freebsd-dev/sys/sparc64/pci/psychoreg.h

329 lines
16 KiB
C
Raw Normal View History

/*-
* Copyright (c) 1999 Matthew R. Green
* 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.
*/
/*-
* Copyright (c) 1998, 1999 Eduardo E. 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: psychoreg.h,v 1.14 2008/05/30 02:29:37 mrg Exp
*
* $FreeBSD$
*/
#ifndef _SPARC64_PCI_PSYCHOREG_H_
#define _SPARC64_PCI_PSYCHOREG_H_
/*
* Sun4u PCI definitions. Here's where we deal w/the machine
* dependencies of Psycho and the PCI controller on the UltraIIi.
*
* All PCI registers are bit-swapped, however they are not byte-swapped.
* This means that they must be accessed using little-endian access modes,
* either map the pages little-endian or use little-endian ASIs.
*
* PSYCHO implements two PCI buses, A and B.
*/
#define PSYCHO_NINTR 6
#define PSYCHO_NRANGE 4
/*
* Psycho register offsets
*
* NB: FFB0 and FFB1 intr map regs also appear at 0x6000 and 0x8000
* respectively.
*/
#define PSR_UPA_PORTID 0x0000 /* UPA port ID register */
#define PSR_UPA_CONFIG 0x0008 /* UPA config register */
#define PSR_CS 0x0010 /* PSYCHO control/status register */
#define PSR_ECCC 0x0020 /* ECC control register */
#define PSR_UE_AFS 0x0030 /* Uncorrectable Error AFSR */
#define PSR_UE_AFA 0x0038 /* Uncorrectable Error AFAR */
#define PSR_CE_AFS 0x0040 /* Correctable Error AFSR */
#define PSR_CE_AFA 0x0048 /* Correctable Error AFAR */
#define PSR_PM_CTL 0x0100 /* Performance monitor control reg */
#define PSR_PM_COUNT 0x0108 /* Performance monitor counter reg */
#define PSR_IOMMU 0x0200 /* IOMMU registers */
#define PSR_PCIA0_INT_MAP 0x0c00 /* PCI bus a slot 0 irq map reg */
#define PSR_PCIA1_INT_MAP 0x0c08 /* PCI bus a slot 1 irq map reg */
#define PSR_PCIA2_INT_MAP 0x0c10 /* PCI bus a slot 2 irq map reg (IIi) */
#define PSR_PCIA3_INT_MAP 0x0c18 /* PCI bus a slot 3 irq map reg (IIi) */
#define PSR_PCIB0_INT_MAP 0x0c20 /* PCI bus b slot 0 irq map reg */
#define PSR_PCIB1_INT_MAP 0x0c28 /* PCI bus b slot 1 irq map reg */
#define PSR_PCIB2_INT_MAP 0x0c30 /* PCI bus b slot 2 irq map reg */
#define PSR_PCIB3_INT_MAP 0x0c38 /* PCI bus b slot 3 irq map reg */
#define PSR_SCSI_INT_MAP 0x1000 /* SCSI interrupt map reg */
#define PSR_ETHER_INT_MAP 0x1008 /* ethernet interrupt map reg */
#define PSR_BPP_INT_MAP 0x1010 /* parallel interrupt map reg */
#define PSR_AUDIOR_INT_MAP 0x1018 /* audio record interrupt map reg */
#define PSR_AUDIOP_INT_MAP 0x1020 /* audio playback interrupt map reg */
#define PSR_POWER_INT_MAP 0x1028 /* power fail interrupt map reg */
#define PSR_SKBDMS_INT_MAP 0x1030 /* serial/kbd/mouse interrupt map reg */
#define PSR_FD_INT_MAP 0x1038 /* floppy interrupt map reg */
#define PSR_SPARE_INT_MAP 0x1040 /* spare interrupt map reg */
#define PSR_KBD_INT_MAP 0x1048 /* kbd [unused] interrupt map reg */
#define PSR_MOUSE_INT_MAP 0x1050 /* mouse [unused] interrupt map reg */
#define PSR_SERIAL_INT_MAP 0x1058 /* second serial interrupt map reg */
#define PSR_TIMER0_INT_MAP 0x1060 /* timer 0 interrupt map reg */
#define PSR_TIMER1_INT_MAP 0x1068 /* timer 1 interrupt map reg */
#define PSR_UE_INT_MAP 0x1070 /* UE interrupt map reg */
#define PSR_CE_INT_MAP 0x1078 /* CE interrupt map reg */
#define PSR_PCIAERR_INT_MAP 0x1080 /* PCI bus a error interrupt map reg */
#define PSR_PCIBERR_INT_MAP 0x1088 /* PCI bus b error interrupt map reg */
#define PSR_PWRMGT_INT_MAP 0x1090 /* power mgmt wake interrupt map reg */
#define PSR_FFB0_INT_MAP 0x1098 /* FFB0 graphics interrupt map reg */
#define PSR_FFB1_INT_MAP 0x10a0 /* FFB1 graphics interrupt map reg */
/* Note: clear interrupt 0 registers are not really used */
#define PSR_PCIA0_INT_CLR 0x1400 /* PCI a slot 0 clear int regs 0..3 */
#define PSR_PCIA1_INT_CLR 0x1420 /* PCI a slot 1 clear int regs 0..3 */
#define PSR_PCIA2_INT_CLR 0x1440 /* PCI a slot 2 clear int regs 0..3 */
#define PSR_PCIA3_INT_CLR 0x1460 /* PCI a slot 3 clear int regs 0..3 */
#define PSR_PCIB0_INT_CLR 0x1480 /* PCI b slot 0 clear int regs 0..3 */
#define PSR_PCIB1_INT_CLR 0x14a0 /* PCI b slot 1 clear int regs 0..3 */
#define PSR_PCIB2_INT_CLR 0x14c0 /* PCI b slot 2 clear int regs 0..3 */
#define PSR_PCIB3_INT_CLR 0x14d0 /* PCI b slot 3 clear int regs 0..3 */
#define PSR_SCSI_INT_CLR 0x1800 /* SCSI clear int reg */
#define PSR_ETHER_INT_CLR 0x1808 /* ethernet clear int reg */
#define PSR_BPP_INT_CLR 0x1810 /* parallel clear int reg */
#define PSR_AUDIOR_INT_CLR 0x1818 /* audio record clear int reg */
#define PSR_AUDIOP_INT_CLR 0x1820 /* audio playback clear int reg */
#define PSR_POWER_INT_CLR 0x1828 /* power fail clear int reg */
#define PSR_SKBDMS_INT_CLR 0x1830 /* serial/kbd/mouse clear int reg */
#define PSR_FD_INT_CLR 0x1838 /* floppy clear int reg */
#define PSR_SPARE_INT_CLR 0x1840 /* spare clear int reg */
#define PSR_KBD_INT_CLR 0x1848 /* kbd [unused] clear int reg */
#define PSR_MOUSE_INT_CLR 0x1850 /* mouse [unused] clear int reg */
#define PSR_SERIAL_INT_CLR 0x1858 /* second serial clear int reg */
#define PSR_TIMER0_INT_CLR 0x1860 /* timer 0 clear int reg */
#define PSR_TIMER1_INT_CLR 0x1868 /* timer 1 clear int reg */
#define PSR_UE_INT_CLR 0x1870 /* UE clear int reg */
#define PSR_CE_INT_CLR 0x1878 /* CE clear int reg */
#define PSR_PCIAERR_INT_CLR 0x1880 /* PCI bus a error clear int reg */
#define PSR_PCIBERR_INT_CLR 0x1888 /* PCI bus b error clear int reg */
#define PSR_PWRMGT_INT_CLR 0x1890 /* power mgmt wake clr interrupt reg */
#define PSR_INTR_RETRY_TIM 0x1a00 /* interrupt retry timer */
#define PSR_TC0 0x1c00 /* timer/counter 0 */
#define PSR_TC1 0x1c10 /* timer/counter 1 */
#define PSR_DMA_WRITE_SYNC 0x1c20 /* PCI DMA write sync register (IIi) */
#define PSR_PCICTL0 0x2000 /* PCICTL registers for 1st Psycho */
#define PSR_PCICTL1 0x4000 /* PCICTL registers for 2nd Psycho */
#define PSR_DMA_SCB_DIAG0 0xa000 /* DMA scoreboard diag reg 0 */
#define PSR_DMA_SCB_DIAG1 0xa008 /* DMA scoreboard diag reg 1 */
#define PSR_IOMMU_SVADIAG 0xa400 /* IOMMU virtual addr diag reg */
#define PSR_IOMMU_TLB_CMP_DIAG 0xa408 /* IOMMU TLB tag compare diag reg */
#define PSR_IOMMU_QUEUE_DIAG 0xa500 /* IOMMU LRU queue diag regs 0..15 */
#define PSR_IOMMU_TLB_TAG_DIAG 0xa580 /* TLB tag diag regs 0..15 */
#define PSR_IOMMU_TLB_DATA_DIAG 0xa600 /* TLB data RAM diag regs 0..15 */
#define PSR_PCI_INT_DIAG 0xa800 /* PCI int state diag reg */
#define PSR_OBIO_INT_DIAG 0xa808 /* OBIO and misc int state diag reg */
#define PSR_STRBUF_DIAG 0xb000 /* Streaming buffer diag regs */
/*
* Here is the rest of the map, which we're not specifying:
*
* 1fe.0100.0000 - 1fe.01ff.ffff PCI configuration space
* 1fe.0100.0000 - 1fe.0100.00ff PCI B configuration header
* 1fe.0101.0000 - 1fe.0101.00ff PCI A configuration header
* 1fe.0200.0000 - 1fe.0200.ffff PCI A I/O space
* 1fe.0201.0000 - 1fe.0201.ffff PCI B I/O space
* 1ff.0000.0000 - 1ff.7fff.ffff PCI A memory space
* 1ff.8000.0000 - 1ff.ffff.ffff PCI B memory space
*
* NB: config and I/O space can use 1-4 byte accesses, not 8 byte
* accesses. Memory space can use any sized accesses.
*
* Note that the SUNW,sabre/SUNW,simba combinations found on the
* Ultra5 and Ultra10 machines uses slightly differrent addresses
* than the above. This is mostly due to the fact that the APB is
* a multi-function PCI device with two PCI bridges, and the U2P is
* two separate PCI bridges. It uses the same PCI configuration
* space, though the configuration header for each PCI bus is
* located differently due to the SUNW,simba PCI busses being
* function 0 and function 1 of the APB, whereas the Psycho's are
* each their own PCI device. The I/O and memory spaces are each
* split into 8 equally sized areas (8x2MB blocks for I/O space,
* and 8x512MB blocks for memory space). These are allocated in to
* either PCI A or PCI B, or neither in the APB's `I/O Address Map
* Register A/B' (0xde) and `Memory Address Map Register A/B' (0xdf)
* registers of each Simba. We must ensure that both of the
* following are correct (the prom should do this for us):
*
* (PCI A Memory Address Map) & (PCI B Memory Address Map) == 0
*
* (PCI A I/O Address Map) & (PCI B I/O Address Map) == 0
*
* 1fe.0100.0000 - 1fe.01ff.ffff PCI configuration space
* 1fe.0100.0800 - 1fe.0100.08ff PCI B configuration header
* 1fe.0100.0900 - 1fe.0100.09ff PCI A configuration header
* 1fe.0200.0000 - 1fe.02ff.ffff PCI I/O space (divided)
* 1ff.0000.0000 - 1ff.ffff.ffff PCI memory space (divided)
*/
/*
* PSR_CS defines:
*
* 63 59 55 50 45 4 3 2 1 0
* +------+------+------+------+--//---+--------+-------+-----+------+
* | IMPL | VERS | MID | IGN | xxx | APCKEN | APERR | IAP | MODE |
* +------+------+------+------+--//---+--------+-------+-----+------+
*
*/
#define PSYCHO_GCSR_IMPL(csr) ((u_int)(((csr) >> 60) & 0xf))
#define PSYCHO_GCSR_VERS(csr) ((u_int)(((csr) >> 56) & 0xf))
#define PSYCHO_GCSR_MID(csr) ((u_int)(((csr) >> 51) & 0x1f))
#define PSYCHO_GCSR_IGN(csr) ((u_int)(((csr) >> 46) & 0x1f))
#define PSYCHO_CSR_APCKEN 8 /* UPA addr parity check enable */
#define PSYCHO_CSR_APERR 4 /* UPA addr parity error */
#define PSYCHO_CSR_IAP 2 /* invert UPA address parity */
#define PSYCHO_CSR_MODE 1 /* UPA/PCI handshake */
/* Offsets into the PSR_PCICTL* register block */
#define PCR_CS 0x0000 /* PCI control/status register */
#define PCR_AFS 0x0010 /* PCI AFSR register */
#define PCR_AFA 0x0018 /* PCI AFAR register */
#define PCR_DIAG 0x0020 /* PCI diagnostic register */
#define PCR_TAS 0x0028 /* PCI target address space reg (IIi) */
#define PCR_STRBUF 0x0800 /* IOMMU streaming buffer registers. */
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
/* INO defines */
#define PSYCHO_MAX_INO 0x3f
/* Device space defines */
#define PSYCHO_CONF_SIZE 0x1000000
#define PSYCHO_CONF_BUS_SHIFT 16
#define PSYCHO_CONF_DEV_SHIFT 11
#define PSYCHO_CONF_FUNC_SHIFT 8
#define PSYCHO_CONF_REG_SHIFT 0
#define PSYCHO_IO_SIZE 0x1000000
#define PSYCHO_MEM_SIZE 0x100000000
#define PSYCHO_CONF_OFF(bus, slot, func, reg) \
(((bus) << PSYCHO_CONF_BUS_SHIFT) | \
((slot) << PSYCHO_CONF_DEV_SHIFT) | \
((func) << PSYCHO_CONF_FUNC_SHIFT) | \
((reg) << PSYCHO_CONF_REG_SHIFT))
/* what the bits mean! */
/*
* PCI [a|b] control/status register
* Note that the Hummingbird/Sabre only has one set of PCI control/status
* registers.
*/
o Changes to psycho_attach(): [1] - Clear the PCI AFSR and status error bits as previous errors still might be indicated. - Set up the PCI control and diagnostic registers according to the capabilities, workarounds, etc of/for specific revisions of the supported bridges. This includes no longer setting Hummingbird-/ Sabre-specific bits in the PCI control register but preserving what the firmware has initialized them to like OpenSolaris does. Previously we were setting these bits according to the example in the Sabre documentation, which I doubt is appropriate for all Sabre based designs and especially not for Hummingbirds. This also includes not enabling bus parking unless the firmware tells us to. - Set the PCI latency timer register as this isn't always done by the firmware. o Remove a redundant argument from psycho_set_intr() and in this function check the return value of bus_setup_intr(). [2] o Let psycho_setup_intr() return ENOMEM instead of 0 when it can't allocate memory for the interrupt wrapper stub and EINVAL instead of 0 if it can't find the interrupt vector in the interrupt map. o Add a workaround for a bug of the Sabre-APB-combination where it doesn't drain DMA write data for devices behind additional PCI-PCI bridges underneath the APB PCI-PCI bridge. This workaround (do things necessary in order to achieve a manual drain when coherency is required) is currently implemented in psycho_setup_intr() and psycho_intr_stub() (for easy MFC'ing) and therefore is only applied for interrupt handlers. This should be moved to psycho(4)-specific bus_dma_tag_create() and bus_dmamap_sync() methods, respectively, once this driver is converted to make use of BUS_GET_DMA_TAG(), so the workaround is also applied for polling(4) callbacks. [3] o Fix some minor style issues. Info from: OpenSolaris [1] Info from: Linux, OpenBSD, OpenSolaris [3] Suggested by: Coverity Prevent (CID 682) [2] MFC after: 1 month
2007-01-08 01:26:47 +00:00
#define PCICTL_SBHERR 0x0000000800000000 /* strm. byte hole error; W1C */
#define PCICTL_SERR 0x0000000400000000 /* SERR asserted; W1C */
o Changes to psycho_attach(): [1] - Clear the PCI AFSR and status error bits as previous errors still might be indicated. - Set up the PCI control and diagnostic registers according to the capabilities, workarounds, etc of/for specific revisions of the supported bridges. This includes no longer setting Hummingbird-/ Sabre-specific bits in the PCI control register but preserving what the firmware has initialized them to like OpenSolaris does. Previously we were setting these bits according to the example in the Sabre documentation, which I doubt is appropriate for all Sabre based designs and especially not for Hummingbirds. This also includes not enabling bus parking unless the firmware tells us to. - Set the PCI latency timer register as this isn't always done by the firmware. o Remove a redundant argument from psycho_set_intr() and in this function check the return value of bus_setup_intr(). [2] o Let psycho_setup_intr() return ENOMEM instead of 0 when it can't allocate memory for the interrupt wrapper stub and EINVAL instead of 0 if it can't find the interrupt vector in the interrupt map. o Add a workaround for a bug of the Sabre-APB-combination where it doesn't drain DMA write data for devices behind additional PCI-PCI bridges underneath the APB PCI-PCI bridge. This workaround (do things necessary in order to achieve a manual drain when coherency is required) is currently implemented in psycho_setup_intr() and psycho_intr_stub() (for easy MFC'ing) and therefore is only applied for interrupt handlers. This should be moved to psycho(4)-specific bus_dma_tag_create() and bus_dmamap_sync() methods, respectively, once this driver is converted to make use of BUS_GET_DMA_TAG(), so the workaround is also applied for polling(4) callbacks. [3] o Fix some minor style issues. Info from: OpenSolaris [1] Info from: Linux, OpenBSD, OpenSolaris [3] Suggested by: Coverity Prevent (CID 682) [2] MFC after: 1 month
2007-01-08 01:26:47 +00:00
#define PCICTL_PCISPEED 0x0000000200000000 /* 0:half 1:full bus speed */
#define PCICTL_ARB_PARK 0x0000000000200000 /* PCI arbitration parking */
o Changes to psycho_attach(): [1] - Clear the PCI AFSR and status error bits as previous errors still might be indicated. - Set up the PCI control and diagnostic registers according to the capabilities, workarounds, etc of/for specific revisions of the supported bridges. This includes no longer setting Hummingbird-/ Sabre-specific bits in the PCI control register but preserving what the firmware has initialized them to like OpenSolaris does. Previously we were setting these bits according to the example in the Sabre documentation, which I doubt is appropriate for all Sabre based designs and especially not for Hummingbirds. This also includes not enabling bus parking unless the firmware tells us to. - Set the PCI latency timer register as this isn't always done by the firmware. o Remove a redundant argument from psycho_set_intr() and in this function check the return value of bus_setup_intr(). [2] o Let psycho_setup_intr() return ENOMEM instead of 0 when it can't allocate memory for the interrupt wrapper stub and EINVAL instead of 0 if it can't find the interrupt vector in the interrupt map. o Add a workaround for a bug of the Sabre-APB-combination where it doesn't drain DMA write data for devices behind additional PCI-PCI bridges underneath the APB PCI-PCI bridge. This workaround (do things necessary in order to achieve a manual drain when coherency is required) is currently implemented in psycho_setup_intr() and psycho_intr_stub() (for easy MFC'ing) and therefore is only applied for interrupt handlers. This should be moved to psycho(4)-specific bus_dma_tag_create() and bus_dmamap_sync() methods, respectively, once this driver is converted to make use of BUS_GET_DMA_TAG(), so the workaround is also applied for polling(4) callbacks. [3] o Fix some minor style issues. Info from: OpenSolaris [1] Info from: Linux, OpenBSD, OpenSolaris [3] Suggested by: Coverity Prevent (CID 682) [2] MFC after: 1 month
2007-01-08 01:26:47 +00:00
#define PCICTL_SBHINTEN 0x0000000000000400 /* strm. byte hole int. en. */
#define PCICTL_WAKEUPEN 0x0000000000000200 /* power mgmt. wakeup enable */
#define PCICTL_ERRINTEN 0x0000000000000100 /* PCI error interrupt enable */
o Changes to psycho_attach(): [1] - Clear the PCI AFSR and status error bits as previous errors still might be indicated. - Set up the PCI control and diagnostic registers according to the capabilities, workarounds, etc of/for specific revisions of the supported bridges. This includes no longer setting Hummingbird-/ Sabre-specific bits in the PCI control register but preserving what the firmware has initialized them to like OpenSolaris does. Previously we were setting these bits according to the example in the Sabre documentation, which I doubt is appropriate for all Sabre based designs and especially not for Hummingbirds. This also includes not enabling bus parking unless the firmware tells us to. - Set the PCI latency timer register as this isn't always done by the firmware. o Remove a redundant argument from psycho_set_intr() and in this function check the return value of bus_setup_intr(). [2] o Let psycho_setup_intr() return ENOMEM instead of 0 when it can't allocate memory for the interrupt wrapper stub and EINVAL instead of 0 if it can't find the interrupt vector in the interrupt map. o Add a workaround for a bug of the Sabre-APB-combination where it doesn't drain DMA write data for devices behind additional PCI-PCI bridges underneath the APB PCI-PCI bridge. This workaround (do things necessary in order to achieve a manual drain when coherency is required) is currently implemented in psycho_setup_intr() and psycho_intr_stub() (for easy MFC'ing) and therefore is only applied for interrupt handlers. This should be moved to psycho(4)-specific bus_dma_tag_create() and bus_dmamap_sync() methods, respectively, once this driver is converted to make use of BUS_GET_DMA_TAG(), so the workaround is also applied for polling(4) callbacks. [3] o Fix some minor style issues. Info from: OpenSolaris [1] Info from: Linux, OpenBSD, OpenSolaris [3] Suggested by: Coverity Prevent (CID 682) [2] MFC after: 1 month
2007-01-08 01:26:47 +00:00
#define PCICTL_ARB_4 0x000000000000000f /* DVMA arb. 4 PCI slots mask */
#define PCICTL_ARB_6 0x000000000000003f /* DVMA arb. 6 PCI slots mask */
/* The following are Hummingbird/Sabre only. */
#define PCICTL_MRLM 0x0000001000000000 /* Memory Read Line/Multiple */
#define PCICTL_CPU_PRIO 0x0000000000100000 /* CPU extra arb. prio. en. */
#define PCICTL_ARB_PRIO 0x00000000000f0000 /* PCI extra arb. prio. en. */
#define PCICTL_RTRYWAIT 0x0000000000000080 /* 0:wait 1:retry DMA write */
o Changes to psycho_attach(): [1] - Clear the PCI AFSR and status error bits as previous errors still might be indicated. - Set up the PCI control and diagnostic registers according to the capabilities, workarounds, etc of/for specific revisions of the supported bridges. This includes no longer setting Hummingbird-/ Sabre-specific bits in the PCI control register but preserving what the firmware has initialized them to like OpenSolaris does. Previously we were setting these bits according to the example in the Sabre documentation, which I doubt is appropriate for all Sabre based designs and especially not for Hummingbirds. This also includes not enabling bus parking unless the firmware tells us to. - Set the PCI latency timer register as this isn't always done by the firmware. o Remove a redundant argument from psycho_set_intr() and in this function check the return value of bus_setup_intr(). [2] o Let psycho_setup_intr() return ENOMEM instead of 0 when it can't allocate memory for the interrupt wrapper stub and EINVAL instead of 0 if it can't find the interrupt vector in the interrupt map. o Add a workaround for a bug of the Sabre-APB-combination where it doesn't drain DMA write data for devices behind additional PCI-PCI bridges underneath the APB PCI-PCI bridge. This workaround (do things necessary in order to achieve a manual drain when coherency is required) is currently implemented in psycho_setup_intr() and psycho_intr_stub() (for easy MFC'ing) and therefore is only applied for interrupt handlers. This should be moved to psycho(4)-specific bus_dma_tag_create() and bus_dmamap_sync() methods, respectively, once this driver is converted to make use of BUS_GET_DMA_TAG(), so the workaround is also applied for polling(4) callbacks. [3] o Fix some minor style issues. Info from: OpenSolaris [1] Info from: Linux, OpenBSD, OpenSolaris [3] Suggested by: Coverity Prevent (CID 682) [2] MFC after: 1 month
2007-01-08 01:26:47 +00:00
/* Uncorrectable error asynchronous fault status register */
#define UEAFSR_BLK (1UL << 23) /* Error caused by block transaction */
#define UEAFSR_P_DTE (1UL << 56) /* Pri. DVMA translation error */
#define UEAFSR_S_DTE (1UL << 57) /* Sec. DVMA translation error */
#define UEAFSR_S_DWR (1UL << 58) /* Sec. error during DVMA write */
#define UEAFSR_S_DRD (1UL << 59) /* Sec. error during DVMA read */
#define UEAFSR_S_PIO (1UL << 60) /* Sec. error during PIO access */
#define UEAFSR_P_DWR (1UL << 61) /* Pri. error during DVMA write */
#define UEAFSR_P_DRD (1UL << 62) /* Pri. error during DVMA read */
#define UEAFSR_P_PIO (1UL << 63) /* Pri. error during PIO access */
o Changes to psycho_attach(): [1] - Clear the PCI AFSR and status error bits as previous errors still might be indicated. - Set up the PCI control and diagnostic registers according to the capabilities, workarounds, etc of/for specific revisions of the supported bridges. This includes no longer setting Hummingbird-/ Sabre-specific bits in the PCI control register but preserving what the firmware has initialized them to like OpenSolaris does. Previously we were setting these bits according to the example in the Sabre documentation, which I doubt is appropriate for all Sabre based designs and especially not for Hummingbirds. This also includes not enabling bus parking unless the firmware tells us to. - Set the PCI latency timer register as this isn't always done by the firmware. o Remove a redundant argument from psycho_set_intr() and in this function check the return value of bus_setup_intr(). [2] o Let psycho_setup_intr() return ENOMEM instead of 0 when it can't allocate memory for the interrupt wrapper stub and EINVAL instead of 0 if it can't find the interrupt vector in the interrupt map. o Add a workaround for a bug of the Sabre-APB-combination where it doesn't drain DMA write data for devices behind additional PCI-PCI bridges underneath the APB PCI-PCI bridge. This workaround (do things necessary in order to achieve a manual drain when coherency is required) is currently implemented in psycho_setup_intr() and psycho_intr_stub() (for easy MFC'ing) and therefore is only applied for interrupt handlers. This should be moved to psycho(4)-specific bus_dma_tag_create() and bus_dmamap_sync() methods, respectively, once this driver is converted to make use of BUS_GET_DMA_TAG(), so the workaround is also applied for polling(4) callbacks. [3] o Fix some minor style issues. Info from: OpenSolaris [1] Info from: Linux, OpenBSD, OpenSolaris [3] Suggested by: Coverity Prevent (CID 682) [2] MFC after: 1 month
2007-01-08 01:26:47 +00:00
/* Correctable error asynchronous fault status register */
#define CEAFSR_BLK (1UL << 23) /* Error caused by block transaction */
#define CEAFSR_S_DWR (1UL << 58) /* Sec. error caused by DVMA write */
#define CEAFSR_S_DRD (1UL << 59) /* Sec. error caused by DVMA read */
#define CEAFSR_S_PIO (1UL << 60) /* Sec. error caused by PIO access */
#define CEAFSR_P_DWR (1UL << 61) /* Pri. error caused by DVMA write */
#define CEAFSR_P_DRD (1UL << 62) /* Pri. error caused by DVMA read */
#define CEAFSR_P_PIO (1UL << 63) /* Pri. error caused by PIO access */
o Changes to psycho_attach(): [1] - Clear the PCI AFSR and status error bits as previous errors still might be indicated. - Set up the PCI control and diagnostic registers according to the capabilities, workarounds, etc of/for specific revisions of the supported bridges. This includes no longer setting Hummingbird-/ Sabre-specific bits in the PCI control register but preserving what the firmware has initialized them to like OpenSolaris does. Previously we were setting these bits according to the example in the Sabre documentation, which I doubt is appropriate for all Sabre based designs and especially not for Hummingbirds. This also includes not enabling bus parking unless the firmware tells us to. - Set the PCI latency timer register as this isn't always done by the firmware. o Remove a redundant argument from psycho_set_intr() and in this function check the return value of bus_setup_intr(). [2] o Let psycho_setup_intr() return ENOMEM instead of 0 when it can't allocate memory for the interrupt wrapper stub and EINVAL instead of 0 if it can't find the interrupt vector in the interrupt map. o Add a workaround for a bug of the Sabre-APB-combination where it doesn't drain DMA write data for devices behind additional PCI-PCI bridges underneath the APB PCI-PCI bridge. This workaround (do things necessary in order to achieve a manual drain when coherency is required) is currently implemented in psycho_setup_intr() and psycho_intr_stub() (for easy MFC'ing) and therefore is only applied for interrupt handlers. This should be moved to psycho(4)-specific bus_dma_tag_create() and bus_dmamap_sync() methods, respectively, once this driver is converted to make use of BUS_GET_DMA_TAG(), so the workaround is also applied for polling(4) callbacks. [3] o Fix some minor style issues. Info from: OpenSolaris [1] Info from: Linux, OpenBSD, OpenSolaris [3] Suggested by: Coverity Prevent (CID 682) [2] MFC after: 1 month
2007-01-08 01:26:47 +00:00
/* PCI asynchronous fault status register */
#define PCIAFSR_P_MA (1UL << 63) /* Pri. master abort */
#define PCIAFSR_P_TA (1UL << 62) /* Pri. target abort */
#define PCIAFSR_P_RTRY (1UL << 61) /* Pri. excessive retries */
#define PCIAFSR_P_RERR (1UL << 60) /* Pri. parity error */
#define PCIAFSR_S_MA (1UL << 59) /* Sec. master abort */
#define PCIAFSR_S_TA (1UL << 58) /* Sec. target abort */
#define PCIAFSR_S_RTRY (1UL << 57) /* Sec. excessive retries */
#define PCIAFSR_S_RERR (1UL << 56) /* Sec. parity error */
#define PCIAFSR_BMASK (0xffffUL << 32)/* Bytemask of failed pri. transfer */
#define PCIAFSR_BLK (1UL << 31) /* failed pri. transfer was block r/w */
#define PCIAFSR_MID (0x3eUL << 25) /* UPA MID causing error transaction */
/* PCI diagnostic register */
#define DIAG_RTRY_DIS 0x0000000000000040 /* dis. retry limit */
#define DIAG_ISYNC_DIS 0x0000000000000020 /* dis. DMA write / int sync */
#define DIAG_DWSYNC_DIS 0x0000000000000010 /* dis. DMA write / PIO sync */
/* Definitions for the target address space register */
#define PCITAS_ADDR_SHIFT 29
/* Definitions for the Psycho configuration space */
#define PCS_DEVICE 0 /* Device number of Psycho CS entry */
#define PCS_FUNC 0 /* Function number of Psycho CS entry */
/* Non-Standard registers in the configration space */
#define PCSR_SECBUS 0x40 /* Secondary bus number register */
#define PCSR_SUBBUS 0x41 /* Subordinate bus number register */
- 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
/* Width of the physical addresses the IOMMU translates to */
#define PSYCHO_IOMMU_BITS 41
#define SABRE_IOMMU_BITS 34
#endif /* !_SPARC64_PCI_PSYCHOREG_H_ */