freebsd-skq/sys/dev/aic7xxx/aic79xx.h

1535 lines
48 KiB
C
Raw Normal View History

/*
* Core definitions and data structures shareable across OS platforms.
*
* Copyright (c) 1994-2002 Justin T. Gibbs.
* Copyright (c) 2000-2002 Adaptec Inc.
* 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,
* without modification.
* 2. Redistributions in binary form must reproduce at minimum a disclaimer
* substantially similar to the "NO WARRANTY" disclaimer below
* ("Disclaimer") and any redistribution must be conditioned upon
* including a substantially similar Disclaimer requirement for further
* binary redistribution.
* 3. Neither the names of the above-listed copyright holders nor the names
* of any contributors may be used to endorse or promote products derived
* from this software without specific prior written permission.
*
* Alternatively, this software may be distributed under the terms of the
* GNU General Public License ("GPL") version 2 as published by the Free
* Software Foundation.
*
* NO WARRANTY
* THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
* "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
* LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTIBILITY AND FITNESS FOR
* A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
* HOLDERS OR CONTRIBUTORS BE LIABLE FOR 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 DAMAGES.
*
* $Id: aic79xx.h,v 1.11 2003/05/26 21:10:58 gibbs Exp $
*
* $FreeBSD$
*/
#ifndef _AIC79XX_H_
#define _AIC79XX_H_
/* Register Definitions */
#include "aic79xx_reg.h"
/************************* Forward Declarations *******************************/
struct ahd_platform_data;
struct scb_platform_data;
/****************************** Useful Macros *********************************/
#ifndef MAX
#define MAX(a,b) (((a) > (b)) ? (a) : (b))
#endif
#ifndef MIN
#define MIN(a,b) (((a) < (b)) ? (a) : (b))
#endif
#ifndef TRUE
#define TRUE 1
#endif
#ifndef FALSE
#define FALSE 0
#endif
#define NUM_ELEMENTS(array) (sizeof(array) / sizeof(*array))
#define ALL_CHANNELS '\0'
#define ALL_TARGETS_MASK 0xFFFF
#define INITIATOR_WILDCARD (~0)
#define SCB_LIST_NULL 0xFF00
#define SCB_LIST_NULL_LE (ahd_htole16(SCB_LIST_NULL))
#define QOUTFIFO_ENTRY_VALID 0x8000
#define QOUTFIFO_ENTRY_VALID_LE (ahd_htole16(0x8000))
#define SCBID_IS_NULL(scbid) (((scbid) & 0xFF00 ) == SCB_LIST_NULL)
#define SCSIID_TARGET(ahd, scsiid) \
(((scsiid) & TID) >> TID_SHIFT)
#define SCSIID_OUR_ID(scsiid) \
((scsiid) & OID)
#define SCSIID_CHANNEL(ahd, scsiid) ('A')
#define SCB_IS_SCSIBUS_B(ahd, scb) (0)
#define SCB_GET_OUR_ID(scb) \
SCSIID_OUR_ID((scb)->hscb->scsiid)
#define SCB_GET_TARGET(ahd, scb) \
SCSIID_TARGET((ahd), (scb)->hscb->scsiid)
#define SCB_GET_CHANNEL(ahd, scb) \
SCSIID_CHANNEL(ahd, (scb)->hscb->scsiid)
#define SCB_GET_LUN(scb) \
((scb)->hscb->lun)
#define SCB_GET_TARGET_OFFSET(ahd, scb) \
SCB_GET_TARGET(ahd, scb)
#define SCB_GET_TARGET_MASK(ahd, scb) \
(0x01 << (SCB_GET_TARGET_OFFSET(ahd, scb)))
aic79xx.c: Implement the SCB_SILENT flag. This is useful for hushing up the driver during DV or other operations that we expect to cause transmission errors. The messages will still print if the SHOW_MASKED_ERRORS debug option is enabled. Save and restore the NEGOADDR address when setting new transfer settings. The sequencer performs lookups in the negotiation table too and it expects NEGOADDR to remain consistent across pause/unpause sessions. Consistently use "offset" instead of "period" to determine if we are running sync or not. Add a SHOW_MESSAGES diagnostic for when we assert ATN during message processing. Print out IU, QAS, and RTI features when showing transfer options. Limit the syncrate after all option conformance changes have taken place in ahd_devlimited_syncrate. Changes in options may change the final syncrate we accept. Keep a copy of the hs_mailbox in our softc so that we can perform read/modify/write operations on the hs_mailbox without having to pause the sequencer to read the last written value. Use the ENINT_COALESS flag in the hs_mailbox to toggle interrupt coalessing. Add entrypoints for enabling interrupt coalessing and setting both a timeout (how long to wait for commands to be coalessed) and a maximum commands to coaless value. Add a statistics timer that decides when to enable or disable interrupt coalessing based on load. Add a routine, ahd_reset_cmds_pending() which is used to update the CMDS_PENDING sequencer variable whenever error recovery compeltes SCBs without notifying the sequencer. Since ahd_reset_cmds_pending is called during ahd_unpause() only if we've aborted SCBs, its call to ahd_flush_qoutfifo should not cause recursion through ahd_run_qoutfifo(). A panic has been added to ensure that this recursion does not occur. In ahd_search_qinfifo, update the CMDS_PENDING sequencer variable directly. ahd_search_qinififo can be called in situations where using ahd_reset_cmds_pending() might cause recursion. Since we can safely determine the exact number to reduce CMDS_PENDING by in this scenario without running the qoutfifo, the manual update is sufficient. Clean up diagnostics. Add ahd_flush_qoutfifo() which will run the qoutfifo as well as complete any commands sitting on the sequencer's COMPLETE_SCB lists or the good status FIFO. Use this routine in several places that did similar things in an add-hoc, but incomplete, fashion. A call to this routine was also added to ahd_abort_scbs() to close a race. In ahd_pause_and_flushwork() only return once selections are safely disabled. Flush all completed commands via ahd_flush_qoutfifo(). Remove "Now packetized" diagnostic now that this information is incorperated into the actual negotiation messages that are displayed. When forcing renegotiation, don't clober the current ppr_options. Much of the driver uses this information to determine if we are currently packetized or not. Remove some stray spaces at column 1 in ahd_set_tags. When complaining about getting a host message loop request with no pending messages, print out the SCB_CONTROL register down on the card. Modify the ahd_sent_msg() routine to handle a search for an outgoing identify message. Use this to detect a msg reject on an identify message which typically indicates that the target thought we were packetized. Force a renegotiation in this case. In ahd_search_qinfifo(), wait more effectively for SCB DMA activities to cease. We also disable SCB fetch operations since we are about to change the qinfifo and any fetch in progress will likely be invalidated. In ahd_qinfifo_count(), fix the qinfifo empty case. In ahd_dump_card_state(), print out CCSCBCTL in the correct mode. If we are a narrow controller, don't set the current width to unknown when forcing a future negotiation. This just confuses the code into attempting a wide negotiation on a narrow bus. Add support for task management function completions. Modify ahd_handle_devreset so that it can handle lun resets in addition to target resets. Use ahd_handle_devreset for lun and target reset task management functions. Handle the abort task TMF race case better. We now wait until any current selections are over and then set the TMF back to zero. This should cause the sequencer to ignore the abort TMF completion should it occur. Correct a bug in the illegal phase handler that caused us to drop down to narrow when handling the unexpected command phase case after 3rd party reset of a packetized device. Indicate the features, bugs, and flags set in the softc that are used to control firmware patch download when booting verbose. aic79xx.h: Add coalessing and HS_MAILBOX fields. Add per-softc variables for the stats "daemon". Add a debug option for interrupt coalessing activities. Add two new softc flags: o AHD_UPDATE_PEND_CMDS Run ahd_reset_cmds_pending() on the next unpause. o AHD_RUNNING_QOUTFIFO Used to catch recursion through ahd_run_qoutfifo(). aic79xx.reg: Correct register addresses related to the software timer and the DFDBCTL register. Add constants paramaterizing the software timer. Add scratch ram locations for storing interrupt coalessing tunables. Break INTMASK in SEQITNCTL out into INTMASK1 and INTMASK2. In at least the REV A, these are writable bits. We make use of that for a swtimer workaround in the sequencer. Since HS_MAILBOX autoclears, provide a sequencer variable to store its contents. Add SEQINT codes for handling task management completions. aic79xx.seq: Correct ignore wide residue processing check for a wide negotiation being in effect. We must be in the SCSI register window in order to access the negotiation table. Use the software timer and a commands completed count to implement interrupt coalessing. The command complete is deferred until either the maximum command threshold or a the expiration of a command deferral timer. If we have more SCBs to complete to the host (sitting in COMPLETE_SCB lists), always try to coaless them up to our coalessing limit. If coalessing is enabled, but we have fewer commands oustanting than the host's min coalessing limit, complete the command immediately. Add code to track the number of commands outstanding. Commands are outstanding from the time they are placed into the execution queue until the DMA to post completion is setup. Add a workaround for intvec_2 interrupts on the H2A4. In H2A4, the mode pointer is not saved for intvec2, but is restored on iret. This can lead to the restoration of a bogus mode ptr. Manually clear the intmask bits and do a normal return to compensate. We use intvec_2 to track interrupt coalessing timeouts. Since we cannot disable the swtimer's countdown, simply mask its interrupt once we no longer care about it firing. In idle_loop_cchan, update LOCAL_HS_MAILBOX everytime we are notified of an HS_MAILBOX update via the HS_MAILBOX_ACT bit in QOFF_CTLSTA. We have to use a local copy of persistant portions of the HS_MAILBOX as the mailbox auto-clears on any read. Move the test for the cfg4istat interrupt up an instruction to hopefully close a race between the next outgoing selection and our disabling of selections. Add a missing ret to the last instruction in load_overrun_buf. Add notifications to the host of task management completions as well as the completions for commands that completed successfully before their corresponding TMF could be sent. Hold a critical section during select-out processing until we have a fully identified connection. This removes a race condition with the legacy abort handler. Correct a few spelling errors in some comments. aic79xx_inline.h: Call ahd_reset_cmds_pending() in ahd_unpause if required. Update cmdcmplt interrupt statistics in our interrupt handler. Allow callers to ahd_send_scb() to set the task management function. aic79xx_pci.c: Disable SERR and pause the controller prior to performing our mmapped I/O test. The U320 controllers do not support "auto-access-pause". aic79xx_osm.c: Set the task management function now that ahd_send_scb() doesn't do it for us. We also perform a lun reset in response to BDR requests to packetized devices.
2003-01-20 20:17:35 +00:00
#ifdef AHD_DEBUG
#define SCB_IS_SILENT(scb) \
((ahd_debug & AHD_SHOW_MASKED_ERRORS) == 0 \
&& (((scb)->flags & SCB_SILENT) != 0))
#else
#define SCB_IS_SILENT(scb) \
(((scb)->flags & SCB_SILENT) != 0)
#endif
/*
* TCLs have the following format: TTTTLLLLLLLL
*/
#define TCL_TARGET_OFFSET(tcl) \
((((tcl) >> 4) & TID) >> 4)
#define TCL_LUN(tcl) \
(tcl & (AHD_NUM_LUNS - 1))
#define BUILD_TCL(scsiid, lun) \
((lun) | (((scsiid) & TID) << 4))
#define BUILD_TCL_RAW(target, channel, lun) \
((lun) | ((target) << 8))
#define SCB_GET_TAG(scb) \
ahd_le16toh(scb->hscb->tag)
#ifndef AHD_TARGET_MODE
#undef AHD_TMODE_ENABLE
#define AHD_TMODE_ENABLE 0
#endif
#define AHD_BUILD_COL_IDX(target, lun) \
(((lun) << 4) | target)
#define AHD_GET_SCB_COL_IDX(ahd, scb) \
((SCB_GET_LUN(scb) << 4) | SCB_GET_TARGET(ahd, scb))
#define AHD_SET_SCB_COL_IDX(scb, col_idx) \
do { \
(scb)->hscb->scsiid = ((col_idx) << TID_SHIFT) & TID; \
(scb)->hscb->lun = ((col_idx) >> 4) & (AHD_NUM_LUNS_NONPKT-1); \
} while (0)
#define AHD_COPY_SCB_COL_IDX(dst, src) \
do { \
dst->hscb->scsiid = src->hscb->scsiid; \
dst->hscb->lun = src->hscb->lun; \
} while (0)
#define AHD_NEVER_COL_IDX 0xFFFF
/**************************** Driver Constants ********************************/
/*
* The maximum number of supported targets.
*/
#define AHD_NUM_TARGETS 16
/*
* The maximum number of supported luns.
* The identify message only supports 64 luns in non-packetized transfers.
* You can have 2^64 luns when information unit transfers are enabled,
* but until we see a need to support that many, we support 256.
*/
#define AHD_NUM_LUNS_NONPKT 64
#define AHD_NUM_LUNS 256
/*
* The maximum transfer per S/G segment.
*/
#define AHD_MAXTRANSFER_SIZE 0x00ffffff /* limited by 24bit counter */
/*
* The maximum amount of SCB storage in hardware on a controller.
* This value represents an upper bound. Due to software design,
* we may not be able to use this number.
*/
#define AHD_SCB_MAX 512
/*
* The maximum number of concurrent transactions supported per driver instance.
* Sequencer Control Blocks (SCBs) store per-transaction information.
*/
#define AHD_MAX_QUEUE AHD_SCB_MAX
/*
* Define the size of our QIN and QOUT FIFOs. They must be a power of 2
Correct spelling errors. Switch to handling bad SCSI status as a sequencer interrupt instead of having the kernel proccess these failures via the completion queue. This is done because: o The old scheme required us to pause the sequencer and clear critical sections for each SCB. It seems that these pause actions, if coincident with a sequencer FIFO interrupt, would result in a FIFO interrupt getting lost or directing to the wrong FIFO. This caused hangs when the driver was stressed under high "queue full" loads. o The completion code assumed that it was always called with the sequencer running. This may not be the case in timeout processing where completions occur manually via ahd_pause_and_flushwork(). o With this scheme, the extra expense of clearing critical sections is avoided since the sequencer will only self pause once all pending selections have cleared and it is not in a critical section. aic79xx.c Add code to handle the new BAD_SCB_STATUS sequencer interrupt code. This just redirects the SCB through the already existing ahd_complete_scb() code path. Remove code in ahd_handle_scsi_status() that paused the sequencer, made sure that no selections where pending, and cleared critical sections. Bad status SCBs are now only processed when all of these conditions are true. aic79xx.reg: Add the BAD_SCB_STATUS sequencer interrupt code. aic79xx.seq: When completing an SCB upload to the host, if we are doing this because the SCB contains non-zero SCSI status, defer completing the SCB until there are no pending selection events. When completing these SCBs, use the new BAD_SCB_STATUS sequencer interrupt. For all other uploaded SCBs (currently only for underruns), the SCB is completed via the normal done queue. Additionally, keep the SCB that is currently being uploaded on the COMPLETE_DMA_SCB list until the dma is completed, not just until the DMA is started. This ensures that the DMA is restarted properly should the host disable the DMA transfer for some reason. In our RevA workaround for Maxtor drives, guard against the host pausing us while trying to pause I/O until the first data-valid REQ by clearing the current snapshot so that we can tell if the transfer has completed prior to us noticing the REQINIT status. In cfg4data_intr, shave off an instruction before getting the data path running by adding an entrypoint to the overrun handler to also increment the FIFO use count. In the overrun handler, be sure to clear our LONGJMP address in both exit paths. Perform a few sequencer optimizations. aic79xx.c: Print the full path from the SCB when a packetized status overrun occurs. Remove references to LONGJMP_SCB which is being removed from firmware usage. Print the new SCB_FIFO_USE_COUNT field in the per-SCB section of ahd_dump_card_state(). The SCB_TAG field is now re-used by the sequencer, so it no longer makes sense to reference this field in the kernel driver. aic79xx.h: Re-arrange fields in the hardware SCB from largest size type to smallest. This makes it easier to move fields without changing field alignment. The hardware scb tag field is now down near the "spare" portion of the SCB to facilitate reuse by the sequencer. aic79xx.reg: Remove LONGJMP_ADDR. Rearrange SCB fields to match aic79xx.h. Add SCB_FIFO_USE_COUNT as the first byte of the SCB_TAG field. aic79xx.seq: Add a per-SCB "Fifos in use count" field and use it to determine when it is safe (all data posted) to deliver status back to the host. The old method involved polling one or both FIFOs to verify that the current task did not have pending data. This makes running down the GSFIFO very cheap, so we will empty the GSFIFO in one idle loop pass in all cases. Use this simplification of the completion process to prune down the data FIFO teardown sequencer for packetized transfers. Much more code is now shared between the data residual and transfer complete cases. Correct some issues in the packetized status handler. It used to be possible to CLRCHN our FIFO before status had fully transferred to the host. We also failed to handle NONPACKREQ phases that could occur should a CRC error occur during transmission of the status data packet. Correct a few big endian issues: aic79xx.c: aic79xx_inline.h: aic79xx_pci.c: aic79xx_osm.c: o Always get the SCB's tag via the SCB_GET_TAG acccessor o Add missing use of byte swapping macros when touching hscb fields. o Don't double swap SEEPROM data when it is printed. Correct a big-endian bug. We cannot assign a o When assigning a 32bit LE variable to a 64bit LE variable, we must be explict about how the words of the 64bit LE variable are initialized. Cast to (uint32_t*) to do this. aic79xx.c: In ahd_clear_critical_section(), hit CRLSCSIINT after restoring the interrupt masks to avoid what appears to be a glitch on SCSIINT. Any real SCSIINT status will be persistent and will immidiately reset SCSIINT. This clear should only get rid of spurious SCSIINTs. This glitch was the cause of the "Unexpected PKT busfree" status that occurred under high queue full loads Call ahd_fini_scbdata() after shutdown so that any ahd_chip_init() routine that might access SCB data will not access free'd memory. Reset the bus on an IOERR since the chip doesn't seem to reset to the new voltage level without this. Change offset calculation for scatter gather maps so that the calculation is correct if an integral multiple of sg lists does not fit in the allocation size. Adjust bus dma tag for data buffers based on 39BIT addressing flag in our softc. Use the QFREEZE count to simplify ahd_pause_and_flushworkd(). We can thus rely on the sequencer eventually clearing ENSELO. In ahd_abort_scbs(), fix a bug that could potentially corrupt sequencer state. The saved SCB was being restored in the SCSI mode instead of the saved mode. It turns out that the SCB did not need to be saved at all as the scbptr is already restored by all subroutines called during this function that modify that register. aic79xx.c: aic79xx.h: aic79xx_pci.c: Add support for parsing the seeprom vital product data. The VPD data are currently unused. aic79xx.h: aic79xx.seq: aic79xx_pci.c: Add a firmware workaround to make the LED blink brighter during packetized operations on the H2A. aic79xx_inline.h: The host does not use timer interrupts, so don't gate our decision on whether or not to unpause the sequencer on whether or not a timer interrupt is pending.
2003-05-04 00:20:07 +00:00
* in size and accommodate as many transactions as can be queued concurrently.
*/
#define AHD_QIN_SIZE AHD_MAX_QUEUE
#define AHD_QOUT_SIZE AHD_MAX_QUEUE
#define AHD_QIN_WRAP(x) ((x) & (AHD_QIN_SIZE-1))
/*
* The maximum amount of SCB storage we allocate in host memory.
*/
#define AHD_SCB_MAX_ALLOC AHD_MAX_QUEUE
/*
* Ring Buffer of incoming target commands.
* We allocate 256 to simplify the logic in the sequencer
* by using the natural wrap point of an 8bit counter.
*/
#define AHD_TMODE_CMDS 256
/* Reset line assertion time in us */
#define AHD_BUSRESET_DELAY 25
/******************* Chip Characteristics/Operating Settings *****************/
/*
* Chip Type
* The chip order is from least sophisticated to most sophisticated.
*/
typedef enum {
AHD_NONE = 0x0000,
AHD_CHIPID_MASK = 0x00FF,
AHD_AIC7901 = 0x0001,
AHD_AIC7902 = 0x0002,
AHD_AIC7901A = 0x0003,
AHD_PCI = 0x0100, /* Bus type PCI */
AHD_PCIX = 0x0200, /* Bus type PCIX */
AHD_BUS_MASK = 0x0F00
} ahd_chip;
/*
* Features available in each chip type.
*/
typedef enum {
AHD_FENONE = 0x00000,
AHD_WIDE = 0x00001,/* Wide Channel */
AHD_MULTI_FUNC = 0x00100,/* Multi-Function/Channel Device */
AHD_TARGETMODE = 0x01000,/* Has tested target mode support */
AHD_MULTIROLE = 0x02000,/* Space for two roles at a time */
AHD_RTI = 0x04000,/* Retained Training Support */
AHD_NEW_IOCELL_OPTS = 0x08000,/* More Signal knobs in the IOCELL */
AHD_NEW_DFCNTRL_OPTS = 0x10000,/* SCSIENWRDIS bit */
AHD_REMOVABLE = 0x00000,/* Hot-Swap supported - None so far*/
AHD_AIC7901_FE = AHD_FENONE,
AHD_AIC7902_FE = AHD_MULTI_FUNC
} ahd_feature;
/*
* Bugs in the silicon that we work around in software.
*/
typedef enum {
AHD_BUGNONE = 0x0000,
/*
* Rev A hardware fails to update LAST/CURR/NEXTSCB
* correctly in certain packetized selection cases.
*/
AHD_SENT_SCB_UPDATE_BUG = 0x0001,
/* The wrong SCB is accessed to check the abort pending bit. */
AHD_ABORT_LQI_BUG = 0x0002,
/* Packetized bitbucket crosses packet boundaries. */
AHD_PKT_BITBUCKET_BUG = 0x0004,
/* The selection timer runs twice as long as its setting. */
AHD_LONG_SETIMO_BUG = 0x0008,
/* The Non-LQ CRC error status is delayed until phase change. */
AHD_NLQICRC_DELAYED_BUG = 0x0010,
/* The chip must be reset for all outgoing bus resets. */
AHD_SCSIRST_BUG = 0x0020,
/* Some PCIX fields must be saved and restored across chip reset. */
AHD_PCIX_CHIPRST_BUG = 0x0040,
/* MMAPIO is not functional in PCI-X mode. */
AHD_PCIX_MMAPIO_BUG = 0x0080,
2003-03-06 23:58:34 +00:00
/* Reads to SCBRAM fail to reset the discard timer. */
AHD_PCIX_SCBRAM_RD_BUG = 0x0100,
/* Bug workarounds that can be disabled on non-PCIX busses. */
AHD_PCIX_BUG_MASK = AHD_PCIX_CHIPRST_BUG
2003-03-06 23:58:34 +00:00
| AHD_PCIX_MMAPIO_BUG
| AHD_PCIX_SCBRAM_RD_BUG,
/*
* LQOSTOP0 status set even for forced selections with ATN
* to perform non-packetized message delivery.
*/
2003-03-06 23:58:34 +00:00
AHD_LQO_ATNO_BUG = 0x0200,
/* FIFO auto-flush does not always trigger. */
2003-03-06 23:58:34 +00:00
AHD_AUTOFLUSH_BUG = 0x0400,
/* The CLRLQO registers are not self-clearing. */
2003-03-06 23:58:34 +00:00
AHD_CLRLQO_AUTOCLR_BUG = 0x0800,
/* The PACKETIZED status bit refers to the previous connection. */
2003-03-06 23:58:34 +00:00
AHD_PKTIZED_STATUS_BUG = 0x1000,
/* "Short Luns" are not placed into outgoing LQ packets correctly. */
2003-03-06 23:58:34 +00:00
AHD_PKT_LUN_BUG = 0x2000,
/*
* Only the FIFO allocated to the non-packetized connection may
* be in use during a non-packetzied connection.
*/
2003-03-06 23:58:34 +00:00
AHD_NONPACKFIFO_BUG = 0x4000,
/*
* Writing to a DFF SCBPTR register may fail if concurent with
* a hardware write to the other DFF SCBPTR register. This is
* not currently a concern in our sequencer since all chips with
* this bug have the AHD_NONPACKFIFO_BUG and all writes of concern
* occur in non-packetized connections.
*/
2003-03-06 23:58:34 +00:00
AHD_MDFF_WSCBPTR_BUG = 0x8000,
/* SGHADDR updates are slow. */
2003-03-06 23:58:34 +00:00
AHD_REG_SLOW_SETTLE_BUG = 0x10000,
/*
* Changing the MODE_PTR coincident with an interrupt that
* switches to a different mode will cause the interrupt to
* be in the mode written outside of interrupt context.
*/
2003-03-06 23:58:34 +00:00
AHD_SET_MODE_BUG = 0x20000,
/* Non-packetized busfree revision does not work. */
2003-03-06 23:58:34 +00:00
AHD_BUSFREEREV_BUG = 0x40000,
/*
* Paced transfers are indicated with a non-standard PPR
* option bit in the neg table, 160MHz is indicated by
* sync factor 0x7, and the offset if off by a factor of 2.
*/
2003-03-06 23:58:34 +00:00
AHD_PACED_NEGTABLE_BUG = 0x80000,
/* LQOOVERRUN false positives. */
2003-03-06 23:58:34 +00:00
AHD_LQOOVERRUN_BUG = 0x100000,
/*
* Controller write to INTSTAT will lose to a host
* write to CLRINT.
*/
2003-03-06 23:58:34 +00:00
AHD_INTCOLLISION_BUG = 0x200000,
aic79xx.c: Clear the LQICRC_NLQ status should it pop up after we have already handled the SCSIPERR. During some streaming operations this status can be delayed until the stream ends. Without this change, the driver would complain about a "Missing case in ahd_handle_scsiint". In the LQOBUSFREE handler... Don't return the LQOMGR back to the idle state until after we have cleaned up ENSELO and any status related to this selection. The last thing we need is the LQO manager starting another select-out before we have updated the execution queue. It is not clear whether the LQOMGR would, or would not start a new selection early. Make sure ENSELO is off prior to clearing SELDO by flushing device writes. Move assignment of the next target SCB pointer inside of an if to make the code clearer. The effect is the same. Dump card state in both "Unexpected PKT busfree" paths. In ahd_reset(), set the chip to SCSI mode before reading SXFRCTL1. That register only exists in the SCSI mode. Also set the mode explicitly to the SCSI mode after chip reset due to paranoia. Re-arrange code so that SXFRCTL1 is restored as quickly after the chip reset as possible. S/G structurs must be 8byte aligned. Make this official by saying so in our DMA tag. Disable CIO bus stretch on MDFFSTAT if SHVALID is about to come true. This can cause a CIO bus lockup if a PCI or PCI-X error occurs while the stretch is occurring - the host cannot service the PCI-X error since the CIO bus is locked out and SHVALID will never resolve. The stretch was added in the Rev B to simplify the wait for SHVALID to resolve, but the code to do this in the open source sequencer is so simple it was never removed. Consistently use MAX_OFFSET for the user max syncrate set from non-volatile storage. This ensures that the offset does not conflict with AH?_OFFSET_UNKNOWN. Have ahd_pause_and_flushwork set the mode to ensure that it has access to the registers it checks. Also modify the checking of intstat so that the check against 0xFF can actually succeed if the INT_PEND mask is something other than 0xFF. Although there are no cardbus U320 controllers, this check may be needed to recover from a hot-plug PCI removal that occurs without informing the driver. Fix a typo. sg_prefetch_cnt -> sg_prefetch_align. This fixes an infinite loop at card initialization if the cacheline size is 0. aic79xx.h: Add AHD_EARLY_REQ_BUG bug flag. Fix spelling errors. Include the CDB's length just after the CDB pointer in the DMA'ed CDB case. Change AH?_OFFSET_UNKNOWN to 0xFF. This is a value that the curr->offset can never be, unlike '0' which we previously used. This fixes code that only checks for a non-zero offset to determine if a sync negotiation is required since it will fire in the unknown case even if the goal is async. aic79xx.reg: Add comments for LQISTAT bits indicating their names in the 7902 data book. We use slightly different and more descriptive names in the firmware. Fix spelling errors. Include the CDB's length just after the CDB pointer in the DMA'ed CDB case. aic79xx.seq: Update comments regarding rundown of the GSFIFO to reflect reality. Fix spelling errors. Since we use an 8byte address and 1 byte length, shorten the size of a block move for the legacy DMA'ed CDB case from 11 to 9 bytes. Remove code that, assuming the abort pending feature worked, would set MK_MESSAGE in the SCB's control byte on completion to catch invalid reselections. Since we don't see interrupts for completed selections, this status update could occur prior to us noticing the SELDO. The "select-out" queue logic will get confused by the MK_MESSAGE bit being set as this is used to catch packatized connections where we select-out with ATN. Since the abort pending feature doesn't work on any released controllers yet, this code was never executed. Add support for the AHD_EARLY_REQ_BUG. Don't ignore persistent REQ assertions just because they were asserted within the bus settle delay window. This allows us to tolerate devices like the GEM318 that violate the SCSI spec. Remove unintentional settnig of SG_CACHE_AVAIL. Writing this bit should have no effect, but who knows... On the Rev A, we must wait for HDMAENACK before loading additional segments to avoid clobbering the address of the first segment in the S/G FIFO. This resolves data-corruption issues with certain IBM (now Hitachi) and Fujitsu U320 drives. Rearrange calc_residual to avoid an extra jmp instruction. On RevA Silicon, if the target returns us to data-out after we have already trained for data-out, it is possible for us to transition the free running clock to data-valid before the required 100ns P1 setup time (8 P1 assertions in fast-160 mode). This will only happen if this L-Q is a continuation of a data transfer for which we have already prefetched data into our FIFO (LQ/Data followed by LQ/Data for the same write transaction). This can cause some target implementations to miss the first few data transfers on the bus. We detect this situation by noticing that this is the first data transfer after an LQ (LQIWORKONLQ true), that the data transfer is a continuation of a transfer already setup in our FIFO (SAVEPTRS interrupt), and that the transaction is a write (DIRECTION set in DFCNTRL). The delay is performed by disabling SCSIEN until we see the first REQ from the target. Only compile in snapshot savepointers handler for RevA silicon where it is enabled. Handle the cfg4icmd packetized interrupt. We just need to load the address and count, start the DMA, and CLRCHN once the transfer is complete. Fix an oversight in the overrun handler for packetized status operations. We need to wait for either CTXTDONE or an overrun when checking for an overrun. The previous code did not wait and thus could decide that no overrun had occurred even though an overrun will occur on the next data-valid req. Add some comment to this section for clarity. Use LAST_SEG_DONE instead of LASTSDONE for testing transfer completion in the packetized status case. LASTSDONE may come up more quickly since it only records completion on the SCSI side, but since LAST_SEG_DONE is used everywhere else (and needs to be), this is less confusing. Add a missing invalidation of the longjmp address in the non-pack handler. This code needs additional review. aic79xx_inline.h: Fix spelling error. aic79xx_osm.c: Set the cdb length for CDBs dma'ed from host memory. Add a comment indicating that, should CAM start supporting cdbs larger than 16bytes, the driver could store the CDB in the status buffer. aic79xx_pci.c: Add a table entry for the 39320A. Added a missing comma to an error string table. Fix spelling errors.
2003-02-27 23:23:16 +00:00
/*
* The GEM318 violates the SCSI spec by not waiting
* the mandated bus settle delay between phase changes
* in some situations. Some aic79xx chip revs. are more
* strict in this regard and will treat REQ assertions
* that fall within the bus settle delay window as
* glitches. This flag tells the firmware to tolerate
* early REQ assertions.
*/
Correct spelling errors. Switch to handling bad SCSI status as a sequencer interrupt instead of having the kernel proccess these failures via the completion queue. This is done because: o The old scheme required us to pause the sequencer and clear critical sections for each SCB. It seems that these pause actions, if coincident with a sequencer FIFO interrupt, would result in a FIFO interrupt getting lost or directing to the wrong FIFO. This caused hangs when the driver was stressed under high "queue full" loads. o The completion code assumed that it was always called with the sequencer running. This may not be the case in timeout processing where completions occur manually via ahd_pause_and_flushwork(). o With this scheme, the extra expense of clearing critical sections is avoided since the sequencer will only self pause once all pending selections have cleared and it is not in a critical section. aic79xx.c Add code to handle the new BAD_SCB_STATUS sequencer interrupt code. This just redirects the SCB through the already existing ahd_complete_scb() code path. Remove code in ahd_handle_scsi_status() that paused the sequencer, made sure that no selections where pending, and cleared critical sections. Bad status SCBs are now only processed when all of these conditions are true. aic79xx.reg: Add the BAD_SCB_STATUS sequencer interrupt code. aic79xx.seq: When completing an SCB upload to the host, if we are doing this because the SCB contains non-zero SCSI status, defer completing the SCB until there are no pending selection events. When completing these SCBs, use the new BAD_SCB_STATUS sequencer interrupt. For all other uploaded SCBs (currently only for underruns), the SCB is completed via the normal done queue. Additionally, keep the SCB that is currently being uploaded on the COMPLETE_DMA_SCB list until the dma is completed, not just until the DMA is started. This ensures that the DMA is restarted properly should the host disable the DMA transfer for some reason. In our RevA workaround for Maxtor drives, guard against the host pausing us while trying to pause I/O until the first data-valid REQ by clearing the current snapshot so that we can tell if the transfer has completed prior to us noticing the REQINIT status. In cfg4data_intr, shave off an instruction before getting the data path running by adding an entrypoint to the overrun handler to also increment the FIFO use count. In the overrun handler, be sure to clear our LONGJMP address in both exit paths. Perform a few sequencer optimizations. aic79xx.c: Print the full path from the SCB when a packetized status overrun occurs. Remove references to LONGJMP_SCB which is being removed from firmware usage. Print the new SCB_FIFO_USE_COUNT field in the per-SCB section of ahd_dump_card_state(). The SCB_TAG field is now re-used by the sequencer, so it no longer makes sense to reference this field in the kernel driver. aic79xx.h: Re-arrange fields in the hardware SCB from largest size type to smallest. This makes it easier to move fields without changing field alignment. The hardware scb tag field is now down near the "spare" portion of the SCB to facilitate reuse by the sequencer. aic79xx.reg: Remove LONGJMP_ADDR. Rearrange SCB fields to match aic79xx.h. Add SCB_FIFO_USE_COUNT as the first byte of the SCB_TAG field. aic79xx.seq: Add a per-SCB "Fifos in use count" field and use it to determine when it is safe (all data posted) to deliver status back to the host. The old method involved polling one or both FIFOs to verify that the current task did not have pending data. This makes running down the GSFIFO very cheap, so we will empty the GSFIFO in one idle loop pass in all cases. Use this simplification of the completion process to prune down the data FIFO teardown sequencer for packetized transfers. Much more code is now shared between the data residual and transfer complete cases. Correct some issues in the packetized status handler. It used to be possible to CLRCHN our FIFO before status had fully transferred to the host. We also failed to handle NONPACKREQ phases that could occur should a CRC error occur during transmission of the status data packet. Correct a few big endian issues: aic79xx.c: aic79xx_inline.h: aic79xx_pci.c: aic79xx_osm.c: o Always get the SCB's tag via the SCB_GET_TAG acccessor o Add missing use of byte swapping macros when touching hscb fields. o Don't double swap SEEPROM data when it is printed. Correct a big-endian bug. We cannot assign a o When assigning a 32bit LE variable to a 64bit LE variable, we must be explict about how the words of the 64bit LE variable are initialized. Cast to (uint32_t*) to do this. aic79xx.c: In ahd_clear_critical_section(), hit CRLSCSIINT after restoring the interrupt masks to avoid what appears to be a glitch on SCSIINT. Any real SCSIINT status will be persistent and will immidiately reset SCSIINT. This clear should only get rid of spurious SCSIINTs. This glitch was the cause of the "Unexpected PKT busfree" status that occurred under high queue full loads Call ahd_fini_scbdata() after shutdown so that any ahd_chip_init() routine that might access SCB data will not access free'd memory. Reset the bus on an IOERR since the chip doesn't seem to reset to the new voltage level without this. Change offset calculation for scatter gather maps so that the calculation is correct if an integral multiple of sg lists does not fit in the allocation size. Adjust bus dma tag for data buffers based on 39BIT addressing flag in our softc. Use the QFREEZE count to simplify ahd_pause_and_flushworkd(). We can thus rely on the sequencer eventually clearing ENSELO. In ahd_abort_scbs(), fix a bug that could potentially corrupt sequencer state. The saved SCB was being restored in the SCSI mode instead of the saved mode. It turns out that the SCB did not need to be saved at all as the scbptr is already restored by all subroutines called during this function that modify that register. aic79xx.c: aic79xx.h: aic79xx_pci.c: Add support for parsing the seeprom vital product data. The VPD data are currently unused. aic79xx.h: aic79xx.seq: aic79xx_pci.c: Add a firmware workaround to make the LED blink brighter during packetized operations on the H2A. aic79xx_inline.h: The host does not use timer interrupts, so don't gate our decision on whether or not to unpause the sequencer on whether or not a timer interrupt is pending.
2003-05-04 00:20:07 +00:00
AHD_EARLY_REQ_BUG = 0x400000,
/*
* The LED does not stay on long enough in packetized modes.
*/
AHD_FAINT_LED_BUG = 0x800000
} ahd_bug;
/*
* Configuration specific settings.
* The driver determines these settings by probing the
* chip/controller's configuration.
*/
typedef enum {
AHD_FNONE = 0x00000,
Correct spelling errors. Switch to handling bad SCSI status as a sequencer interrupt instead of having the kernel proccess these failures via the completion queue. This is done because: o The old scheme required us to pause the sequencer and clear critical sections for each SCB. It seems that these pause actions, if coincident with a sequencer FIFO interrupt, would result in a FIFO interrupt getting lost or directing to the wrong FIFO. This caused hangs when the driver was stressed under high "queue full" loads. o The completion code assumed that it was always called with the sequencer running. This may not be the case in timeout processing where completions occur manually via ahd_pause_and_flushwork(). o With this scheme, the extra expense of clearing critical sections is avoided since the sequencer will only self pause once all pending selections have cleared and it is not in a critical section. aic79xx.c Add code to handle the new BAD_SCB_STATUS sequencer interrupt code. This just redirects the SCB through the already existing ahd_complete_scb() code path. Remove code in ahd_handle_scsi_status() that paused the sequencer, made sure that no selections where pending, and cleared critical sections. Bad status SCBs are now only processed when all of these conditions are true. aic79xx.reg: Add the BAD_SCB_STATUS sequencer interrupt code. aic79xx.seq: When completing an SCB upload to the host, if we are doing this because the SCB contains non-zero SCSI status, defer completing the SCB until there are no pending selection events. When completing these SCBs, use the new BAD_SCB_STATUS sequencer interrupt. For all other uploaded SCBs (currently only for underruns), the SCB is completed via the normal done queue. Additionally, keep the SCB that is currently being uploaded on the COMPLETE_DMA_SCB list until the dma is completed, not just until the DMA is started. This ensures that the DMA is restarted properly should the host disable the DMA transfer for some reason. In our RevA workaround for Maxtor drives, guard against the host pausing us while trying to pause I/O until the first data-valid REQ by clearing the current snapshot so that we can tell if the transfer has completed prior to us noticing the REQINIT status. In cfg4data_intr, shave off an instruction before getting the data path running by adding an entrypoint to the overrun handler to also increment the FIFO use count. In the overrun handler, be sure to clear our LONGJMP address in both exit paths. Perform a few sequencer optimizations. aic79xx.c: Print the full path from the SCB when a packetized status overrun occurs. Remove references to LONGJMP_SCB which is being removed from firmware usage. Print the new SCB_FIFO_USE_COUNT field in the per-SCB section of ahd_dump_card_state(). The SCB_TAG field is now re-used by the sequencer, so it no longer makes sense to reference this field in the kernel driver. aic79xx.h: Re-arrange fields in the hardware SCB from largest size type to smallest. This makes it easier to move fields without changing field alignment. The hardware scb tag field is now down near the "spare" portion of the SCB to facilitate reuse by the sequencer. aic79xx.reg: Remove LONGJMP_ADDR. Rearrange SCB fields to match aic79xx.h. Add SCB_FIFO_USE_COUNT as the first byte of the SCB_TAG field. aic79xx.seq: Add a per-SCB "Fifos in use count" field and use it to determine when it is safe (all data posted) to deliver status back to the host. The old method involved polling one or both FIFOs to verify that the current task did not have pending data. This makes running down the GSFIFO very cheap, so we will empty the GSFIFO in one idle loop pass in all cases. Use this simplification of the completion process to prune down the data FIFO teardown sequencer for packetized transfers. Much more code is now shared between the data residual and transfer complete cases. Correct some issues in the packetized status handler. It used to be possible to CLRCHN our FIFO before status had fully transferred to the host. We also failed to handle NONPACKREQ phases that could occur should a CRC error occur during transmission of the status data packet. Correct a few big endian issues: aic79xx.c: aic79xx_inline.h: aic79xx_pci.c: aic79xx_osm.c: o Always get the SCB's tag via the SCB_GET_TAG acccessor o Add missing use of byte swapping macros when touching hscb fields. o Don't double swap SEEPROM data when it is printed. Correct a big-endian bug. We cannot assign a o When assigning a 32bit LE variable to a 64bit LE variable, we must be explict about how the words of the 64bit LE variable are initialized. Cast to (uint32_t*) to do this. aic79xx.c: In ahd_clear_critical_section(), hit CRLSCSIINT after restoring the interrupt masks to avoid what appears to be a glitch on SCSIINT. Any real SCSIINT status will be persistent and will immidiately reset SCSIINT. This clear should only get rid of spurious SCSIINTs. This glitch was the cause of the "Unexpected PKT busfree" status that occurred under high queue full loads Call ahd_fini_scbdata() after shutdown so that any ahd_chip_init() routine that might access SCB data will not access free'd memory. Reset the bus on an IOERR since the chip doesn't seem to reset to the new voltage level without this. Change offset calculation for scatter gather maps so that the calculation is correct if an integral multiple of sg lists does not fit in the allocation size. Adjust bus dma tag for data buffers based on 39BIT addressing flag in our softc. Use the QFREEZE count to simplify ahd_pause_and_flushworkd(). We can thus rely on the sequencer eventually clearing ENSELO. In ahd_abort_scbs(), fix a bug that could potentially corrupt sequencer state. The saved SCB was being restored in the SCSI mode instead of the saved mode. It turns out that the SCB did not need to be saved at all as the scbptr is already restored by all subroutines called during this function that modify that register. aic79xx.c: aic79xx.h: aic79xx_pci.c: Add support for parsing the seeprom vital product data. The VPD data are currently unused. aic79xx.h: aic79xx.seq: aic79xx_pci.c: Add a firmware workaround to make the LED blink brighter during packetized operations on the H2A. aic79xx_inline.h: The host does not use timer interrupts, so don't gate our decision on whether or not to unpause the sequencer on whether or not a timer interrupt is pending.
2003-05-04 00:20:07 +00:00
AHD_BOOT_CHANNEL = 0x00001,/* We were set as the boot channel. */
AHD_USEDEFAULTS = 0x00004,/*
* For cards without an seeprom
* or a BIOS to initialize the chip's
* SRAM, we use the default target
* settings.
*/
AHD_SEQUENCER_DEBUG = 0x00008,
AHD_RESET_BUS_A = 0x00010,
AHD_EXTENDED_TRANS_A = 0x00020,
AHD_TERM_ENB_A = 0x00040,
AHD_SPCHK_ENB_A = 0x00080,
AHD_STPWLEVEL_A = 0x00100,
AHD_INITIATORROLE = 0x00200,/*
* Allow initiator operations on
* this controller.
*/
AHD_TARGETROLE = 0x00400,/*
* Allow target operations on this
* controller.
*/
AHD_RESOURCE_SHORTAGE = 0x00800,
AHD_TQINFIFO_BLOCKED = 0x01000,/* Blocked waiting for ATIOs */
AHD_INT50_SPEEDFLEX = 0x02000,/*
* Internal 50pin connector
* sits behind an aic3860
*/
AHD_BIOS_ENABLED = 0x04000,
AHD_ALL_INTERRUPTS = 0x08000,
AHD_39BIT_ADDRESSING = 0x10000,/* Use 39 bit addressing scheme. */
AHD_64BIT_ADDRESSING = 0x20000,/* Use 64 bit addressing scheme. */
AHD_CURRENT_SENSING = 0x40000,
AHD_SCB_CONFIG_USED = 0x80000,/* No SEEPROM but SCB had info. */
AHD_HP_BOARD = 0x100000,
aic79xx.c: Implement the SCB_SILENT flag. This is useful for hushing up the driver during DV or other operations that we expect to cause transmission errors. The messages will still print if the SHOW_MASKED_ERRORS debug option is enabled. Save and restore the NEGOADDR address when setting new transfer settings. The sequencer performs lookups in the negotiation table too and it expects NEGOADDR to remain consistent across pause/unpause sessions. Consistently use "offset" instead of "period" to determine if we are running sync or not. Add a SHOW_MESSAGES diagnostic for when we assert ATN during message processing. Print out IU, QAS, and RTI features when showing transfer options. Limit the syncrate after all option conformance changes have taken place in ahd_devlimited_syncrate. Changes in options may change the final syncrate we accept. Keep a copy of the hs_mailbox in our softc so that we can perform read/modify/write operations on the hs_mailbox without having to pause the sequencer to read the last written value. Use the ENINT_COALESS flag in the hs_mailbox to toggle interrupt coalessing. Add entrypoints for enabling interrupt coalessing and setting both a timeout (how long to wait for commands to be coalessed) and a maximum commands to coaless value. Add a statistics timer that decides when to enable or disable interrupt coalessing based on load. Add a routine, ahd_reset_cmds_pending() which is used to update the CMDS_PENDING sequencer variable whenever error recovery compeltes SCBs without notifying the sequencer. Since ahd_reset_cmds_pending is called during ahd_unpause() only if we've aborted SCBs, its call to ahd_flush_qoutfifo should not cause recursion through ahd_run_qoutfifo(). A panic has been added to ensure that this recursion does not occur. In ahd_search_qinfifo, update the CMDS_PENDING sequencer variable directly. ahd_search_qinififo can be called in situations where using ahd_reset_cmds_pending() might cause recursion. Since we can safely determine the exact number to reduce CMDS_PENDING by in this scenario without running the qoutfifo, the manual update is sufficient. Clean up diagnostics. Add ahd_flush_qoutfifo() which will run the qoutfifo as well as complete any commands sitting on the sequencer's COMPLETE_SCB lists or the good status FIFO. Use this routine in several places that did similar things in an add-hoc, but incomplete, fashion. A call to this routine was also added to ahd_abort_scbs() to close a race. In ahd_pause_and_flushwork() only return once selections are safely disabled. Flush all completed commands via ahd_flush_qoutfifo(). Remove "Now packetized" diagnostic now that this information is incorperated into the actual negotiation messages that are displayed. When forcing renegotiation, don't clober the current ppr_options. Much of the driver uses this information to determine if we are currently packetized or not. Remove some stray spaces at column 1 in ahd_set_tags. When complaining about getting a host message loop request with no pending messages, print out the SCB_CONTROL register down on the card. Modify the ahd_sent_msg() routine to handle a search for an outgoing identify message. Use this to detect a msg reject on an identify message which typically indicates that the target thought we were packetized. Force a renegotiation in this case. In ahd_search_qinfifo(), wait more effectively for SCB DMA activities to cease. We also disable SCB fetch operations since we are about to change the qinfifo and any fetch in progress will likely be invalidated. In ahd_qinfifo_count(), fix the qinfifo empty case. In ahd_dump_card_state(), print out CCSCBCTL in the correct mode. If we are a narrow controller, don't set the current width to unknown when forcing a future negotiation. This just confuses the code into attempting a wide negotiation on a narrow bus. Add support for task management function completions. Modify ahd_handle_devreset so that it can handle lun resets in addition to target resets. Use ahd_handle_devreset for lun and target reset task management functions. Handle the abort task TMF race case better. We now wait until any current selections are over and then set the TMF back to zero. This should cause the sequencer to ignore the abort TMF completion should it occur. Correct a bug in the illegal phase handler that caused us to drop down to narrow when handling the unexpected command phase case after 3rd party reset of a packetized device. Indicate the features, bugs, and flags set in the softc that are used to control firmware patch download when booting verbose. aic79xx.h: Add coalessing and HS_MAILBOX fields. Add per-softc variables for the stats "daemon". Add a debug option for interrupt coalessing activities. Add two new softc flags: o AHD_UPDATE_PEND_CMDS Run ahd_reset_cmds_pending() on the next unpause. o AHD_RUNNING_QOUTFIFO Used to catch recursion through ahd_run_qoutfifo(). aic79xx.reg: Correct register addresses related to the software timer and the DFDBCTL register. Add constants paramaterizing the software timer. Add scratch ram locations for storing interrupt coalessing tunables. Break INTMASK in SEQITNCTL out into INTMASK1 and INTMASK2. In at least the REV A, these are writable bits. We make use of that for a swtimer workaround in the sequencer. Since HS_MAILBOX autoclears, provide a sequencer variable to store its contents. Add SEQINT codes for handling task management completions. aic79xx.seq: Correct ignore wide residue processing check for a wide negotiation being in effect. We must be in the SCSI register window in order to access the negotiation table. Use the software timer and a commands completed count to implement interrupt coalessing. The command complete is deferred until either the maximum command threshold or a the expiration of a command deferral timer. If we have more SCBs to complete to the host (sitting in COMPLETE_SCB lists), always try to coaless them up to our coalessing limit. If coalessing is enabled, but we have fewer commands oustanting than the host's min coalessing limit, complete the command immediately. Add code to track the number of commands outstanding. Commands are outstanding from the time they are placed into the execution queue until the DMA to post completion is setup. Add a workaround for intvec_2 interrupts on the H2A4. In H2A4, the mode pointer is not saved for intvec2, but is restored on iret. This can lead to the restoration of a bogus mode ptr. Manually clear the intmask bits and do a normal return to compensate. We use intvec_2 to track interrupt coalessing timeouts. Since we cannot disable the swtimer's countdown, simply mask its interrupt once we no longer care about it firing. In idle_loop_cchan, update LOCAL_HS_MAILBOX everytime we are notified of an HS_MAILBOX update via the HS_MAILBOX_ACT bit in QOFF_CTLSTA. We have to use a local copy of persistant portions of the HS_MAILBOX as the mailbox auto-clears on any read. Move the test for the cfg4istat interrupt up an instruction to hopefully close a race between the next outgoing selection and our disabling of selections. Add a missing ret to the last instruction in load_overrun_buf. Add notifications to the host of task management completions as well as the completions for commands that completed successfully before their corresponding TMF could be sent. Hold a critical section during select-out processing until we have a fully identified connection. This removes a race condition with the legacy abort handler. Correct a few spelling errors in some comments. aic79xx_inline.h: Call ahd_reset_cmds_pending() in ahd_unpause if required. Update cmdcmplt interrupt statistics in our interrupt handler. Allow callers to ahd_send_scb() to set the task management function. aic79xx_pci.c: Disable SERR and pause the controller prior to performing our mmapped I/O test. The U320 controllers do not support "auto-access-pause". aic79xx_osm.c: Set the task management function now that ahd_send_scb() doesn't do it for us. We also perform a lun reset in response to BDR requests to packetized devices.
2003-01-20 20:17:35 +00:00
AHD_RESET_POLL_ACTIVE = 0x200000,
AHD_UPDATE_PEND_CMDS = 0x400000,
AHD_RUNNING_QOUTFIFO = 0x800000
} ahd_flag;
/************************* Hardware SCB Definition ***************************/
/*
* The driver keeps up to MAX_SCB scb structures per card in memory. The SCB
Correct spelling errors. Switch to handling bad SCSI status as a sequencer interrupt instead of having the kernel proccess these failures via the completion queue. This is done because: o The old scheme required us to pause the sequencer and clear critical sections for each SCB. It seems that these pause actions, if coincident with a sequencer FIFO interrupt, would result in a FIFO interrupt getting lost or directing to the wrong FIFO. This caused hangs when the driver was stressed under high "queue full" loads. o The completion code assumed that it was always called with the sequencer running. This may not be the case in timeout processing where completions occur manually via ahd_pause_and_flushwork(). o With this scheme, the extra expense of clearing critical sections is avoided since the sequencer will only self pause once all pending selections have cleared and it is not in a critical section. aic79xx.c Add code to handle the new BAD_SCB_STATUS sequencer interrupt code. This just redirects the SCB through the already existing ahd_complete_scb() code path. Remove code in ahd_handle_scsi_status() that paused the sequencer, made sure that no selections where pending, and cleared critical sections. Bad status SCBs are now only processed when all of these conditions are true. aic79xx.reg: Add the BAD_SCB_STATUS sequencer interrupt code. aic79xx.seq: When completing an SCB upload to the host, if we are doing this because the SCB contains non-zero SCSI status, defer completing the SCB until there are no pending selection events. When completing these SCBs, use the new BAD_SCB_STATUS sequencer interrupt. For all other uploaded SCBs (currently only for underruns), the SCB is completed via the normal done queue. Additionally, keep the SCB that is currently being uploaded on the COMPLETE_DMA_SCB list until the dma is completed, not just until the DMA is started. This ensures that the DMA is restarted properly should the host disable the DMA transfer for some reason. In our RevA workaround for Maxtor drives, guard against the host pausing us while trying to pause I/O until the first data-valid REQ by clearing the current snapshot so that we can tell if the transfer has completed prior to us noticing the REQINIT status. In cfg4data_intr, shave off an instruction before getting the data path running by adding an entrypoint to the overrun handler to also increment the FIFO use count. In the overrun handler, be sure to clear our LONGJMP address in both exit paths. Perform a few sequencer optimizations. aic79xx.c: Print the full path from the SCB when a packetized status overrun occurs. Remove references to LONGJMP_SCB which is being removed from firmware usage. Print the new SCB_FIFO_USE_COUNT field in the per-SCB section of ahd_dump_card_state(). The SCB_TAG field is now re-used by the sequencer, so it no longer makes sense to reference this field in the kernel driver. aic79xx.h: Re-arrange fields in the hardware SCB from largest size type to smallest. This makes it easier to move fields without changing field alignment. The hardware scb tag field is now down near the "spare" portion of the SCB to facilitate reuse by the sequencer. aic79xx.reg: Remove LONGJMP_ADDR. Rearrange SCB fields to match aic79xx.h. Add SCB_FIFO_USE_COUNT as the first byte of the SCB_TAG field. aic79xx.seq: Add a per-SCB "Fifos in use count" field and use it to determine when it is safe (all data posted) to deliver status back to the host. The old method involved polling one or both FIFOs to verify that the current task did not have pending data. This makes running down the GSFIFO very cheap, so we will empty the GSFIFO in one idle loop pass in all cases. Use this simplification of the completion process to prune down the data FIFO teardown sequencer for packetized transfers. Much more code is now shared between the data residual and transfer complete cases. Correct some issues in the packetized status handler. It used to be possible to CLRCHN our FIFO before status had fully transferred to the host. We also failed to handle NONPACKREQ phases that could occur should a CRC error occur during transmission of the status data packet. Correct a few big endian issues: aic79xx.c: aic79xx_inline.h: aic79xx_pci.c: aic79xx_osm.c: o Always get the SCB's tag via the SCB_GET_TAG acccessor o Add missing use of byte swapping macros when touching hscb fields. o Don't double swap SEEPROM data when it is printed. Correct a big-endian bug. We cannot assign a o When assigning a 32bit LE variable to a 64bit LE variable, we must be explict about how the words of the 64bit LE variable are initialized. Cast to (uint32_t*) to do this. aic79xx.c: In ahd_clear_critical_section(), hit CRLSCSIINT after restoring the interrupt masks to avoid what appears to be a glitch on SCSIINT. Any real SCSIINT status will be persistent and will immidiately reset SCSIINT. This clear should only get rid of spurious SCSIINTs. This glitch was the cause of the "Unexpected PKT busfree" status that occurred under high queue full loads Call ahd_fini_scbdata() after shutdown so that any ahd_chip_init() routine that might access SCB data will not access free'd memory. Reset the bus on an IOERR since the chip doesn't seem to reset to the new voltage level without this. Change offset calculation for scatter gather maps so that the calculation is correct if an integral multiple of sg lists does not fit in the allocation size. Adjust bus dma tag for data buffers based on 39BIT addressing flag in our softc. Use the QFREEZE count to simplify ahd_pause_and_flushworkd(). We can thus rely on the sequencer eventually clearing ENSELO. In ahd_abort_scbs(), fix a bug that could potentially corrupt sequencer state. The saved SCB was being restored in the SCSI mode instead of the saved mode. It turns out that the SCB did not need to be saved at all as the scbptr is already restored by all subroutines called during this function that modify that register. aic79xx.c: aic79xx.h: aic79xx_pci.c: Add support for parsing the seeprom vital product data. The VPD data are currently unused. aic79xx.h: aic79xx.seq: aic79xx_pci.c: Add a firmware workaround to make the LED blink brighter during packetized operations on the H2A. aic79xx_inline.h: The host does not use timer interrupts, so don't gate our decision on whether or not to unpause the sequencer on whether or not a timer interrupt is pending.
2003-05-04 00:20:07 +00:00
* consists of a "hardware SCB" mirroring the fields available on the card
* and additional information the kernel stores for each transaction.
*
* To minimize space utilization, a portion of the hardware scb stores
* different data during different portions of a SCSI transaction.
* As initialized by the host driver for the initiator role, this area
* contains the SCSI cdb (or a pointer to the cdb) to be executed. After
* the cdb has been presented to the target, this area serves to store
* residual transfer information and the SCSI status byte.
* For the target role, the contents of this area do not change, but
* still serve a different purpose than for the initiator role. See
* struct target_data for details.
*/
/*
* Status information embedded in the shared poriton of
* an SCB after passing the cdb to the target. The kernel
* driver will only read this data for transactions that
* complete abnormally.
*/
struct initiator_status {
uint32_t residual_datacnt; /* Residual in the current S/G seg */
uint32_t residual_sgptr; /* The next S/G for this transfer */
uint8_t scsi_status; /* Standard SCSI status byte */
};
struct target_status {
uint32_t residual_datacnt; /* Residual in the current S/G seg */
uint32_t residual_sgptr; /* The next S/G for this transfer */
uint8_t scsi_status; /* SCSI status to give to initiator */
uint8_t target_phases; /* Bitmap of phases to execute */
uint8_t data_phase; /* Data-In or Data-Out */
uint8_t initiator_tag; /* Initiator's transaction tag */
};
/*
* Initiator mode SCB shared data area.
* If the embedded CDB is 12 bytes or less, we embed
* the sense buffer address in the SCB. This allows
aic79xx.c: Clear the LQICRC_NLQ status should it pop up after we have already handled the SCSIPERR. During some streaming operations this status can be delayed until the stream ends. Without this change, the driver would complain about a "Missing case in ahd_handle_scsiint". In the LQOBUSFREE handler... Don't return the LQOMGR back to the idle state until after we have cleaned up ENSELO and any status related to this selection. The last thing we need is the LQO manager starting another select-out before we have updated the execution queue. It is not clear whether the LQOMGR would, or would not start a new selection early. Make sure ENSELO is off prior to clearing SELDO by flushing device writes. Move assignment of the next target SCB pointer inside of an if to make the code clearer. The effect is the same. Dump card state in both "Unexpected PKT busfree" paths. In ahd_reset(), set the chip to SCSI mode before reading SXFRCTL1. That register only exists in the SCSI mode. Also set the mode explicitly to the SCSI mode after chip reset due to paranoia. Re-arrange code so that SXFRCTL1 is restored as quickly after the chip reset as possible. S/G structurs must be 8byte aligned. Make this official by saying so in our DMA tag. Disable CIO bus stretch on MDFFSTAT if SHVALID is about to come true. This can cause a CIO bus lockup if a PCI or PCI-X error occurs while the stretch is occurring - the host cannot service the PCI-X error since the CIO bus is locked out and SHVALID will never resolve. The stretch was added in the Rev B to simplify the wait for SHVALID to resolve, but the code to do this in the open source sequencer is so simple it was never removed. Consistently use MAX_OFFSET for the user max syncrate set from non-volatile storage. This ensures that the offset does not conflict with AH?_OFFSET_UNKNOWN. Have ahd_pause_and_flushwork set the mode to ensure that it has access to the registers it checks. Also modify the checking of intstat so that the check against 0xFF can actually succeed if the INT_PEND mask is something other than 0xFF. Although there are no cardbus U320 controllers, this check may be needed to recover from a hot-plug PCI removal that occurs without informing the driver. Fix a typo. sg_prefetch_cnt -> sg_prefetch_align. This fixes an infinite loop at card initialization if the cacheline size is 0. aic79xx.h: Add AHD_EARLY_REQ_BUG bug flag. Fix spelling errors. Include the CDB's length just after the CDB pointer in the DMA'ed CDB case. Change AH?_OFFSET_UNKNOWN to 0xFF. This is a value that the curr->offset can never be, unlike '0' which we previously used. This fixes code that only checks for a non-zero offset to determine if a sync negotiation is required since it will fire in the unknown case even if the goal is async. aic79xx.reg: Add comments for LQISTAT bits indicating their names in the 7902 data book. We use slightly different and more descriptive names in the firmware. Fix spelling errors. Include the CDB's length just after the CDB pointer in the DMA'ed CDB case. aic79xx.seq: Update comments regarding rundown of the GSFIFO to reflect reality. Fix spelling errors. Since we use an 8byte address and 1 byte length, shorten the size of a block move for the legacy DMA'ed CDB case from 11 to 9 bytes. Remove code that, assuming the abort pending feature worked, would set MK_MESSAGE in the SCB's control byte on completion to catch invalid reselections. Since we don't see interrupts for completed selections, this status update could occur prior to us noticing the SELDO. The "select-out" queue logic will get confused by the MK_MESSAGE bit being set as this is used to catch packatized connections where we select-out with ATN. Since the abort pending feature doesn't work on any released controllers yet, this code was never executed. Add support for the AHD_EARLY_REQ_BUG. Don't ignore persistent REQ assertions just because they were asserted within the bus settle delay window. This allows us to tolerate devices like the GEM318 that violate the SCSI spec. Remove unintentional settnig of SG_CACHE_AVAIL. Writing this bit should have no effect, but who knows... On the Rev A, we must wait for HDMAENACK before loading additional segments to avoid clobbering the address of the first segment in the S/G FIFO. This resolves data-corruption issues with certain IBM (now Hitachi) and Fujitsu U320 drives. Rearrange calc_residual to avoid an extra jmp instruction. On RevA Silicon, if the target returns us to data-out after we have already trained for data-out, it is possible for us to transition the free running clock to data-valid before the required 100ns P1 setup time (8 P1 assertions in fast-160 mode). This will only happen if this L-Q is a continuation of a data transfer for which we have already prefetched data into our FIFO (LQ/Data followed by LQ/Data for the same write transaction). This can cause some target implementations to miss the first few data transfers on the bus. We detect this situation by noticing that this is the first data transfer after an LQ (LQIWORKONLQ true), that the data transfer is a continuation of a transfer already setup in our FIFO (SAVEPTRS interrupt), and that the transaction is a write (DIRECTION set in DFCNTRL). The delay is performed by disabling SCSIEN until we see the first REQ from the target. Only compile in snapshot savepointers handler for RevA silicon where it is enabled. Handle the cfg4icmd packetized interrupt. We just need to load the address and count, start the DMA, and CLRCHN once the transfer is complete. Fix an oversight in the overrun handler for packetized status operations. We need to wait for either CTXTDONE or an overrun when checking for an overrun. The previous code did not wait and thus could decide that no overrun had occurred even though an overrun will occur on the next data-valid req. Add some comment to this section for clarity. Use LAST_SEG_DONE instead of LASTSDONE for testing transfer completion in the packetized status case. LASTSDONE may come up more quickly since it only records completion on the SCSI side, but since LAST_SEG_DONE is used everywhere else (and needs to be), this is less confusing. Add a missing invalidation of the longjmp address in the non-pack handler. This code needs additional review. aic79xx_inline.h: Fix spelling error. aic79xx_osm.c: Set the cdb length for CDBs dma'ed from host memory. Add a comment indicating that, should CAM start supporting cdbs larger than 16bytes, the driver could store the CDB in the status buffer. aic79xx_pci.c: Add a table entry for the 39320A. Added a missing comma to an error string table. Fix spelling errors.
2003-02-27 23:23:16 +00:00
* us to retrieve sense information without interrupting
* the host in packetized mode.
*/
typedef uint32_t sense_addr_t;
#define MAX_CDB_LEN 16
#define MAX_CDB_LEN_WITH_SENSE_ADDR (MAX_CDB_LEN - sizeof(sense_addr_t))
union initiator_data {
aic79xx.c: Clear the LQICRC_NLQ status should it pop up after we have already handled the SCSIPERR. During some streaming operations this status can be delayed until the stream ends. Without this change, the driver would complain about a "Missing case in ahd_handle_scsiint". In the LQOBUSFREE handler... Don't return the LQOMGR back to the idle state until after we have cleaned up ENSELO and any status related to this selection. The last thing we need is the LQO manager starting another select-out before we have updated the execution queue. It is not clear whether the LQOMGR would, or would not start a new selection early. Make sure ENSELO is off prior to clearing SELDO by flushing device writes. Move assignment of the next target SCB pointer inside of an if to make the code clearer. The effect is the same. Dump card state in both "Unexpected PKT busfree" paths. In ahd_reset(), set the chip to SCSI mode before reading SXFRCTL1. That register only exists in the SCSI mode. Also set the mode explicitly to the SCSI mode after chip reset due to paranoia. Re-arrange code so that SXFRCTL1 is restored as quickly after the chip reset as possible. S/G structurs must be 8byte aligned. Make this official by saying so in our DMA tag. Disable CIO bus stretch on MDFFSTAT if SHVALID is about to come true. This can cause a CIO bus lockup if a PCI or PCI-X error occurs while the stretch is occurring - the host cannot service the PCI-X error since the CIO bus is locked out and SHVALID will never resolve. The stretch was added in the Rev B to simplify the wait for SHVALID to resolve, but the code to do this in the open source sequencer is so simple it was never removed. Consistently use MAX_OFFSET for the user max syncrate set from non-volatile storage. This ensures that the offset does not conflict with AH?_OFFSET_UNKNOWN. Have ahd_pause_and_flushwork set the mode to ensure that it has access to the registers it checks. Also modify the checking of intstat so that the check against 0xFF can actually succeed if the INT_PEND mask is something other than 0xFF. Although there are no cardbus U320 controllers, this check may be needed to recover from a hot-plug PCI removal that occurs without informing the driver. Fix a typo. sg_prefetch_cnt -> sg_prefetch_align. This fixes an infinite loop at card initialization if the cacheline size is 0. aic79xx.h: Add AHD_EARLY_REQ_BUG bug flag. Fix spelling errors. Include the CDB's length just after the CDB pointer in the DMA'ed CDB case. Change AH?_OFFSET_UNKNOWN to 0xFF. This is a value that the curr->offset can never be, unlike '0' which we previously used. This fixes code that only checks for a non-zero offset to determine if a sync negotiation is required since it will fire in the unknown case even if the goal is async. aic79xx.reg: Add comments for LQISTAT bits indicating their names in the 7902 data book. We use slightly different and more descriptive names in the firmware. Fix spelling errors. Include the CDB's length just after the CDB pointer in the DMA'ed CDB case. aic79xx.seq: Update comments regarding rundown of the GSFIFO to reflect reality. Fix spelling errors. Since we use an 8byte address and 1 byte length, shorten the size of a block move for the legacy DMA'ed CDB case from 11 to 9 bytes. Remove code that, assuming the abort pending feature worked, would set MK_MESSAGE in the SCB's control byte on completion to catch invalid reselections. Since we don't see interrupts for completed selections, this status update could occur prior to us noticing the SELDO. The "select-out" queue logic will get confused by the MK_MESSAGE bit being set as this is used to catch packatized connections where we select-out with ATN. Since the abort pending feature doesn't work on any released controllers yet, this code was never executed. Add support for the AHD_EARLY_REQ_BUG. Don't ignore persistent REQ assertions just because they were asserted within the bus settle delay window. This allows us to tolerate devices like the GEM318 that violate the SCSI spec. Remove unintentional settnig of SG_CACHE_AVAIL. Writing this bit should have no effect, but who knows... On the Rev A, we must wait for HDMAENACK before loading additional segments to avoid clobbering the address of the first segment in the S/G FIFO. This resolves data-corruption issues with certain IBM (now Hitachi) and Fujitsu U320 drives. Rearrange calc_residual to avoid an extra jmp instruction. On RevA Silicon, if the target returns us to data-out after we have already trained for data-out, it is possible for us to transition the free running clock to data-valid before the required 100ns P1 setup time (8 P1 assertions in fast-160 mode). This will only happen if this L-Q is a continuation of a data transfer for which we have already prefetched data into our FIFO (LQ/Data followed by LQ/Data for the same write transaction). This can cause some target implementations to miss the first few data transfers on the bus. We detect this situation by noticing that this is the first data transfer after an LQ (LQIWORKONLQ true), that the data transfer is a continuation of a transfer already setup in our FIFO (SAVEPTRS interrupt), and that the transaction is a write (DIRECTION set in DFCNTRL). The delay is performed by disabling SCSIEN until we see the first REQ from the target. Only compile in snapshot savepointers handler for RevA silicon where it is enabled. Handle the cfg4icmd packetized interrupt. We just need to load the address and count, start the DMA, and CLRCHN once the transfer is complete. Fix an oversight in the overrun handler for packetized status operations. We need to wait for either CTXTDONE or an overrun when checking for an overrun. The previous code did not wait and thus could decide that no overrun had occurred even though an overrun will occur on the next data-valid req. Add some comment to this section for clarity. Use LAST_SEG_DONE instead of LASTSDONE for testing transfer completion in the packetized status case. LASTSDONE may come up more quickly since it only records completion on the SCSI side, but since LAST_SEG_DONE is used everywhere else (and needs to be), this is less confusing. Add a missing invalidation of the longjmp address in the non-pack handler. This code needs additional review. aic79xx_inline.h: Fix spelling error. aic79xx_osm.c: Set the cdb length for CDBs dma'ed from host memory. Add a comment indicating that, should CAM start supporting cdbs larger than 16bytes, the driver could store the CDB in the status buffer. aic79xx_pci.c: Add a table entry for the 39320A. Added a missing comma to an error string table. Fix spelling errors.
2003-02-27 23:23:16 +00:00
struct {
uint64_t cdbptr;
uint8_t cdblen;
} cdb_from_host;
uint8_t cdb[MAX_CDB_LEN];
struct {
uint8_t cdb[MAX_CDB_LEN_WITH_SENSE_ADDR];
sense_addr_t sense_addr;
} cdb_plus_saddr;
};
/*
* Target mode version of the shared data SCB segment.
*/
struct target_data {
uint32_t spare[2];
uint8_t scsi_status; /* SCSI status to give to initiator */
uint8_t target_phases; /* Bitmap of phases to execute */
uint8_t data_phase; /* Data-In or Data-Out */
uint8_t initiator_tag; /* Initiator's transaction tag */
};
struct hardware_scb {
/*0*/ union {
union initiator_data idata;
struct target_data tdata;
struct initiator_status istatus;
struct target_status tstatus;
} shared_data;
/*
* A word about residuals.
* The scb is presented to the sequencer with the dataptr and datacnt
* fields initialized to the contents of the first S/G element to
* transfer. The sgptr field is initialized to the bus address for
* the S/G element that follows the first in the in core S/G array
* or'ed with the SG_FULL_RESID flag. Sgptr may point to an invalid
* S/G entry for this transfer (single S/G element transfer with the
* first elements address and length preloaded in the dataptr/datacnt
* fields). If no transfer is to occur, sgptr is set to SG_LIST_NULL.
* The SG_FULL_RESID flag ensures that the residual will be correctly
* noted even if no data transfers occur. Once the data phase is entered,
* the residual sgptr and datacnt are loaded from the sgptr and the
* datacnt fields. After each S/G element's dataptr and length are
* loaded into the hardware, the residual sgptr is advanced. After
* each S/G element is expired, its datacnt field is checked to see
* if the LAST_SEG flag is set. If so, SG_LIST_NULL is set in the
* residual sg ptr and the transfer is considered complete. If the
* sequencer determines that there is a residual in the tranfer, or
* there is non-zero status, it will set the SG_STATUS_VALID flag in
* sgptr and dma the scb back into host memory. To sumarize:
*
* Sequencer:
* o A residual has occurred if SG_FULL_RESID is set in sgptr,
* or residual_sgptr does not have SG_LIST_NULL set.
*
* o We are transfering the last segment if residual_datacnt has
* the SG_LAST_SEG flag set.
*
* Host:
* o A residual can only have occurred if a completed scb has the
* SG_STATUS_VALID flag set. Inspection of the SCSI status field,
* the residual_datacnt, and the residual_sgptr field will tell
* for sure.
*
* o residual_sgptr and sgptr refer to the "next" sg entry
* and so may point beyond the last valid sg entry for the
* transfer.
*/
#define SG_PTR_MASK 0xFFFFFFF8
/*16*/ uint16_t tag; /* Reused by Sequencer. */
/*18*/ uint8_t control; /* See SCB_CONTROL in aic79xx.reg for details */
/*19*/ uint8_t scsiid; /*
* Selection out Id
* Our Id (bits 0-3) Their ID (bits 4-7)
*/
/*20*/ uint8_t lun;
/*21*/ uint8_t task_attribute;
/*22*/ uint8_t cdb_len;
/*23*/ uint8_t task_management;
/*24*/ uint64_t dataptr;
/*32*/ uint32_t datacnt; /* Byte 3 is spare. */
/*36*/ uint32_t sgptr;
/*40*/ uint32_t hscb_busaddr;
/*44*/ uint32_t next_hscb_busaddr;
Correct spelling errors. Switch to handling bad SCSI status as a sequencer interrupt instead of having the kernel proccess these failures via the completion queue. This is done because: o The old scheme required us to pause the sequencer and clear critical sections for each SCB. It seems that these pause actions, if coincident with a sequencer FIFO interrupt, would result in a FIFO interrupt getting lost or directing to the wrong FIFO. This caused hangs when the driver was stressed under high "queue full" loads. o The completion code assumed that it was always called with the sequencer running. This may not be the case in timeout processing where completions occur manually via ahd_pause_and_flushwork(). o With this scheme, the extra expense of clearing critical sections is avoided since the sequencer will only self pause once all pending selections have cleared and it is not in a critical section. aic79xx.c Add code to handle the new BAD_SCB_STATUS sequencer interrupt code. This just redirects the SCB through the already existing ahd_complete_scb() code path. Remove code in ahd_handle_scsi_status() that paused the sequencer, made sure that no selections where pending, and cleared critical sections. Bad status SCBs are now only processed when all of these conditions are true. aic79xx.reg: Add the BAD_SCB_STATUS sequencer interrupt code. aic79xx.seq: When completing an SCB upload to the host, if we are doing this because the SCB contains non-zero SCSI status, defer completing the SCB until there are no pending selection events. When completing these SCBs, use the new BAD_SCB_STATUS sequencer interrupt. For all other uploaded SCBs (currently only for underruns), the SCB is completed via the normal done queue. Additionally, keep the SCB that is currently being uploaded on the COMPLETE_DMA_SCB list until the dma is completed, not just until the DMA is started. This ensures that the DMA is restarted properly should the host disable the DMA transfer for some reason. In our RevA workaround for Maxtor drives, guard against the host pausing us while trying to pause I/O until the first data-valid REQ by clearing the current snapshot so that we can tell if the transfer has completed prior to us noticing the REQINIT status. In cfg4data_intr, shave off an instruction before getting the data path running by adding an entrypoint to the overrun handler to also increment the FIFO use count. In the overrun handler, be sure to clear our LONGJMP address in both exit paths. Perform a few sequencer optimizations. aic79xx.c: Print the full path from the SCB when a packetized status overrun occurs. Remove references to LONGJMP_SCB which is being removed from firmware usage. Print the new SCB_FIFO_USE_COUNT field in the per-SCB section of ahd_dump_card_state(). The SCB_TAG field is now re-used by the sequencer, so it no longer makes sense to reference this field in the kernel driver. aic79xx.h: Re-arrange fields in the hardware SCB from largest size type to smallest. This makes it easier to move fields without changing field alignment. The hardware scb tag field is now down near the "spare" portion of the SCB to facilitate reuse by the sequencer. aic79xx.reg: Remove LONGJMP_ADDR. Rearrange SCB fields to match aic79xx.h. Add SCB_FIFO_USE_COUNT as the first byte of the SCB_TAG field. aic79xx.seq: Add a per-SCB "Fifos in use count" field and use it to determine when it is safe (all data posted) to deliver status back to the host. The old method involved polling one or both FIFOs to verify that the current task did not have pending data. This makes running down the GSFIFO very cheap, so we will empty the GSFIFO in one idle loop pass in all cases. Use this simplification of the completion process to prune down the data FIFO teardown sequencer for packetized transfers. Much more code is now shared between the data residual and transfer complete cases. Correct some issues in the packetized status handler. It used to be possible to CLRCHN our FIFO before status had fully transferred to the host. We also failed to handle NONPACKREQ phases that could occur should a CRC error occur during transmission of the status data packet. Correct a few big endian issues: aic79xx.c: aic79xx_inline.h: aic79xx_pci.c: aic79xx_osm.c: o Always get the SCB's tag via the SCB_GET_TAG acccessor o Add missing use of byte swapping macros when touching hscb fields. o Don't double swap SEEPROM data when it is printed. Correct a big-endian bug. We cannot assign a o When assigning a 32bit LE variable to a 64bit LE variable, we must be explict about how the words of the 64bit LE variable are initialized. Cast to (uint32_t*) to do this. aic79xx.c: In ahd_clear_critical_section(), hit CRLSCSIINT after restoring the interrupt masks to avoid what appears to be a glitch on SCSIINT. Any real SCSIINT status will be persistent and will immidiately reset SCSIINT. This clear should only get rid of spurious SCSIINTs. This glitch was the cause of the "Unexpected PKT busfree" status that occurred under high queue full loads Call ahd_fini_scbdata() after shutdown so that any ahd_chip_init() routine that might access SCB data will not access free'd memory. Reset the bus on an IOERR since the chip doesn't seem to reset to the new voltage level without this. Change offset calculation for scatter gather maps so that the calculation is correct if an integral multiple of sg lists does not fit in the allocation size. Adjust bus dma tag for data buffers based on 39BIT addressing flag in our softc. Use the QFREEZE count to simplify ahd_pause_and_flushworkd(). We can thus rely on the sequencer eventually clearing ENSELO. In ahd_abort_scbs(), fix a bug that could potentially corrupt sequencer state. The saved SCB was being restored in the SCSI mode instead of the saved mode. It turns out that the SCB did not need to be saved at all as the scbptr is already restored by all subroutines called during this function that modify that register. aic79xx.c: aic79xx.h: aic79xx_pci.c: Add support for parsing the seeprom vital product data. The VPD data are currently unused. aic79xx.h: aic79xx.seq: aic79xx_pci.c: Add a firmware workaround to make the LED blink brighter during packetized operations on the H2A. aic79xx_inline.h: The host does not use timer interrupts, so don't gate our decision on whether or not to unpause the sequencer on whether or not a timer interrupt is pending.
2003-05-04 00:20:07 +00:00
/********** Long lun field only downloaded for full 8 byte lun support ********/
/*48*/ uint8_t pkt_long_lun[8];
/******* Fields below are not Downloaded (Sequencer may use for scratch) ******/
/*56*/ uint8_t spare[8];
};
/************************ Kernel SCB Definitions ******************************/
/*
* Some fields of the SCB are OS dependent. Here we collect the
* definitions for elements that all OS platforms need to include
* in there SCB definition.
*/
/*
* Definition of a scatter/gather element as transfered to the controller.
* The aic7xxx chips only support a 24bit length. We use the top byte of
* the length to store additional address bits and a flag to indicate
* that a given segment terminates the transfer. This gives us an
* addressable range of 512GB on machines with 64bit PCI or with chips
* that can support dual address cycles on 32bit PCI busses.
*/
struct ahd_dma_seg {
uint32_t addr;
uint32_t len;
#define AHD_DMA_LAST_SEG 0x80000000
#define AHD_SG_HIGH_ADDR_MASK 0x7F000000
#define AHD_SG_LEN_MASK 0x00FFFFFF
};
struct ahd_dma64_seg {
uint64_t addr;
uint32_t len;
uint32_t pad;
};
struct map_node {
bus_dmamap_t dmamap;
bus_addr_t physaddr;
uint8_t *vaddr;
SLIST_ENTRY(map_node) links;
};
/*
* The current state of this SCB.
*/
typedef enum {
SCB_FLAG_NONE = 0x00000,
SCB_TRANSMISSION_ERROR = 0x00001,/*
* We detected a parity or CRC
* error that has effected the
* payload of the command. This
* flag is checked when normal
* status is returned to catch
* the case of a target not
* responding to our attempt
* to report the error.
*/
SCB_OTHERTCL_TIMEOUT = 0x00002,/*
* Another device was active
* during the first timeout for
* this SCB so we gave ourselves
* an additional timeout period
* in case it was hogging the
* bus.
*/
SCB_DEVICE_RESET = 0x00004,
SCB_SENSE = 0x00008,
SCB_CDB32_PTR = 0x00010,
SCB_RECOVERY_SCB = 0x00020,
SCB_AUTO_NEGOTIATE = 0x00040,/* Negotiate to achieve goal. */
SCB_NEGOTIATE = 0x00080,/* Negotiation forced for command. */
SCB_ABORT = 0x00100,
SCB_ACTIVE = 0x00200,
SCB_TARGET_IMMEDIATE = 0x00400,
SCB_PACKETIZED = 0x00800,
SCB_EXPECT_PPR_BUSFREE = 0x01000,
SCB_PKT_SENSE = 0x02000,
SCB_CMDPHASE_ABORT = 0x04000,
aic79xx.c: Implement the SCB_SILENT flag. This is useful for hushing up the driver during DV or other operations that we expect to cause transmission errors. The messages will still print if the SHOW_MASKED_ERRORS debug option is enabled. Save and restore the NEGOADDR address when setting new transfer settings. The sequencer performs lookups in the negotiation table too and it expects NEGOADDR to remain consistent across pause/unpause sessions. Consistently use "offset" instead of "period" to determine if we are running sync or not. Add a SHOW_MESSAGES diagnostic for when we assert ATN during message processing. Print out IU, QAS, and RTI features when showing transfer options. Limit the syncrate after all option conformance changes have taken place in ahd_devlimited_syncrate. Changes in options may change the final syncrate we accept. Keep a copy of the hs_mailbox in our softc so that we can perform read/modify/write operations on the hs_mailbox without having to pause the sequencer to read the last written value. Use the ENINT_COALESS flag in the hs_mailbox to toggle interrupt coalessing. Add entrypoints for enabling interrupt coalessing and setting both a timeout (how long to wait for commands to be coalessed) and a maximum commands to coaless value. Add a statistics timer that decides when to enable or disable interrupt coalessing based on load. Add a routine, ahd_reset_cmds_pending() which is used to update the CMDS_PENDING sequencer variable whenever error recovery compeltes SCBs without notifying the sequencer. Since ahd_reset_cmds_pending is called during ahd_unpause() only if we've aborted SCBs, its call to ahd_flush_qoutfifo should not cause recursion through ahd_run_qoutfifo(). A panic has been added to ensure that this recursion does not occur. In ahd_search_qinfifo, update the CMDS_PENDING sequencer variable directly. ahd_search_qinififo can be called in situations where using ahd_reset_cmds_pending() might cause recursion. Since we can safely determine the exact number to reduce CMDS_PENDING by in this scenario without running the qoutfifo, the manual update is sufficient. Clean up diagnostics. Add ahd_flush_qoutfifo() which will run the qoutfifo as well as complete any commands sitting on the sequencer's COMPLETE_SCB lists or the good status FIFO. Use this routine in several places that did similar things in an add-hoc, but incomplete, fashion. A call to this routine was also added to ahd_abort_scbs() to close a race. In ahd_pause_and_flushwork() only return once selections are safely disabled. Flush all completed commands via ahd_flush_qoutfifo(). Remove "Now packetized" diagnostic now that this information is incorperated into the actual negotiation messages that are displayed. When forcing renegotiation, don't clober the current ppr_options. Much of the driver uses this information to determine if we are currently packetized or not. Remove some stray spaces at column 1 in ahd_set_tags. When complaining about getting a host message loop request with no pending messages, print out the SCB_CONTROL register down on the card. Modify the ahd_sent_msg() routine to handle a search for an outgoing identify message. Use this to detect a msg reject on an identify message which typically indicates that the target thought we were packetized. Force a renegotiation in this case. In ahd_search_qinfifo(), wait more effectively for SCB DMA activities to cease. We also disable SCB fetch operations since we are about to change the qinfifo and any fetch in progress will likely be invalidated. In ahd_qinfifo_count(), fix the qinfifo empty case. In ahd_dump_card_state(), print out CCSCBCTL in the correct mode. If we are a narrow controller, don't set the current width to unknown when forcing a future negotiation. This just confuses the code into attempting a wide negotiation on a narrow bus. Add support for task management function completions. Modify ahd_handle_devreset so that it can handle lun resets in addition to target resets. Use ahd_handle_devreset for lun and target reset task management functions. Handle the abort task TMF race case better. We now wait until any current selections are over and then set the TMF back to zero. This should cause the sequencer to ignore the abort TMF completion should it occur. Correct a bug in the illegal phase handler that caused us to drop down to narrow when handling the unexpected command phase case after 3rd party reset of a packetized device. Indicate the features, bugs, and flags set in the softc that are used to control firmware patch download when booting verbose. aic79xx.h: Add coalessing and HS_MAILBOX fields. Add per-softc variables for the stats "daemon". Add a debug option for interrupt coalessing activities. Add two new softc flags: o AHD_UPDATE_PEND_CMDS Run ahd_reset_cmds_pending() on the next unpause. o AHD_RUNNING_QOUTFIFO Used to catch recursion through ahd_run_qoutfifo(). aic79xx.reg: Correct register addresses related to the software timer and the DFDBCTL register. Add constants paramaterizing the software timer. Add scratch ram locations for storing interrupt coalessing tunables. Break INTMASK in SEQITNCTL out into INTMASK1 and INTMASK2. In at least the REV A, these are writable bits. We make use of that for a swtimer workaround in the sequencer. Since HS_MAILBOX autoclears, provide a sequencer variable to store its contents. Add SEQINT codes for handling task management completions. aic79xx.seq: Correct ignore wide residue processing check for a wide negotiation being in effect. We must be in the SCSI register window in order to access the negotiation table. Use the software timer and a commands completed count to implement interrupt coalessing. The command complete is deferred until either the maximum command threshold or a the expiration of a command deferral timer. If we have more SCBs to complete to the host (sitting in COMPLETE_SCB lists), always try to coaless them up to our coalessing limit. If coalessing is enabled, but we have fewer commands oustanting than the host's min coalessing limit, complete the command immediately. Add code to track the number of commands outstanding. Commands are outstanding from the time they are placed into the execution queue until the DMA to post completion is setup. Add a workaround for intvec_2 interrupts on the H2A4. In H2A4, the mode pointer is not saved for intvec2, but is restored on iret. This can lead to the restoration of a bogus mode ptr. Manually clear the intmask bits and do a normal return to compensate. We use intvec_2 to track interrupt coalessing timeouts. Since we cannot disable the swtimer's countdown, simply mask its interrupt once we no longer care about it firing. In idle_loop_cchan, update LOCAL_HS_MAILBOX everytime we are notified of an HS_MAILBOX update via the HS_MAILBOX_ACT bit in QOFF_CTLSTA. We have to use a local copy of persistant portions of the HS_MAILBOX as the mailbox auto-clears on any read. Move the test for the cfg4istat interrupt up an instruction to hopefully close a race between the next outgoing selection and our disabling of selections. Add a missing ret to the last instruction in load_overrun_buf. Add notifications to the host of task management completions as well as the completions for commands that completed successfully before their corresponding TMF could be sent. Hold a critical section during select-out processing until we have a fully identified connection. This removes a race condition with the legacy abort handler. Correct a few spelling errors in some comments. aic79xx_inline.h: Call ahd_reset_cmds_pending() in ahd_unpause if required. Update cmdcmplt interrupt statistics in our interrupt handler. Allow callers to ahd_send_scb() to set the task management function. aic79xx_pci.c: Disable SERR and pause the controller prior to performing our mmapped I/O test. The U320 controllers do not support "auto-access-pause". aic79xx_osm.c: Set the task management function now that ahd_send_scb() doesn't do it for us. We also perform a lun reset in response to BDR requests to packetized devices.
2003-01-20 20:17:35 +00:00
SCB_ON_COL_LIST = 0x08000,
SCB_SILENT = 0x10000 /*
* Be quiet about transmission type
* errors. They are expected and we
* don't want to upset the user. This
* flag is typically used during DV.
*/
} scb_flag;
struct scb {
struct hardware_scb *hscb;
union {
SLIST_ENTRY(scb) sle;
LIST_ENTRY(scb) le;
TAILQ_ENTRY(scb) tqe;
} links;
union {
SLIST_ENTRY(scb) sle;
LIST_ENTRY(scb) le;
TAILQ_ENTRY(scb) tqe;
} links2;
#define pending_links links2.le
#define collision_links links2.le
struct scb *col_scb;
ahd_io_ctx_t io_ctx;
struct ahd_softc *ahd_softc;
scb_flag flags;
#ifndef __linux__
bus_dmamap_t dmamap;
#endif
struct scb_platform_data *platform_data;
struct map_node *hscb_map;
struct map_node *sg_map;
struct map_node *sense_map;
void *sg_list;
uint8_t *sense_data;
bus_addr_t sg_list_busaddr;
bus_addr_t sense_busaddr;
u_int sg_count;/* How full ahd_dma_seg is */
#define AHD_MAX_LQ_CRC_ERRORS 5
u_int crc_retry_count;
};
TAILQ_HEAD(scb_tailq, scb);
LIST_HEAD(scb_list, scb);
struct scb_data {
/*
* TAILQ of lists of free SCBs grouped by device
* collision domains.
*/
struct scb_tailq free_scbs;
/*
* Per-device lists of SCBs whose tag ID would collide
* with an already active tag on the device.
*/
struct scb_list free_scb_lists[AHD_NUM_TARGETS * AHD_NUM_LUNS_NONPKT];
/*
* SCBs that will not collide with any active device.
*/
struct scb_list any_dev_free_scb_list;
/*
* Mapping from tag to SCB.
*/
struct scb *scbindex[AHD_SCB_MAX];
/*
* "Bus" addresses of our data structures.
*/
bus_dma_tag_t hscb_dmat; /* dmat for our hardware SCB array */
bus_dma_tag_t sg_dmat; /* dmat for our sg segments */
bus_dma_tag_t sense_dmat; /* dmat for our sense buffers */
SLIST_HEAD(, map_node) hscb_maps;
SLIST_HEAD(, map_node) sg_maps;
SLIST_HEAD(, map_node) sense_maps;
int scbs_left; /* unallocated scbs in head map_node */
int sgs_left; /* unallocated sgs in head map_node */
int sense_left; /* unallocated sense in head map_node */
uint16_t numscbs;
uint16_t maxhscbs; /* Number of SCBs on the card */
uint8_t init_level; /*
* How far we've initialized
* this structure.
*/
};
/************************ Target Mode Definitions *****************************/
/*
* Connection desciptor for select-in requests in target mode.
*/
struct target_cmd {
uint8_t scsiid; /* Our ID and the initiator's ID */
uint8_t identify; /* Identify message */
uint8_t bytes[22]; /*
* Bytes contains any additional message
* bytes terminated by 0xFF. The remainder
* is the cdb to execute.
*/
uint8_t cmd_valid; /*
* When a command is complete, the firmware
* will set cmd_valid to all bits set.
* After the host has seen the command,
* the bits are cleared. This allows us
* to just peek at host memory to determine
* if more work is complete. cmd_valid is on
* an 8 byte boundary to simplify setting
* it on aic7880 hardware which only has
* limited direct access to the DMA FIFO.
*/
uint8_t pad[7];
};
/*
* Number of events we can buffer up if we run out
* of immediate notify ccbs.
*/
#define AHD_TMODE_EVENT_BUFFER_SIZE 8
struct ahd_tmode_event {
uint8_t initiator_id;
uint8_t event_type; /* MSG type or EVENT_TYPE_BUS_RESET */
#define EVENT_TYPE_BUS_RESET 0xFF
uint8_t event_arg;
};
/*
* Per enabled lun target mode state.
* As this state is directly influenced by the host OS'es target mode
* environment, we let the OS module define it. Forward declare the
* structure here so we can store arrays of them, etc. in OS neutral
* data structures.
*/
#ifdef AHD_TARGET_MODE
struct ahd_tmode_lstate {
struct cam_path *path;
struct ccb_hdr_slist accept_tios;
struct ccb_hdr_slist immed_notifies;
struct ahd_tmode_event event_buffer[AHD_TMODE_EVENT_BUFFER_SIZE];
uint8_t event_r_idx;
uint8_t event_w_idx;
};
#else
struct ahd_tmode_lstate;
#endif
/******************** Transfer Negotiation Datastructures *********************/
#define AHD_TRANS_CUR 0x01 /* Modify current neogtiation status */
#define AHD_TRANS_ACTIVE 0x03 /* Assume this target is on the bus */
#define AHD_TRANS_GOAL 0x04 /* Modify negotiation goal */
#define AHD_TRANS_USER 0x08 /* Modify user negotiation settings */
#define AHD_PERIOD_10MHz 0x19
#define AHD_WIDTH_UNKNOWN 0xFF
#define AHD_PERIOD_UNKNOWN 0xFF
aic79xx.c: Clear the LQICRC_NLQ status should it pop up after we have already handled the SCSIPERR. During some streaming operations this status can be delayed until the stream ends. Without this change, the driver would complain about a "Missing case in ahd_handle_scsiint". In the LQOBUSFREE handler... Don't return the LQOMGR back to the idle state until after we have cleaned up ENSELO and any status related to this selection. The last thing we need is the LQO manager starting another select-out before we have updated the execution queue. It is not clear whether the LQOMGR would, or would not start a new selection early. Make sure ENSELO is off prior to clearing SELDO by flushing device writes. Move assignment of the next target SCB pointer inside of an if to make the code clearer. The effect is the same. Dump card state in both "Unexpected PKT busfree" paths. In ahd_reset(), set the chip to SCSI mode before reading SXFRCTL1. That register only exists in the SCSI mode. Also set the mode explicitly to the SCSI mode after chip reset due to paranoia. Re-arrange code so that SXFRCTL1 is restored as quickly after the chip reset as possible. S/G structurs must be 8byte aligned. Make this official by saying so in our DMA tag. Disable CIO bus stretch on MDFFSTAT if SHVALID is about to come true. This can cause a CIO bus lockup if a PCI or PCI-X error occurs while the stretch is occurring - the host cannot service the PCI-X error since the CIO bus is locked out and SHVALID will never resolve. The stretch was added in the Rev B to simplify the wait for SHVALID to resolve, but the code to do this in the open source sequencer is so simple it was never removed. Consistently use MAX_OFFSET for the user max syncrate set from non-volatile storage. This ensures that the offset does not conflict with AH?_OFFSET_UNKNOWN. Have ahd_pause_and_flushwork set the mode to ensure that it has access to the registers it checks. Also modify the checking of intstat so that the check against 0xFF can actually succeed if the INT_PEND mask is something other than 0xFF. Although there are no cardbus U320 controllers, this check may be needed to recover from a hot-plug PCI removal that occurs without informing the driver. Fix a typo. sg_prefetch_cnt -> sg_prefetch_align. This fixes an infinite loop at card initialization if the cacheline size is 0. aic79xx.h: Add AHD_EARLY_REQ_BUG bug flag. Fix spelling errors. Include the CDB's length just after the CDB pointer in the DMA'ed CDB case. Change AH?_OFFSET_UNKNOWN to 0xFF. This is a value that the curr->offset can never be, unlike '0' which we previously used. This fixes code that only checks for a non-zero offset to determine if a sync negotiation is required since it will fire in the unknown case even if the goal is async. aic79xx.reg: Add comments for LQISTAT bits indicating their names in the 7902 data book. We use slightly different and more descriptive names in the firmware. Fix spelling errors. Include the CDB's length just after the CDB pointer in the DMA'ed CDB case. aic79xx.seq: Update comments regarding rundown of the GSFIFO to reflect reality. Fix spelling errors. Since we use an 8byte address and 1 byte length, shorten the size of a block move for the legacy DMA'ed CDB case from 11 to 9 bytes. Remove code that, assuming the abort pending feature worked, would set MK_MESSAGE in the SCB's control byte on completion to catch invalid reselections. Since we don't see interrupts for completed selections, this status update could occur prior to us noticing the SELDO. The "select-out" queue logic will get confused by the MK_MESSAGE bit being set as this is used to catch packatized connections where we select-out with ATN. Since the abort pending feature doesn't work on any released controllers yet, this code was never executed. Add support for the AHD_EARLY_REQ_BUG. Don't ignore persistent REQ assertions just because they were asserted within the bus settle delay window. This allows us to tolerate devices like the GEM318 that violate the SCSI spec. Remove unintentional settnig of SG_CACHE_AVAIL. Writing this bit should have no effect, but who knows... On the Rev A, we must wait for HDMAENACK before loading additional segments to avoid clobbering the address of the first segment in the S/G FIFO. This resolves data-corruption issues with certain IBM (now Hitachi) and Fujitsu U320 drives. Rearrange calc_residual to avoid an extra jmp instruction. On RevA Silicon, if the target returns us to data-out after we have already trained for data-out, it is possible for us to transition the free running clock to data-valid before the required 100ns P1 setup time (8 P1 assertions in fast-160 mode). This will only happen if this L-Q is a continuation of a data transfer for which we have already prefetched data into our FIFO (LQ/Data followed by LQ/Data for the same write transaction). This can cause some target implementations to miss the first few data transfers on the bus. We detect this situation by noticing that this is the first data transfer after an LQ (LQIWORKONLQ true), that the data transfer is a continuation of a transfer already setup in our FIFO (SAVEPTRS interrupt), and that the transaction is a write (DIRECTION set in DFCNTRL). The delay is performed by disabling SCSIEN until we see the first REQ from the target. Only compile in snapshot savepointers handler for RevA silicon where it is enabled. Handle the cfg4icmd packetized interrupt. We just need to load the address and count, start the DMA, and CLRCHN once the transfer is complete. Fix an oversight in the overrun handler for packetized status operations. We need to wait for either CTXTDONE or an overrun when checking for an overrun. The previous code did not wait and thus could decide that no overrun had occurred even though an overrun will occur on the next data-valid req. Add some comment to this section for clarity. Use LAST_SEG_DONE instead of LASTSDONE for testing transfer completion in the packetized status case. LASTSDONE may come up more quickly since it only records completion on the SCSI side, but since LAST_SEG_DONE is used everywhere else (and needs to be), this is less confusing. Add a missing invalidation of the longjmp address in the non-pack handler. This code needs additional review. aic79xx_inline.h: Fix spelling error. aic79xx_osm.c: Set the cdb length for CDBs dma'ed from host memory. Add a comment indicating that, should CAM start supporting cdbs larger than 16bytes, the driver could store the CDB in the status buffer. aic79xx_pci.c: Add a table entry for the 39320A. Added a missing comma to an error string table. Fix spelling errors.
2003-02-27 23:23:16 +00:00
#define AHD_OFFSET_UNKNOWN 0xFF
#define AHD_PPR_OPTS_UNKNOWN 0xFF
/*
* Transfer Negotiation Information.
*/
struct ahd_transinfo {
uint8_t protocol_version; /* SCSI Revision level */
uint8_t transport_version; /* SPI Revision level */
uint8_t width; /* Bus width */
uint8_t period; /* Sync rate factor */
uint8_t offset; /* Sync offset */
uint8_t ppr_options; /* Parallel Protocol Request options */
};
/*
* Per-initiator current, goal and user transfer negotiation information. */
struct ahd_initiator_tinfo {
struct ahd_transinfo curr;
struct ahd_transinfo goal;
struct ahd_transinfo user;
};
/*
* Per enabled target ID state.
* Pointers to lun target state as well as sync/wide negotiation information
* for each initiator<->target mapping. For the initiator role we pretend
* that we are the target and the targets are the initiators since the
* negotiation is the same regardless of role.
*/
struct ahd_tmode_tstate {
struct ahd_tmode_lstate* enabled_luns[AHD_NUM_LUNS];
struct ahd_initiator_tinfo transinfo[AHD_NUM_TARGETS];
/*
* Per initiator state bitmasks.
*/
uint16_t auto_negotiate;/* Auto Negotiation Required */
uint16_t discenable; /* Disconnection allowed */
uint16_t tagenable; /* Tagged Queuing allowed */
};
/*
* Points of interest along the negotiated transfer scale.
*/
#define AHD_SYNCRATE_160 0x8
#define AHD_SYNCRATE_PACED 0x8
#define AHD_SYNCRATE_DT 0x9
#define AHD_SYNCRATE_ULTRA2 0xa
#define AHD_SYNCRATE_ULTRA 0xc
#define AHD_SYNCRATE_FAST 0x19
#define AHD_SYNCRATE_MIN_DT AHD_SYNCRATE_FAST
#define AHD_SYNCRATE_SYNC 0x32
#define AHD_SYNCRATE_MIN 0x60
#define AHD_SYNCRATE_ASYNC 0xFF
aic79xx.c: Implement the SCB_SILENT flag. This is useful for hushing up the driver during DV or other operations that we expect to cause transmission errors. The messages will still print if the SHOW_MASKED_ERRORS debug option is enabled. Save and restore the NEGOADDR address when setting new transfer settings. The sequencer performs lookups in the negotiation table too and it expects NEGOADDR to remain consistent across pause/unpause sessions. Consistently use "offset" instead of "period" to determine if we are running sync or not. Add a SHOW_MESSAGES diagnostic for when we assert ATN during message processing. Print out IU, QAS, and RTI features when showing transfer options. Limit the syncrate after all option conformance changes have taken place in ahd_devlimited_syncrate. Changes in options may change the final syncrate we accept. Keep a copy of the hs_mailbox in our softc so that we can perform read/modify/write operations on the hs_mailbox without having to pause the sequencer to read the last written value. Use the ENINT_COALESS flag in the hs_mailbox to toggle interrupt coalessing. Add entrypoints for enabling interrupt coalessing and setting both a timeout (how long to wait for commands to be coalessed) and a maximum commands to coaless value. Add a statistics timer that decides when to enable or disable interrupt coalessing based on load. Add a routine, ahd_reset_cmds_pending() which is used to update the CMDS_PENDING sequencer variable whenever error recovery compeltes SCBs without notifying the sequencer. Since ahd_reset_cmds_pending is called during ahd_unpause() only if we've aborted SCBs, its call to ahd_flush_qoutfifo should not cause recursion through ahd_run_qoutfifo(). A panic has been added to ensure that this recursion does not occur. In ahd_search_qinfifo, update the CMDS_PENDING sequencer variable directly. ahd_search_qinififo can be called in situations where using ahd_reset_cmds_pending() might cause recursion. Since we can safely determine the exact number to reduce CMDS_PENDING by in this scenario without running the qoutfifo, the manual update is sufficient. Clean up diagnostics. Add ahd_flush_qoutfifo() which will run the qoutfifo as well as complete any commands sitting on the sequencer's COMPLETE_SCB lists or the good status FIFO. Use this routine in several places that did similar things in an add-hoc, but incomplete, fashion. A call to this routine was also added to ahd_abort_scbs() to close a race. In ahd_pause_and_flushwork() only return once selections are safely disabled. Flush all completed commands via ahd_flush_qoutfifo(). Remove "Now packetized" diagnostic now that this information is incorperated into the actual negotiation messages that are displayed. When forcing renegotiation, don't clober the current ppr_options. Much of the driver uses this information to determine if we are currently packetized or not. Remove some stray spaces at column 1 in ahd_set_tags. When complaining about getting a host message loop request with no pending messages, print out the SCB_CONTROL register down on the card. Modify the ahd_sent_msg() routine to handle a search for an outgoing identify message. Use this to detect a msg reject on an identify message which typically indicates that the target thought we were packetized. Force a renegotiation in this case. In ahd_search_qinfifo(), wait more effectively for SCB DMA activities to cease. We also disable SCB fetch operations since we are about to change the qinfifo and any fetch in progress will likely be invalidated. In ahd_qinfifo_count(), fix the qinfifo empty case. In ahd_dump_card_state(), print out CCSCBCTL in the correct mode. If we are a narrow controller, don't set the current width to unknown when forcing a future negotiation. This just confuses the code into attempting a wide negotiation on a narrow bus. Add support for task management function completions. Modify ahd_handle_devreset so that it can handle lun resets in addition to target resets. Use ahd_handle_devreset for lun and target reset task management functions. Handle the abort task TMF race case better. We now wait until any current selections are over and then set the TMF back to zero. This should cause the sequencer to ignore the abort TMF completion should it occur. Correct a bug in the illegal phase handler that caused us to drop down to narrow when handling the unexpected command phase case after 3rd party reset of a packetized device. Indicate the features, bugs, and flags set in the softc that are used to control firmware patch download when booting verbose. aic79xx.h: Add coalessing and HS_MAILBOX fields. Add per-softc variables for the stats "daemon". Add a debug option for interrupt coalessing activities. Add two new softc flags: o AHD_UPDATE_PEND_CMDS Run ahd_reset_cmds_pending() on the next unpause. o AHD_RUNNING_QOUTFIFO Used to catch recursion through ahd_run_qoutfifo(). aic79xx.reg: Correct register addresses related to the software timer and the DFDBCTL register. Add constants paramaterizing the software timer. Add scratch ram locations for storing interrupt coalessing tunables. Break INTMASK in SEQITNCTL out into INTMASK1 and INTMASK2. In at least the REV A, these are writable bits. We make use of that for a swtimer workaround in the sequencer. Since HS_MAILBOX autoclears, provide a sequencer variable to store its contents. Add SEQINT codes for handling task management completions. aic79xx.seq: Correct ignore wide residue processing check for a wide negotiation being in effect. We must be in the SCSI register window in order to access the negotiation table. Use the software timer and a commands completed count to implement interrupt coalessing. The command complete is deferred until either the maximum command threshold or a the expiration of a command deferral timer. If we have more SCBs to complete to the host (sitting in COMPLETE_SCB lists), always try to coaless them up to our coalessing limit. If coalessing is enabled, but we have fewer commands oustanting than the host's min coalessing limit, complete the command immediately. Add code to track the number of commands outstanding. Commands are outstanding from the time they are placed into the execution queue until the DMA to post completion is setup. Add a workaround for intvec_2 interrupts on the H2A4. In H2A4, the mode pointer is not saved for intvec2, but is restored on iret. This can lead to the restoration of a bogus mode ptr. Manually clear the intmask bits and do a normal return to compensate. We use intvec_2 to track interrupt coalessing timeouts. Since we cannot disable the swtimer's countdown, simply mask its interrupt once we no longer care about it firing. In idle_loop_cchan, update LOCAL_HS_MAILBOX everytime we are notified of an HS_MAILBOX update via the HS_MAILBOX_ACT bit in QOFF_CTLSTA. We have to use a local copy of persistant portions of the HS_MAILBOX as the mailbox auto-clears on any read. Move the test for the cfg4istat interrupt up an instruction to hopefully close a race between the next outgoing selection and our disabling of selections. Add a missing ret to the last instruction in load_overrun_buf. Add notifications to the host of task management completions as well as the completions for commands that completed successfully before their corresponding TMF could be sent. Hold a critical section during select-out processing until we have a fully identified connection. This removes a race condition with the legacy abort handler. Correct a few spelling errors in some comments. aic79xx_inline.h: Call ahd_reset_cmds_pending() in ahd_unpause if required. Update cmdcmplt interrupt statistics in our interrupt handler. Allow callers to ahd_send_scb() to set the task management function. aic79xx_pci.c: Disable SERR and pause the controller prior to performing our mmapped I/O test. The U320 controllers do not support "auto-access-pause". aic79xx_osm.c: Set the task management function now that ahd_send_scb() doesn't do it for us. We also perform a lun reset in response to BDR requests to packetized devices.
2003-01-20 20:17:35 +00:00
#define AHD_SYNCRATE_MAX AHD_SYNCRATE_160
/* Safe and valid period for async negotiations. */
#define AHD_ASYNC_XFER_PERIOD 0x44
/*
* In RevA, the synctable uses a 120MHz rate for the period
* factor 8 and 160MHz for the period factor 7. The 120MHz
* rate never made it into the official SCSI spec, so we must
* compensate when setting the negotiation table for Rev A
* parts.
*/
#define AHD_SYNCRATE_REVA_120 0x8
#define AHD_SYNCRATE_REVA_160 0x7
/***************************** Lookup Tables **********************************/
/*
* Phase -> name and message out response
* to parity errors in each phase table.
*/
struct ahd_phase_table_entry {
uint8_t phase;
uint8_t mesg_out; /* Message response to parity errors */
char *phasemsg;
};
/************************** Serial EEPROM Format ******************************/
struct seeprom_config {
/*
* Per SCSI ID Configuration Flags
*/
uint16_t device_flags[16]; /* words 0-15 */
#define CFXFER 0x003F /* synchronous transfer rate */
#define CFXFER_ASYNC 0x3F
#define CFQAS 0x0040 /* Negotiate QAS */
#define CFPACKETIZED 0x0080 /* Negotiate Packetized Transfers */
#define CFSTART 0x0100 /* send start unit SCSI command */
#define CFINCBIOS 0x0200 /* include in BIOS scan */
#define CFDISC 0x0400 /* enable disconnection */
#define CFMULTILUNDEV 0x0800 /* Probe multiple luns in BIOS scan */
#define CFWIDEB 0x1000 /* wide bus device */
#define CFHOSTMANAGED 0x8000 /* Managed by a RAID controller */
/*
* BIOS Control Bits
*/
uint16_t bios_control; /* word 16 */
#define CFSUPREM 0x0001 /* support all removeable drives */
#define CFSUPREMB 0x0002 /* support removeable boot drives */
#define CFBIOSSTATE 0x000C /* BIOS Action State */
#define CFBS_DISABLED 0x00
#define CFBS_ENABLED 0x04
#define CFBS_DISABLED_SCAN 0x08
#define CFENABLEDV 0x0010 /* Perform Domain Validation */
#define CFCTRL_A 0x0020 /* BIOS displays Ctrl-A message */
#define CFSPARITY 0x0040 /* SCSI parity */
#define CFEXTEND 0x0080 /* extended translation enabled */
#define CFBOOTCD 0x0100 /* Support Bootable CD-ROM */
#define CFMSG_LEVEL 0x0600 /* BIOS Message Level */
#define CFMSG_VERBOSE 0x0000
#define CFMSG_SILENT 0x0200
#define CFMSG_DIAG 0x0400
#define CFRESETB 0x0800 /* reset SCSI bus at boot */
/* UNUSED 0xf000 */
/*
* Host Adapter Control Bits
*/
uint16_t adapter_control; /* word 17 */
#define CFAUTOTERM 0x0001 /* Perform Auto termination */
#define CFSTERM 0x0002 /* SCSI low byte termination */
#define CFWSTERM 0x0004 /* SCSI high byte termination */
#define CFSEAUTOTERM 0x0008 /* Ultra2 Perform secondary Auto Term*/
#define CFSELOWTERM 0x0010 /* Ultra2 secondary low term */
#define CFSEHIGHTERM 0x0020 /* Ultra2 secondary high term */
#define CFSTPWLEVEL 0x0040 /* Termination level control */
#define CFBIOSAUTOTERM 0x0080 /* Perform Auto termination */
#define CFTERM_MENU 0x0100 /* BIOS displays termination menu */
#define CFCLUSTERENB 0x8000 /* Cluster Enable */
/*
* Bus Release Time, Host Adapter ID
*/
uint16_t brtime_id; /* word 18 */
#define CFSCSIID 0x000f /* host adapter SCSI ID */
/* UNUSED 0x00f0 */
#define CFBRTIME 0xff00 /* bus release time/PCI Latency Time */
/*
* Maximum targets
*/
uint16_t max_targets; /* word 19 */
#define CFMAXTARG 0x00ff /* maximum targets */
#define CFBOOTLUN 0x0f00 /* Lun to boot from */
#define CFBOOTID 0xf000 /* Target to boot from */
uint16_t res_1[10]; /* words 20-29 */
uint16_t signature; /* BIOS Signature */
#define CFSIGNATURE 0x400
uint16_t checksum; /* word 31 */
};
Correct spelling errors. Switch to handling bad SCSI status as a sequencer interrupt instead of having the kernel proccess these failures via the completion queue. This is done because: o The old scheme required us to pause the sequencer and clear critical sections for each SCB. It seems that these pause actions, if coincident with a sequencer FIFO interrupt, would result in a FIFO interrupt getting lost or directing to the wrong FIFO. This caused hangs when the driver was stressed under high "queue full" loads. o The completion code assumed that it was always called with the sequencer running. This may not be the case in timeout processing where completions occur manually via ahd_pause_and_flushwork(). o With this scheme, the extra expense of clearing critical sections is avoided since the sequencer will only self pause once all pending selections have cleared and it is not in a critical section. aic79xx.c Add code to handle the new BAD_SCB_STATUS sequencer interrupt code. This just redirects the SCB through the already existing ahd_complete_scb() code path. Remove code in ahd_handle_scsi_status() that paused the sequencer, made sure that no selections where pending, and cleared critical sections. Bad status SCBs are now only processed when all of these conditions are true. aic79xx.reg: Add the BAD_SCB_STATUS sequencer interrupt code. aic79xx.seq: When completing an SCB upload to the host, if we are doing this because the SCB contains non-zero SCSI status, defer completing the SCB until there are no pending selection events. When completing these SCBs, use the new BAD_SCB_STATUS sequencer interrupt. For all other uploaded SCBs (currently only for underruns), the SCB is completed via the normal done queue. Additionally, keep the SCB that is currently being uploaded on the COMPLETE_DMA_SCB list until the dma is completed, not just until the DMA is started. This ensures that the DMA is restarted properly should the host disable the DMA transfer for some reason. In our RevA workaround for Maxtor drives, guard against the host pausing us while trying to pause I/O until the first data-valid REQ by clearing the current snapshot so that we can tell if the transfer has completed prior to us noticing the REQINIT status. In cfg4data_intr, shave off an instruction before getting the data path running by adding an entrypoint to the overrun handler to also increment the FIFO use count. In the overrun handler, be sure to clear our LONGJMP address in both exit paths. Perform a few sequencer optimizations. aic79xx.c: Print the full path from the SCB when a packetized status overrun occurs. Remove references to LONGJMP_SCB which is being removed from firmware usage. Print the new SCB_FIFO_USE_COUNT field in the per-SCB section of ahd_dump_card_state(). The SCB_TAG field is now re-used by the sequencer, so it no longer makes sense to reference this field in the kernel driver. aic79xx.h: Re-arrange fields in the hardware SCB from largest size type to smallest. This makes it easier to move fields without changing field alignment. The hardware scb tag field is now down near the "spare" portion of the SCB to facilitate reuse by the sequencer. aic79xx.reg: Remove LONGJMP_ADDR. Rearrange SCB fields to match aic79xx.h. Add SCB_FIFO_USE_COUNT as the first byte of the SCB_TAG field. aic79xx.seq: Add a per-SCB "Fifos in use count" field and use it to determine when it is safe (all data posted) to deliver status back to the host. The old method involved polling one or both FIFOs to verify that the current task did not have pending data. This makes running down the GSFIFO very cheap, so we will empty the GSFIFO in one idle loop pass in all cases. Use this simplification of the completion process to prune down the data FIFO teardown sequencer for packetized transfers. Much more code is now shared between the data residual and transfer complete cases. Correct some issues in the packetized status handler. It used to be possible to CLRCHN our FIFO before status had fully transferred to the host. We also failed to handle NONPACKREQ phases that could occur should a CRC error occur during transmission of the status data packet. Correct a few big endian issues: aic79xx.c: aic79xx_inline.h: aic79xx_pci.c: aic79xx_osm.c: o Always get the SCB's tag via the SCB_GET_TAG acccessor o Add missing use of byte swapping macros when touching hscb fields. o Don't double swap SEEPROM data when it is printed. Correct a big-endian bug. We cannot assign a o When assigning a 32bit LE variable to a 64bit LE variable, we must be explict about how the words of the 64bit LE variable are initialized. Cast to (uint32_t*) to do this. aic79xx.c: In ahd_clear_critical_section(), hit CRLSCSIINT after restoring the interrupt masks to avoid what appears to be a glitch on SCSIINT. Any real SCSIINT status will be persistent and will immidiately reset SCSIINT. This clear should only get rid of spurious SCSIINTs. This glitch was the cause of the "Unexpected PKT busfree" status that occurred under high queue full loads Call ahd_fini_scbdata() after shutdown so that any ahd_chip_init() routine that might access SCB data will not access free'd memory. Reset the bus on an IOERR since the chip doesn't seem to reset to the new voltage level without this. Change offset calculation for scatter gather maps so that the calculation is correct if an integral multiple of sg lists does not fit in the allocation size. Adjust bus dma tag for data buffers based on 39BIT addressing flag in our softc. Use the QFREEZE count to simplify ahd_pause_and_flushworkd(). We can thus rely on the sequencer eventually clearing ENSELO. In ahd_abort_scbs(), fix a bug that could potentially corrupt sequencer state. The saved SCB was being restored in the SCSI mode instead of the saved mode. It turns out that the SCB did not need to be saved at all as the scbptr is already restored by all subroutines called during this function that modify that register. aic79xx.c: aic79xx.h: aic79xx_pci.c: Add support for parsing the seeprom vital product data. The VPD data are currently unused. aic79xx.h: aic79xx.seq: aic79xx_pci.c: Add a firmware workaround to make the LED blink brighter during packetized operations on the H2A. aic79xx_inline.h: The host does not use timer interrupts, so don't gate our decision on whether or not to unpause the sequencer on whether or not a timer interrupt is pending.
2003-05-04 00:20:07 +00:00
/*
* Vital Product Data used during POST and by the BIOS.
*/
struct vpd_config {
uint8_t bios_flags;
#define VPDMASTERBIOS 0x0001
#define VPDBOOTHOST 0x0002
uint8_t reserved_1[21];
uint8_t resource_type;
uint8_t resource_len[2];
uint8_t resource_data[8];
uint8_t vpd_tag;
uint16_t vpd_len;
uint8_t vpd_keyword[2];
uint8_t length;
uint8_t revision;
uint8_t device_flags;
uint8_t termnation_menus[2];
uint8_t fifo_threshold;
uint8_t end_tag;
uint8_t vpd_checksum;
uint16_t default_target_flags;
uint16_t default_bios_flags;
uint16_t default_ctrl_flags;
uint8_t default_irq;
uint8_t pci_lattime;
uint8_t max_target;
uint8_t boot_lun;
uint16_t signature;
uint8_t reserved_2;
uint8_t checksum;
uint8_t reserved_3[4];
};
/****************************** Flexport Logic ********************************/
#define FLXADDR_TERMCTL 0x0
#define FLX_TERMCTL_ENSECHIGH 0x8
#define FLX_TERMCTL_ENSECLOW 0x4
#define FLX_TERMCTL_ENPRIHIGH 0x2
#define FLX_TERMCTL_ENPRILOW 0x1
#define FLXADDR_ROMSTAT_CURSENSECTL 0x1
#define FLX_ROMSTAT_SEECFG 0xF0
#define FLX_ROMSTAT_EECFG 0x0F
#define FLX_ROMSTAT_SEE_93C66 0x00
#define FLX_ROMSTAT_SEE_NONE 0xF0
#define FLX_ROMSTAT_EE_512x8 0x0
#define FLX_ROMSTAT_EE_1MBx8 0x1
#define FLX_ROMSTAT_EE_2MBx8 0x2
#define FLX_ROMSTAT_EE_4MBx8 0x3
#define FLX_ROMSTAT_EE_16MBx8 0x4
#define CURSENSE_ENB 0x1
#define FLXADDR_FLEXSTAT 0x2
#define FLX_FSTAT_BUSY 0x1
#define FLXADDR_CURRENT_STAT 0x4
#define FLX_CSTAT_SEC_HIGH 0xC0
#define FLX_CSTAT_SEC_LOW 0x30
#define FLX_CSTAT_PRI_HIGH 0x0C
#define FLX_CSTAT_PRI_LOW 0x03
#define FLX_CSTAT_MASK 0x03
#define FLX_CSTAT_SHIFT 2
#define FLX_CSTAT_OKAY 0x0
#define FLX_CSTAT_OVER 0x1
#define FLX_CSTAT_UNDER 0x2
#define FLX_CSTAT_INVALID 0x3
int ahd_read_seeprom(struct ahd_softc *ahd, uint16_t *buf,
Correct spelling errors. Switch to handling bad SCSI status as a sequencer interrupt instead of having the kernel proccess these failures via the completion queue. This is done because: o The old scheme required us to pause the sequencer and clear critical sections for each SCB. It seems that these pause actions, if coincident with a sequencer FIFO interrupt, would result in a FIFO interrupt getting lost or directing to the wrong FIFO. This caused hangs when the driver was stressed under high "queue full" loads. o The completion code assumed that it was always called with the sequencer running. This may not be the case in timeout processing where completions occur manually via ahd_pause_and_flushwork(). o With this scheme, the extra expense of clearing critical sections is avoided since the sequencer will only self pause once all pending selections have cleared and it is not in a critical section. aic79xx.c Add code to handle the new BAD_SCB_STATUS sequencer interrupt code. This just redirects the SCB through the already existing ahd_complete_scb() code path. Remove code in ahd_handle_scsi_status() that paused the sequencer, made sure that no selections where pending, and cleared critical sections. Bad status SCBs are now only processed when all of these conditions are true. aic79xx.reg: Add the BAD_SCB_STATUS sequencer interrupt code. aic79xx.seq: When completing an SCB upload to the host, if we are doing this because the SCB contains non-zero SCSI status, defer completing the SCB until there are no pending selection events. When completing these SCBs, use the new BAD_SCB_STATUS sequencer interrupt. For all other uploaded SCBs (currently only for underruns), the SCB is completed via the normal done queue. Additionally, keep the SCB that is currently being uploaded on the COMPLETE_DMA_SCB list until the dma is completed, not just until the DMA is started. This ensures that the DMA is restarted properly should the host disable the DMA transfer for some reason. In our RevA workaround for Maxtor drives, guard against the host pausing us while trying to pause I/O until the first data-valid REQ by clearing the current snapshot so that we can tell if the transfer has completed prior to us noticing the REQINIT status. In cfg4data_intr, shave off an instruction before getting the data path running by adding an entrypoint to the overrun handler to also increment the FIFO use count. In the overrun handler, be sure to clear our LONGJMP address in both exit paths. Perform a few sequencer optimizations. aic79xx.c: Print the full path from the SCB when a packetized status overrun occurs. Remove references to LONGJMP_SCB which is being removed from firmware usage. Print the new SCB_FIFO_USE_COUNT field in the per-SCB section of ahd_dump_card_state(). The SCB_TAG field is now re-used by the sequencer, so it no longer makes sense to reference this field in the kernel driver. aic79xx.h: Re-arrange fields in the hardware SCB from largest size type to smallest. This makes it easier to move fields without changing field alignment. The hardware scb tag field is now down near the "spare" portion of the SCB to facilitate reuse by the sequencer. aic79xx.reg: Remove LONGJMP_ADDR. Rearrange SCB fields to match aic79xx.h. Add SCB_FIFO_USE_COUNT as the first byte of the SCB_TAG field. aic79xx.seq: Add a per-SCB "Fifos in use count" field and use it to determine when it is safe (all data posted) to deliver status back to the host. The old method involved polling one or both FIFOs to verify that the current task did not have pending data. This makes running down the GSFIFO very cheap, so we will empty the GSFIFO in one idle loop pass in all cases. Use this simplification of the completion process to prune down the data FIFO teardown sequencer for packetized transfers. Much more code is now shared between the data residual and transfer complete cases. Correct some issues in the packetized status handler. It used to be possible to CLRCHN our FIFO before status had fully transferred to the host. We also failed to handle NONPACKREQ phases that could occur should a CRC error occur during transmission of the status data packet. Correct a few big endian issues: aic79xx.c: aic79xx_inline.h: aic79xx_pci.c: aic79xx_osm.c: o Always get the SCB's tag via the SCB_GET_TAG acccessor o Add missing use of byte swapping macros when touching hscb fields. o Don't double swap SEEPROM data when it is printed. Correct a big-endian bug. We cannot assign a o When assigning a 32bit LE variable to a 64bit LE variable, we must be explict about how the words of the 64bit LE variable are initialized. Cast to (uint32_t*) to do this. aic79xx.c: In ahd_clear_critical_section(), hit CRLSCSIINT after restoring the interrupt masks to avoid what appears to be a glitch on SCSIINT. Any real SCSIINT status will be persistent and will immidiately reset SCSIINT. This clear should only get rid of spurious SCSIINTs. This glitch was the cause of the "Unexpected PKT busfree" status that occurred under high queue full loads Call ahd_fini_scbdata() after shutdown so that any ahd_chip_init() routine that might access SCB data will not access free'd memory. Reset the bus on an IOERR since the chip doesn't seem to reset to the new voltage level without this. Change offset calculation for scatter gather maps so that the calculation is correct if an integral multiple of sg lists does not fit in the allocation size. Adjust bus dma tag for data buffers based on 39BIT addressing flag in our softc. Use the QFREEZE count to simplify ahd_pause_and_flushworkd(). We can thus rely on the sequencer eventually clearing ENSELO. In ahd_abort_scbs(), fix a bug that could potentially corrupt sequencer state. The saved SCB was being restored in the SCSI mode instead of the saved mode. It turns out that the SCB did not need to be saved at all as the scbptr is already restored by all subroutines called during this function that modify that register. aic79xx.c: aic79xx.h: aic79xx_pci.c: Add support for parsing the seeprom vital product data. The VPD data are currently unused. aic79xx.h: aic79xx.seq: aic79xx_pci.c: Add a firmware workaround to make the LED blink brighter during packetized operations on the H2A. aic79xx_inline.h: The host does not use timer interrupts, so don't gate our decision on whether or not to unpause the sequencer on whether or not a timer interrupt is pending.
2003-05-04 00:20:07 +00:00
u_int start_addr, u_int count, int bstream);
int ahd_write_seeprom(struct ahd_softc *ahd, uint16_t *buf,
u_int start_addr, u_int count);
int ahd_wait_seeprom(struct ahd_softc *ahd);
Correct spelling errors. Switch to handling bad SCSI status as a sequencer interrupt instead of having the kernel proccess these failures via the completion queue. This is done because: o The old scheme required us to pause the sequencer and clear critical sections for each SCB. It seems that these pause actions, if coincident with a sequencer FIFO interrupt, would result in a FIFO interrupt getting lost or directing to the wrong FIFO. This caused hangs when the driver was stressed under high "queue full" loads. o The completion code assumed that it was always called with the sequencer running. This may not be the case in timeout processing where completions occur manually via ahd_pause_and_flushwork(). o With this scheme, the extra expense of clearing critical sections is avoided since the sequencer will only self pause once all pending selections have cleared and it is not in a critical section. aic79xx.c Add code to handle the new BAD_SCB_STATUS sequencer interrupt code. This just redirects the SCB through the already existing ahd_complete_scb() code path. Remove code in ahd_handle_scsi_status() that paused the sequencer, made sure that no selections where pending, and cleared critical sections. Bad status SCBs are now only processed when all of these conditions are true. aic79xx.reg: Add the BAD_SCB_STATUS sequencer interrupt code. aic79xx.seq: When completing an SCB upload to the host, if we are doing this because the SCB contains non-zero SCSI status, defer completing the SCB until there are no pending selection events. When completing these SCBs, use the new BAD_SCB_STATUS sequencer interrupt. For all other uploaded SCBs (currently only for underruns), the SCB is completed via the normal done queue. Additionally, keep the SCB that is currently being uploaded on the COMPLETE_DMA_SCB list until the dma is completed, not just until the DMA is started. This ensures that the DMA is restarted properly should the host disable the DMA transfer for some reason. In our RevA workaround for Maxtor drives, guard against the host pausing us while trying to pause I/O until the first data-valid REQ by clearing the current snapshot so that we can tell if the transfer has completed prior to us noticing the REQINIT status. In cfg4data_intr, shave off an instruction before getting the data path running by adding an entrypoint to the overrun handler to also increment the FIFO use count. In the overrun handler, be sure to clear our LONGJMP address in both exit paths. Perform a few sequencer optimizations. aic79xx.c: Print the full path from the SCB when a packetized status overrun occurs. Remove references to LONGJMP_SCB which is being removed from firmware usage. Print the new SCB_FIFO_USE_COUNT field in the per-SCB section of ahd_dump_card_state(). The SCB_TAG field is now re-used by the sequencer, so it no longer makes sense to reference this field in the kernel driver. aic79xx.h: Re-arrange fields in the hardware SCB from largest size type to smallest. This makes it easier to move fields without changing field alignment. The hardware scb tag field is now down near the "spare" portion of the SCB to facilitate reuse by the sequencer. aic79xx.reg: Remove LONGJMP_ADDR. Rearrange SCB fields to match aic79xx.h. Add SCB_FIFO_USE_COUNT as the first byte of the SCB_TAG field. aic79xx.seq: Add a per-SCB "Fifos in use count" field and use it to determine when it is safe (all data posted) to deliver status back to the host. The old method involved polling one or both FIFOs to verify that the current task did not have pending data. This makes running down the GSFIFO very cheap, so we will empty the GSFIFO in one idle loop pass in all cases. Use this simplification of the completion process to prune down the data FIFO teardown sequencer for packetized transfers. Much more code is now shared between the data residual and transfer complete cases. Correct some issues in the packetized status handler. It used to be possible to CLRCHN our FIFO before status had fully transferred to the host. We also failed to handle NONPACKREQ phases that could occur should a CRC error occur during transmission of the status data packet. Correct a few big endian issues: aic79xx.c: aic79xx_inline.h: aic79xx_pci.c: aic79xx_osm.c: o Always get the SCB's tag via the SCB_GET_TAG acccessor o Add missing use of byte swapping macros when touching hscb fields. o Don't double swap SEEPROM data when it is printed. Correct a big-endian bug. We cannot assign a o When assigning a 32bit LE variable to a 64bit LE variable, we must be explict about how the words of the 64bit LE variable are initialized. Cast to (uint32_t*) to do this. aic79xx.c: In ahd_clear_critical_section(), hit CRLSCSIINT after restoring the interrupt masks to avoid what appears to be a glitch on SCSIINT. Any real SCSIINT status will be persistent and will immidiately reset SCSIINT. This clear should only get rid of spurious SCSIINTs. This glitch was the cause of the "Unexpected PKT busfree" status that occurred under high queue full loads Call ahd_fini_scbdata() after shutdown so that any ahd_chip_init() routine that might access SCB data will not access free'd memory. Reset the bus on an IOERR since the chip doesn't seem to reset to the new voltage level without this. Change offset calculation for scatter gather maps so that the calculation is correct if an integral multiple of sg lists does not fit in the allocation size. Adjust bus dma tag for data buffers based on 39BIT addressing flag in our softc. Use the QFREEZE count to simplify ahd_pause_and_flushworkd(). We can thus rely on the sequencer eventually clearing ENSELO. In ahd_abort_scbs(), fix a bug that could potentially corrupt sequencer state. The saved SCB was being restored in the SCSI mode instead of the saved mode. It turns out that the SCB did not need to be saved at all as the scbptr is already restored by all subroutines called during this function that modify that register. aic79xx.c: aic79xx.h: aic79xx_pci.c: Add support for parsing the seeprom vital product data. The VPD data are currently unused. aic79xx.h: aic79xx.seq: aic79xx_pci.c: Add a firmware workaround to make the LED blink brighter during packetized operations on the H2A. aic79xx_inline.h: The host does not use timer interrupts, so don't gate our decision on whether or not to unpause the sequencer on whether or not a timer interrupt is pending.
2003-05-04 00:20:07 +00:00
int ahd_verify_vpd_cksum(struct vpd_config *vpd);
int ahd_verify_cksum(struct seeprom_config *sc);
int ahd_acquire_seeprom(struct ahd_softc *ahd);
void ahd_release_seeprom(struct ahd_softc *ahd);
/**************************** Message Buffer *********************************/
typedef enum {
MSG_FLAG_NONE = 0x00,
MSG_FLAG_EXPECT_PPR_BUSFREE = 0x01,
MSG_FLAG_IU_REQ_CHANGED = 0x02,
MSG_FLAG_EXPECT_IDE_BUSFREE = 0x04,
MSG_FLAG_EXPECT_QASREJ_BUSFREE = 0x08,
MSG_FLAG_PACKETIZED = 0x10
} ahd_msg_flags;
typedef enum {
MSG_TYPE_NONE = 0x00,
MSG_TYPE_INITIATOR_MSGOUT = 0x01,
MSG_TYPE_INITIATOR_MSGIN = 0x02,
MSG_TYPE_TARGET_MSGOUT = 0x03,
MSG_TYPE_TARGET_MSGIN = 0x04
} ahd_msg_type;
typedef enum {
MSGLOOP_IN_PROG,
MSGLOOP_MSGCOMPLETE,
MSGLOOP_TERMINATED
} msg_loop_stat;
/*********************** Software Configuration Structure *********************/
struct ahd_suspend_channel_state {
uint8_t scsiseq;
uint8_t sxfrctl0;
uint8_t sxfrctl1;
uint8_t simode0;
uint8_t simode1;
uint8_t seltimer;
uint8_t seqctl;
};
struct ahd_suspend_state {
struct ahd_suspend_channel_state channel[2];
uint8_t optionmode;
uint8_t dscommand0;
uint8_t dspcistatus;
/* hsmailbox */
uint8_t crccontrol1;
uint8_t scbbaddr;
/* Host and sequencer SCB counts */
uint8_t dff_thrsh;
uint8_t *scratch_ram;
uint8_t *btt;
};
typedef void (*ahd_bus_intr_t)(struct ahd_softc *);
typedef enum {
AHD_MODE_DFF0,
AHD_MODE_DFF1,
AHD_MODE_CCHAN,
AHD_MODE_SCSI,
AHD_MODE_CFG,
AHD_MODE_UNKNOWN
} ahd_mode;
#define AHD_MK_MSK(x) (0x01 << (x))
#define AHD_MODE_DFF0_MSK AHD_MK_MSK(AHD_MODE_DFF0)
#define AHD_MODE_DFF1_MSK AHD_MK_MSK(AHD_MODE_DFF1)
#define AHD_MODE_CCHAN_MSK AHD_MK_MSK(AHD_MODE_CCHAN)
#define AHD_MODE_SCSI_MSK AHD_MK_MSK(AHD_MODE_SCSI)
#define AHD_MODE_CFG_MSK AHD_MK_MSK(AHD_MODE_CFG)
#define AHD_MODE_UNKNOWN_MSK AHD_MK_MSK(AHD_MODE_UNKNOWN)
#define AHD_MODE_ANY_MSK (~0)
typedef uint8_t ahd_mode_state;
typedef void ahd_callback_t (void *);
struct ahd_softc {
bus_space_tag_t tags[2];
bus_space_handle_t bshs[2];
#ifndef __linux__
bus_dma_tag_t buffer_dmat; /* dmat for buffer I/O */
#endif
struct scb_data scb_data;
struct hardware_scb *next_queued_hscb;
/*
* SCBs that have been sent to the controller
*/
LIST_HEAD(, scb) pending_scbs;
/*
* Current register window mode information.
*/
ahd_mode dst_mode;
ahd_mode src_mode;
/*
* Saved register window mode information
* used for restore on next unpause.
*/
ahd_mode saved_dst_mode;
ahd_mode saved_src_mode;
/*
* Platform specific data.
*/
struct ahd_platform_data *platform_data;
/*
* Platform specific device information.
*/
ahd_dev_softc_t dev_softc;
/*
* Bus specific device information.
*/
ahd_bus_intr_t bus_intr;
/*
* Target mode related state kept on a per enabled lun basis.
* Targets that are not enabled will have null entries.
* As an initiator, we keep one target entry for our initiator
* ID to store our sync/wide transfer settings.
*/
struct ahd_tmode_tstate *enabled_targets[AHD_NUM_TARGETS];
/*
* The black hole device responsible for handling requests for
* disabled luns on enabled targets.
*/
struct ahd_tmode_lstate *black_hole;
/*
* Device instance currently on the bus awaiting a continue TIO
* for a command that was not given the disconnect priveledge.
*/
struct ahd_tmode_lstate *pending_device;
/*
* Timer handles for timer driven callbacks.
*/
ahd_timer_t reset_timer;
aic79xx.c: Implement the SCB_SILENT flag. This is useful for hushing up the driver during DV or other operations that we expect to cause transmission errors. The messages will still print if the SHOW_MASKED_ERRORS debug option is enabled. Save and restore the NEGOADDR address when setting new transfer settings. The sequencer performs lookups in the negotiation table too and it expects NEGOADDR to remain consistent across pause/unpause sessions. Consistently use "offset" instead of "period" to determine if we are running sync or not. Add a SHOW_MESSAGES diagnostic for when we assert ATN during message processing. Print out IU, QAS, and RTI features when showing transfer options. Limit the syncrate after all option conformance changes have taken place in ahd_devlimited_syncrate. Changes in options may change the final syncrate we accept. Keep a copy of the hs_mailbox in our softc so that we can perform read/modify/write operations on the hs_mailbox without having to pause the sequencer to read the last written value. Use the ENINT_COALESS flag in the hs_mailbox to toggle interrupt coalessing. Add entrypoints for enabling interrupt coalessing and setting both a timeout (how long to wait for commands to be coalessed) and a maximum commands to coaless value. Add a statistics timer that decides when to enable or disable interrupt coalessing based on load. Add a routine, ahd_reset_cmds_pending() which is used to update the CMDS_PENDING sequencer variable whenever error recovery compeltes SCBs without notifying the sequencer. Since ahd_reset_cmds_pending is called during ahd_unpause() only if we've aborted SCBs, its call to ahd_flush_qoutfifo should not cause recursion through ahd_run_qoutfifo(). A panic has been added to ensure that this recursion does not occur. In ahd_search_qinfifo, update the CMDS_PENDING sequencer variable directly. ahd_search_qinififo can be called in situations where using ahd_reset_cmds_pending() might cause recursion. Since we can safely determine the exact number to reduce CMDS_PENDING by in this scenario without running the qoutfifo, the manual update is sufficient. Clean up diagnostics. Add ahd_flush_qoutfifo() which will run the qoutfifo as well as complete any commands sitting on the sequencer's COMPLETE_SCB lists or the good status FIFO. Use this routine in several places that did similar things in an add-hoc, but incomplete, fashion. A call to this routine was also added to ahd_abort_scbs() to close a race. In ahd_pause_and_flushwork() only return once selections are safely disabled. Flush all completed commands via ahd_flush_qoutfifo(). Remove "Now packetized" diagnostic now that this information is incorperated into the actual negotiation messages that are displayed. When forcing renegotiation, don't clober the current ppr_options. Much of the driver uses this information to determine if we are currently packetized or not. Remove some stray spaces at column 1 in ahd_set_tags. When complaining about getting a host message loop request with no pending messages, print out the SCB_CONTROL register down on the card. Modify the ahd_sent_msg() routine to handle a search for an outgoing identify message. Use this to detect a msg reject on an identify message which typically indicates that the target thought we were packetized. Force a renegotiation in this case. In ahd_search_qinfifo(), wait more effectively for SCB DMA activities to cease. We also disable SCB fetch operations since we are about to change the qinfifo and any fetch in progress will likely be invalidated. In ahd_qinfifo_count(), fix the qinfifo empty case. In ahd_dump_card_state(), print out CCSCBCTL in the correct mode. If we are a narrow controller, don't set the current width to unknown when forcing a future negotiation. This just confuses the code into attempting a wide negotiation on a narrow bus. Add support for task management function completions. Modify ahd_handle_devreset so that it can handle lun resets in addition to target resets. Use ahd_handle_devreset for lun and target reset task management functions. Handle the abort task TMF race case better. We now wait until any current selections are over and then set the TMF back to zero. This should cause the sequencer to ignore the abort TMF completion should it occur. Correct a bug in the illegal phase handler that caused us to drop down to narrow when handling the unexpected command phase case after 3rd party reset of a packetized device. Indicate the features, bugs, and flags set in the softc that are used to control firmware patch download when booting verbose. aic79xx.h: Add coalessing and HS_MAILBOX fields. Add per-softc variables for the stats "daemon". Add a debug option for interrupt coalessing activities. Add two new softc flags: o AHD_UPDATE_PEND_CMDS Run ahd_reset_cmds_pending() on the next unpause. o AHD_RUNNING_QOUTFIFO Used to catch recursion through ahd_run_qoutfifo(). aic79xx.reg: Correct register addresses related to the software timer and the DFDBCTL register. Add constants paramaterizing the software timer. Add scratch ram locations for storing interrupt coalessing tunables. Break INTMASK in SEQITNCTL out into INTMASK1 and INTMASK2. In at least the REV A, these are writable bits. We make use of that for a swtimer workaround in the sequencer. Since HS_MAILBOX autoclears, provide a sequencer variable to store its contents. Add SEQINT codes for handling task management completions. aic79xx.seq: Correct ignore wide residue processing check for a wide negotiation being in effect. We must be in the SCSI register window in order to access the negotiation table. Use the software timer and a commands completed count to implement interrupt coalessing. The command complete is deferred until either the maximum command threshold or a the expiration of a command deferral timer. If we have more SCBs to complete to the host (sitting in COMPLETE_SCB lists), always try to coaless them up to our coalessing limit. If coalessing is enabled, but we have fewer commands oustanting than the host's min coalessing limit, complete the command immediately. Add code to track the number of commands outstanding. Commands are outstanding from the time they are placed into the execution queue until the DMA to post completion is setup. Add a workaround for intvec_2 interrupts on the H2A4. In H2A4, the mode pointer is not saved for intvec2, but is restored on iret. This can lead to the restoration of a bogus mode ptr. Manually clear the intmask bits and do a normal return to compensate. We use intvec_2 to track interrupt coalessing timeouts. Since we cannot disable the swtimer's countdown, simply mask its interrupt once we no longer care about it firing. In idle_loop_cchan, update LOCAL_HS_MAILBOX everytime we are notified of an HS_MAILBOX update via the HS_MAILBOX_ACT bit in QOFF_CTLSTA. We have to use a local copy of persistant portions of the HS_MAILBOX as the mailbox auto-clears on any read. Move the test for the cfg4istat interrupt up an instruction to hopefully close a race between the next outgoing selection and our disabling of selections. Add a missing ret to the last instruction in load_overrun_buf. Add notifications to the host of task management completions as well as the completions for commands that completed successfully before their corresponding TMF could be sent. Hold a critical section during select-out processing until we have a fully identified connection. This removes a race condition with the legacy abort handler. Correct a few spelling errors in some comments. aic79xx_inline.h: Call ahd_reset_cmds_pending() in ahd_unpause if required. Update cmdcmplt interrupt statistics in our interrupt handler. Allow callers to ahd_send_scb() to set the task management function. aic79xx_pci.c: Disable SERR and pause the controller prior to performing our mmapped I/O test. The U320 controllers do not support "auto-access-pause". aic79xx_osm.c: Set the task management function now that ahd_send_scb() doesn't do it for us. We also perform a lun reset in response to BDR requests to packetized devices.
2003-01-20 20:17:35 +00:00
ahd_timer_t stat_timer;
/*
* Statistics.
*/
#define AHD_STAT_UPDATE_US 250000 /* 250ms */
#define AHD_STAT_BUCKETS 4
u_int cmdcmplt_bucket;
uint32_t cmdcmplt_counts[AHD_STAT_BUCKETS];
uint32_t cmdcmplt_total;
/*
* Card characteristics
*/
ahd_chip chip;
ahd_feature features;
ahd_bug bugs;
ahd_flag flags;
struct seeprom_config *seep_config;
/* Values to store in the SEQCTL register for pause and unpause */
uint8_t unpause;
uint8_t pause;
/* Command Queues */
uint16_t qoutfifonext;
uint16_t qoutfifonext_valid_tag;
uint16_t qinfifonext;
uint16_t qinfifo[AHD_SCB_MAX];
uint16_t *qoutfifo;
/* Critical Section Data */
struct cs *critical_sections;
u_int num_critical_sections;
/* Buffer for handling packetized bitbucket. */
uint8_t *overrun_buf;
/* Links for chaining softcs */
TAILQ_ENTRY(ahd_softc) links;
/* Channel Names ('A', 'B', etc.) */
char channel;
/* Initiator Bus ID */
uint8_t our_id;
/*
* Target incoming command FIFO.
*/
struct target_cmd *targetcmds;
uint8_t tqinfifonext;
aic79xx.c: Implement the SCB_SILENT flag. This is useful for hushing up the driver during DV or other operations that we expect to cause transmission errors. The messages will still print if the SHOW_MASKED_ERRORS debug option is enabled. Save and restore the NEGOADDR address when setting new transfer settings. The sequencer performs lookups in the negotiation table too and it expects NEGOADDR to remain consistent across pause/unpause sessions. Consistently use "offset" instead of "period" to determine if we are running sync or not. Add a SHOW_MESSAGES diagnostic for when we assert ATN during message processing. Print out IU, QAS, and RTI features when showing transfer options. Limit the syncrate after all option conformance changes have taken place in ahd_devlimited_syncrate. Changes in options may change the final syncrate we accept. Keep a copy of the hs_mailbox in our softc so that we can perform read/modify/write operations on the hs_mailbox without having to pause the sequencer to read the last written value. Use the ENINT_COALESS flag in the hs_mailbox to toggle interrupt coalessing. Add entrypoints for enabling interrupt coalessing and setting both a timeout (how long to wait for commands to be coalessed) and a maximum commands to coaless value. Add a statistics timer that decides when to enable or disable interrupt coalessing based on load. Add a routine, ahd_reset_cmds_pending() which is used to update the CMDS_PENDING sequencer variable whenever error recovery compeltes SCBs without notifying the sequencer. Since ahd_reset_cmds_pending is called during ahd_unpause() only if we've aborted SCBs, its call to ahd_flush_qoutfifo should not cause recursion through ahd_run_qoutfifo(). A panic has been added to ensure that this recursion does not occur. In ahd_search_qinfifo, update the CMDS_PENDING sequencer variable directly. ahd_search_qinififo can be called in situations where using ahd_reset_cmds_pending() might cause recursion. Since we can safely determine the exact number to reduce CMDS_PENDING by in this scenario without running the qoutfifo, the manual update is sufficient. Clean up diagnostics. Add ahd_flush_qoutfifo() which will run the qoutfifo as well as complete any commands sitting on the sequencer's COMPLETE_SCB lists or the good status FIFO. Use this routine in several places that did similar things in an add-hoc, but incomplete, fashion. A call to this routine was also added to ahd_abort_scbs() to close a race. In ahd_pause_and_flushwork() only return once selections are safely disabled. Flush all completed commands via ahd_flush_qoutfifo(). Remove "Now packetized" diagnostic now that this information is incorperated into the actual negotiation messages that are displayed. When forcing renegotiation, don't clober the current ppr_options. Much of the driver uses this information to determine if we are currently packetized or not. Remove some stray spaces at column 1 in ahd_set_tags. When complaining about getting a host message loop request with no pending messages, print out the SCB_CONTROL register down on the card. Modify the ahd_sent_msg() routine to handle a search for an outgoing identify message. Use this to detect a msg reject on an identify message which typically indicates that the target thought we were packetized. Force a renegotiation in this case. In ahd_search_qinfifo(), wait more effectively for SCB DMA activities to cease. We also disable SCB fetch operations since we are about to change the qinfifo and any fetch in progress will likely be invalidated. In ahd_qinfifo_count(), fix the qinfifo empty case. In ahd_dump_card_state(), print out CCSCBCTL in the correct mode. If we are a narrow controller, don't set the current width to unknown when forcing a future negotiation. This just confuses the code into attempting a wide negotiation on a narrow bus. Add support for task management function completions. Modify ahd_handle_devreset so that it can handle lun resets in addition to target resets. Use ahd_handle_devreset for lun and target reset task management functions. Handle the abort task TMF race case better. We now wait until any current selections are over and then set the TMF back to zero. This should cause the sequencer to ignore the abort TMF completion should it occur. Correct a bug in the illegal phase handler that caused us to drop down to narrow when handling the unexpected command phase case after 3rd party reset of a packetized device. Indicate the features, bugs, and flags set in the softc that are used to control firmware patch download when booting verbose. aic79xx.h: Add coalessing and HS_MAILBOX fields. Add per-softc variables for the stats "daemon". Add a debug option for interrupt coalessing activities. Add two new softc flags: o AHD_UPDATE_PEND_CMDS Run ahd_reset_cmds_pending() on the next unpause. o AHD_RUNNING_QOUTFIFO Used to catch recursion through ahd_run_qoutfifo(). aic79xx.reg: Correct register addresses related to the software timer and the DFDBCTL register. Add constants paramaterizing the software timer. Add scratch ram locations for storing interrupt coalessing tunables. Break INTMASK in SEQITNCTL out into INTMASK1 and INTMASK2. In at least the REV A, these are writable bits. We make use of that for a swtimer workaround in the sequencer. Since HS_MAILBOX autoclears, provide a sequencer variable to store its contents. Add SEQINT codes for handling task management completions. aic79xx.seq: Correct ignore wide residue processing check for a wide negotiation being in effect. We must be in the SCSI register window in order to access the negotiation table. Use the software timer and a commands completed count to implement interrupt coalessing. The command complete is deferred until either the maximum command threshold or a the expiration of a command deferral timer. If we have more SCBs to complete to the host (sitting in COMPLETE_SCB lists), always try to coaless them up to our coalessing limit. If coalessing is enabled, but we have fewer commands oustanting than the host's min coalessing limit, complete the command immediately. Add code to track the number of commands outstanding. Commands are outstanding from the time they are placed into the execution queue until the DMA to post completion is setup. Add a workaround for intvec_2 interrupts on the H2A4. In H2A4, the mode pointer is not saved for intvec2, but is restored on iret. This can lead to the restoration of a bogus mode ptr. Manually clear the intmask bits and do a normal return to compensate. We use intvec_2 to track interrupt coalessing timeouts. Since we cannot disable the swtimer's countdown, simply mask its interrupt once we no longer care about it firing. In idle_loop_cchan, update LOCAL_HS_MAILBOX everytime we are notified of an HS_MAILBOX update via the HS_MAILBOX_ACT bit in QOFF_CTLSTA. We have to use a local copy of persistant portions of the HS_MAILBOX as the mailbox auto-clears on any read. Move the test for the cfg4istat interrupt up an instruction to hopefully close a race between the next outgoing selection and our disabling of selections. Add a missing ret to the last instruction in load_overrun_buf. Add notifications to the host of task management completions as well as the completions for commands that completed successfully before their corresponding TMF could be sent. Hold a critical section during select-out processing until we have a fully identified connection. This removes a race condition with the legacy abort handler. Correct a few spelling errors in some comments. aic79xx_inline.h: Call ahd_reset_cmds_pending() in ahd_unpause if required. Update cmdcmplt interrupt statistics in our interrupt handler. Allow callers to ahd_send_scb() to set the task management function. aic79xx_pci.c: Disable SERR and pause the controller prior to performing our mmapped I/O test. The U320 controllers do not support "auto-access-pause". aic79xx_osm.c: Set the task management function now that ahd_send_scb() doesn't do it for us. We also perform a lun reset in response to BDR requests to packetized devices.
2003-01-20 20:17:35 +00:00
/*
* Cached verson of the hs_mailbox so we can avoid
* pausing the sequencer during mailbox updates.
*/
uint8_t hs_mailbox;
/*
* Incoming and outgoing message handling.
*/
uint8_t send_msg_perror;
ahd_msg_flags msg_flags;
ahd_msg_type msg_type;
uint8_t msgout_buf[12];/* Message we are sending */
uint8_t msgin_buf[12];/* Message we are receiving */
u_int msgout_len; /* Length of message to send */
u_int msgout_index; /* Current index in msgout */
u_int msgin_index; /* Current index in msgin */
/*
* Mapping information for data structures shared
* between the sequencer and kernel.
*/
bus_dma_tag_t parent_dmat;
bus_dma_tag_t shared_data_dmat;
bus_dmamap_t shared_data_dmamap;
bus_addr_t shared_data_busaddr;
/* Information saved through suspend/resume cycles */
struct ahd_suspend_state suspend_state;
/* Number of enabled target mode device on this card */
u_int enabled_luns;
/* Initialization level of this data structure */
u_int init_level;
/* PCI cacheline size. */
u_int pci_cachesize;
/* IO Cell Parameters */
uint8_t iocell_opts[AHD_NUM_PER_DEV_ANNEXCOLS];
u_int stack_size;
uint16_t *saved_stack;
/* Per-Unit descriptive information */
const char *description;
const char *bus_description;
char *name;
int unit;
/* Selection Timer settings */
int seltime;
aic79xx.c: Implement the SCB_SILENT flag. This is useful for hushing up the driver during DV or other operations that we expect to cause transmission errors. The messages will still print if the SHOW_MASKED_ERRORS debug option is enabled. Save and restore the NEGOADDR address when setting new transfer settings. The sequencer performs lookups in the negotiation table too and it expects NEGOADDR to remain consistent across pause/unpause sessions. Consistently use "offset" instead of "period" to determine if we are running sync or not. Add a SHOW_MESSAGES diagnostic for when we assert ATN during message processing. Print out IU, QAS, and RTI features when showing transfer options. Limit the syncrate after all option conformance changes have taken place in ahd_devlimited_syncrate. Changes in options may change the final syncrate we accept. Keep a copy of the hs_mailbox in our softc so that we can perform read/modify/write operations on the hs_mailbox without having to pause the sequencer to read the last written value. Use the ENINT_COALESS flag in the hs_mailbox to toggle interrupt coalessing. Add entrypoints for enabling interrupt coalessing and setting both a timeout (how long to wait for commands to be coalessed) and a maximum commands to coaless value. Add a statistics timer that decides when to enable or disable interrupt coalessing based on load. Add a routine, ahd_reset_cmds_pending() which is used to update the CMDS_PENDING sequencer variable whenever error recovery compeltes SCBs without notifying the sequencer. Since ahd_reset_cmds_pending is called during ahd_unpause() only if we've aborted SCBs, its call to ahd_flush_qoutfifo should not cause recursion through ahd_run_qoutfifo(). A panic has been added to ensure that this recursion does not occur. In ahd_search_qinfifo, update the CMDS_PENDING sequencer variable directly. ahd_search_qinififo can be called in situations where using ahd_reset_cmds_pending() might cause recursion. Since we can safely determine the exact number to reduce CMDS_PENDING by in this scenario without running the qoutfifo, the manual update is sufficient. Clean up diagnostics. Add ahd_flush_qoutfifo() which will run the qoutfifo as well as complete any commands sitting on the sequencer's COMPLETE_SCB lists or the good status FIFO. Use this routine in several places that did similar things in an add-hoc, but incomplete, fashion. A call to this routine was also added to ahd_abort_scbs() to close a race. In ahd_pause_and_flushwork() only return once selections are safely disabled. Flush all completed commands via ahd_flush_qoutfifo(). Remove "Now packetized" diagnostic now that this information is incorperated into the actual negotiation messages that are displayed. When forcing renegotiation, don't clober the current ppr_options. Much of the driver uses this information to determine if we are currently packetized or not. Remove some stray spaces at column 1 in ahd_set_tags. When complaining about getting a host message loop request with no pending messages, print out the SCB_CONTROL register down on the card. Modify the ahd_sent_msg() routine to handle a search for an outgoing identify message. Use this to detect a msg reject on an identify message which typically indicates that the target thought we were packetized. Force a renegotiation in this case. In ahd_search_qinfifo(), wait more effectively for SCB DMA activities to cease. We also disable SCB fetch operations since we are about to change the qinfifo and any fetch in progress will likely be invalidated. In ahd_qinfifo_count(), fix the qinfifo empty case. In ahd_dump_card_state(), print out CCSCBCTL in the correct mode. If we are a narrow controller, don't set the current width to unknown when forcing a future negotiation. This just confuses the code into attempting a wide negotiation on a narrow bus. Add support for task management function completions. Modify ahd_handle_devreset so that it can handle lun resets in addition to target resets. Use ahd_handle_devreset for lun and target reset task management functions. Handle the abort task TMF race case better. We now wait until any current selections are over and then set the TMF back to zero. This should cause the sequencer to ignore the abort TMF completion should it occur. Correct a bug in the illegal phase handler that caused us to drop down to narrow when handling the unexpected command phase case after 3rd party reset of a packetized device. Indicate the features, bugs, and flags set in the softc that are used to control firmware patch download when booting verbose. aic79xx.h: Add coalessing and HS_MAILBOX fields. Add per-softc variables for the stats "daemon". Add a debug option for interrupt coalessing activities. Add two new softc flags: o AHD_UPDATE_PEND_CMDS Run ahd_reset_cmds_pending() on the next unpause. o AHD_RUNNING_QOUTFIFO Used to catch recursion through ahd_run_qoutfifo(). aic79xx.reg: Correct register addresses related to the software timer and the DFDBCTL register. Add constants paramaterizing the software timer. Add scratch ram locations for storing interrupt coalessing tunables. Break INTMASK in SEQITNCTL out into INTMASK1 and INTMASK2. In at least the REV A, these are writable bits. We make use of that for a swtimer workaround in the sequencer. Since HS_MAILBOX autoclears, provide a sequencer variable to store its contents. Add SEQINT codes for handling task management completions. aic79xx.seq: Correct ignore wide residue processing check for a wide negotiation being in effect. We must be in the SCSI register window in order to access the negotiation table. Use the software timer and a commands completed count to implement interrupt coalessing. The command complete is deferred until either the maximum command threshold or a the expiration of a command deferral timer. If we have more SCBs to complete to the host (sitting in COMPLETE_SCB lists), always try to coaless them up to our coalessing limit. If coalessing is enabled, but we have fewer commands oustanting than the host's min coalessing limit, complete the command immediately. Add code to track the number of commands outstanding. Commands are outstanding from the time they are placed into the execution queue until the DMA to post completion is setup. Add a workaround for intvec_2 interrupts on the H2A4. In H2A4, the mode pointer is not saved for intvec2, but is restored on iret. This can lead to the restoration of a bogus mode ptr. Manually clear the intmask bits and do a normal return to compensate. We use intvec_2 to track interrupt coalessing timeouts. Since we cannot disable the swtimer's countdown, simply mask its interrupt once we no longer care about it firing. In idle_loop_cchan, update LOCAL_HS_MAILBOX everytime we are notified of an HS_MAILBOX update via the HS_MAILBOX_ACT bit in QOFF_CTLSTA. We have to use a local copy of persistant portions of the HS_MAILBOX as the mailbox auto-clears on any read. Move the test for the cfg4istat interrupt up an instruction to hopefully close a race between the next outgoing selection and our disabling of selections. Add a missing ret to the last instruction in load_overrun_buf. Add notifications to the host of task management completions as well as the completions for commands that completed successfully before their corresponding TMF could be sent. Hold a critical section during select-out processing until we have a fully identified connection. This removes a race condition with the legacy abort handler. Correct a few spelling errors in some comments. aic79xx_inline.h: Call ahd_reset_cmds_pending() in ahd_unpause if required. Update cmdcmplt interrupt statistics in our interrupt handler. Allow callers to ahd_send_scb() to set the task management function. aic79xx_pci.c: Disable SERR and pause the controller prior to performing our mmapped I/O test. The U320 controllers do not support "auto-access-pause". aic79xx_osm.c: Set the task management function now that ahd_send_scb() doesn't do it for us. We also perform a lun reset in response to BDR requests to packetized devices.
2003-01-20 20:17:35 +00:00
/*
* Interrupt coalescing settings.
aic79xx.c: Implement the SCB_SILENT flag. This is useful for hushing up the driver during DV or other operations that we expect to cause transmission errors. The messages will still print if the SHOW_MASKED_ERRORS debug option is enabled. Save and restore the NEGOADDR address when setting new transfer settings. The sequencer performs lookups in the negotiation table too and it expects NEGOADDR to remain consistent across pause/unpause sessions. Consistently use "offset" instead of "period" to determine if we are running sync or not. Add a SHOW_MESSAGES diagnostic for when we assert ATN during message processing. Print out IU, QAS, and RTI features when showing transfer options. Limit the syncrate after all option conformance changes have taken place in ahd_devlimited_syncrate. Changes in options may change the final syncrate we accept. Keep a copy of the hs_mailbox in our softc so that we can perform read/modify/write operations on the hs_mailbox without having to pause the sequencer to read the last written value. Use the ENINT_COALESS flag in the hs_mailbox to toggle interrupt coalessing. Add entrypoints for enabling interrupt coalessing and setting both a timeout (how long to wait for commands to be coalessed) and a maximum commands to coaless value. Add a statistics timer that decides when to enable or disable interrupt coalessing based on load. Add a routine, ahd_reset_cmds_pending() which is used to update the CMDS_PENDING sequencer variable whenever error recovery compeltes SCBs without notifying the sequencer. Since ahd_reset_cmds_pending is called during ahd_unpause() only if we've aborted SCBs, its call to ahd_flush_qoutfifo should not cause recursion through ahd_run_qoutfifo(). A panic has been added to ensure that this recursion does not occur. In ahd_search_qinfifo, update the CMDS_PENDING sequencer variable directly. ahd_search_qinififo can be called in situations where using ahd_reset_cmds_pending() might cause recursion. Since we can safely determine the exact number to reduce CMDS_PENDING by in this scenario without running the qoutfifo, the manual update is sufficient. Clean up diagnostics. Add ahd_flush_qoutfifo() which will run the qoutfifo as well as complete any commands sitting on the sequencer's COMPLETE_SCB lists or the good status FIFO. Use this routine in several places that did similar things in an add-hoc, but incomplete, fashion. A call to this routine was also added to ahd_abort_scbs() to close a race. In ahd_pause_and_flushwork() only return once selections are safely disabled. Flush all completed commands via ahd_flush_qoutfifo(). Remove "Now packetized" diagnostic now that this information is incorperated into the actual negotiation messages that are displayed. When forcing renegotiation, don't clober the current ppr_options. Much of the driver uses this information to determine if we are currently packetized or not. Remove some stray spaces at column 1 in ahd_set_tags. When complaining about getting a host message loop request with no pending messages, print out the SCB_CONTROL register down on the card. Modify the ahd_sent_msg() routine to handle a search for an outgoing identify message. Use this to detect a msg reject on an identify message which typically indicates that the target thought we were packetized. Force a renegotiation in this case. In ahd_search_qinfifo(), wait more effectively for SCB DMA activities to cease. We also disable SCB fetch operations since we are about to change the qinfifo and any fetch in progress will likely be invalidated. In ahd_qinfifo_count(), fix the qinfifo empty case. In ahd_dump_card_state(), print out CCSCBCTL in the correct mode. If we are a narrow controller, don't set the current width to unknown when forcing a future negotiation. This just confuses the code into attempting a wide negotiation on a narrow bus. Add support for task management function completions. Modify ahd_handle_devreset so that it can handle lun resets in addition to target resets. Use ahd_handle_devreset for lun and target reset task management functions. Handle the abort task TMF race case better. We now wait until any current selections are over and then set the TMF back to zero. This should cause the sequencer to ignore the abort TMF completion should it occur. Correct a bug in the illegal phase handler that caused us to drop down to narrow when handling the unexpected command phase case after 3rd party reset of a packetized device. Indicate the features, bugs, and flags set in the softc that are used to control firmware patch download when booting verbose. aic79xx.h: Add coalessing and HS_MAILBOX fields. Add per-softc variables for the stats "daemon". Add a debug option for interrupt coalessing activities. Add two new softc flags: o AHD_UPDATE_PEND_CMDS Run ahd_reset_cmds_pending() on the next unpause. o AHD_RUNNING_QOUTFIFO Used to catch recursion through ahd_run_qoutfifo(). aic79xx.reg: Correct register addresses related to the software timer and the DFDBCTL register. Add constants paramaterizing the software timer. Add scratch ram locations for storing interrupt coalessing tunables. Break INTMASK in SEQITNCTL out into INTMASK1 and INTMASK2. In at least the REV A, these are writable bits. We make use of that for a swtimer workaround in the sequencer. Since HS_MAILBOX autoclears, provide a sequencer variable to store its contents. Add SEQINT codes for handling task management completions. aic79xx.seq: Correct ignore wide residue processing check for a wide negotiation being in effect. We must be in the SCSI register window in order to access the negotiation table. Use the software timer and a commands completed count to implement interrupt coalessing. The command complete is deferred until either the maximum command threshold or a the expiration of a command deferral timer. If we have more SCBs to complete to the host (sitting in COMPLETE_SCB lists), always try to coaless them up to our coalessing limit. If coalessing is enabled, but we have fewer commands oustanting than the host's min coalessing limit, complete the command immediately. Add code to track the number of commands outstanding. Commands are outstanding from the time they are placed into the execution queue until the DMA to post completion is setup. Add a workaround for intvec_2 interrupts on the H2A4. In H2A4, the mode pointer is not saved for intvec2, but is restored on iret. This can lead to the restoration of a bogus mode ptr. Manually clear the intmask bits and do a normal return to compensate. We use intvec_2 to track interrupt coalessing timeouts. Since we cannot disable the swtimer's countdown, simply mask its interrupt once we no longer care about it firing. In idle_loop_cchan, update LOCAL_HS_MAILBOX everytime we are notified of an HS_MAILBOX update via the HS_MAILBOX_ACT bit in QOFF_CTLSTA. We have to use a local copy of persistant portions of the HS_MAILBOX as the mailbox auto-clears on any read. Move the test for the cfg4istat interrupt up an instruction to hopefully close a race between the next outgoing selection and our disabling of selections. Add a missing ret to the last instruction in load_overrun_buf. Add notifications to the host of task management completions as well as the completions for commands that completed successfully before their corresponding TMF could be sent. Hold a critical section during select-out processing until we have a fully identified connection. This removes a race condition with the legacy abort handler. Correct a few spelling errors in some comments. aic79xx_inline.h: Call ahd_reset_cmds_pending() in ahd_unpause if required. Update cmdcmplt interrupt statistics in our interrupt handler. Allow callers to ahd_send_scb() to set the task management function. aic79xx_pci.c: Disable SERR and pause the controller prior to performing our mmapped I/O test. The U320 controllers do not support "auto-access-pause". aic79xx_osm.c: Set the task management function now that ahd_send_scb() doesn't do it for us. We also perform a lun reset in response to BDR requests to packetized devices.
2003-01-20 20:17:35 +00:00
*/
#define AHD_INT_COALESCING_TIMER_DEFAULT 250 /*us*/
#define AHD_INT_COALESCING_MAXCMDS_DEFAULT 10
#define AHD_INT_COALESCING_MAXCMDS_MAX 127
#define AHD_INT_COALESCING_MINCMDS_DEFAULT 5
#define AHD_INT_COALESCING_MINCMDS_MAX 127
#define AHD_INT_COALESCING_THRESHOLD_DEFAULT 2000
#define AHD_INT_COALESCING_STOP_THRESHOLD_DEFAULT 1000
u_int int_coalescing_timer;
u_int int_coalescing_maxcmds;
u_int int_coalescing_mincmds;
u_int int_coalescing_threshold;
u_int int_coalescing_stop_threshold;
aic79xx.c: Implement the SCB_SILENT flag. This is useful for hushing up the driver during DV or other operations that we expect to cause transmission errors. The messages will still print if the SHOW_MASKED_ERRORS debug option is enabled. Save and restore the NEGOADDR address when setting new transfer settings. The sequencer performs lookups in the negotiation table too and it expects NEGOADDR to remain consistent across pause/unpause sessions. Consistently use "offset" instead of "period" to determine if we are running sync or not. Add a SHOW_MESSAGES diagnostic for when we assert ATN during message processing. Print out IU, QAS, and RTI features when showing transfer options. Limit the syncrate after all option conformance changes have taken place in ahd_devlimited_syncrate. Changes in options may change the final syncrate we accept. Keep a copy of the hs_mailbox in our softc so that we can perform read/modify/write operations on the hs_mailbox without having to pause the sequencer to read the last written value. Use the ENINT_COALESS flag in the hs_mailbox to toggle interrupt coalessing. Add entrypoints for enabling interrupt coalessing and setting both a timeout (how long to wait for commands to be coalessed) and a maximum commands to coaless value. Add a statistics timer that decides when to enable or disable interrupt coalessing based on load. Add a routine, ahd_reset_cmds_pending() which is used to update the CMDS_PENDING sequencer variable whenever error recovery compeltes SCBs without notifying the sequencer. Since ahd_reset_cmds_pending is called during ahd_unpause() only if we've aborted SCBs, its call to ahd_flush_qoutfifo should not cause recursion through ahd_run_qoutfifo(). A panic has been added to ensure that this recursion does not occur. In ahd_search_qinfifo, update the CMDS_PENDING sequencer variable directly. ahd_search_qinififo can be called in situations where using ahd_reset_cmds_pending() might cause recursion. Since we can safely determine the exact number to reduce CMDS_PENDING by in this scenario without running the qoutfifo, the manual update is sufficient. Clean up diagnostics. Add ahd_flush_qoutfifo() which will run the qoutfifo as well as complete any commands sitting on the sequencer's COMPLETE_SCB lists or the good status FIFO. Use this routine in several places that did similar things in an add-hoc, but incomplete, fashion. A call to this routine was also added to ahd_abort_scbs() to close a race. In ahd_pause_and_flushwork() only return once selections are safely disabled. Flush all completed commands via ahd_flush_qoutfifo(). Remove "Now packetized" diagnostic now that this information is incorperated into the actual negotiation messages that are displayed. When forcing renegotiation, don't clober the current ppr_options. Much of the driver uses this information to determine if we are currently packetized or not. Remove some stray spaces at column 1 in ahd_set_tags. When complaining about getting a host message loop request with no pending messages, print out the SCB_CONTROL register down on the card. Modify the ahd_sent_msg() routine to handle a search for an outgoing identify message. Use this to detect a msg reject on an identify message which typically indicates that the target thought we were packetized. Force a renegotiation in this case. In ahd_search_qinfifo(), wait more effectively for SCB DMA activities to cease. We also disable SCB fetch operations since we are about to change the qinfifo and any fetch in progress will likely be invalidated. In ahd_qinfifo_count(), fix the qinfifo empty case. In ahd_dump_card_state(), print out CCSCBCTL in the correct mode. If we are a narrow controller, don't set the current width to unknown when forcing a future negotiation. This just confuses the code into attempting a wide negotiation on a narrow bus. Add support for task management function completions. Modify ahd_handle_devreset so that it can handle lun resets in addition to target resets. Use ahd_handle_devreset for lun and target reset task management functions. Handle the abort task TMF race case better. We now wait until any current selections are over and then set the TMF back to zero. This should cause the sequencer to ignore the abort TMF completion should it occur. Correct a bug in the illegal phase handler that caused us to drop down to narrow when handling the unexpected command phase case after 3rd party reset of a packetized device. Indicate the features, bugs, and flags set in the softc that are used to control firmware patch download when booting verbose. aic79xx.h: Add coalessing and HS_MAILBOX fields. Add per-softc variables for the stats "daemon". Add a debug option for interrupt coalessing activities. Add two new softc flags: o AHD_UPDATE_PEND_CMDS Run ahd_reset_cmds_pending() on the next unpause. o AHD_RUNNING_QOUTFIFO Used to catch recursion through ahd_run_qoutfifo(). aic79xx.reg: Correct register addresses related to the software timer and the DFDBCTL register. Add constants paramaterizing the software timer. Add scratch ram locations for storing interrupt coalessing tunables. Break INTMASK in SEQITNCTL out into INTMASK1 and INTMASK2. In at least the REV A, these are writable bits. We make use of that for a swtimer workaround in the sequencer. Since HS_MAILBOX autoclears, provide a sequencer variable to store its contents. Add SEQINT codes for handling task management completions. aic79xx.seq: Correct ignore wide residue processing check for a wide negotiation being in effect. We must be in the SCSI register window in order to access the negotiation table. Use the software timer and a commands completed count to implement interrupt coalessing. The command complete is deferred until either the maximum command threshold or a the expiration of a command deferral timer. If we have more SCBs to complete to the host (sitting in COMPLETE_SCB lists), always try to coaless them up to our coalessing limit. If coalessing is enabled, but we have fewer commands oustanting than the host's min coalessing limit, complete the command immediately. Add code to track the number of commands outstanding. Commands are outstanding from the time they are placed into the execution queue until the DMA to post completion is setup. Add a workaround for intvec_2 interrupts on the H2A4. In H2A4, the mode pointer is not saved for intvec2, but is restored on iret. This can lead to the restoration of a bogus mode ptr. Manually clear the intmask bits and do a normal return to compensate. We use intvec_2 to track interrupt coalessing timeouts. Since we cannot disable the swtimer's countdown, simply mask its interrupt once we no longer care about it firing. In idle_loop_cchan, update LOCAL_HS_MAILBOX everytime we are notified of an HS_MAILBOX update via the HS_MAILBOX_ACT bit in QOFF_CTLSTA. We have to use a local copy of persistant portions of the HS_MAILBOX as the mailbox auto-clears on any read. Move the test for the cfg4istat interrupt up an instruction to hopefully close a race between the next outgoing selection and our disabling of selections. Add a missing ret to the last instruction in load_overrun_buf. Add notifications to the host of task management completions as well as the completions for commands that completed successfully before their corresponding TMF could be sent. Hold a critical section during select-out processing until we have a fully identified connection. This removes a race condition with the legacy abort handler. Correct a few spelling errors in some comments. aic79xx_inline.h: Call ahd_reset_cmds_pending() in ahd_unpause if required. Update cmdcmplt interrupt statistics in our interrupt handler. Allow callers to ahd_send_scb() to set the task management function. aic79xx_pci.c: Disable SERR and pause the controller prior to performing our mmapped I/O test. The U320 controllers do not support "auto-access-pause". aic79xx_osm.c: Set the task management function now that ahd_send_scb() doesn't do it for us. We also perform a lun reset in response to BDR requests to packetized devices.
2003-01-20 20:17:35 +00:00
uint16_t user_discenable;/* Disconnection allowed */
uint16_t user_tagenable;/* Tagged Queuing allowed */
};
TAILQ_HEAD(ahd_softc_tailq, ahd_softc);
extern struct ahd_softc_tailq ahd_tailq;
/*************************** IO Cell Configuration ****************************/
#define AHD_PRECOMP_SLEW_INDEX \
(AHD_ANNEXCOL_PRECOMP_SLEW - AHD_ANNEXCOL_PER_DEV0)
#define AHD_AMPLITUDE_INDEX \
(AHD_ANNEXCOL_AMPLITUDE - AHD_ANNEXCOL_PER_DEV0)
#define AHD_SET_SLEWRATE(ahd, new_slew) \
do { \
(ahd)->iocell_opts[AHD_PRECOMP_SLEW_INDEX] &= ~AHD_SLEWRATE_MASK; \
(ahd)->iocell_opts[AHD_PRECOMP_SLEW_INDEX] |= \
(((new_slew) << AHD_SLEWRATE_SHIFT) & AHD_SLEWRATE_MASK); \
} while (0)
#define AHD_SET_PRECOMP(ahd, new_pcomp) \
do { \
(ahd)->iocell_opts[AHD_PRECOMP_SLEW_INDEX] &= ~AHD_PRECOMP_MASK; \
(ahd)->iocell_opts[AHD_PRECOMP_SLEW_INDEX] |= \
(((new_pcomp) << AHD_PRECOMP_SHIFT) & AHD_PRECOMP_MASK); \
} while (0)
#define AHD_SET_AMPLITUDE(ahd, new_amp) \
do { \
(ahd)->iocell_opts[AHD_AMPLITUDE_INDEX] &= ~AHD_AMPLITUDE_MASK; \
(ahd)->iocell_opts[AHD_AMPLITUDE_INDEX] |= \
(((new_amp) << AHD_AMPLITUDE_SHIFT) & AHD_AMPLITUDE_MASK); \
} while (0)
/************************ Active Device Information ***************************/
typedef enum {
ROLE_UNKNOWN,
ROLE_INITIATOR,
ROLE_TARGET
} role_t;
struct ahd_devinfo {
int our_scsiid;
int target_offset;
uint16_t target_mask;
u_int target;
u_int lun;
char channel;
role_t role; /*
* Only guaranteed to be correct if not
* in the busfree state.
*/
};
/****************************** PCI Structures ********************************/
#define AHD_PCI_IOADDR0 PCIR_MAPS /* I/O BAR*/
#define AHD_PCI_MEMADDR (PCIR_MAPS + 4) /* Memory BAR */
#define AHD_PCI_IOADDR1 (PCIR_MAPS + 12)/* Second I/O BAR */
typedef int (ahd_device_setup_t)(struct ahd_softc *);
struct ahd_pci_identity {
uint64_t full_id;
uint64_t id_mask;
char *name;
ahd_device_setup_t *setup;
};
extern struct ahd_pci_identity ahd_pci_ident_table [];
extern const u_int ahd_num_pci_devs;
/***************************** VL/EISA Declarations ***************************/
struct aic7770_identity {
uint32_t full_id;
uint32_t id_mask;
char *name;
ahd_device_setup_t *setup;
};
extern struct aic7770_identity aic7770_ident_table [];
extern const int ahd_num_aic7770_devs;
#define AHD_EISA_SLOT_OFFSET 0xc00
#define AHD_EISA_IOSIZE 0x100
/*************************** Function Declarations ****************************/
/******************************************************************************/
aic79xx.c: Implement the SCB_SILENT flag. This is useful for hushing up the driver during DV or other operations that we expect to cause transmission errors. The messages will still print if the SHOW_MASKED_ERRORS debug option is enabled. Save and restore the NEGOADDR address when setting new transfer settings. The sequencer performs lookups in the negotiation table too and it expects NEGOADDR to remain consistent across pause/unpause sessions. Consistently use "offset" instead of "period" to determine if we are running sync or not. Add a SHOW_MESSAGES diagnostic for when we assert ATN during message processing. Print out IU, QAS, and RTI features when showing transfer options. Limit the syncrate after all option conformance changes have taken place in ahd_devlimited_syncrate. Changes in options may change the final syncrate we accept. Keep a copy of the hs_mailbox in our softc so that we can perform read/modify/write operations on the hs_mailbox without having to pause the sequencer to read the last written value. Use the ENINT_COALESS flag in the hs_mailbox to toggle interrupt coalessing. Add entrypoints for enabling interrupt coalessing and setting both a timeout (how long to wait for commands to be coalessed) and a maximum commands to coaless value. Add a statistics timer that decides when to enable or disable interrupt coalessing based on load. Add a routine, ahd_reset_cmds_pending() which is used to update the CMDS_PENDING sequencer variable whenever error recovery compeltes SCBs without notifying the sequencer. Since ahd_reset_cmds_pending is called during ahd_unpause() only if we've aborted SCBs, its call to ahd_flush_qoutfifo should not cause recursion through ahd_run_qoutfifo(). A panic has been added to ensure that this recursion does not occur. In ahd_search_qinfifo, update the CMDS_PENDING sequencer variable directly. ahd_search_qinififo can be called in situations where using ahd_reset_cmds_pending() might cause recursion. Since we can safely determine the exact number to reduce CMDS_PENDING by in this scenario without running the qoutfifo, the manual update is sufficient. Clean up diagnostics. Add ahd_flush_qoutfifo() which will run the qoutfifo as well as complete any commands sitting on the sequencer's COMPLETE_SCB lists or the good status FIFO. Use this routine in several places that did similar things in an add-hoc, but incomplete, fashion. A call to this routine was also added to ahd_abort_scbs() to close a race. In ahd_pause_and_flushwork() only return once selections are safely disabled. Flush all completed commands via ahd_flush_qoutfifo(). Remove "Now packetized" diagnostic now that this information is incorperated into the actual negotiation messages that are displayed. When forcing renegotiation, don't clober the current ppr_options. Much of the driver uses this information to determine if we are currently packetized or not. Remove some stray spaces at column 1 in ahd_set_tags. When complaining about getting a host message loop request with no pending messages, print out the SCB_CONTROL register down on the card. Modify the ahd_sent_msg() routine to handle a search for an outgoing identify message. Use this to detect a msg reject on an identify message which typically indicates that the target thought we were packetized. Force a renegotiation in this case. In ahd_search_qinfifo(), wait more effectively for SCB DMA activities to cease. We also disable SCB fetch operations since we are about to change the qinfifo and any fetch in progress will likely be invalidated. In ahd_qinfifo_count(), fix the qinfifo empty case. In ahd_dump_card_state(), print out CCSCBCTL in the correct mode. If we are a narrow controller, don't set the current width to unknown when forcing a future negotiation. This just confuses the code into attempting a wide negotiation on a narrow bus. Add support for task management function completions. Modify ahd_handle_devreset so that it can handle lun resets in addition to target resets. Use ahd_handle_devreset for lun and target reset task management functions. Handle the abort task TMF race case better. We now wait until any current selections are over and then set the TMF back to zero. This should cause the sequencer to ignore the abort TMF completion should it occur. Correct a bug in the illegal phase handler that caused us to drop down to narrow when handling the unexpected command phase case after 3rd party reset of a packetized device. Indicate the features, bugs, and flags set in the softc that are used to control firmware patch download when booting verbose. aic79xx.h: Add coalessing and HS_MAILBOX fields. Add per-softc variables for the stats "daemon". Add a debug option for interrupt coalessing activities. Add two new softc flags: o AHD_UPDATE_PEND_CMDS Run ahd_reset_cmds_pending() on the next unpause. o AHD_RUNNING_QOUTFIFO Used to catch recursion through ahd_run_qoutfifo(). aic79xx.reg: Correct register addresses related to the software timer and the DFDBCTL register. Add constants paramaterizing the software timer. Add scratch ram locations for storing interrupt coalessing tunables. Break INTMASK in SEQITNCTL out into INTMASK1 and INTMASK2. In at least the REV A, these are writable bits. We make use of that for a swtimer workaround in the sequencer. Since HS_MAILBOX autoclears, provide a sequencer variable to store its contents. Add SEQINT codes for handling task management completions. aic79xx.seq: Correct ignore wide residue processing check for a wide negotiation being in effect. We must be in the SCSI register window in order to access the negotiation table. Use the software timer and a commands completed count to implement interrupt coalessing. The command complete is deferred until either the maximum command threshold or a the expiration of a command deferral timer. If we have more SCBs to complete to the host (sitting in COMPLETE_SCB lists), always try to coaless them up to our coalessing limit. If coalessing is enabled, but we have fewer commands oustanting than the host's min coalessing limit, complete the command immediately. Add code to track the number of commands outstanding. Commands are outstanding from the time they are placed into the execution queue until the DMA to post completion is setup. Add a workaround for intvec_2 interrupts on the H2A4. In H2A4, the mode pointer is not saved for intvec2, but is restored on iret. This can lead to the restoration of a bogus mode ptr. Manually clear the intmask bits and do a normal return to compensate. We use intvec_2 to track interrupt coalessing timeouts. Since we cannot disable the swtimer's countdown, simply mask its interrupt once we no longer care about it firing. In idle_loop_cchan, update LOCAL_HS_MAILBOX everytime we are notified of an HS_MAILBOX update via the HS_MAILBOX_ACT bit in QOFF_CTLSTA. We have to use a local copy of persistant portions of the HS_MAILBOX as the mailbox auto-clears on any read. Move the test for the cfg4istat interrupt up an instruction to hopefully close a race between the next outgoing selection and our disabling of selections. Add a missing ret to the last instruction in load_overrun_buf. Add notifications to the host of task management completions as well as the completions for commands that completed successfully before their corresponding TMF could be sent. Hold a critical section during select-out processing until we have a fully identified connection. This removes a race condition with the legacy abort handler. Correct a few spelling errors in some comments. aic79xx_inline.h: Call ahd_reset_cmds_pending() in ahd_unpause if required. Update cmdcmplt interrupt statistics in our interrupt handler. Allow callers to ahd_send_scb() to set the task management function. aic79xx_pci.c: Disable SERR and pause the controller prior to performing our mmapped I/O test. The U320 controllers do not support "auto-access-pause". aic79xx_osm.c: Set the task management function now that ahd_send_scb() doesn't do it for us. We also perform a lun reset in response to BDR requests to packetized devices.
2003-01-20 20:17:35 +00:00
void ahd_reset_cmds_pending(struct ahd_softc *ahd);
u_int ahd_find_busy_tcl(struct ahd_softc *ahd, u_int tcl);
void ahd_busy_tcl(struct ahd_softc *ahd,
u_int tcl, u_int busyid);
static __inline void ahd_unbusy_tcl(struct ahd_softc *ahd, u_int tcl);
static __inline void
ahd_unbusy_tcl(struct ahd_softc *ahd, u_int tcl)
{
ahd_busy_tcl(ahd, tcl, SCB_LIST_NULL);
}
/***************************** PCI Front End *********************************/
struct ahd_pci_identity *ahd_find_pci_device(ahd_dev_softc_t);
int ahd_pci_config(struct ahd_softc *,
struct ahd_pci_identity *);
int ahd_pci_test_register_access(struct ahd_softc *);
/************************** SCB and SCB queue management **********************/
int ahd_probe_scbs(struct ahd_softc *);
void ahd_qinfifo_requeue_tail(struct ahd_softc *ahd,
struct scb *scb);
int ahd_match_scb(struct ahd_softc *ahd, struct scb *scb,
int target, char channel, int lun,
u_int tag, role_t role);
/****************************** Initialization ********************************/
struct ahd_softc *ahd_alloc(void *platform_arg, char *name);
int ahd_softc_init(struct ahd_softc *);
void ahd_controller_info(struct ahd_softc *ahd, char *buf);
int ahd_init(struct ahd_softc *ahd);
int ahd_default_config(struct ahd_softc *ahd);
Correct spelling errors. Switch to handling bad SCSI status as a sequencer interrupt instead of having the kernel proccess these failures via the completion queue. This is done because: o The old scheme required us to pause the sequencer and clear critical sections for each SCB. It seems that these pause actions, if coincident with a sequencer FIFO interrupt, would result in a FIFO interrupt getting lost or directing to the wrong FIFO. This caused hangs when the driver was stressed under high "queue full" loads. o The completion code assumed that it was always called with the sequencer running. This may not be the case in timeout processing where completions occur manually via ahd_pause_and_flushwork(). o With this scheme, the extra expense of clearing critical sections is avoided since the sequencer will only self pause once all pending selections have cleared and it is not in a critical section. aic79xx.c Add code to handle the new BAD_SCB_STATUS sequencer interrupt code. This just redirects the SCB through the already existing ahd_complete_scb() code path. Remove code in ahd_handle_scsi_status() that paused the sequencer, made sure that no selections where pending, and cleared critical sections. Bad status SCBs are now only processed when all of these conditions are true. aic79xx.reg: Add the BAD_SCB_STATUS sequencer interrupt code. aic79xx.seq: When completing an SCB upload to the host, if we are doing this because the SCB contains non-zero SCSI status, defer completing the SCB until there are no pending selection events. When completing these SCBs, use the new BAD_SCB_STATUS sequencer interrupt. For all other uploaded SCBs (currently only for underruns), the SCB is completed via the normal done queue. Additionally, keep the SCB that is currently being uploaded on the COMPLETE_DMA_SCB list until the dma is completed, not just until the DMA is started. This ensures that the DMA is restarted properly should the host disable the DMA transfer for some reason. In our RevA workaround for Maxtor drives, guard against the host pausing us while trying to pause I/O until the first data-valid REQ by clearing the current snapshot so that we can tell if the transfer has completed prior to us noticing the REQINIT status. In cfg4data_intr, shave off an instruction before getting the data path running by adding an entrypoint to the overrun handler to also increment the FIFO use count. In the overrun handler, be sure to clear our LONGJMP address in both exit paths. Perform a few sequencer optimizations. aic79xx.c: Print the full path from the SCB when a packetized status overrun occurs. Remove references to LONGJMP_SCB which is being removed from firmware usage. Print the new SCB_FIFO_USE_COUNT field in the per-SCB section of ahd_dump_card_state(). The SCB_TAG field is now re-used by the sequencer, so it no longer makes sense to reference this field in the kernel driver. aic79xx.h: Re-arrange fields in the hardware SCB from largest size type to smallest. This makes it easier to move fields without changing field alignment. The hardware scb tag field is now down near the "spare" portion of the SCB to facilitate reuse by the sequencer. aic79xx.reg: Remove LONGJMP_ADDR. Rearrange SCB fields to match aic79xx.h. Add SCB_FIFO_USE_COUNT as the first byte of the SCB_TAG field. aic79xx.seq: Add a per-SCB "Fifos in use count" field and use it to determine when it is safe (all data posted) to deliver status back to the host. The old method involved polling one or both FIFOs to verify that the current task did not have pending data. This makes running down the GSFIFO very cheap, so we will empty the GSFIFO in one idle loop pass in all cases. Use this simplification of the completion process to prune down the data FIFO teardown sequencer for packetized transfers. Much more code is now shared between the data residual and transfer complete cases. Correct some issues in the packetized status handler. It used to be possible to CLRCHN our FIFO before status had fully transferred to the host. We also failed to handle NONPACKREQ phases that could occur should a CRC error occur during transmission of the status data packet. Correct a few big endian issues: aic79xx.c: aic79xx_inline.h: aic79xx_pci.c: aic79xx_osm.c: o Always get the SCB's tag via the SCB_GET_TAG acccessor o Add missing use of byte swapping macros when touching hscb fields. o Don't double swap SEEPROM data when it is printed. Correct a big-endian bug. We cannot assign a o When assigning a 32bit LE variable to a 64bit LE variable, we must be explict about how the words of the 64bit LE variable are initialized. Cast to (uint32_t*) to do this. aic79xx.c: In ahd_clear_critical_section(), hit CRLSCSIINT after restoring the interrupt masks to avoid what appears to be a glitch on SCSIINT. Any real SCSIINT status will be persistent and will immidiately reset SCSIINT. This clear should only get rid of spurious SCSIINTs. This glitch was the cause of the "Unexpected PKT busfree" status that occurred under high queue full loads Call ahd_fini_scbdata() after shutdown so that any ahd_chip_init() routine that might access SCB data will not access free'd memory. Reset the bus on an IOERR since the chip doesn't seem to reset to the new voltage level without this. Change offset calculation for scatter gather maps so that the calculation is correct if an integral multiple of sg lists does not fit in the allocation size. Adjust bus dma tag for data buffers based on 39BIT addressing flag in our softc. Use the QFREEZE count to simplify ahd_pause_and_flushworkd(). We can thus rely on the sequencer eventually clearing ENSELO. In ahd_abort_scbs(), fix a bug that could potentially corrupt sequencer state. The saved SCB was being restored in the SCSI mode instead of the saved mode. It turns out that the SCB did not need to be saved at all as the scbptr is already restored by all subroutines called during this function that modify that register. aic79xx.c: aic79xx.h: aic79xx_pci.c: Add support for parsing the seeprom vital product data. The VPD data are currently unused. aic79xx.h: aic79xx.seq: aic79xx_pci.c: Add a firmware workaround to make the LED blink brighter during packetized operations on the H2A. aic79xx_inline.h: The host does not use timer interrupts, so don't gate our decision on whether or not to unpause the sequencer on whether or not a timer interrupt is pending.
2003-05-04 00:20:07 +00:00
int ahd_parse_vpddata(struct ahd_softc *ahd,
struct vpd_config *vpd);
int ahd_parse_cfgdata(struct ahd_softc *ahd,
struct seeprom_config *sc);
void ahd_intr_enable(struct ahd_softc *ahd, int enable);
void ahd_update_coalescing_values(struct ahd_softc *ahd,
aic79xx.c: Implement the SCB_SILENT flag. This is useful for hushing up the driver during DV or other operations that we expect to cause transmission errors. The messages will still print if the SHOW_MASKED_ERRORS debug option is enabled. Save and restore the NEGOADDR address when setting new transfer settings. The sequencer performs lookups in the negotiation table too and it expects NEGOADDR to remain consistent across pause/unpause sessions. Consistently use "offset" instead of "period" to determine if we are running sync or not. Add a SHOW_MESSAGES diagnostic for when we assert ATN during message processing. Print out IU, QAS, and RTI features when showing transfer options. Limit the syncrate after all option conformance changes have taken place in ahd_devlimited_syncrate. Changes in options may change the final syncrate we accept. Keep a copy of the hs_mailbox in our softc so that we can perform read/modify/write operations on the hs_mailbox without having to pause the sequencer to read the last written value. Use the ENINT_COALESS flag in the hs_mailbox to toggle interrupt coalessing. Add entrypoints for enabling interrupt coalessing and setting both a timeout (how long to wait for commands to be coalessed) and a maximum commands to coaless value. Add a statistics timer that decides when to enable or disable interrupt coalessing based on load. Add a routine, ahd_reset_cmds_pending() which is used to update the CMDS_PENDING sequencer variable whenever error recovery compeltes SCBs without notifying the sequencer. Since ahd_reset_cmds_pending is called during ahd_unpause() only if we've aborted SCBs, its call to ahd_flush_qoutfifo should not cause recursion through ahd_run_qoutfifo(). A panic has been added to ensure that this recursion does not occur. In ahd_search_qinfifo, update the CMDS_PENDING sequencer variable directly. ahd_search_qinififo can be called in situations where using ahd_reset_cmds_pending() might cause recursion. Since we can safely determine the exact number to reduce CMDS_PENDING by in this scenario without running the qoutfifo, the manual update is sufficient. Clean up diagnostics. Add ahd_flush_qoutfifo() which will run the qoutfifo as well as complete any commands sitting on the sequencer's COMPLETE_SCB lists or the good status FIFO. Use this routine in several places that did similar things in an add-hoc, but incomplete, fashion. A call to this routine was also added to ahd_abort_scbs() to close a race. In ahd_pause_and_flushwork() only return once selections are safely disabled. Flush all completed commands via ahd_flush_qoutfifo(). Remove "Now packetized" diagnostic now that this information is incorperated into the actual negotiation messages that are displayed. When forcing renegotiation, don't clober the current ppr_options. Much of the driver uses this information to determine if we are currently packetized or not. Remove some stray spaces at column 1 in ahd_set_tags. When complaining about getting a host message loop request with no pending messages, print out the SCB_CONTROL register down on the card. Modify the ahd_sent_msg() routine to handle a search for an outgoing identify message. Use this to detect a msg reject on an identify message which typically indicates that the target thought we were packetized. Force a renegotiation in this case. In ahd_search_qinfifo(), wait more effectively for SCB DMA activities to cease. We also disable SCB fetch operations since we are about to change the qinfifo and any fetch in progress will likely be invalidated. In ahd_qinfifo_count(), fix the qinfifo empty case. In ahd_dump_card_state(), print out CCSCBCTL in the correct mode. If we are a narrow controller, don't set the current width to unknown when forcing a future negotiation. This just confuses the code into attempting a wide negotiation on a narrow bus. Add support for task management function completions. Modify ahd_handle_devreset so that it can handle lun resets in addition to target resets. Use ahd_handle_devreset for lun and target reset task management functions. Handle the abort task TMF race case better. We now wait until any current selections are over and then set the TMF back to zero. This should cause the sequencer to ignore the abort TMF completion should it occur. Correct a bug in the illegal phase handler that caused us to drop down to narrow when handling the unexpected command phase case after 3rd party reset of a packetized device. Indicate the features, bugs, and flags set in the softc that are used to control firmware patch download when booting verbose. aic79xx.h: Add coalessing and HS_MAILBOX fields. Add per-softc variables for the stats "daemon". Add a debug option for interrupt coalessing activities. Add two new softc flags: o AHD_UPDATE_PEND_CMDS Run ahd_reset_cmds_pending() on the next unpause. o AHD_RUNNING_QOUTFIFO Used to catch recursion through ahd_run_qoutfifo(). aic79xx.reg: Correct register addresses related to the software timer and the DFDBCTL register. Add constants paramaterizing the software timer. Add scratch ram locations for storing interrupt coalessing tunables. Break INTMASK in SEQITNCTL out into INTMASK1 and INTMASK2. In at least the REV A, these are writable bits. We make use of that for a swtimer workaround in the sequencer. Since HS_MAILBOX autoclears, provide a sequencer variable to store its contents. Add SEQINT codes for handling task management completions. aic79xx.seq: Correct ignore wide residue processing check for a wide negotiation being in effect. We must be in the SCSI register window in order to access the negotiation table. Use the software timer and a commands completed count to implement interrupt coalessing. The command complete is deferred until either the maximum command threshold or a the expiration of a command deferral timer. If we have more SCBs to complete to the host (sitting in COMPLETE_SCB lists), always try to coaless them up to our coalessing limit. If coalessing is enabled, but we have fewer commands oustanting than the host's min coalessing limit, complete the command immediately. Add code to track the number of commands outstanding. Commands are outstanding from the time they are placed into the execution queue until the DMA to post completion is setup. Add a workaround for intvec_2 interrupts on the H2A4. In H2A4, the mode pointer is not saved for intvec2, but is restored on iret. This can lead to the restoration of a bogus mode ptr. Manually clear the intmask bits and do a normal return to compensate. We use intvec_2 to track interrupt coalessing timeouts. Since we cannot disable the swtimer's countdown, simply mask its interrupt once we no longer care about it firing. In idle_loop_cchan, update LOCAL_HS_MAILBOX everytime we are notified of an HS_MAILBOX update via the HS_MAILBOX_ACT bit in QOFF_CTLSTA. We have to use a local copy of persistant portions of the HS_MAILBOX as the mailbox auto-clears on any read. Move the test for the cfg4istat interrupt up an instruction to hopefully close a race between the next outgoing selection and our disabling of selections. Add a missing ret to the last instruction in load_overrun_buf. Add notifications to the host of task management completions as well as the completions for commands that completed successfully before their corresponding TMF could be sent. Hold a critical section during select-out processing until we have a fully identified connection. This removes a race condition with the legacy abort handler. Correct a few spelling errors in some comments. aic79xx_inline.h: Call ahd_reset_cmds_pending() in ahd_unpause if required. Update cmdcmplt interrupt statistics in our interrupt handler. Allow callers to ahd_send_scb() to set the task management function. aic79xx_pci.c: Disable SERR and pause the controller prior to performing our mmapped I/O test. The U320 controllers do not support "auto-access-pause". aic79xx_osm.c: Set the task management function now that ahd_send_scb() doesn't do it for us. We also perform a lun reset in response to BDR requests to packetized devices.
2003-01-20 20:17:35 +00:00
u_int timer,
u_int maxcmds,
u_int mincmds);
void ahd_enable_coalescing(struct ahd_softc *ahd,
aic79xx.c: Implement the SCB_SILENT flag. This is useful for hushing up the driver during DV or other operations that we expect to cause transmission errors. The messages will still print if the SHOW_MASKED_ERRORS debug option is enabled. Save and restore the NEGOADDR address when setting new transfer settings. The sequencer performs lookups in the negotiation table too and it expects NEGOADDR to remain consistent across pause/unpause sessions. Consistently use "offset" instead of "period" to determine if we are running sync or not. Add a SHOW_MESSAGES diagnostic for when we assert ATN during message processing. Print out IU, QAS, and RTI features when showing transfer options. Limit the syncrate after all option conformance changes have taken place in ahd_devlimited_syncrate. Changes in options may change the final syncrate we accept. Keep a copy of the hs_mailbox in our softc so that we can perform read/modify/write operations on the hs_mailbox without having to pause the sequencer to read the last written value. Use the ENINT_COALESS flag in the hs_mailbox to toggle interrupt coalessing. Add entrypoints for enabling interrupt coalessing and setting both a timeout (how long to wait for commands to be coalessed) and a maximum commands to coaless value. Add a statistics timer that decides when to enable or disable interrupt coalessing based on load. Add a routine, ahd_reset_cmds_pending() which is used to update the CMDS_PENDING sequencer variable whenever error recovery compeltes SCBs without notifying the sequencer. Since ahd_reset_cmds_pending is called during ahd_unpause() only if we've aborted SCBs, its call to ahd_flush_qoutfifo should not cause recursion through ahd_run_qoutfifo(). A panic has been added to ensure that this recursion does not occur. In ahd_search_qinfifo, update the CMDS_PENDING sequencer variable directly. ahd_search_qinififo can be called in situations where using ahd_reset_cmds_pending() might cause recursion. Since we can safely determine the exact number to reduce CMDS_PENDING by in this scenario without running the qoutfifo, the manual update is sufficient. Clean up diagnostics. Add ahd_flush_qoutfifo() which will run the qoutfifo as well as complete any commands sitting on the sequencer's COMPLETE_SCB lists or the good status FIFO. Use this routine in several places that did similar things in an add-hoc, but incomplete, fashion. A call to this routine was also added to ahd_abort_scbs() to close a race. In ahd_pause_and_flushwork() only return once selections are safely disabled. Flush all completed commands via ahd_flush_qoutfifo(). Remove "Now packetized" diagnostic now that this information is incorperated into the actual negotiation messages that are displayed. When forcing renegotiation, don't clober the current ppr_options. Much of the driver uses this information to determine if we are currently packetized or not. Remove some stray spaces at column 1 in ahd_set_tags. When complaining about getting a host message loop request with no pending messages, print out the SCB_CONTROL register down on the card. Modify the ahd_sent_msg() routine to handle a search for an outgoing identify message. Use this to detect a msg reject on an identify message which typically indicates that the target thought we were packetized. Force a renegotiation in this case. In ahd_search_qinfifo(), wait more effectively for SCB DMA activities to cease. We also disable SCB fetch operations since we are about to change the qinfifo and any fetch in progress will likely be invalidated. In ahd_qinfifo_count(), fix the qinfifo empty case. In ahd_dump_card_state(), print out CCSCBCTL in the correct mode. If we are a narrow controller, don't set the current width to unknown when forcing a future negotiation. This just confuses the code into attempting a wide negotiation on a narrow bus. Add support for task management function completions. Modify ahd_handle_devreset so that it can handle lun resets in addition to target resets. Use ahd_handle_devreset for lun and target reset task management functions. Handle the abort task TMF race case better. We now wait until any current selections are over and then set the TMF back to zero. This should cause the sequencer to ignore the abort TMF completion should it occur. Correct a bug in the illegal phase handler that caused us to drop down to narrow when handling the unexpected command phase case after 3rd party reset of a packetized device. Indicate the features, bugs, and flags set in the softc that are used to control firmware patch download when booting verbose. aic79xx.h: Add coalessing and HS_MAILBOX fields. Add per-softc variables for the stats "daemon". Add a debug option for interrupt coalessing activities. Add two new softc flags: o AHD_UPDATE_PEND_CMDS Run ahd_reset_cmds_pending() on the next unpause. o AHD_RUNNING_QOUTFIFO Used to catch recursion through ahd_run_qoutfifo(). aic79xx.reg: Correct register addresses related to the software timer and the DFDBCTL register. Add constants paramaterizing the software timer. Add scratch ram locations for storing interrupt coalessing tunables. Break INTMASK in SEQITNCTL out into INTMASK1 and INTMASK2. In at least the REV A, these are writable bits. We make use of that for a swtimer workaround in the sequencer. Since HS_MAILBOX autoclears, provide a sequencer variable to store its contents. Add SEQINT codes for handling task management completions. aic79xx.seq: Correct ignore wide residue processing check for a wide negotiation being in effect. We must be in the SCSI register window in order to access the negotiation table. Use the software timer and a commands completed count to implement interrupt coalessing. The command complete is deferred until either the maximum command threshold or a the expiration of a command deferral timer. If we have more SCBs to complete to the host (sitting in COMPLETE_SCB lists), always try to coaless them up to our coalessing limit. If coalessing is enabled, but we have fewer commands oustanting than the host's min coalessing limit, complete the command immediately. Add code to track the number of commands outstanding. Commands are outstanding from the time they are placed into the execution queue until the DMA to post completion is setup. Add a workaround for intvec_2 interrupts on the H2A4. In H2A4, the mode pointer is not saved for intvec2, but is restored on iret. This can lead to the restoration of a bogus mode ptr. Manually clear the intmask bits and do a normal return to compensate. We use intvec_2 to track interrupt coalessing timeouts. Since we cannot disable the swtimer's countdown, simply mask its interrupt once we no longer care about it firing. In idle_loop_cchan, update LOCAL_HS_MAILBOX everytime we are notified of an HS_MAILBOX update via the HS_MAILBOX_ACT bit in QOFF_CTLSTA. We have to use a local copy of persistant portions of the HS_MAILBOX as the mailbox auto-clears on any read. Move the test for the cfg4istat interrupt up an instruction to hopefully close a race between the next outgoing selection and our disabling of selections. Add a missing ret to the last instruction in load_overrun_buf. Add notifications to the host of task management completions as well as the completions for commands that completed successfully before their corresponding TMF could be sent. Hold a critical section during select-out processing until we have a fully identified connection. This removes a race condition with the legacy abort handler. Correct a few spelling errors in some comments. aic79xx_inline.h: Call ahd_reset_cmds_pending() in ahd_unpause if required. Update cmdcmplt interrupt statistics in our interrupt handler. Allow callers to ahd_send_scb() to set the task management function. aic79xx_pci.c: Disable SERR and pause the controller prior to performing our mmapped I/O test. The U320 controllers do not support "auto-access-pause". aic79xx_osm.c: Set the task management function now that ahd_send_scb() doesn't do it for us. We also perform a lun reset in response to BDR requests to packetized devices.
2003-01-20 20:17:35 +00:00
int enable);
void ahd_pause_and_flushwork(struct ahd_softc *ahd);
int ahd_suspend(struct ahd_softc *ahd);
int ahd_resume(struct ahd_softc *ahd);
void ahd_softc_insert(struct ahd_softc *);
struct ahd_softc *ahd_find_softc(struct ahd_softc *ahd);
void ahd_set_unit(struct ahd_softc *, int);
void ahd_set_name(struct ahd_softc *, char *);
struct scb *ahd_get_scb(struct ahd_softc *ahd, u_int col_idx);
void ahd_free_scb(struct ahd_softc *ahd, struct scb *scb);
void ahd_alloc_scbs(struct ahd_softc *ahd);
void ahd_free(struct ahd_softc *ahd);
int ahd_reset(struct ahd_softc *ahd);
void ahd_shutdown(void *arg);
int ahd_write_flexport(struct ahd_softc *ahd,
u_int addr, u_int value);
int ahd_read_flexport(struct ahd_softc *ahd, u_int addr,
uint8_t *value);
int ahd_wait_flexport(struct ahd_softc *ahd);
/*************************** Interrupt Services *******************************/
void ahd_pci_intr(struct ahd_softc *ahd);
void ahd_clear_intstat(struct ahd_softc *ahd);
aic79xx.c: Implement the SCB_SILENT flag. This is useful for hushing up the driver during DV or other operations that we expect to cause transmission errors. The messages will still print if the SHOW_MASKED_ERRORS debug option is enabled. Save and restore the NEGOADDR address when setting new transfer settings. The sequencer performs lookups in the negotiation table too and it expects NEGOADDR to remain consistent across pause/unpause sessions. Consistently use "offset" instead of "period" to determine if we are running sync or not. Add a SHOW_MESSAGES diagnostic for when we assert ATN during message processing. Print out IU, QAS, and RTI features when showing transfer options. Limit the syncrate after all option conformance changes have taken place in ahd_devlimited_syncrate. Changes in options may change the final syncrate we accept. Keep a copy of the hs_mailbox in our softc so that we can perform read/modify/write operations on the hs_mailbox without having to pause the sequencer to read the last written value. Use the ENINT_COALESS flag in the hs_mailbox to toggle interrupt coalessing. Add entrypoints for enabling interrupt coalessing and setting both a timeout (how long to wait for commands to be coalessed) and a maximum commands to coaless value. Add a statistics timer that decides when to enable or disable interrupt coalessing based on load. Add a routine, ahd_reset_cmds_pending() which is used to update the CMDS_PENDING sequencer variable whenever error recovery compeltes SCBs without notifying the sequencer. Since ahd_reset_cmds_pending is called during ahd_unpause() only if we've aborted SCBs, its call to ahd_flush_qoutfifo should not cause recursion through ahd_run_qoutfifo(). A panic has been added to ensure that this recursion does not occur. In ahd_search_qinfifo, update the CMDS_PENDING sequencer variable directly. ahd_search_qinififo can be called in situations where using ahd_reset_cmds_pending() might cause recursion. Since we can safely determine the exact number to reduce CMDS_PENDING by in this scenario without running the qoutfifo, the manual update is sufficient. Clean up diagnostics. Add ahd_flush_qoutfifo() which will run the qoutfifo as well as complete any commands sitting on the sequencer's COMPLETE_SCB lists or the good status FIFO. Use this routine in several places that did similar things in an add-hoc, but incomplete, fashion. A call to this routine was also added to ahd_abort_scbs() to close a race. In ahd_pause_and_flushwork() only return once selections are safely disabled. Flush all completed commands via ahd_flush_qoutfifo(). Remove "Now packetized" diagnostic now that this information is incorperated into the actual negotiation messages that are displayed. When forcing renegotiation, don't clober the current ppr_options. Much of the driver uses this information to determine if we are currently packetized or not. Remove some stray spaces at column 1 in ahd_set_tags. When complaining about getting a host message loop request with no pending messages, print out the SCB_CONTROL register down on the card. Modify the ahd_sent_msg() routine to handle a search for an outgoing identify message. Use this to detect a msg reject on an identify message which typically indicates that the target thought we were packetized. Force a renegotiation in this case. In ahd_search_qinfifo(), wait more effectively for SCB DMA activities to cease. We also disable SCB fetch operations since we are about to change the qinfifo and any fetch in progress will likely be invalidated. In ahd_qinfifo_count(), fix the qinfifo empty case. In ahd_dump_card_state(), print out CCSCBCTL in the correct mode. If we are a narrow controller, don't set the current width to unknown when forcing a future negotiation. This just confuses the code into attempting a wide negotiation on a narrow bus. Add support for task management function completions. Modify ahd_handle_devreset so that it can handle lun resets in addition to target resets. Use ahd_handle_devreset for lun and target reset task management functions. Handle the abort task TMF race case better. We now wait until any current selections are over and then set the TMF back to zero. This should cause the sequencer to ignore the abort TMF completion should it occur. Correct a bug in the illegal phase handler that caused us to drop down to narrow when handling the unexpected command phase case after 3rd party reset of a packetized device. Indicate the features, bugs, and flags set in the softc that are used to control firmware patch download when booting verbose. aic79xx.h: Add coalessing and HS_MAILBOX fields. Add per-softc variables for the stats "daemon". Add a debug option for interrupt coalessing activities. Add two new softc flags: o AHD_UPDATE_PEND_CMDS Run ahd_reset_cmds_pending() on the next unpause. o AHD_RUNNING_QOUTFIFO Used to catch recursion through ahd_run_qoutfifo(). aic79xx.reg: Correct register addresses related to the software timer and the DFDBCTL register. Add constants paramaterizing the software timer. Add scratch ram locations for storing interrupt coalessing tunables. Break INTMASK in SEQITNCTL out into INTMASK1 and INTMASK2. In at least the REV A, these are writable bits. We make use of that for a swtimer workaround in the sequencer. Since HS_MAILBOX autoclears, provide a sequencer variable to store its contents. Add SEQINT codes for handling task management completions. aic79xx.seq: Correct ignore wide residue processing check for a wide negotiation being in effect. We must be in the SCSI register window in order to access the negotiation table. Use the software timer and a commands completed count to implement interrupt coalessing. The command complete is deferred until either the maximum command threshold or a the expiration of a command deferral timer. If we have more SCBs to complete to the host (sitting in COMPLETE_SCB lists), always try to coaless them up to our coalessing limit. If coalessing is enabled, but we have fewer commands oustanting than the host's min coalessing limit, complete the command immediately. Add code to track the number of commands outstanding. Commands are outstanding from the time they are placed into the execution queue until the DMA to post completion is setup. Add a workaround for intvec_2 interrupts on the H2A4. In H2A4, the mode pointer is not saved for intvec2, but is restored on iret. This can lead to the restoration of a bogus mode ptr. Manually clear the intmask bits and do a normal return to compensate. We use intvec_2 to track interrupt coalessing timeouts. Since we cannot disable the swtimer's countdown, simply mask its interrupt once we no longer care about it firing. In idle_loop_cchan, update LOCAL_HS_MAILBOX everytime we are notified of an HS_MAILBOX update via the HS_MAILBOX_ACT bit in QOFF_CTLSTA. We have to use a local copy of persistant portions of the HS_MAILBOX as the mailbox auto-clears on any read. Move the test for the cfg4istat interrupt up an instruction to hopefully close a race between the next outgoing selection and our disabling of selections. Add a missing ret to the last instruction in load_overrun_buf. Add notifications to the host of task management completions as well as the completions for commands that completed successfully before their corresponding TMF could be sent. Hold a critical section during select-out processing until we have a fully identified connection. This removes a race condition with the legacy abort handler. Correct a few spelling errors in some comments. aic79xx_inline.h: Call ahd_reset_cmds_pending() in ahd_unpause if required. Update cmdcmplt interrupt statistics in our interrupt handler. Allow callers to ahd_send_scb() to set the task management function. aic79xx_pci.c: Disable SERR and pause the controller prior to performing our mmapped I/O test. The U320 controllers do not support "auto-access-pause". aic79xx_osm.c: Set the task management function now that ahd_send_scb() doesn't do it for us. We also perform a lun reset in response to BDR requests to packetized devices.
2003-01-20 20:17:35 +00:00
void ahd_flush_qoutfifo(struct ahd_softc *ahd);
void ahd_run_qoutfifo(struct ahd_softc *ahd);
#ifdef AHD_TARGET_MODE
void ahd_run_tqinfifo(struct ahd_softc *ahd, int paused);
#endif
void ahd_handle_hwerrint(struct ahd_softc *ahd);
void ahd_handle_seqint(struct ahd_softc *ahd, u_int intstat);
void ahd_handle_scsiint(struct ahd_softc *ahd,
u_int intstat);
void ahd_clear_critical_section(struct ahd_softc *ahd);
/***************************** Error Recovery *********************************/
typedef enum {
SEARCH_COMPLETE,
SEARCH_COUNT,
SEARCH_REMOVE,
SEARCH_PRINT
} ahd_search_action;
int ahd_search_qinfifo(struct ahd_softc *ahd, int target,
char channel, int lun, u_int tag,
role_t role, uint32_t status,
ahd_search_action action);
int ahd_search_disc_list(struct ahd_softc *ahd, int target,
char channel, int lun, u_int tag,
int stop_on_first, int remove,
int save_state);
void ahd_freeze_devq(struct ahd_softc *ahd, struct scb *scb);
int ahd_reset_channel(struct ahd_softc *ahd, char channel,
int initiate_reset);
int ahd_abort_scbs(struct ahd_softc *ahd, int target,
char channel, int lun, u_int tag,
role_t role, uint32_t status);
void ahd_restart(struct ahd_softc *ahd);
void ahd_clear_fifo(struct ahd_softc *ahd, u_int fifo);
void ahd_handle_scb_status(struct ahd_softc *ahd,
struct scb *scb);
void ahd_handle_scsi_status(struct ahd_softc *ahd,
struct scb *scb);
void ahd_calc_residual(struct ahd_softc *ahd,
struct scb *scb);
/*************************** Utility Functions ********************************/
struct ahd_phase_table_entry*
ahd_lookup_phase_entry(int phase);
void ahd_compile_devinfo(struct ahd_devinfo *devinfo,
u_int our_id, u_int target,
u_int lun, char channel,
role_t role);
/************************** Transfer Negotiation ******************************/
void ahd_find_syncrate(struct ahd_softc *ahd, u_int *period,
u_int *ppr_options, u_int maxsync);
void ahd_validate_offset(struct ahd_softc *ahd,
struct ahd_initiator_tinfo *tinfo,
u_int period, u_int *offset,
int wide, role_t role);
void ahd_validate_width(struct ahd_softc *ahd,
struct ahd_initiator_tinfo *tinfo,
u_int *bus_width,
role_t role);
/*
* Negotiation types. These are used to qualify if we should renegotiate
* even if our goal and current transport parameters are identical.
*/
typedef enum {
AHD_NEG_TO_GOAL, /* Renegotiate only if goal and curr differ. */
AHD_NEG_IF_NON_ASYNC, /* Renegotiate so long as goal is non-async. */
AHD_NEG_ALWAYS /* Renegotiat even if goal is async. */
} ahd_neg_type;
int ahd_update_neg_request(struct ahd_softc*,
struct ahd_devinfo*,
struct ahd_tmode_tstate*,
struct ahd_initiator_tinfo*,
ahd_neg_type);
void ahd_set_width(struct ahd_softc *ahd,
struct ahd_devinfo *devinfo,
u_int width, u_int type, int paused);
void ahd_set_syncrate(struct ahd_softc *ahd,
struct ahd_devinfo *devinfo,
u_int period, u_int offset,
u_int ppr_options,
u_int type, int paused);
typedef enum {
AHD_QUEUE_NONE,
AHD_QUEUE_BASIC,
AHD_QUEUE_TAGGED
} ahd_queue_alg;
void ahd_set_tags(struct ahd_softc *ahd,
struct ahd_devinfo *devinfo,
ahd_queue_alg alg);
/**************************** Target Mode *************************************/
#ifdef AHD_TARGET_MODE
void ahd_send_lstate_events(struct ahd_softc *,
struct ahd_tmode_lstate *);
void ahd_handle_en_lun(struct ahd_softc *ahd,
struct cam_sim *sim, union ccb *ccb);
cam_status ahd_find_tmode_devs(struct ahd_softc *ahd,
struct cam_sim *sim, union ccb *ccb,
struct ahd_tmode_tstate **tstate,
struct ahd_tmode_lstate **lstate,
int notfound_failure);
#ifndef AHD_TMODE_ENABLE
#define AHD_TMODE_ENABLE 0
#endif
#endif
/******************************* Debug ***************************************/
#ifdef AHD_DEBUG
extern uint32_t ahd_debug;
#define AHD_SHOW_MISC 0x00001
#define AHD_SHOW_SENSE 0x00002
#define AHD_SHOW_RECOVERY 0x00004
#define AHD_DUMP_SEEPROM 0x00008
#define AHD_SHOW_TERMCTL 0x00010
#define AHD_SHOW_MEMORY 0x00020
#define AHD_SHOW_MESSAGES 0x00040
#define AHD_SHOW_MODEPTR 0x00080
#define AHD_SHOW_SELTO 0x00100
#define AHD_SHOW_FIFOS 0x00200
#define AHD_SHOW_QFULL 0x00400
#define AHD_SHOW_DV 0x00800
#define AHD_SHOW_MASKED_ERRORS 0x01000
#define AHD_SHOW_QUEUE 0x02000
#define AHD_SHOW_TQIN 0x04000
#define AHD_SHOW_SG 0x08000
#define AHD_SHOW_INT_COALESCING 0x10000
aic79xx.c: Implement the SCB_SILENT flag. This is useful for hushing up the driver during DV or other operations that we expect to cause transmission errors. The messages will still print if the SHOW_MASKED_ERRORS debug option is enabled. Save and restore the NEGOADDR address when setting new transfer settings. The sequencer performs lookups in the negotiation table too and it expects NEGOADDR to remain consistent across pause/unpause sessions. Consistently use "offset" instead of "period" to determine if we are running sync or not. Add a SHOW_MESSAGES diagnostic for when we assert ATN during message processing. Print out IU, QAS, and RTI features when showing transfer options. Limit the syncrate after all option conformance changes have taken place in ahd_devlimited_syncrate. Changes in options may change the final syncrate we accept. Keep a copy of the hs_mailbox in our softc so that we can perform read/modify/write operations on the hs_mailbox without having to pause the sequencer to read the last written value. Use the ENINT_COALESS flag in the hs_mailbox to toggle interrupt coalessing. Add entrypoints for enabling interrupt coalessing and setting both a timeout (how long to wait for commands to be coalessed) and a maximum commands to coaless value. Add a statistics timer that decides when to enable or disable interrupt coalessing based on load. Add a routine, ahd_reset_cmds_pending() which is used to update the CMDS_PENDING sequencer variable whenever error recovery compeltes SCBs without notifying the sequencer. Since ahd_reset_cmds_pending is called during ahd_unpause() only if we've aborted SCBs, its call to ahd_flush_qoutfifo should not cause recursion through ahd_run_qoutfifo(). A panic has been added to ensure that this recursion does not occur. In ahd_search_qinfifo, update the CMDS_PENDING sequencer variable directly. ahd_search_qinififo can be called in situations where using ahd_reset_cmds_pending() might cause recursion. Since we can safely determine the exact number to reduce CMDS_PENDING by in this scenario without running the qoutfifo, the manual update is sufficient. Clean up diagnostics. Add ahd_flush_qoutfifo() which will run the qoutfifo as well as complete any commands sitting on the sequencer's COMPLETE_SCB lists or the good status FIFO. Use this routine in several places that did similar things in an add-hoc, but incomplete, fashion. A call to this routine was also added to ahd_abort_scbs() to close a race. In ahd_pause_and_flushwork() only return once selections are safely disabled. Flush all completed commands via ahd_flush_qoutfifo(). Remove "Now packetized" diagnostic now that this information is incorperated into the actual negotiation messages that are displayed. When forcing renegotiation, don't clober the current ppr_options. Much of the driver uses this information to determine if we are currently packetized or not. Remove some stray spaces at column 1 in ahd_set_tags. When complaining about getting a host message loop request with no pending messages, print out the SCB_CONTROL register down on the card. Modify the ahd_sent_msg() routine to handle a search for an outgoing identify message. Use this to detect a msg reject on an identify message which typically indicates that the target thought we were packetized. Force a renegotiation in this case. In ahd_search_qinfifo(), wait more effectively for SCB DMA activities to cease. We also disable SCB fetch operations since we are about to change the qinfifo and any fetch in progress will likely be invalidated. In ahd_qinfifo_count(), fix the qinfifo empty case. In ahd_dump_card_state(), print out CCSCBCTL in the correct mode. If we are a narrow controller, don't set the current width to unknown when forcing a future negotiation. This just confuses the code into attempting a wide negotiation on a narrow bus. Add support for task management function completions. Modify ahd_handle_devreset so that it can handle lun resets in addition to target resets. Use ahd_handle_devreset for lun and target reset task management functions. Handle the abort task TMF race case better. We now wait until any current selections are over and then set the TMF back to zero. This should cause the sequencer to ignore the abort TMF completion should it occur. Correct a bug in the illegal phase handler that caused us to drop down to narrow when handling the unexpected command phase case after 3rd party reset of a packetized device. Indicate the features, bugs, and flags set in the softc that are used to control firmware patch download when booting verbose. aic79xx.h: Add coalessing and HS_MAILBOX fields. Add per-softc variables for the stats "daemon". Add a debug option for interrupt coalessing activities. Add two new softc flags: o AHD_UPDATE_PEND_CMDS Run ahd_reset_cmds_pending() on the next unpause. o AHD_RUNNING_QOUTFIFO Used to catch recursion through ahd_run_qoutfifo(). aic79xx.reg: Correct register addresses related to the software timer and the DFDBCTL register. Add constants paramaterizing the software timer. Add scratch ram locations for storing interrupt coalessing tunables. Break INTMASK in SEQITNCTL out into INTMASK1 and INTMASK2. In at least the REV A, these are writable bits. We make use of that for a swtimer workaround in the sequencer. Since HS_MAILBOX autoclears, provide a sequencer variable to store its contents. Add SEQINT codes for handling task management completions. aic79xx.seq: Correct ignore wide residue processing check for a wide negotiation being in effect. We must be in the SCSI register window in order to access the negotiation table. Use the software timer and a commands completed count to implement interrupt coalessing. The command complete is deferred until either the maximum command threshold or a the expiration of a command deferral timer. If we have more SCBs to complete to the host (sitting in COMPLETE_SCB lists), always try to coaless them up to our coalessing limit. If coalessing is enabled, but we have fewer commands oustanting than the host's min coalessing limit, complete the command immediately. Add code to track the number of commands outstanding. Commands are outstanding from the time they are placed into the execution queue until the DMA to post completion is setup. Add a workaround for intvec_2 interrupts on the H2A4. In H2A4, the mode pointer is not saved for intvec2, but is restored on iret. This can lead to the restoration of a bogus mode ptr. Manually clear the intmask bits and do a normal return to compensate. We use intvec_2 to track interrupt coalessing timeouts. Since we cannot disable the swtimer's countdown, simply mask its interrupt once we no longer care about it firing. In idle_loop_cchan, update LOCAL_HS_MAILBOX everytime we are notified of an HS_MAILBOX update via the HS_MAILBOX_ACT bit in QOFF_CTLSTA. We have to use a local copy of persistant portions of the HS_MAILBOX as the mailbox auto-clears on any read. Move the test for the cfg4istat interrupt up an instruction to hopefully close a race between the next outgoing selection and our disabling of selections. Add a missing ret to the last instruction in load_overrun_buf. Add notifications to the host of task management completions as well as the completions for commands that completed successfully before their corresponding TMF could be sent. Hold a critical section during select-out processing until we have a fully identified connection. This removes a race condition with the legacy abort handler. Correct a few spelling errors in some comments. aic79xx_inline.h: Call ahd_reset_cmds_pending() in ahd_unpause if required. Update cmdcmplt interrupt statistics in our interrupt handler. Allow callers to ahd_send_scb() to set the task management function. aic79xx_pci.c: Disable SERR and pause the controller prior to performing our mmapped I/O test. The U320 controllers do not support "auto-access-pause". aic79xx_osm.c: Set the task management function now that ahd_send_scb() doesn't do it for us. We also perform a lun reset in response to BDR requests to packetized devices.
2003-01-20 20:17:35 +00:00
#define AHD_DEBUG_SEQUENCER 0x20000
#endif
void ahd_print_scb(struct scb *scb);
void ahd_print_devinfo(struct ahd_softc *ahd,
struct ahd_devinfo *devinfo);
void ahd_dump_sglist(struct scb *scb);
void ahd_dump_all_cards_state(void);
void ahd_dump_card_state(struct ahd_softc *ahd);
int ahd_print_register(ahd_reg_parse_entry_t *table,
u_int num_entries,
const char *name,
u_int address,
u_int value,
u_int *cur_column,
u_int wrap_point);
void ahd_dump_scbs(struct ahd_softc *ahd);
#endif /* _AIC79XX_H_ */