pcib_softc and the SYSCTL reporting nodes set up.
2011-10-02 23:22:38 +00:00
struct sbus_softc *sc;
struct bus_space_tag *tag;
int i;
Make sparc64 compatible with NEW_PCIB and enable it: - Implement bus_adjust_resource() methods as far as necessary and in non-PCI bridge drivers as far as feasible without rototilling them. - As NEW_PCIB does a layering violation by activating resources at layers above pci(4) without previously bubbling up their allocation there, move the assignment of bus tags and handles from the bus_alloc_resource() to the bus_activate_resource() methods like at least the other NEW_PCIB enabled architectures do. This is somewhat unfortunate as previously sparc64 (ab)used resource activation to indicate whether SYS_RES_MEMORY resources should be mapped into KVA, which is only necessary if their going to be accessed via the pointer returned from rman_get_virtual() but not for bus_space(9) as the later always uses physical access on sparc64. Besides wasting KVA if we always map in SYS_RES_MEMORY resources, a driver also may deliberately not map them in if the firmware already has done so, possibly in a special way. So in order to still allow a driver to decide whether a SYS_RES_MEMORY resource should be mapped into KVA we let it indicate that by calling bus_space_map(9) with BUS_SPACE_MAP_LINEAR as actually documented in the bus_space(9) page. This is implemented by allocating a separate bus tag per SYS_RES_MEMORY resource and passing the resource via the previously unused bus tag cookie so we later on can call rman_set_virtual() in sparc64_bus_mem_map(). As a side effect this now also allows to actually indicate that a SYS_RES_MEMORY resource should be mapped in as cacheable and/or read-only via BUS_SPACE_MAP_CACHEABLE and BUS_SPACE_MAP_READONLY respectively. - Do some minor cleanup like taking advantage of rman_init_from_resource(), factor out the common part of bus tag allocation into a newly added sparc64_alloc_bus_tag(), hook up some missing newbus methods and replace some homegrown versions with the generic counterparts etc. - While at it, let apb_attach() (which can't use the generic NEW_PCIB code as APB bridges just don't have the base and limit registers implemented) regarding the config space registers cached in pcib_softc and the SYSCTL reporting nodes set up.
2011-10-02 23:22:38 +00:00
switch (type) {
case SYS_RES_IRQ:
return (bus_generic_activate_resource(bus, child, type, rid,
r));
case SYS_RES_MEMORY:
sc = device_get_softc(bus);
for (i = 0; i < sc->sc_nrange; i++) {
if (rman_is_region_manager(r,
&sc->sc_rd[i].rd_rman) != 0) {
tag = sparc64_alloc_bus_tag(r,
rman_get_bustag(sc->sc_sysio_res),
SBUS_BUS_SPACE, NULL);
if (tag == NULL)
return (ENOMEM);
rman_set_bustag(r, tag);
rman_set_bushandle(r,
sc->sc_rd[i].rd_bushandle +
rman_get_start(r));
return (rman_activate_resource(r));
}
}
/* FALLTHROUGH */
default:
return (EINVAL);
}
}
static int
Make sparc64 compatible with NEW_PCIB and enable it: - Implement bus_adjust_resource() methods as far as necessary and in non-PCI bridge drivers as far as feasible without rototilling them. - As NEW_PCIB does a layering violation by activating resources at layers above pci(4) without previously bubbling up their allocation there, move the assignment of bus tags and handles from the bus_alloc_resource() to the bus_activate_resource() methods like at least the other NEW_PCIB enabled architectures do. This is somewhat unfortunate as previously sparc64 (ab)used resource activation to indicate whether SYS_RES_MEMORY resources should be mapped into KVA, which is only necessary if their going to be accessed via the pointer returned from rman_get_virtual() but not for bus_space(9) as the later always uses physical access on sparc64. Besides wasting KVA if we always map in SYS_RES_MEMORY resources, a driver also may deliberately not map them in if the firmware already has done so, possibly in a special way. So in order to still allow a driver to decide whether a SYS_RES_MEMORY resource should be mapped into KVA we let it indicate that by calling bus_space_map(9) with BUS_SPACE_MAP_LINEAR as actually documented in the bus_space(9) page. This is implemented by allocating a separate bus tag per SYS_RES_MEMORY resource and passing the resource via the previously unused bus tag cookie so we later on can call rman_set_virtual() in sparc64_bus_mem_map(). As a side effect this now also allows to actually indicate that a SYS_RES_MEMORY resource should be mapped in as cacheable and/or read-only via BUS_SPACE_MAP_CACHEABLE and BUS_SPACE_MAP_READONLY respectively. - Do some minor cleanup like taking advantage of rman_init_from_resource(), factor out the common part of bus tag allocation into a newly added sparc64_alloc_bus_tag(), hook up some missing newbus methods and replace some homegrown versions with the generic counterparts etc. - While at it, let apb_attach() (which can't use the generic NEW_PCIB code as APB bridges just don't have the base and limit registers implemented) regarding the config space registers cached in pcib_softc and the SYSCTL reporting nodes set up.
2011-10-02 23:22:38 +00:00
sbus_adjust_resource(device_t bus, device_t child, int type,
struct resource *r, u_long start, u_long end)
{
Make sparc64 compatible with NEW_PCIB and enable it: - Implement bus_adjust_resource() methods as far as necessary and in non-PCI bridge drivers as far as feasible without rototilling them. - As NEW_PCIB does a layering violation by activating resources at layers above pci(4) without previously bubbling up their allocation there, move the assignment of bus tags and handles from the bus_alloc_resource() to the bus_activate_resource() methods like at least the other NEW_PCIB enabled architectures do. This is somewhat unfortunate as previously sparc64 (ab)used resource activation to indicate whether SYS_RES_MEMORY resources should be mapped into KVA, which is only necessary if their going to be accessed via the pointer returned from rman_get_virtual() but not for bus_space(9) as the later always uses physical access on sparc64. Besides wasting KVA if we always map in SYS_RES_MEMORY resources, a driver also may deliberately not map them in if the firmware already has done so, possibly in a special way. So in order to still allow a driver to decide whether a SYS_RES_MEMORY resource should be mapped into KVA we let it indicate that by calling bus_space_map(9) with BUS_SPACE_MAP_LINEAR as actually documented in the bus_space(9) page. This is implemented by allocating a separate bus tag per SYS_RES_MEMORY resource and passing the resource via the previously unused bus tag cookie so we later on can call rman_set_virtual() in sparc64_bus_mem_map(). As a side effect this now also allows to actually indicate that a SYS_RES_MEMORY resource should be mapped in as cacheable and/or read-only via BUS_SPACE_MAP_CACHEABLE and BUS_SPACE_MAP_READONLY respectively. - Do some minor cleanup like taking advantage of rman_init_from_resource(), factor out the common part of bus tag allocation into a newly added sparc64_alloc_bus_tag(), hook up some missing newbus methods and replace some homegrown versions with the generic counterparts etc. - While at it, let apb_attach() (which can't use the generic NEW_PCIB code as APB bridges just don't have the base and limit registers implemented) regarding the config space registers cached in pcib_softc and the SYSCTL reporting nodes set up.
2011-10-02 23:22:38 +00:00
struct sbus_softc *sc;
int i;
if (type == SYS_RES_MEMORY) {
Make sparc64 compatible with NEW_PCIB and enable it: - Implement bus_adjust_resource() methods as far as necessary and in non-PCI bridge drivers as far as feasible without rototilling them. - As NEW_PCIB does a layering violation by activating resources at layers above pci(4) without previously bubbling up their allocation there, move the assignment of bus tags and handles from the bus_alloc_resource() to the bus_activate_resource() methods like at least the other NEW_PCIB enabled architectures do. This is somewhat unfortunate as previously sparc64 (ab)used resource activation to indicate whether SYS_RES_MEMORY resources should be mapped into KVA, which is only necessary if their going to be accessed via the pointer returned from rman_get_virtual() but not for bus_space(9) as the later always uses physical access on sparc64. Besides wasting KVA if we always map in SYS_RES_MEMORY resources, a driver also may deliberately not map them in if the firmware already has done so, possibly in a special way. So in order to still allow a driver to decide whether a SYS_RES_MEMORY resource should be mapped into KVA we let it indicate that by calling bus_space_map(9) with BUS_SPACE_MAP_LINEAR as actually documented in the bus_space(9) page. This is implemented by allocating a separate bus tag per SYS_RES_MEMORY resource and passing the resource via the previously unused bus tag cookie so we later on can call rman_set_virtual() in sparc64_bus_mem_map(). As a side effect this now also allows to actually indicate that a SYS_RES_MEMORY resource should be mapped in as cacheable and/or read-only via BUS_SPACE_MAP_CACHEABLE and BUS_SPACE_MAP_READONLY respectively. - Do some minor cleanup like taking advantage of rman_init_from_resource(), factor out the common part of bus tag allocation into a newly added sparc64_alloc_bus_tag(), hook up some missing newbus methods and replace some homegrown versions with the generic counterparts etc. - While at it, let apb_attach() (which can't use the generic NEW_PCIB code as APB bridges just don't have the base and limit registers implemented) regarding the config space registers cached in pcib_softc and the SYSCTL reporting nodes set up.
2011-10-02 23:22:38 +00:00
sc = device_get_softc(bus);
for (i = 0; i < sc->sc_nrange; i++)
if (rman_is_region_manager(r,
&sc->sc_rd[i].rd_rman) != 0)
return (rman_adjust_resource(r, start, end));
return (EINVAL);
}
Make sparc64 compatible with NEW_PCIB and enable it: - Implement bus_adjust_resource() methods as far as necessary and in non-PCI bridge drivers as far as feasible without rototilling them. - As NEW_PCIB does a layering violation by activating resources at layers above pci(4) without previously bubbling up their allocation there, move the assignment of bus tags and handles from the bus_alloc_resource() to the bus_activate_resource() methods like at least the other NEW_PCIB enabled architectures do. This is somewhat unfortunate as previously sparc64 (ab)used resource activation to indicate whether SYS_RES_MEMORY resources should be mapped into KVA, which is only necessary if their going to be accessed via the pointer returned from rman_get_virtual() but not for bus_space(9) as the later always uses physical access on sparc64. Besides wasting KVA if we always map in SYS_RES_MEMORY resources, a driver also may deliberately not map them in if the firmware already has done so, possibly in a special way. So in order to still allow a driver to decide whether a SYS_RES_MEMORY resource should be mapped into KVA we let it indicate that by calling bus_space_map(9) with BUS_SPACE_MAP_LINEAR as actually documented in the bus_space(9) page. This is implemented by allocating a separate bus tag per SYS_RES_MEMORY resource and passing the resource via the previously unused bus tag cookie so we later on can call rman_set_virtual() in sparc64_bus_mem_map(). As a side effect this now also allows to actually indicate that a SYS_RES_MEMORY resource should be mapped in as cacheable and/or read-only via BUS_SPACE_MAP_CACHEABLE and BUS_SPACE_MAP_READONLY respectively. - Do some minor cleanup like taking advantage of rman_init_from_resource(), factor out the common part of bus tag allocation into a newly added sparc64_alloc_bus_tag(), hook up some missing newbus methods and replace some homegrown versions with the generic counterparts etc. - While at it, let apb_attach() (which can't use the generic NEW_PCIB code as APB bridges just don't have the base and limit registers implemented) regarding the config space registers cached in pcib_softc and the SYSCTL reporting nodes set up.
2011-10-02 23:22:38 +00:00
return (bus_generic_adjust_resource(bus, child, type, r, start, end));
}
static int
sbus_release_resource(device_t bus, device_t child, int type, int rid,
struct resource *r)
{
struct resource_list *rl;
struct resource_list_entry *rle;
int error, passthrough;
passthrough = (device_get_parent(child) != bus);
rl = BUS_GET_RESOURCE_LIST(bus, child);
Make sparc64 compatible with NEW_PCIB and enable it: - Implement bus_adjust_resource() methods as far as necessary and in non-PCI bridge drivers as far as feasible without rototilling them. - As NEW_PCIB does a layering violation by activating resources at layers above pci(4) without previously bubbling up their allocation there, move the assignment of bus tags and handles from the bus_alloc_resource() to the bus_activate_resource() methods like at least the other NEW_PCIB enabled architectures do. This is somewhat unfortunate as previously sparc64 (ab)used resource activation to indicate whether SYS_RES_MEMORY resources should be mapped into KVA, which is only necessary if their going to be accessed via the pointer returned from rman_get_virtual() but not for bus_space(9) as the later always uses physical access on sparc64. Besides wasting KVA if we always map in SYS_RES_MEMORY resources, a driver also may deliberately not map them in if the firmware already has done so, possibly in a special way. So in order to still allow a driver to decide whether a SYS_RES_MEMORY resource should be mapped into KVA we let it indicate that by calling bus_space_map(9) with BUS_SPACE_MAP_LINEAR as actually documented in the bus_space(9) page. This is implemented by allocating a separate bus tag per SYS_RES_MEMORY resource and passing the resource via the previously unused bus tag cookie so we later on can call rman_set_virtual() in sparc64_bus_mem_map(). As a side effect this now also allows to actually indicate that a SYS_RES_MEMORY resource should be mapped in as cacheable and/or read-only via BUS_SPACE_MAP_CACHEABLE and BUS_SPACE_MAP_READONLY respectively. - Do some minor cleanup like taking advantage of rman_init_from_resource(), factor out the common part of bus tag allocation into a newly added sparc64_alloc_bus_tag(), hook up some missing newbus methods and replace some homegrown versions with the generic counterparts etc. - While at it, let apb_attach() (which can't use the generic NEW_PCIB code as APB bridges just don't have the base and limit registers implemented) regarding the config space registers cached in pcib_softc and the SYSCTL reporting nodes set up.
2011-10-02 23:22:38 +00:00
if (type == SYS_RES_MEMORY) {
if ((rman_get_flags(r) & RF_ACTIVE) != 0) {
error = bus_deactivate_resource(child, type, rid, r);
if (error)
return (error);
}
error = rman_release_resource(r);
if (error != 0)
return (error);
Make sparc64 compatible with NEW_PCIB and enable it: - Implement bus_adjust_resource() methods as far as necessary and in non-PCI bridge drivers as far as feasible without rototilling them. - As NEW_PCIB does a layering violation by activating resources at layers above pci(4) without previously bubbling up their allocation there, move the assignment of bus tags and handles from the bus_alloc_resource() to the bus_activate_resource() methods like at least the other NEW_PCIB enabled architectures do. This is somewhat unfortunate as previously sparc64 (ab)used resource activation to indicate whether SYS_RES_MEMORY resources should be mapped into KVA, which is only necessary if their going to be accessed via the pointer returned from rman_get_virtual() but not for bus_space(9) as the later always uses physical access on sparc64. Besides wasting KVA if we always map in SYS_RES_MEMORY resources, a driver also may deliberately not map them in if the firmware already has done so, possibly in a special way. So in order to still allow a driver to decide whether a SYS_RES_MEMORY resource should be mapped into KVA we let it indicate that by calling bus_space_map(9) with BUS_SPACE_MAP_LINEAR as actually documented in the bus_space(9) page. This is implemented by allocating a separate bus tag per SYS_RES_MEMORY resource and passing the resource via the previously unused bus tag cookie so we later on can call rman_set_virtual() in sparc64_bus_mem_map(). As a side effect this now also allows to actually indicate that a SYS_RES_MEMORY resource should be mapped in as cacheable and/or read-only via BUS_SPACE_MAP_CACHEABLE and BUS_SPACE_MAP_READONLY respectively. - Do some minor cleanup like taking advantage of rman_init_from_resource(), factor out the common part of bus tag allocation into a newly added sparc64_alloc_bus_tag(), hook up some missing newbus methods and replace some homegrown versions with the generic counterparts etc. - While at it, let apb_attach() (which can't use the generic NEW_PCIB code as APB bridges just don't have the base and limit registers implemented) regarding the config space registers cached in pcib_softc and the SYSCTL reporting nodes set up.
2011-10-02 23:22:38 +00:00
if (!passthrough) {
rle = resource_list_find(rl, type, rid);
KASSERT(rle != NULL,
("%s: resource entry not found!", __func__));
KASSERT(rle->res != NULL,
("%s: resource entry is not busy", __func__));
rle->res = NULL;
}
return (0);
}
Make sparc64 compatible with NEW_PCIB and enable it: - Implement bus_adjust_resource() methods as far as necessary and in non-PCI bridge drivers as far as feasible without rototilling them. - As NEW_PCIB does a layering violation by activating resources at layers above pci(4) without previously bubbling up their allocation there, move the assignment of bus tags and handles from the bus_alloc_resource() to the bus_activate_resource() methods like at least the other NEW_PCIB enabled architectures do. This is somewhat unfortunate as previously sparc64 (ab)used resource activation to indicate whether SYS_RES_MEMORY resources should be mapped into KVA, which is only necessary if their going to be accessed via the pointer returned from rman_get_virtual() but not for bus_space(9) as the later always uses physical access on sparc64. Besides wasting KVA if we always map in SYS_RES_MEMORY resources, a driver also may deliberately not map them in if the firmware already has done so, possibly in a special way. So in order to still allow a driver to decide whether a SYS_RES_MEMORY resource should be mapped into KVA we let it indicate that by calling bus_space_map(9) with BUS_SPACE_MAP_LINEAR as actually documented in the bus_space(9) page. This is implemented by allocating a separate bus tag per SYS_RES_MEMORY resource and passing the resource via the previously unused bus tag cookie so we later on can call rman_set_virtual() in sparc64_bus_mem_map(). As a side effect this now also allows to actually indicate that a SYS_RES_MEMORY resource should be mapped in as cacheable and/or read-only via BUS_SPACE_MAP_CACHEABLE and BUS_SPACE_MAP_READONLY respectively. - Do some minor cleanup like taking advantage of rman_init_from_resource(), factor out the common part of bus tag allocation into a newly added sparc64_alloc_bus_tag(), hook up some missing newbus methods and replace some homegrown versions with the generic counterparts etc. - While at it, let apb_attach() (which can't use the generic NEW_PCIB code as APB bridges just don't have the base and limit registers implemented) regarding the config space registers cached in pcib_softc and the SYSCTL reporting nodes set up.
2011-10-02 23:22:38 +00:00
return (resource_list_release(rl, bus, child, type, rid, r));
}
Rototill the sparc64 nexus(4) (actually this brings in the code the sun4v nexus(4) in turn is based on): o Change nexus(4) to manage the resources of its children so the respective device drivers don't need to figure them out of OFW themselves. o Change nexus(4) to provide the ofw_bus KOBJ interface instead of using IVARs for supplying the OFW node and the subset of standard properties of its children. Together with the previous change this also allows to fully take advantage of newbus in that drivers like fhc(4), which attach on multiple parent busses, no longer require different bus front-ends as obtaining the OFW node and properties as well as resource allocation works the same for all supported busses. As such this change also is part 4/4 of allowing creator(4) to work in USIII-based machines as it allows this driver to attach on both nexus(4) and upa(4). On the other hand removing these IVARs breaks API compatibility with the powerpc nexus(4) but which isn't that bad as a) sparc64 currently doesn't share any device driver hanging off of nexus(4) with powerpc and b) they were no longer compatible regarding OFW-related extensions at the pci(4) level since quite some time. o Provide bus_get_dma_tag methods in nexus(4) and its children in order to handle DMA tags in a hierarchical way and get rid of the sparc64_root_dma_tag kludge. Together with the previous two items this changes also allows to completely get rid of the nexus(4) IVAR interface. It also includes: - pushing the constraints previously specified by the nexus_dmatag down into the DMA tags of psycho(4) and sbus(4) as it's their IOMMUs which induce these restrictions (and nothing at the nexus(4) or anything that would warrant specifying them there), - fixing some obviously wrong constraints of the psycho(4) and sbus(4) DMA tags, which happened to not actually be used with the sparc64_root_dma_tag kludge in place and therefore didn't cause problems so far, - replacing magic constants for constraints with macros as far as it is obvious as to where they come from. This doesn't include taking advantage of the newbus way to get the parent DMA tags implemented by this change in order to divorce the IOTSBs of the PCI and SBus IOMMUs or for implementing the workaround for the DMA sync bug in Sabre (and Tomatillo) bridges, yet, though. o Get rid of the notion that nexus(4) (mostly) reflects an UPA bus by replacing ofw_upa.h and with ofw_nexus.h (which was repo-copied from ofw_upa.h) and renaming its content, which actually applies to all of Fireplane/Safari, JBus and UPA (in the host bus case), as appropriate. o Just use M_DEVBUF instead of a separate M_NEXUS malloc type for allocating the device info for the children of nexus(4). This is done in order to not need to export M_NEXUS when deriving drivers for subordinate busses from the nexus(4) class. o Use the DEFINE_CLASS_0() macro to declare the nexus(4) driver so we can derive subclasses from it. o Const'ify the nexus_excl_name and nexus_excl_type arrays as well as add 'associations' and 'rsc', which are pseudo-devices without resources and therefore of no real interest for nexus(4), to the former. o Let the nexus(4) device memory rman manage the entire 64-bit address space instead of just the UPA_MEMSTART to UPA_MEMEND subregion as Fireplane/Safari- and JBus-based machines use multiple ranges, which can't be as easily divided as in the case of UPA (limiting the address space only served for sanity checking anyway). o Use M_WAITOK instead of M_NOWAIT when allocating the device info for children of nexus(4) in order to give one less opportunity for adding devices to nexus(4) to fail. o While adapting the drivers affected by the above nexus(4) changes, change them to take advantage of rman_get_rid() instead of caching the RIDs assigned to allocated resources, now that the RIDs of resources are correctly set. o In iommu(4) and nexus(4) replace hard-coded functions names, which actually became outdated in several places, in panic strings and status massages with __func__. [1] o Use driver_filter_t in prototypes where appropriate. o Add my copyright to creator(4), fhc(4), nexus(4), psycho(4) and sbus(4) as I changed considerable amounts of these drivers as well as added a bunch of new features, workarounds for silicon bugs etc. o Fix some white space nits. Due to lack of access to Exx00 hardware, these changes, i.e. central(4) and fhc(4), couldn't be runtime tested on such a machine. Exx00 are currently reported to panic before trying to attach nexus(4) anyway though. PR: 76052 [1] Approved by: re (kensmith)
2007-03-07 21:13:51 +00:00
static bus_dma_tag_t
sbus_get_dma_tag(device_t bus, device_t child)
{
struct sbus_softc *sc;
sc = device_get_softc(bus);
return (sc->sc_cdmatag);
}
static const struct ofw_bus_devinfo *
sbus_get_devinfo(device_t bus, device_t child)
{
struct sbus_devinfo *sdi;
sdi = device_get_ivars(child);
return (&sdi->sdi_obdinfo);
}
/*
* Handle an overtemp situation.
*
* SPARCs have temperature sensors which generate interrupts
* if the machine's temperature exceeds a certain threshold.
* This handles the interrupt and powers off the machine.
* The same needs to be done to PCI controller drivers.
*/
static void
sbus_overtemp(void *arg)
{
o Revamp the sparc64 interrupt code in order to be able to interface with the INTR_FILTER-enabled MI code. Basically this consists of registering an interrupt controller (of which there can be multiple and optionally different ones either per host-to-foo bridge or shared amongst host-to-foo bridges in any one machine) along with an interrupt vector as specific argument for all the interrupt vectors used by a given host-to-foo bridge (roughly similar to registering interrupt sources on amd64 and i386), providing functions to enable, clear and disable the interrupts of the children beneath the bridge. This also includes: - No longer entering a critical section in tl0_intr() and tl1_intr() for executing interrupt handlers but rather let the handlers enter it themselves so in the case of intr_event_handle() we don't enter a nested critical section. - Adding infrastructure for binding delivery of interrupt vectors to specific CPUs which later on can be interfaced with the code from amd64/i386 for binding interrupts to specific CPUs. - Getting rid of the wrapper hack introduced along the lines of the API changes for INTR_FILTER which as a side-effect caused interrupts associated with ithread handlers only to get the elevated priority of those associated with filters ("fast handlers") (this removes the hack also in the non-INTR_FILTER case). - Disabling (by not clearing) an interrupt in the interrupt controller until all associated handlers have been executed, which is crucial for the typical locking strategy of NIC drivers in order to work correctly in case of shared interrupts. This was a more or less theoretical problem on sparc64 though, as shared interrupts are rather uncommon there except for the on-board SCCs and UARTs. Note that due to the behavior of at least of some of the interrupt controllers used on sparc64 an enable+EOI instead of a disable+EOI approach (as implied by the INTR_FILTER MI code and implemented on other architectures) is used as the latter can cause lost interrupts or in the worst case interrupt starvation. o Correct a typo in sbus_alloc_resource() which caused (pass-through) allocations to only work down to the grandchildren of the bus, which wasn't a real problem so far as we don't support any devices which are great-grandchildren or greater of a U2S bridge, yet. o In fhc(4) use bus_{read,write}_4() instead of bus_space_{read,write}_4() in order to get rid of sc_bh and sc_bt in the fhc_softc. Also get rid of some other unneeded members in fhc_softc. Reviewed by: marcel (earlier version) Approved by: re (kensmith)
2007-09-06 19:16:30 +00:00
static int shutdown;
o Revamp the sparc64 interrupt code in order to be able to interface with the INTR_FILTER-enabled MI code. Basically this consists of registering an interrupt controller (of which there can be multiple and optionally different ones either per host-to-foo bridge or shared amongst host-to-foo bridges in any one machine) along with an interrupt vector as specific argument for all the interrupt vectors used by a given host-to-foo bridge (roughly similar to registering interrupt sources on amd64 and i386), providing functions to enable, clear and disable the interrupts of the children beneath the bridge. This also includes: - No longer entering a critical section in tl0_intr() and tl1_intr() for executing interrupt handlers but rather let the handlers enter it themselves so in the case of intr_event_handle() we don't enter a nested critical section. - Adding infrastructure for binding delivery of interrupt vectors to specific CPUs which later on can be interfaced with the code from amd64/i386 for binding interrupts to specific CPUs. - Getting rid of the wrapper hack introduced along the lines of the API changes for INTR_FILTER which as a side-effect caused interrupts associated with ithread handlers only to get the elevated priority of those associated with filters ("fast handlers") (this removes the hack also in the non-INTR_FILTER case). - Disabling (by not clearing) an interrupt in the interrupt controller until all associated handlers have been executed, which is crucial for the typical locking strategy of NIC drivers in order to work correctly in case of shared interrupts. This was a more or less theoretical problem on sparc64 though, as shared interrupts are rather uncommon there except for the on-board SCCs and UARTs. Note that due to the behavior of at least of some of the interrupt controllers used on sparc64 an enable+EOI instead of a disable+EOI approach (as implied by the INTR_FILTER MI code and implemented on other architectures) is used as the latter can cause lost interrupts or in the worst case interrupt starvation. o Correct a typo in sbus_alloc_resource() which caused (pass-through) allocations to only work down to the grandchildren of the bus, which wasn't a real problem so far as we don't support any devices which are great-grandchildren or greater of a U2S bridge, yet. o In fhc(4) use bus_{read,write}_4() instead of bus_space_{read,write}_4() in order to get rid of sc_bh and sc_bt in the fhc_softc. Also get rid of some other unneeded members in fhc_softc. Reviewed by: marcel (earlier version) Approved by: re (kensmith)
2007-09-06 19:16:30 +00:00
/* As the interrupt is cleared we may be called multiple times. */
if (shutdown != 0)
return;
shutdown++;
printf("DANGER: OVER TEMPERATURE detected\nShutting down NOW.\n");
shutdown_nice(RB_POWEROFF);
}
/* Try to shut down in time in case of power failure. */
static void
sbus_pwrfail(void *arg)
{
o Revamp the sparc64 interrupt code in order to be able to interface with the INTR_FILTER-enabled MI code. Basically this consists of registering an interrupt controller (of which there can be multiple and optionally different ones either per host-to-foo bridge or shared amongst host-to-foo bridges in any one machine) along with an interrupt vector as specific argument for all the interrupt vectors used by a given host-to-foo bridge (roughly similar to registering interrupt sources on amd64 and i386), providing functions to enable, clear and disable the interrupts of the children beneath the bridge. This also includes: - No longer entering a critical section in tl0_intr() and tl1_intr() for executing interrupt handlers but rather let the handlers enter it themselves so in the case of intr_event_handle() we don't enter a nested critical section. - Adding infrastructure for binding delivery of interrupt vectors to specific CPUs which later on can be interfaced with the code from amd64/i386 for binding interrupts to specific CPUs. - Getting rid of the wrapper hack introduced along the lines of the API changes for INTR_FILTER which as a side-effect caused interrupts associated with ithread handlers only to get the elevated priority of those associated with filters ("fast handlers") (this removes the hack also in the non-INTR_FILTER case). - Disabling (by not clearing) an interrupt in the interrupt controller until all associated handlers have been executed, which is crucial for the typical locking strategy of NIC drivers in order to work correctly in case of shared interrupts. This was a more or less theoretical problem on sparc64 though, as shared interrupts are rather uncommon there except for the on-board SCCs and UARTs. Note that due to the behavior of at least of some of the interrupt controllers used on sparc64 an enable+EOI instead of a disable+EOI approach (as implied by the INTR_FILTER MI code and implemented on other architectures) is used as the latter can cause lost interrupts or in the worst case interrupt starvation. o Correct a typo in sbus_alloc_resource() which caused (pass-through) allocations to only work down to the grandchildren of the bus, which wasn't a real problem so far as we don't support any devices which are great-grandchildren or greater of a U2S bridge, yet. o In fhc(4) use bus_{read,write}_4() instead of bus_space_{read,write}_4() in order to get rid of sc_bh and sc_bt in the fhc_softc. Also get rid of some other unneeded members in fhc_softc. Reviewed by: marcel (earlier version) Approved by: re (kensmith)
2007-09-06 19:16:30 +00:00
static int shutdown;
o Revamp the sparc64 interrupt code in order to be able to interface with the INTR_FILTER-enabled MI code. Basically this consists of registering an interrupt controller (of which there can be multiple and optionally different ones either per host-to-foo bridge or shared amongst host-to-foo bridges in any one machine) along with an interrupt vector as specific argument for all the interrupt vectors used by a given host-to-foo bridge (roughly similar to registering interrupt sources on amd64 and i386), providing functions to enable, clear and disable the interrupts of the children beneath the bridge. This also includes: - No longer entering a critical section in tl0_intr() and tl1_intr() for executing interrupt handlers but rather let the handlers enter it themselves so in the case of intr_event_handle() we don't enter a nested critical section. - Adding infrastructure for binding delivery of interrupt vectors to specific CPUs which later on can be interfaced with the code from amd64/i386 for binding interrupts to specific CPUs. - Getting rid of the wrapper hack introduced along the lines of the API changes for INTR_FILTER which as a side-effect caused interrupts associated with ithread handlers only to get the elevated priority of those associated with filters ("fast handlers") (this removes the hack also in the non-INTR_FILTER case). - Disabling (by not clearing) an interrupt in the interrupt controller until all associated handlers have been executed, which is crucial for the typical locking strategy of NIC drivers in order to work correctly in case of shared interrupts. This was a more or less theoretical problem on sparc64 though, as shared interrupts are rather uncommon there except for the on-board SCCs and UARTs. Note that due to the behavior of at least of some of the interrupt controllers used on sparc64 an enable+EOI instead of a disable+EOI approach (as implied by the INTR_FILTER MI code and implemented on other architectures) is used as the latter can cause lost interrupts or in the worst case interrupt starvation. o Correct a typo in sbus_alloc_resource() which caused (pass-through) allocations to only work down to the grandchildren of the bus, which wasn't a real problem so far as we don't support any devices which are great-grandchildren or greater of a U2S bridge, yet. o In fhc(4) use bus_{read,write}_4() instead of bus_space_{read,write}_4() in order to get rid of sc_bh and sc_bt in the fhc_softc. Also get rid of some other unneeded members in fhc_softc. Reviewed by: marcel (earlier version) Approved by: re (kensmith)
2007-09-06 19:16:30 +00:00
/* As the interrupt is cleared we may be called multiple times. */
if (shutdown != 0)
return;
shutdown++;
printf("Power failure detected\nShutting down NOW.\n");
shutdown_nice(0);
}
static int
sbus_print_res(struct sbus_devinfo *sdi)
- Introduce an ofw_bus kobj-interface for retrieving the OFW node and a subset ("compatible", "device_type", "model" and "name") of the standard properties in drivers for devices on Open Firmware supported busses. The standard properties "reg", "interrupts" und "address" are not covered by this interface because they are only of interest in the respective bridge code. There's a remaining standard property "status" which is unclear how to support properly but which also isn't used in FreeBSD at present. This ofw_bus kobj-interface allows to replace the various (ebus_get_node(), ofw_pci_get_node(), etc.) and partially inconsistent (central_get_type() vs. sbus_get_device_type(), etc.) existing IVAR ones with a common one. This in turn allows to simplify and remove code-duplication in drivers for devices that can hang off of more than one OFW supported bus. - Convert the sparc64 Central, EBus, FHC, PCI and SBus bus drivers and the drivers for their children to use the ofw_bus kobj-interface. The IVAR- interfaces of the Central, EBus and FHC are entirely replaced by this. The PCI bus driver used its own kobj-interface and now also uses the ofw_bus one. The IVARs special to the SBus, e.g. for retrieving the burst size, remain. Beware: this causes an ABI-breakage for modules of drivers which used the IVAR-interfaces, i.e. esp(4), hme(4), isp(4) and uart(4), which need to be recompiled. The style-inconsistencies introduced in some of the bus drivers will be fixed by tmm@ in a generic clean-up of the respective drivers later (he requested to add the changes in the "new" style). - Convert the powerpc MacIO bus driver and the drivers for its children to use the ofw_bus kobj-interface. This invloves removing the IVARs related to the "reg" property which were unused and a leftover from the NetBSD origini of the code. There's no ABI-breakage caused by this because none of these driver are currently built as modules. There are other powerpc bus drivers which can be converted to the ofw_bus kobj-interface, e.g. the PCI bus driver, which should be done together with converting powerpc to use the OFW PCI code from sparc64. - Make the SBus and FHC front-end of zs(4) and the sparc64 eeprom(4) take advantage of the ofw_bus kobj-interface and simplify them a bit. Reviewed by: grehan, tmm Approved by: re (scottl) Discussed with: tmm Tested with: Sun AX1105, AXe, Ultra 2, Ultra 60; PPC cross-build on i386
2004-08-12 17:41:33 +00:00
{
int rv;
- Introduce an ofw_bus kobj-interface for retrieving the OFW node and a subset ("compatible", "device_type", "model" and "name") of the standard properties in drivers for devices on Open Firmware supported busses. The standard properties "reg", "interrupts" und "address" are not covered by this interface because they are only of interest in the respective bridge code. There's a remaining standard property "status" which is unclear how to support properly but which also isn't used in FreeBSD at present. This ofw_bus kobj-interface allows to replace the various (ebus_get_node(), ofw_pci_get_node(), etc.) and partially inconsistent (central_get_type() vs. sbus_get_device_type(), etc.) existing IVAR ones with a common one. This in turn allows to simplify and remove code-duplication in drivers for devices that can hang off of more than one OFW supported bus. - Convert the sparc64 Central, EBus, FHC, PCI and SBus bus drivers and the drivers for their children to use the ofw_bus kobj-interface. The IVAR- interfaces of the Central, EBus and FHC are entirely replaced by this. The PCI bus driver used its own kobj-interface and now also uses the ofw_bus one. The IVARs special to the SBus, e.g. for retrieving the burst size, remain. Beware: this causes an ABI-breakage for modules of drivers which used the IVAR-interfaces, i.e. esp(4), hme(4), isp(4) and uart(4), which need to be recompiled. The style-inconsistencies introduced in some of the bus drivers will be fixed by tmm@ in a generic clean-up of the respective drivers later (he requested to add the changes in the "new" style). - Convert the powerpc MacIO bus driver and the drivers for its children to use the ofw_bus kobj-interface. This invloves removing the IVARs related to the "reg" property which were unused and a leftover from the NetBSD origini of the code. There's no ABI-breakage caused by this because none of these driver are currently built as modules. There are other powerpc bus drivers which can be converted to the ofw_bus kobj-interface, e.g. the PCI bus driver, which should be done together with converting powerpc to use the OFW PCI code from sparc64. - Make the SBus and FHC front-end of zs(4) and the sparc64 eeprom(4) take advantage of the ofw_bus kobj-interface and simplify them a bit. Reviewed by: grehan, tmm Approved by: re (scottl) Discussed with: tmm Tested with: Sun AX1105, AXe, Ultra 2, Ultra 60; PPC cross-build on i386
2004-08-12 17:41:33 +00:00
rv = 0;
rv += resource_list_print_type(&sdi->sdi_rl, "mem", SYS_RES_MEMORY,
"%#lx");
rv += resource_list_print_type(&sdi->sdi_rl, "irq", SYS_RES_IRQ,
"%ld");
return (rv);
- Introduce an ofw_bus kobj-interface for retrieving the OFW node and a subset ("compatible", "device_type", "model" and "name") of the standard properties in drivers for devices on Open Firmware supported busses. The standard properties "reg", "interrupts" und "address" are not covered by this interface because they are only of interest in the respective bridge code. There's a remaining standard property "status" which is unclear how to support properly but which also isn't used in FreeBSD at present. This ofw_bus kobj-interface allows to replace the various (ebus_get_node(), ofw_pci_get_node(), etc.) and partially inconsistent (central_get_type() vs. sbus_get_device_type(), etc.) existing IVAR ones with a common one. This in turn allows to simplify and remove code-duplication in drivers for devices that can hang off of more than one OFW supported bus. - Convert the sparc64 Central, EBus, FHC, PCI and SBus bus drivers and the drivers for their children to use the ofw_bus kobj-interface. The IVAR- interfaces of the Central, EBus and FHC are entirely replaced by this. The PCI bus driver used its own kobj-interface and now also uses the ofw_bus one. The IVARs special to the SBus, e.g. for retrieving the burst size, remain. Beware: this causes an ABI-breakage for modules of drivers which used the IVAR-interfaces, i.e. esp(4), hme(4), isp(4) and uart(4), which need to be recompiled. The style-inconsistencies introduced in some of the bus drivers will be fixed by tmm@ in a generic clean-up of the respective drivers later (he requested to add the changes in the "new" style). - Convert the powerpc MacIO bus driver and the drivers for its children to use the ofw_bus kobj-interface. This invloves removing the IVARs related to the "reg" property which were unused and a leftover from the NetBSD origini of the code. There's no ABI-breakage caused by this because none of these driver are currently built as modules. There are other powerpc bus drivers which can be converted to the ofw_bus kobj-interface, e.g. the PCI bus driver, which should be done together with converting powerpc to use the OFW PCI code from sparc64. - Make the SBus and FHC front-end of zs(4) and the sparc64 eeprom(4) take advantage of the ofw_bus kobj-interface and simplify them a bit. Reviewed by: grehan, tmm Approved by: re (scottl) Discussed with: tmm Tested with: Sun AX1105, AXe, Ultra 2, Ultra 60; PPC cross-build on i386
2004-08-12 17:41:33 +00:00
}