Aren
/
NetBSD
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
NetBSD/sys/dev/ic/isp_netbsd.c

817 lines
20 KiB
C
Raw Normal View History

Apply nearly all of Bill Sommerfeld's -Wformat patches. The one change I didn't take is the %llu format- I can't have a common across multiple platform module assume a %ll argument capability- which really pointed out that I shouldn't be trying to *print* something which could long long.
2000-12-29 01:27:46 +03:00
/* $NetBSD: isp_netbsd.c,v 1.37 2000/12/28 22:27:47 mjacob Exp $ */
Add a maintenance note. Change all splbio's to use the ISP_LOCK/ISP_UNLOCK or ISP_ILOCK/ISP_IUNLOCK macros.
2000-08-14 11:08:12 +04:00
/*
* This driver, which is contained in NetBSD in the files:
*
* sys/dev/ic/isp.c
Fix pathnames in comment.
2000-12-23 04:37:57 +03:00
* sys/dev/ic/isp_inline.h
* sys/dev/ic/isp_netbsd.c
* sys/dev/ic/isp_netbsd.h
* sys/dev/ic/isp_target.c
* sys/dev/ic/isp_target.h
* sys/dev/ic/isp_tpublic.h
* sys/dev/ic/ispmbox.h
* sys/dev/ic/ispreg.h
* sys/dev/ic/ispvar.h
Add a maintenance note. Change all splbio's to use the ISP_LOCK/ISP_UNLOCK or ISP_ILOCK/ISP_IUNLOCK macros.
2000-08-14 11:08:12 +04:00
* sys/microcode/isp/asm_sbus.h
* sys/microcode/isp/asm_1040.h
* sys/microcode/isp/asm_1080.h
* sys/microcode/isp/asm_12160.h
* sys/microcode/isp/asm_2100.h
* sys/microcode/isp/asm_2200.h
* sys/pci/isp_pci.c
* sys/sbus/isp_sbus.c
*
* Is being actively maintained by Matthew Jacob (mjacob@netbsd.org).
* This driver also is shared source with FreeBSD, OpenBSD, Linux, Solaris,
* Linux versions. This tends to be an interesting maintenance problem.
*
* Please coordinate with Matthew Jacob on changes you wish to make here.
*/
NetBSD OS specific routines and definitions.
1998-07-15 23:44:04 +04:00
/*
* Platform (NetBSD) dependent common attachment code for Qlogic adapters.
Change to approved NASA/Ames copyright. Add Fabric support. Fix SCCLUN support. Add code that tries to track LoopID shifting.
1999-07-06 00:31:35 +04:00
* Matthew Jacob <mjacob@nas.nasa.gov>
*/
/*
* Copyright (C) 1997, 1998, 1999 National Aeronautics & Space Administration
NetBSD OS specific routines and definitions.
1998-07-15 23:44:04 +04:00
* 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
Change to approved NASA/Ames copyright. Add Fabric support. Fix SCCLUN support. Add code that tries to track LoopID shifting.
1999-07-06 00:31:35 +04:00
* notice, this list of conditions and the following disclaimer.
NetBSD OS specific routines and definitions.
1998-07-15 23:44:04 +04:00
* 2. Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
* 3. The name of the author may not be used to endorse or promote products
Change to approved NASA/Ames copyright. Add Fabric support. Fix SCCLUN support. Add code that tries to track LoopID shifting.
1999-07-06 00:31:35 +04:00
* derived from this software without specific prior written permission
NetBSD OS specific routines and definitions.
1998-07-15 23:44:04 +04:00
*
Change to approved NASA/Ames copyright. Add Fabric support. Fix SCCLUN support. Add code that tries to track LoopID shifting.
1999-07-06 00:31:35 +04:00
* THIS SOFTWARE IS PROVIDED BY THE AUTHOR ``AS IS'' AND ANY EXPRESS OR
* IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
* OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
* IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR ANY DIRECT, INDIRECT,
* INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
* NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
* DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
* THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
* (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
* THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
NetBSD OS specific routines and definitions.
1998-07-15 23:44:04 +04:00
*/
#include <dev/ic/isp_netbsd.h>
Add in an ioctl entry point so scsictl mediated bus resets will work. Redo how we start commands- do a 'slow' start function which then looks to see when we're done the configuration process at which point it *then* enables sync/wide mode. Set the max openings amount to the true max openings- not a synthetic. Add a timeout driven command requeue function so that Loop Down events well freeze things until a later point in time where they might be restarted.
1999-10-14 06:31:11 +04:00
#include <sys/scsiio.h>
NetBSD OS specific routines and definitions.
1998-07-15 23:44:04 +04:00
Gah. Bad port from FreeBSD code (thanks Robert Elz) for setting up default timeout for watchdogging commands.
2000-07-07 07:14:53 +04:00
/*
* Set a timeout for the watchdogging of a command.
*
* The dimensional analysis is
*
* milliseconds * (seconds/millisecond) * (ticks/second) = ticks
*
* =
*
* (milliseconds / 1000) * hz = ticks
*
*
* For timeouts less than 1 second, we'll get zero. Because of this, and
* because we want to establish *our* timeout to be longer than what the
* firmware might do, we just add 3 seconds at the back end.
*/
#define _XT(xs) ((((xs)->timeout/1000) * hz) + (3 * hz))
Redo watchdogs to handle cases of false death of commands- time each command now.. Get rid of SCCLUn stuff. Use an isp_done routine to handle cases of watchdog and isp_done racing to completion.
2000-07-06 02:23:05 +04:00
NetBSD OS specific routines and definitions.
1998-07-15 23:44:04 +04:00
static void ispminphys __P((struct buf *));
Core version 2.0 (platform version 1.0) rewrite of ISP driver. Some interace cleanups, some new common functions. The major impact that will be noticeable right away is that if you boot with not Fibre connected to the FC cards, you no longer hang indefinitely.
2000-08-02 03:55:09 +04:00
static int32_t ispcmd __P((XS_T *));
Add in an ioctl entry point so scsictl mediated bus resets will work. Redo how we start commands- do a 'slow' start function which then looks to see when we're done the configuration process at which point it *then* enables sync/wide mode. Set the max openings amount to the true max openings- not a synthetic. Add a timeout driven command requeue function so that Loop Down events well freeze things until a later point in time where they might be restarted.
1999-10-14 06:31:11 +04:00
static int
ispioctl __P((struct scsipi_link *, u_long, caddr_t, int, struct proc *));
NetBSD OS specific routines and definitions.
1998-07-15 23:44:04 +04:00
static struct scsipi_device isp_dev = { NULL, NULL, NULL, NULL };
Core version 2.0 (platform version 1.0) rewrite of ISP driver. Some interace cleanups, some new common functions. The major impact that will be noticeable right away is that if you boot with not Fibre connected to the FC cards, you no longer hang indefinitely.
2000-08-02 03:55:09 +04:00
static int isp_polled_cmd __P((struct ispsoftc *, XS_T *));
Redo watchdogs to handle cases of false death of commands- time each command now.. Get rid of SCCLUn stuff. Use an isp_done routine to handle cases of watchdog and isp_done racing to completion.
2000-07-06 02:23:05 +04:00
static void isp_dog __P((void *));
Add in an ioctl entry point so scsictl mediated bus resets will work. Redo how we start commands- do a 'slow' start function which then looks to see when we're done the configuration process at which point it *then* enables sync/wide mode. Set the max openings amount to the true max openings- not a synthetic. Add a timeout driven command requeue function so that Loop Down events well freeze things until a later point in time where they might be restarted.
1999-10-14 06:31:11 +04:00
static void isp_command_requeue __P((void *));
Do SCSI Bus resets in this layer (now). Don't do it for Fibre Channel yet (we get LOOP DOWN events, and we'll hang on that at this time). Add other isp_async cases- ISPASYNC_LOOP_DOWN and ISPASYNC_LOOP_UP. DOWN will cause internal queuing until UP, whereupon a timeout will fire up any pending xfers. It doesn't really keep commands from getting destroyed by loop down events, but at least minimizes the damage. This was much easier to implement with CAM.
1999-02-09 03:42:22 +03:00
static void isp_internal_restart __P((void *));
NetBSD OS specific routines and definitions.
1998-07-15 23:44:04 +04:00
/*
* Complete attachment of hardware, include subdevices.
*/
void
isp_attach(isp)
struct ispsoftc *isp;
{
Adapt to the new scsipi_adapter interface.
1998-11-20 00:43:00 +03:00
isp->isp_osinfo._adapter.scsipi_minphys = ispminphys;
Add in an ioctl entry point so scsictl mediated bus resets will work. Redo how we start commands- do a 'slow' start function which then looks to see when we're done the configuration process at which point it *then* enables sync/wide mode. Set the max openings amount to the true max openings- not a synthetic. Add a timeout driven command requeue function so that Loop Down events well freeze things until a later point in time where they might be restarted.
1999-10-14 06:31:11 +04:00
isp->isp_osinfo._adapter.scsipi_ioctl = ispioctl;
Remove ispcmd_slow routine- SCBUSACCEL routine will set/clear tag/wide/sync. Make some changes about where some things sit in the softc.
2000-08-09 02:58:30 +04:00
isp->isp_osinfo._adapter.scsipi_cmd = ispcmd;
Adapt to the new scsipi_adapter interface.
1998-11-20 00:43:00 +03:00
NetBSD OS specific routines and definitions.
1998-07-15 23:44:04 +04:00
isp->isp_state = ISP_RUNSTATE;
Dual LVD (1280) support.
1999-12-16 08:35:42 +03:00
isp->isp_osinfo._link.scsipi_scsi.channel =
(IS_DUALBUS(isp))? 0 : SCSI_CHANNEL_ONLY_ONE;
NetBSD OS specific routines and definitions.
1998-07-15 23:44:04 +04:00
isp->isp_osinfo._link.adapter_softc = isp;
isp->isp_osinfo._link.device = &isp_dev;
Adapt to the new scsipi_adapter interface.
1998-11-20 00:43:00 +03:00
isp->isp_osinfo._link.adapter = &isp->isp_osinfo._adapter;
Add in an ioctl entry point so scsictl mediated bus resets will work. Redo how we start commands- do a 'slow' start function which then looks to see when we're done the configuration process at which point it *then* enables sync/wide mode. Set the max openings amount to the true max openings- not a synthetic. Add a timeout driven command requeue function so that Loop Down events well freeze things until a later point in time where they might be restarted.
1999-10-14 06:31:11 +04:00
isp->isp_osinfo._link.openings = isp->isp_maxcmds;
Core version 2.0 (platform version 1.0) rewrite of ISP driver. Some interace cleanups, some new common functions. The major impact that will be noticeable right away is that if you boot with not Fibre connected to the FC cards, you no longer hang indefinitely.
2000-08-02 03:55:09 +04:00
/*
* Until the midlayer is fixed to use REPORT LUNS, limit to 8 luns.
*/
isp->isp_osinfo._link.scsipi_scsi.max_lun =
(isp->isp_maxluns < 7)? isp->isp_maxluns - 1 : 7;
TAILQ_INIT(&isp->isp_osinfo.waitq); /* The 2nd bus will share.. */
NetBSD OS specific routines and definitions.
1998-07-15 23:44:04 +04:00
Change to approved NASA/Ames copyright. Add Fabric support. Fix SCCLUN support. Add code that tries to track LoopID shifting.
1999-07-06 00:31:35 +04:00
if (IS_FC(isp)) {
NetBSD OS specific routines and definitions.
1998-07-15 23:44:04 +04:00
isp->isp_osinfo._link.scsipi_scsi.max_target = MAX_FC_TARG-1;
} else {
Do a fairly large internal restructuring to accomodate dual-bus host adapters (e.g., the 1240). Include the new 1080/1240 NVRAM layout reading code. Some moderately significant mailbox changes were necessary also to accomodate a second channel.
1999-05-12 22:59:23 +04:00
sdparam *sdp = isp->isp_param;
NetBSD OS specific routines and definitions.
1998-07-15 23:44:04 +04:00
isp->isp_osinfo._link.scsipi_scsi.max_target = MAX_TARGETS-1;
isp->isp_osinfo._link.scsipi_scsi.adapter_target =
Do a fairly large internal restructuring to accomodate dual-bus host adapters (e.g., the 1240). Include the new 1080/1240 NVRAM layout reading code. Some moderately significant mailbox changes were necessary also to accomodate a second channel.
1999-05-12 22:59:23 +04:00
sdp->isp_initiator_id;
Add in an ioctl entry point so scsictl mediated bus resets will work. Redo how we start commands- do a 'slow' start function which then looks to see when we're done the configuration process at which point it *then* enables sync/wide mode. Set the max openings amount to the true max openings- not a synthetic. Add a timeout driven command requeue function so that Loop Down events well freeze things until a later point in time where they might be restarted.
1999-10-14 06:31:11 +04:00
isp->isp_osinfo.discovered[0] = 1 << sdp->isp_initiator_id;
Dual LVD (1280) support.
1999-12-16 08:35:42 +03:00
if (IS_DUALBUS(isp)) {
Do a fairly large internal restructuring to accomodate dual-bus host adapters (e.g., the 1240). Include the new 1080/1240 NVRAM layout reading code. Some moderately significant mailbox changes were necessary also to accomodate a second channel.
1999-05-12 22:59:23 +04:00
isp->isp_osinfo._link_b = isp->isp_osinfo._link;
sdp++;
some comments added
1999-10-17 05:23:21 +04:00
isp->isp_osinfo.discovered[1] =
1 << sdp->isp_initiator_id;
Do a fairly large internal restructuring to accomodate dual-bus host adapters (e.g., the 1240). Include the new 1080/1240 NVRAM layout reading code. Some moderately significant mailbox changes were necessary also to accomodate a second channel.
1999-05-12 22:59:23 +04:00
isp->isp_osinfo._link_b.scsipi_scsi.adapter_target =
sdp->isp_initiator_id;
isp->isp_osinfo._link_b.scsipi_scsi.channel = 1;
Redo watchdogs to handle cases of false death of commands- time each command now.. Get rid of SCCLUn stuff. Use an isp_done routine to handle cases of watchdog and isp_done racing to completion.
2000-07-06 02:23:05 +04:00
isp->isp_osinfo._link_b.scsipi_scsi.max_lun =
isp->isp_osinfo._link.scsipi_scsi.max_lun;
Do a fairly large internal restructuring to accomodate dual-bus host adapters (e.g., the 1240). Include the new 1080/1240 NVRAM layout reading code. Some moderately significant mailbox changes were necessary also to accomodate a second channel.
1999-05-12 22:59:23 +04:00
}
NetBSD OS specific routines and definitions.
1998-07-15 23:44:04 +04:00
}
isp->isp_osinfo._link.type = BUS_SCSI;
clean up headers; move uninit/watch to outer layers
1998-12-28 22:10:43 +03:00
Do SCSI Bus resets in this layer (now). Don't do it for Fibre Channel yet (we get LOOP DOWN events, and we'll hang on that at this time). Add other isp_async cases- ISPASYNC_LOOP_DOWN and ISPASYNC_LOOP_UP. DOWN will cause internal queuing until UP, whereupon a timeout will fire up any pending xfers. It doesn't really keep commands from getting destroyed by loop down events, but at least minimizes the damage. This was much easier to implement with CAM.
1999-02-09 03:42:22 +03:00
/*
Update from Matthew Jacob: Correctly account for F-port vs. F-port (no FLOGI_ACC) topologies. Make sure we get a port database entry for the fabric name server. Preserve fabric logins if the device didn't change across fabric or port database changes, or the device has already logged into us (e.g., for target/initiator dual role devices like Veritas SANbox). Propagate class 3 service parameter changes where devices can change roles. Fix all occurrences of setting a sendmarker so that setting it for one bus on dual bus cards doesn't wipe a pending sendmarker for other busses on the same card :-;. Comments added and clarifications made in some of the target mode code.
2000-05-13 20:53:03 +04:00
* Send a SCSI Bus Reset.
Do SCSI Bus resets in this layer (now). Don't do it for Fibre Channel yet (we get LOOP DOWN events, and we'll hang on that at this time). Add other isp_async cases- ISPASYNC_LOOP_DOWN and ISPASYNC_LOOP_UP. DOWN will cause internal queuing until UP, whereupon a timeout will fire up any pending xfers. It doesn't really keep commands from getting destroyed by loop down events, but at least minimizes the damage. This was much easier to implement with CAM.
1999-02-09 03:42:22 +03:00
*/
Do a fairly large internal restructuring to accomodate dual-bus host adapters (e.g., the 1240). Include the new 1080/1240 NVRAM layout reading code. Some moderately significant mailbox changes were necessary also to accomodate a second channel.
1999-05-12 22:59:23 +04:00
if (IS_SCSI(isp)) {
int bus = 0;
Core version 2.0 (platform version 1.0) rewrite of ISP driver. Some interace cleanups, some new common functions. The major impact that will be noticeable right away is that if you boot with not Fibre connected to the FC cards, you no longer hang indefinitely.
2000-08-02 03:55:09 +04:00
ISP_LOCK(isp);
Do a fairly large internal restructuring to accomodate dual-bus host adapters (e.g., the 1240). Include the new 1080/1240 NVRAM layout reading code. Some moderately significant mailbox changes were necessary also to accomodate a second channel.
1999-05-12 22:59:23 +04:00
(void) isp_control(isp, ISPCTL_RESET_BUS, &bus);
Dual LVD (1280) support.
1999-12-16 08:35:42 +03:00
if (IS_DUALBUS(isp)) {
Do a fairly large internal restructuring to accomodate dual-bus host adapters (e.g., the 1240). Include the new 1080/1240 NVRAM layout reading code. Some moderately significant mailbox changes were necessary also to accomodate a second channel.
1999-05-12 22:59:23 +04:00
bus++;
(void) isp_control(isp, ISPCTL_RESET_BUS, &bus);
}
Remove ispcmd_slow routine- SCBUSACCEL routine will set/clear tag/wide/sync. Make some changes about where some things sit in the softc.
2000-08-09 02:58:30 +04:00
ISP_UNLOCK(isp);
(wow- I thought I'd already committed most of this). Clean up some isp_attach time stuff- if ISP2100_FABRIC is defined try *really* hard to make sure that we get the firmware state to FW_READY and see the loop state where the Port Database is ready to be gathered- if we don't do this it's unlikely we will be able to correctly query the nameserver because we won't see that we're on a fabric. Clean up the completely broken and stupid attempts to hot switch the 'slow' start routine out of the way. Sigh. Turn speed announcements into CFGPRINTF functions (available only if DEBUG defined).
1999-12-04 06:04:59 +03:00
} else {
Core version 2.0 (platform version 1.0) rewrite of ISP driver. Some interace cleanups, some new common functions. The major impact that will be noticeable right away is that if you boot with not Fibre connected to the FC cards, you no longer hang indefinitely.
2000-08-02 03:55:09 +04:00
int defid;
(wow- I thought I'd already committed most of this). Clean up some isp_attach time stuff- if ISP2100_FABRIC is defined try *really* hard to make sure that we get the firmware state to FW_READY and see the loop state where the Port Database is ready to be gathered- if we don't do this it's unlikely we will be able to correctly query the nameserver because we won't see that we're on a fabric. Clean up the completely broken and stupid attempts to hot switch the 'slow' start routine out of the way. Sigh. Turn speed announcements into CFGPRINTF functions (available only if DEBUG defined).
1999-12-04 06:04:59 +03:00
fcparam *fcp = isp->isp_param;
delay(2 * 1000000);
Core version 2.0 (platform version 1.0) rewrite of ISP driver. Some interace cleanups, some new common functions. The major impact that will be noticeable right away is that if you boot with not Fibre connected to the FC cards, you no longer hang indefinitely.
2000-08-02 03:55:09 +04:00
defid = MAX_FC_TARG;
ISP_LOCK(isp);
/*
* We probably won't have clock interrupts running,
* so we'll be really short (smoke test, really)
* at this time.
*/
if (isp_control(isp, ISPCTL_FCLINK_TEST, NULL)) {
(void) isp_control(isp, ISPCTL_PDB_SYNC, NULL);
(wow- I thought I'd already committed most of this). Clean up some isp_attach time stuff- if ISP2100_FABRIC is defined try *really* hard to make sure that we get the firmware state to FW_READY and see the loop state where the Port Database is ready to be gathered- if we don't do this it's unlikely we will be able to correctly query the nameserver because we won't see that we're on a fabric. Clean up the completely broken and stupid attempts to hot switch the 'slow' start routine out of the way. Sigh. Turn speed announcements into CFGPRINTF functions (available only if DEBUG defined).
1999-12-04 06:04:59 +03:00
if (fcp->isp_fwstate == FW_READY &&
fcp->isp_loopstate >= LOOP_PDB_RCVD) {
Core version 2.0 (platform version 1.0) rewrite of ISP driver. Some interace cleanups, some new common functions. The major impact that will be noticeable right away is that if you boot with not Fibre connected to the FC cards, you no longer hang indefinitely.
2000-08-02 03:55:09 +04:00
defid = fcp->isp_loopid;
(wow- I thought I'd already committed most of this). Clean up some isp_attach time stuff- if ISP2100_FABRIC is defined try *really* hard to make sure that we get the firmware state to FW_READY and see the loop state where the Port Database is ready to be gathered- if we don't do this it's unlikely we will be able to correctly query the nameserver because we won't see that we're on a fabric. Clean up the completely broken and stupid attempts to hot switch the 'slow' start routine out of the way. Sigh. Turn speed announcements into CFGPRINTF functions (available only if DEBUG defined).
1999-12-04 06:04:59 +03:00
}
}
Core version 2.0 (platform version 1.0) rewrite of ISP driver. Some interace cleanups, some new common functions. The major impact that will be noticeable right away is that if you boot with not Fibre connected to the FC cards, you no longer hang indefinitely.
2000-08-02 03:55:09 +04:00
ISP_UNLOCK(isp);
isp->isp_osinfo._link.scsipi_scsi.adapter_target = defid;
Update the driver with some infrastructure for the 1080. Fix an embarrassing clock botch bug. Additional infrastructure for PDB change stuff.
1999-03-17 09:15:47 +03:00
}
Do SCSI Bus resets in this layer (now). Don't do it for Fibre Channel yet (we get LOOP DOWN events, and we'll hang on that at this time). Add other isp_async cases- ISPASYNC_LOOP_DOWN and ISPASYNC_LOOP_UP. DOWN will cause internal queuing until UP, whereupon a timeout will fire up any pending xfers. It doesn't really keep commands from getting destroyed by loop down events, but at least minimizes the damage. This was much easier to implement with CAM.
1999-02-09 03:42:22 +03:00
Core version 2.0 (platform version 1.0) rewrite of ISP driver. Some interace cleanups, some new common functions. The major impact that will be noticeable right away is that if you boot with not Fibre connected to the FC cards, you no longer hang indefinitely.
2000-08-02 03:55:09 +04:00
/*
* After this point, we'll be doing the new configuration
* schema which allows interrups, so we can do tsleep/wakeup
* for mailbox stuff at that point.
*/
isp->isp_osinfo.no_mbox_ints = 0;
clean up headers; move uninit/watch to outer layers
1998-12-28 22:10:43 +03:00
/*
* And attach children (if any).
*/
NetBSD OS specific routines and definitions.
1998-07-15 23:44:04 +04:00
config_found((void *)isp, &isp->isp_osinfo._link, scsiprint);
Dual LVD (1280) support.
1999-12-16 08:35:42 +03:00
if (IS_DUALBUS(isp)) {
Do a fairly large internal restructuring to accomodate dual-bus host adapters (e.g., the 1240). Include the new 1080/1240 NVRAM layout reading code. Some moderately significant mailbox changes were necessary also to accomodate a second channel.
1999-05-12 22:59:23 +04:00
config_found((void *)isp, &isp->isp_osinfo._link_b, scsiprint);
}
NetBSD OS specific routines and definitions.
1998-07-15 23:44:04 +04:00
}
/*
* minphys our xfers
*
* Unfortunately, the buffer pointer describes the target device- not the
* adapter device, so we can't use the pointer to find out what kind of
* adapter we are and adjust accordingly.
*/
static void
ispminphys(bp)
struct buf *bp;
{
/*
* XX: Only the 1020 has a 24 bit limit.
*/
if (bp->b_bcount >= (1 << 24)) {
bp->b_bcount = (1 << 24);
}
minphys(bp);
}
Add in an ioctl entry point so scsictl mediated bus resets will work. Redo how we start commands- do a 'slow' start function which then looks to see when we're done the configuration process at which point it *then* enables sync/wide mode. Set the max openings amount to the true max openings- not a synthetic. Add a timeout driven command requeue function so that Loop Down events well freeze things until a later point in time where they might be restarted.
1999-10-14 06:31:11 +04:00
static int
ispioctl(sc_link, cmd, addr, flag, p)
struct scsipi_link *sc_link;
u_long cmd;
caddr_t addr;
int flag;
struct proc *p;
{
struct ispsoftc *isp = sc_link->adapter_softc;
int s, chan, retval = ENOTTY;
Remove ispcmd_slow routine- SCBUSACCEL routine will set/clear tag/wide/sync. Make some changes about where some things sit in the softc.
2000-08-09 02:58:30 +04:00
chan = (sc_link->scsipi_scsi.channel == SCSI_CHANNEL_ONLY_ONE)? 0 :
sc_link->scsipi_scsi.channel;
Add in an ioctl entry point so scsictl mediated bus resets will work. Redo how we start commands- do a 'slow' start function which then looks to see when we're done the configuration process at which point it *then* enables sync/wide mode. Set the max openings amount to the true max openings- not a synthetic. Add a timeout driven command requeue function so that Loop Down events well freeze things until a later point in time where they might be restarted.
1999-10-14 06:31:11 +04:00
switch (cmd) {
Remove ispcmd_slow routine- SCBUSACCEL routine will set/clear tag/wide/sync. Make some changes about where some things sit in the softc.
2000-08-09 02:58:30 +04:00
case SCBUSACCEL:
{
struct scbusaccel_args *sp = (struct scbusaccel_args *)addr;
if (IS_SCSI(isp) && sp->sa_lun == 0) {
int dflags = 0;
sdparam *sdp = SDPARAM(isp);
sdp += chan;
if (sp->sa_flags & SC_ACCEL_TAGS)
dflags |= DPARM_TQING;
if (sp->sa_flags & SC_ACCEL_WIDE)
dflags |= DPARM_WIDE;
if (sp->sa_flags & SC_ACCEL_SYNC)
dflags |= DPARM_SYNC;
s = splbio();
sdp->isp_devparam[sp->sa_target].dev_flags |= dflags;
dflags = sdp->isp_devparam[sp->sa_target].dev_flags;
sdp->isp_devparam[sp->sa_target].dev_update = 1;
isp->isp_update |= (1 << chan);
splx(s);
isp_prt(isp, ISP_LOGDEBUG1,
"ispioctl: device flags 0x%x for %d.%d.X",
dflags, chan, sp->sa_target);
}
retval = 0;
break;
}
Add in an ioctl entry point so scsictl mediated bus resets will work. Redo how we start commands- do a 'slow' start function which then looks to see when we're done the configuration process at which point it *then* enables sync/wide mode. Set the max openings amount to the true max openings- not a synthetic. Add a timeout driven command requeue function so that Loop Down events well freeze things until a later point in time where they might be restarted.
1999-10-14 06:31:11 +04:00
case SCBUSIORESET:
s = splbio();
if (isp_control(isp, ISPCTL_RESET_BUS, &chan))
retval = EIO;
else
retval = 0;
(void) splx(s);
break;
default:
break;
}
return (retval);
}
static int32_t
ispcmd(xs)
Core version 2.0 (platform version 1.0) rewrite of ISP driver. Some interace cleanups, some new common functions. The major impact that will be noticeable right away is that if you boot with not Fibre connected to the FC cards, you no longer hang indefinitely.
2000-08-02 03:55:09 +04:00
XS_T *xs;
Add in an ioctl entry point so scsictl mediated bus resets will work. Redo how we start commands- do a 'slow' start function which then looks to see when we're done the configuration process at which point it *then* enables sync/wide mode. Set the max openings amount to the true max openings- not a synthetic. Add a timeout driven command requeue function so that Loop Down events well freeze things until a later point in time where they might be restarted.
1999-10-14 06:31:11 +04:00
{
struct ispsoftc *isp;
int result, s;
isp = XS_ISP(xs);
s = splbio();
Update the driver with some infrastructure for the 1080. Fix an embarrassing clock botch bug. Additional infrastructure for PDB change stuff.
1999-03-17 09:15:47 +03:00
if (isp->isp_state < ISP_RUNSTATE) {
DISABLE_INTS(isp);
isp_init(isp);
if (isp->isp_state != ISP_INITSTATE) {
ENABLE_INTS(isp);
(void) splx(s);
XS_SETERR(xs, HBA_BOTCH);
Add in an ioctl entry point so scsictl mediated bus resets will work. Redo how we start commands- do a 'slow' start function which then looks to see when we're done the configuration process at which point it *then* enables sync/wide mode. Set the max openings amount to the true max openings- not a synthetic. Add a timeout driven command requeue function so that Loop Down events well freeze things until a later point in time where they might be restarted.
1999-10-14 06:31:11 +04:00
return (COMPLETE);
Update the driver with some infrastructure for the 1080. Fix an embarrassing clock botch bug. Additional infrastructure for PDB change stuff.
1999-03-17 09:15:47 +03:00
}
isp->isp_state = ISP_RUNSTATE;
ENABLE_INTS(isp);
}
Do SCSI Bus resets in this layer (now). Don't do it for Fibre Channel yet (we get LOOP DOWN events, and we'll hang on that at this time). Add other isp_async cases- ISPASYNC_LOOP_DOWN and ISPASYNC_LOOP_UP. DOWN will cause internal queuing until UP, whereupon a timeout will fire up any pending xfers. It doesn't really keep commands from getting destroyed by loop down events, but at least minimizes the damage. This was much easier to implement with CAM.
1999-02-09 03:42:22 +03:00
/*
* Check for queue blockage...
*/
if (isp->isp_osinfo.blocked) {
Update for SCSIPI changes. Note that capabilities update is disabled until Matt Jacob has a change to update the driver for the latest firmware, etc. since update device parameters seems to fail once the system is really up and running, and eventually causes the controller to wedge. This may be due to a firmware bug. Per discussion with Matt.
1999-10-01 03:06:18 +04:00
if (xs->xs_control & XS_CTL_POLL) {
Do SCSI Bus resets in this layer (now). Don't do it for Fibre Channel yet (we get LOOP DOWN events, and we'll hang on that at this time). Add other isp_async cases- ISPASYNC_LOOP_DOWN and ISPASYNC_LOOP_UP. DOWN will cause internal queuing until UP, whereupon a timeout will fire up any pending xfers. It doesn't really keep commands from getting destroyed by loop down events, but at least minimizes the damage. This was much easier to implement with CAM.
1999-02-09 03:42:22 +03:00
xs->error = XS_DRIVER_STUFFUP;
splx(s);
return (TRY_AGAIN_LATER);
}
TAILQ_INSERT_TAIL(&isp->isp_osinfo.waitq, xs, adapter_q);
splx(s);
Add in an ioctl entry point so scsictl mediated bus resets will work. Redo how we start commands- do a 'slow' start function which then looks to see when we're done the configuration process at which point it *then* enables sync/wide mode. Set the max openings amount to the true max openings- not a synthetic. Add a timeout driven command requeue function so that Loop Down events well freeze things until a later point in time where they might be restarted.
1999-10-14 06:31:11 +04:00
return (SUCCESSFULLY_QUEUED);
Do SCSI Bus resets in this layer (now). Don't do it for Fibre Channel yet (we get LOOP DOWN events, and we'll hang on that at this time). Add other isp_async cases- ISPASYNC_LOOP_DOWN and ISPASYNC_LOOP_UP. DOWN will cause internal queuing until UP, whereupon a timeout will fire up any pending xfers. It doesn't really keep commands from getting destroyed by loop down events, but at least minimizes the damage. This was much easier to implement with CAM.
1999-02-09 03:42:22 +03:00
}
Redo watchdogs to handle cases of false death of commands- time each command now.. Get rid of SCCLUn stuff. Use an isp_done routine to handle cases of watchdog and isp_done racing to completion.
2000-07-06 02:23:05 +04:00
if (xs->xs_control & XS_CTL_POLL) {
Core version 2.0 (platform version 1.0) rewrite of ISP driver. Some interace cleanups, some new common functions. The major impact that will be noticeable right away is that if you boot with not Fibre connected to the FC cards, you no longer hang indefinitely.
2000-08-02 03:55:09 +04:00
volatile u_int8_t ombi = isp->isp_osinfo.no_mbox_ints;
isp->isp_osinfo.no_mbox_ints = 1;
Redo watchdogs to handle cases of false death of commands- time each command now.. Get rid of SCCLUn stuff. Use an isp_done routine to handle cases of watchdog and isp_done racing to completion.
2000-07-06 02:23:05 +04:00
result = isp_polled_cmd(isp, xs);
Core version 2.0 (platform version 1.0) rewrite of ISP driver. Some interace cleanups, some new common functions. The major impact that will be noticeable right away is that if you boot with not Fibre connected to the FC cards, you no longer hang indefinitely.
2000-08-02 03:55:09 +04:00
isp->isp_osinfo.no_mbox_ints = ombi;
Do SCSI Bus resets in this layer (now). Don't do it for Fibre Channel yet (we get LOOP DOWN events, and we'll hang on that at this time). Add other isp_async cases- ISPASYNC_LOOP_DOWN and ISPASYNC_LOOP_UP. DOWN will cause internal queuing until UP, whereupon a timeout will fire up any pending xfers. It doesn't really keep commands from getting destroyed by loop down events, but at least minimizes the damage. This was much easier to implement with CAM.
1999-02-09 03:42:22 +03:00
(void) splx(s);
add case for going from probetime to runtime speeds
1998-09-08 11:19:58 +04:00
return (result);
NetBSD OS specific routines and definitions.
1998-07-15 23:44:04 +04:00
}
Core version 2.0 (platform version 1.0) rewrite of ISP driver. Some interace cleanups, some new common functions. The major impact that will be noticeable right away is that if you boot with not Fibre connected to the FC cards, you no longer hang indefinitely.
2000-08-02 03:55:09 +04:00
result = isp_start(xs);
Add in commented-out hiwater measurement. Remove egregious older bug which had us refusing to log into fabric devices unless they were NL ports. Whuff.
2000-10-16 09:15:55 +04:00
#if 0
{
static int na[16] = { 0 };
if (na[isp->isp_unit] < isp->isp_nactive) {
isp_prt(isp, ISP_LOGALL, "active hiwater %d", isp->isp_nactive);
na[isp->isp_unit] = isp->isp_nactive;
}
}
#endif
Redo watchdogs to handle cases of false death of commands- time each command now.. Get rid of SCCLUn stuff. Use an isp_done routine to handle cases of watchdog and isp_done racing to completion.
2000-07-06 02:23:05 +04:00
switch (result) {
case CMD_QUEUED:
result = SUCCESSFULLY_QUEUED;
if (xs->timeout) {
callout_reset(&xs->xs_callout, _XT(xs), isp_dog, xs);
}
break;
case CMD_EAGAIN:
result = TRY_AGAIN_LATER;
break;
case CMD_RQLATER:
result = SUCCESSFULLY_QUEUED;
callout_reset(&xs->xs_callout, hz, isp_command_requeue, xs);
break;
case CMD_COMPLETE:
result = COMPLETE;
break;
}
(void) splx(s);
return (result);
}
static int
isp_polled_cmd(isp, xs)
struct ispsoftc *isp;
Core version 2.0 (platform version 1.0) rewrite of ISP driver. Some interace cleanups, some new common functions. The major impact that will be noticeable right away is that if you boot with not Fibre connected to the FC cards, you no longer hang indefinitely.
2000-08-02 03:55:09 +04:00
XS_T *xs;
Redo watchdogs to handle cases of false death of commands- time each command now.. Get rid of SCCLUn stuff. Use an isp_done routine to handle cases of watchdog and isp_done racing to completion.
2000-07-06 02:23:05 +04:00
{
int result;
int infinite = 0, mswait;
Core version 2.0 (platform version 1.0) rewrite of ISP driver. Some interace cleanups, some new common functions. The major impact that will be noticeable right away is that if you boot with not Fibre connected to the FC cards, you no longer hang indefinitely.
2000-08-02 03:55:09 +04:00
result = isp_start(xs);
Redo watchdogs to handle cases of false death of commands- time each command now.. Get rid of SCCLUn stuff. Use an isp_done routine to handle cases of watchdog and isp_done racing to completion.
2000-07-06 02:23:05 +04:00
Add in an ioctl entry point so scsictl mediated bus resets will work. Redo how we start commands- do a 'slow' start function which then looks to see when we're done the configuration process at which point it *then* enables sync/wide mode. Set the max openings amount to the true max openings- not a synthetic. Add a timeout driven command requeue function so that Loop Down events well freeze things until a later point in time where they might be restarted.
1999-10-14 06:31:11 +04:00
switch (result) {
case CMD_QUEUED:
result = SUCCESSFULLY_QUEUED;
break;
case CMD_RQLATER:
case CMD_EAGAIN:
if (XS_NOERR(xs)) {
xs->error = XS_DRIVER_STUFFUP;
}
result = TRY_AGAIN_LATER;
break;
case CMD_COMPLETE:
result = COMPLETE;
break;
}
Redo watchdogs to handle cases of false death of commands- time each command now.. Get rid of SCCLUn stuff. Use an isp_done routine to handle cases of watchdog and isp_done racing to completion.
2000-07-06 02:23:05 +04:00
if (result != SUCCESSFULLY_QUEUED) {
return (result);
}
NetBSD OS specific routines and definitions.
1998-07-15 23:44:04 +04:00
/*
* If we can't use interrupts, poll on completion.
*/
Redo watchdogs to handle cases of false death of commands- time each command now.. Get rid of SCCLUn stuff. Use an isp_done routine to handle cases of watchdog and isp_done racing to completion.
2000-07-06 02:23:05 +04:00
if ((mswait = XS_TIME(xs)) == 0)
infinite = 1;
while (mswait || infinite) {
if (isp_intr((void *)isp)) {
if (XS_CMD_DONE_P(xs)) {
break;
NetBSD OS specific routines and definitions.
1998-07-15 23:44:04 +04:00
}
}
Core version 2.0 (platform version 1.0) rewrite of ISP driver. Some interace cleanups, some new common functions. The major impact that will be noticeable right away is that if you boot with not Fibre connected to the FC cards, you no longer hang indefinitely.
2000-08-02 03:55:09 +04:00
USEC_DELAY(1000);
Redo watchdogs to handle cases of false death of commands- time each command now.. Get rid of SCCLUn stuff. Use an isp_done routine to handle cases of watchdog and isp_done racing to completion.
2000-07-06 02:23:05 +04:00
mswait -= 1;
NetBSD OS specific routines and definitions.
1998-07-15 23:44:04 +04:00
}
Redo watchdogs to handle cases of false death of commands- time each command now.. Get rid of SCCLUn stuff. Use an isp_done routine to handle cases of watchdog and isp_done racing to completion.
2000-07-06 02:23:05 +04:00
/*
* If no other error occurred but we didn't finish,
* something bad happened.
*/
if (XS_CMD_DONE_P(xs) == 0) {
if (isp_control(isp, ISPCTL_ABORT_CMD, xs)) {
Core version 2.0 (platform version 1.0) rewrite of ISP driver. Some interace cleanups, some new common functions. The major impact that will be noticeable right away is that if you boot with not Fibre connected to the FC cards, you no longer hang indefinitely.
2000-08-02 03:55:09 +04:00
isp_reinit(isp);
Redo watchdogs to handle cases of false death of commands- time each command now.. Get rid of SCCLUn stuff. Use an isp_done routine to handle cases of watchdog and isp_done racing to completion.
2000-07-06 02:23:05 +04:00
}
if (XS_NOERR(xs)) {
XS_SETERR(xs, HBA_BOTCH);
}
}
result = COMPLETE;
Add in an ioctl entry point so scsictl mediated bus resets will work. Redo how we start commands- do a 'slow' start function which then looks to see when we're done the configuration process at which point it *then* enables sync/wide mode. Set the max openings amount to the true max openings- not a synthetic. Add a timeout driven command requeue function so that Loop Down events well freeze things until a later point in time where they might be restarted.
1999-10-14 06:31:11 +04:00
return (result);
NetBSD OS specific routines and definitions.
1998-07-15 23:44:04 +04:00
}
Redo watchdogs to handle cases of false death of commands- time each command now.. Get rid of SCCLUn stuff. Use an isp_done routine to handle cases of watchdog and isp_done racing to completion.
2000-07-06 02:23:05 +04:00
void
isp_done(xs)
Core version 2.0 (platform version 1.0) rewrite of ISP driver. Some interace cleanups, some new common functions. The major impact that will be noticeable right away is that if you boot with not Fibre connected to the FC cards, you no longer hang indefinitely.
2000-08-02 03:55:09 +04:00
XS_T *xs;
NetBSD OS specific routines and definitions.
1998-07-15 23:44:04 +04:00
{
Redo watchdogs to handle cases of false death of commands- time each command now.. Get rid of SCCLUn stuff. Use an isp_done routine to handle cases of watchdog and isp_done racing to completion.
2000-07-06 02:23:05 +04:00
XS_CMD_S_DONE(xs);
if (XS_CMD_WDOG_P(xs) == 0) {
struct ispsoftc *isp = XS_ISP(xs);
callout_stop(&xs->xs_callout);
if (XS_CMD_GRACE_P(xs)) {
Core version 2.0 (platform version 1.0) rewrite of ISP driver. Some interace cleanups, some new common functions. The major impact that will be noticeable right away is that if you boot with not Fibre connected to the FC cards, you no longer hang indefinitely.
2000-08-02 03:55:09 +04:00
isp_prt(isp, ISP_LOGDEBUG1,
"finished command on borrowed time");
NetBSD OS specific routines and definitions.
1998-07-15 23:44:04 +04:00
}
Redo watchdogs to handle cases of false death of commands- time each command now.. Get rid of SCCLUn stuff. Use an isp_done routine to handle cases of watchdog and isp_done racing to completion.
2000-07-06 02:23:05 +04:00
XS_CMD_S_CLEAR(xs);
scsipi_done(xs);
NetBSD OS specific routines and definitions.
1998-07-15 23:44:04 +04:00
}
}
clean up headers; move uninit/watch to outer layers
1998-12-28 22:10:43 +03:00
static void
Redo watchdogs to handle cases of false death of commands- time each command now.. Get rid of SCCLUn stuff. Use an isp_done routine to handle cases of watchdog and isp_done racing to completion.
2000-07-06 02:23:05 +04:00
isp_dog(arg)
clean up headers; move uninit/watch to outer layers
1998-12-28 22:10:43 +03:00
void *arg;
{
Core version 2.0 (platform version 1.0) rewrite of ISP driver. Some interace cleanups, some new common functions. The major impact that will be noticeable right away is that if you boot with not Fibre connected to the FC cards, you no longer hang indefinitely.
2000-08-02 03:55:09 +04:00
XS_T *xs = arg;
Redo watchdogs to handle cases of false death of commands- time each command now.. Get rid of SCCLUn stuff. Use an isp_done routine to handle cases of watchdog and isp_done racing to completion.
2000-07-06 02:23:05 +04:00
struct ispsoftc *isp = XS_ISP(xs);
u_int32_t handle;
clean up headers; move uninit/watch to outer layers
1998-12-28 22:10:43 +03:00
Add a maintenance note. Change all splbio's to use the ISP_LOCK/ISP_UNLOCK or ISP_ILOCK/ISP_IUNLOCK macros.
2000-08-14 11:08:12 +04:00
ISP_ILOCK(isp);
clean up headers; move uninit/watch to outer layers
1998-12-28 22:10:43 +03:00
/*
Redo watchdogs to handle cases of false death of commands- time each command now.. Get rid of SCCLUn stuff. Use an isp_done routine to handle cases of watchdog and isp_done racing to completion.
2000-07-06 02:23:05 +04:00
* We've decided this command is dead. Make sure we're not trying
* to kill a command that's already dead by getting it's handle and
* and seeing whether it's still alive.
clean up headers; move uninit/watch to outer layers
1998-12-28 22:10:43 +03:00
*/
Redo watchdogs to handle cases of false death of commands- time each command now.. Get rid of SCCLUn stuff. Use an isp_done routine to handle cases of watchdog and isp_done racing to completion.
2000-07-06 02:23:05 +04:00
handle = isp_find_handle(isp, xs);
if (handle) {
u_int16_t r, r1, i;
if (XS_CMD_DONE_P(xs)) {
Core version 2.0 (platform version 1.0) rewrite of ISP driver. Some interace cleanups, some new common functions. The major impact that will be noticeable right away is that if you boot with not Fibre connected to the FC cards, you no longer hang indefinitely.
2000-08-02 03:55:09 +04:00
isp_prt(isp, ISP_LOGDEBUG1,
"watchdog found done cmd (handle 0x%x)", handle);
Add a maintenance note. Change all splbio's to use the ISP_LOCK/ISP_UNLOCK or ISP_ILOCK/ISP_IUNLOCK macros.
2000-08-14 11:08:12 +04:00
ISP_IUNLOCK(isp);
Redo watchdogs to handle cases of false death of commands- time each command now.. Get rid of SCCLUn stuff. Use an isp_done routine to handle cases of watchdog and isp_done racing to completion.
2000-07-06 02:23:05 +04:00
return;
clean up headers; move uninit/watch to outer layers
1998-12-28 22:10:43 +03:00
}
Redo watchdogs to handle cases of false death of commands- time each command now.. Get rid of SCCLUn stuff. Use an isp_done routine to handle cases of watchdog and isp_done racing to completion.
2000-07-06 02:23:05 +04:00
if (XS_CMD_WDOG_P(xs)) {
Core version 2.0 (platform version 1.0) rewrite of ISP driver. Some interace cleanups, some new common functions. The major impact that will be noticeable right away is that if you boot with not Fibre connected to the FC cards, you no longer hang indefinitely.
2000-08-02 03:55:09 +04:00
isp_prt(isp, ISP_LOGDEBUG1,
"recursive watchdog (handle 0x%x)", handle);
Add a maintenance note. Change all splbio's to use the ISP_LOCK/ISP_UNLOCK or ISP_ILOCK/ISP_IUNLOCK macros.
2000-08-14 11:08:12 +04:00
ISP_IUNLOCK(isp);
Redo watchdogs to handle cases of false death of commands- time each command now.. Get rid of SCCLUn stuff. Use an isp_done routine to handle cases of watchdog and isp_done racing to completion.
2000-07-06 02:23:05 +04:00
return;
clean up headers; move uninit/watch to outer layers
1998-12-28 22:10:43 +03:00
}
Redo watchdogs to handle cases of false death of commands- time each command now.. Get rid of SCCLUn stuff. Use an isp_done routine to handle cases of watchdog and isp_done racing to completion.
2000-07-06 02:23:05 +04:00
XS_CMD_S_WDOG(xs);
i = 0;
do {
r = ISP_READ(isp, BIU_ISR);
Core version 2.0 (platform version 1.0) rewrite of ISP driver. Some interace cleanups, some new common functions. The major impact that will be noticeable right away is that if you boot with not Fibre connected to the FC cards, you no longer hang indefinitely.
2000-08-02 03:55:09 +04:00
USEC_DELAY(1);
Redo watchdogs to handle cases of false death of commands- time each command now.. Get rid of SCCLUn stuff. Use an isp_done routine to handle cases of watchdog and isp_done racing to completion.
2000-07-06 02:23:05 +04:00
r1 = ISP_READ(isp, BIU_ISR);
} while (r != r1 && ++i < 1000);
if (INT_PENDING(isp, r) && isp_intr(isp) && XS_CMD_DONE_P(xs)) {
Core version 2.0 (platform version 1.0) rewrite of ISP driver. Some interace cleanups, some new common functions. The major impact that will be noticeable right away is that if you boot with not Fibre connected to the FC cards, you no longer hang indefinitely.
2000-08-02 03:55:09 +04:00
isp_prt(isp, ISP_LOGDEBUG1, "watchdog cleanup (%x, %x)",
handle, r);
Redo watchdogs to handle cases of false death of commands- time each command now.. Get rid of SCCLUn stuff. Use an isp_done routine to handle cases of watchdog and isp_done racing to completion.
2000-07-06 02:23:05 +04:00
XS_CMD_C_WDOG(xs);
isp_done(xs);
} else if (XS_CMD_GRACE_P(xs)) {
Core version 2.0 (platform version 1.0) rewrite of ISP driver. Some interace cleanups, some new common functions. The major impact that will be noticeable right away is that if you boot with not Fibre connected to the FC cards, you no longer hang indefinitely.
2000-08-02 03:55:09 +04:00
isp_prt(isp, ISP_LOGDEBUG1, "watchdog timeout (%x, %x)",
handle, r);
Redo watchdogs to handle cases of false death of commands- time each command now.. Get rid of SCCLUn stuff. Use an isp_done routine to handle cases of watchdog and isp_done racing to completion.
2000-07-06 02:23:05 +04:00
/*
* Make sure the command is *really* dead before we
* release the handle (and DMA resources) for reuse.
*/
(void) isp_control(isp, ISPCTL_ABORT_CMD, arg);
/*
* After this point, the comamnd is really dead.
*/
if (XS_XFRLEN(xs)) {
ISP_DMAFREE(isp, xs, handle);
}
isp_destroy_handle(isp, handle);
XS_SETERR(xs, XS_TIMEOUT);
XS_CMD_S_CLEAR(xs);
isp_done(xs);
} else {
u_int16_t iptr, optr;
ispreq_t *mp;
Core version 2.0 (platform version 1.0) rewrite of ISP driver. Some interace cleanups, some new common functions. The major impact that will be noticeable right away is that if you boot with not Fibre connected to the FC cards, you no longer hang indefinitely.
2000-08-02 03:55:09 +04:00
isp_prt(isp, ISP_LOGDEBUG2,
"possible command timeout (%x, %x)", handle, r);
Redo watchdogs to handle cases of false death of commands- time each command now.. Get rid of SCCLUn stuff. Use an isp_done routine to handle cases of watchdog and isp_done racing to completion.
2000-07-06 02:23:05 +04:00
XS_CMD_C_WDOG(xs);
callout_reset(&xs->xs_callout, hz, isp_dog, xs);
if (isp_getrqentry(isp, &iptr, &optr, (void **) &mp)) {
Finally fix this driver to be sensible about the ENDIAN dance. It's not quite simply a question of the Qlogic being little endian and having to have stuff swapped on big endian machines- it also has to do with the fact that the SBus and PCI DMA layouts are wierd with respect to this. At any rate, now finally fixed- works on Mac G4, tested it on a SS10 for sparc, checked on alpha to see if I've broken anything, and as soon as I get another spare afternoon I'll finally install a sparc64 version which should just work (as it'll be like the Mac).
2000-12-09 11:06:31 +03:00
ISP_UNLOCK(isp);
Redo watchdogs to handle cases of false death of commands- time each command now.. Get rid of SCCLUn stuff. Use an isp_done routine to handle cases of watchdog and isp_done racing to completion.
2000-07-06 02:23:05 +04:00
return;
}
XS_CMD_S_GRACE(xs);
MEMZERO((void *) mp, sizeof (*mp));
mp->req_header.rqs_entry_count = 1;
mp->req_header.rqs_entry_type = RQSTYPE_MARKER;
mp->req_modifier = SYNC_ALL;
mp->req_target = XS_CHANNEL(xs) << 7;
ISP_SWIZZLE_REQUEST(isp, mp);
ISP_ADD_REQUEST(isp, iptr);
clean up headers; move uninit/watch to outer layers
1998-12-28 22:10:43 +03:00
}
Add a maintenance note. Change all splbio's to use the ISP_LOCK/ISP_UNLOCK or ISP_ILOCK/ISP_IUNLOCK macros.
2000-08-14 11:08:12 +04:00
} else {
isp_prt(isp, ISP_LOGDEBUG0, "watchdog with no command");
clean up headers; move uninit/watch to outer layers
1998-12-28 22:10:43 +03:00
}
Add a maintenance note. Change all splbio's to use the ISP_LOCK/ISP_UNLOCK or ISP_ILOCK/ISP_IUNLOCK macros.
2000-08-14 11:08:12 +04:00
ISP_IUNLOCK(isp);
clean up headers; move uninit/watch to outer layers
1998-12-28 22:10:43 +03:00
}
/*
* Free any associated resources prior to decommissioning and
* set the card to a known state (so it doesn't wake up and kick
* us when we aren't expecting it to).
*
* Locks are held before coming here.
*/
void
isp_uninit(isp)
struct ispsoftc *isp;
{
Core version 2.0 (platform version 1.0) rewrite of ISP driver. Some interace cleanups, some new common functions. The major impact that will be noticeable right away is that if you boot with not Fibre connected to the FC cards, you no longer hang indefinitely.
2000-08-02 03:55:09 +04:00
isp_lock(isp);
clean up headers; move uninit/watch to outer layers
1998-12-28 22:10:43 +03:00
/*
* Leave with interrupts disabled.
*/
DISABLE_INTS(isp);
Core version 2.0 (platform version 1.0) rewrite of ISP driver. Some interace cleanups, some new common functions. The major impact that will be noticeable right away is that if you boot with not Fibre connected to the FC cards, you no longer hang indefinitely.
2000-08-02 03:55:09 +04:00
isp_unlock(isp);
clean up headers; move uninit/watch to outer layers
1998-12-28 22:10:43 +03:00
}
Implement and use Fast Posting for both parallel && fibre. Redo a bit of the startup code. Implement a call to outer framework function so that asynchronous events can be handled (e.g., speed negotiation, target mode). Roll internal release tags.
1999-01-30 10:31:50 +03:00
Add in an ioctl entry point so scsictl mediated bus resets will work. Redo how we start commands- do a 'slow' start function which then looks to see when we're done the configuration process at which point it *then* enables sync/wide mode. Set the max openings amount to the true max openings- not a synthetic. Add a timeout driven command requeue function so that Loop Down events well freeze things until a later point in time where they might be restarted.
1999-10-14 06:31:11 +04:00
/*
* Restart function for a command to be requeued later.
*/
static void
isp_command_requeue(arg)
void *arg;
{
struct scsipi_xfer *xs = arg;
struct ispsoftc *isp = XS_ISP(xs);
Turn ISP_LOCK/ISP_UNLOCK into ISP_ILOCK/ISP_IUNLOCK macros for routines that are driven from a timeout (oops, they're on the interrupt stack).
2000-12-29 00:37:04 +03:00
ISP_ILOCK(isp);
Remove ispcmd_slow routine- SCBUSACCEL routine will set/clear tag/wide/sync. Make some changes about where some things sit in the softc.
2000-08-09 02:58:30 +04:00
switch (ispcmd(xs)) {
Add in an ioctl entry point so scsictl mediated bus resets will work. Redo how we start commands- do a 'slow' start function which then looks to see when we're done the configuration process at which point it *then* enables sync/wide mode. Set the max openings amount to the true max openings- not a synthetic. Add a timeout driven command requeue function so that Loop Down events well freeze things until a later point in time where they might be restarted.
1999-10-14 06:31:11 +04:00
case SUCCESSFULLY_QUEUED:
Core version 2.0 (platform version 1.0) rewrite of ISP driver. Some interace cleanups, some new common functions. The major impact that will be noticeable right away is that if you boot with not Fibre connected to the FC cards, you no longer hang indefinitely.
2000-08-02 03:55:09 +04:00
isp_prt(isp, ISP_LOGINFO,
"requeued commands for %d.%d", XS_TGT(xs), XS_LUN(xs));
Redo watchdogs to handle cases of false death of commands- time each command now.. Get rid of SCCLUn stuff. Use an isp_done routine to handle cases of watchdog and isp_done racing to completion.
2000-07-06 02:23:05 +04:00
if (xs->timeout) {
callout_reset(&xs->xs_callout, _XT(xs), isp_dog, xs);
}
Add in an ioctl entry point so scsictl mediated bus resets will work. Redo how we start commands- do a 'slow' start function which then looks to see when we're done the configuration process at which point it *then* enables sync/wide mode. Set the max openings amount to the true max openings- not a synthetic. Add a timeout driven command requeue function so that Loop Down events well freeze things until a later point in time where they might be restarted.
1999-10-14 06:31:11 +04:00
break;
case TRY_AGAIN_LATER:
Core version 2.0 (platform version 1.0) rewrite of ISP driver. Some interace cleanups, some new common functions. The major impact that will be noticeable right away is that if you boot with not Fibre connected to the FC cards, you no longer hang indefinitely.
2000-08-02 03:55:09 +04:00
isp_prt(isp, ISP_LOGINFO,
"EAGAIN on requeue for %d.%d", XS_TGT(xs), XS_LUN(xs));
callout_reset(&xs->xs_callout, hz, isp_command_requeue, xs);
break;
Add in an ioctl entry point so scsictl mediated bus resets will work. Redo how we start commands- do a 'slow' start function which then looks to see when we're done the configuration process at which point it *then* enables sync/wide mode. Set the max openings amount to the true max openings- not a synthetic. Add a timeout driven command requeue function so that Loop Down events well freeze things until a later point in time where they might be restarted.
1999-10-14 06:31:11 +04:00
case COMPLETE:
some comments added
1999-10-17 05:23:21 +04:00
/* can only be an error */
Redo watchdogs to handle cases of false death of commands- time each command now.. Get rid of SCCLUn stuff. Use an isp_done routine to handle cases of watchdog and isp_done racing to completion.
2000-07-06 02:23:05 +04:00
XS_CMD_S_DONE(xs);
callout_stop(&xs->xs_callout);
Add in an ioctl entry point so scsictl mediated bus resets will work. Redo how we start commands- do a 'slow' start function which then looks to see when we're done the configuration process at which point it *then* enables sync/wide mode. Set the max openings amount to the true max openings- not a synthetic. Add a timeout driven command requeue function so that Loop Down events well freeze things until a later point in time where they might be restarted.
1999-10-14 06:31:11 +04:00
if (XS_NOERR(xs)) {
XS_SETERR(xs, HBA_BOTCH);
}
scsipi_done(xs);
break;
}
Turn ISP_LOCK/ISP_UNLOCK into ISP_ILOCK/ISP_IUNLOCK macros for routines that are driven from a timeout (oops, they're on the interrupt stack).
2000-12-29 00:37:04 +03:00
ISP_IUNLOCK(isp);
Add in an ioctl entry point so scsictl mediated bus resets will work. Redo how we start commands- do a 'slow' start function which then looks to see when we're done the configuration process at which point it *then* enables sync/wide mode. Set the max openings amount to the true max openings- not a synthetic. Add a timeout driven command requeue function so that Loop Down events well freeze things until a later point in time where they might be restarted.
1999-10-14 06:31:11 +04:00
}
some comments added
1999-10-17 05:23:21 +04:00
Do SCSI Bus resets in this layer (now). Don't do it for Fibre Channel yet (we get LOOP DOWN events, and we'll hang on that at this time). Add other isp_async cases- ISPASYNC_LOOP_DOWN and ISPASYNC_LOOP_UP. DOWN will cause internal queuing until UP, whereupon a timeout will fire up any pending xfers. It doesn't really keep commands from getting destroyed by loop down events, but at least minimizes the damage. This was much easier to implement with CAM.
1999-02-09 03:42:22 +03:00
/*
* Restart function after a LOOP UP event (e.g.),
* done as a timeout for some hysteresis.
*/
static void
isp_internal_restart(arg)
void *arg;
{
struct ispsoftc *isp = arg;
Add a maintenance note. Change all splbio's to use the ISP_LOCK/ISP_UNLOCK or ISP_ILOCK/ISP_IUNLOCK macros.
2000-08-14 11:08:12 +04:00
int result, nrestarted = 0;
Do SCSI Bus resets in this layer (now). Don't do it for Fibre Channel yet (we get LOOP DOWN events, and we'll hang on that at this time). Add other isp_async cases- ISPASYNC_LOOP_DOWN and ISPASYNC_LOOP_UP. DOWN will cause internal queuing until UP, whereupon a timeout will fire up any pending xfers. It doesn't really keep commands from getting destroyed by loop down events, but at least minimizes the damage. This was much easier to implement with CAM.
1999-02-09 03:42:22 +03:00
Turn ISP_LOCK/ISP_UNLOCK into ISP_ILOCK/ISP_IUNLOCK macros for routines that are driven from a timeout (oops, they're on the interrupt stack).
2000-12-29 00:37:04 +03:00
ISP_ILOCK(isp);
Do SCSI Bus resets in this layer (now). Don't do it for Fibre Channel yet (we get LOOP DOWN events, and we'll hang on that at this time). Add other isp_async cases- ISPASYNC_LOOP_DOWN and ISPASYNC_LOOP_UP. DOWN will cause internal queuing until UP, whereupon a timeout will fire up any pending xfers. It doesn't really keep commands from getting destroyed by loop down events, but at least minimizes the damage. This was much easier to implement with CAM.
1999-02-09 03:42:22 +03:00
if (isp->isp_osinfo.blocked == 0) {
struct scsipi_xfer *xs;
while ((xs = TAILQ_FIRST(&isp->isp_osinfo.waitq)) != NULL) {
TAILQ_REMOVE(&isp->isp_osinfo.waitq, xs, adapter_q);
Core version 2.0 (platform version 1.0) rewrite of ISP driver. Some interace cleanups, some new common functions. The major impact that will be noticeable right away is that if you boot with not Fibre connected to the FC cards, you no longer hang indefinitely.
2000-08-02 03:55:09 +04:00
result = isp_start(xs);
Update the driver with some infrastructure for the 1080. Fix an embarrassing clock botch bug. Additional infrastructure for PDB change stuff.
1999-03-17 09:15:47 +03:00
if (result != CMD_QUEUED) {
Core version 2.0 (platform version 1.0) rewrite of ISP driver. Some interace cleanups, some new common functions. The major impact that will be noticeable right away is that if you boot with not Fibre connected to the FC cards, you no longer hang indefinitely.
2000-08-02 03:55:09 +04:00
isp_prt(isp, ISP_LOGERR,
"botched command restart (err=%d)", result);
Redo watchdogs to handle cases of false death of commands- time each command now.. Get rid of SCCLUn stuff. Use an isp_done routine to handle cases of watchdog and isp_done racing to completion.
2000-07-06 02:23:05 +04:00
XS_CMD_S_DONE(xs);
Do SCSI Bus resets in this layer (now). Don't do it for Fibre Channel yet (we get LOOP DOWN events, and we'll hang on that at this time). Add other isp_async cases- ISPASYNC_LOOP_DOWN and ISPASYNC_LOOP_UP. DOWN will cause internal queuing until UP, whereupon a timeout will fire up any pending xfers. It doesn't really keep commands from getting destroyed by loop down events, but at least minimizes the damage. This was much easier to implement with CAM.
1999-02-09 03:42:22 +03:00
if (xs->error == XS_NOERROR)
xs->error = XS_DRIVER_STUFFUP;
Redo watchdogs to handle cases of false death of commands- time each command now.. Get rid of SCCLUn stuff. Use an isp_done routine to handle cases of watchdog and isp_done racing to completion.
2000-07-06 02:23:05 +04:00
callout_stop(&xs->xs_callout);
Do SCSI Bus resets in this layer (now). Don't do it for Fibre Channel yet (we get LOOP DOWN events, and we'll hang on that at this time). Add other isp_async cases- ISPASYNC_LOOP_DOWN and ISPASYNC_LOOP_UP. DOWN will cause internal queuing until UP, whereupon a timeout will fire up any pending xfers. It doesn't really keep commands from getting destroyed by loop down events, but at least minimizes the damage. This was much easier to implement with CAM.
1999-02-09 03:42:22 +03:00
scsipi_done(xs);
Redo watchdogs to handle cases of false death of commands- time each command now.. Get rid of SCCLUn stuff. Use an isp_done routine to handle cases of watchdog and isp_done racing to completion.
2000-07-06 02:23:05 +04:00
} else if (xs->timeout) {
callout_reset(&xs->xs_callout,
_XT(xs), isp_dog, xs);
Do SCSI Bus resets in this layer (now). Don't do it for Fibre Channel yet (we get LOOP DOWN events, and we'll hang on that at this time). Add other isp_async cases- ISPASYNC_LOOP_DOWN and ISPASYNC_LOOP_UP. DOWN will cause internal queuing until UP, whereupon a timeout will fire up any pending xfers. It doesn't really keep commands from getting destroyed by loop down events, but at least minimizes the damage. This was much easier to implement with CAM.
1999-02-09 03:42:22 +03:00
}
nrestarted++;
}
Core version 2.0 (platform version 1.0) rewrite of ISP driver. Some interace cleanups, some new common functions. The major impact that will be noticeable right away is that if you boot with not Fibre connected to the FC cards, you no longer hang indefinitely.
2000-08-02 03:55:09 +04:00
isp_prt(isp, ISP_LOGINFO,
"isp_restart requeued %d commands", nrestarted);
Do SCSI Bus resets in this layer (now). Don't do it for Fibre Channel yet (we get LOOP DOWN events, and we'll hang on that at this time). Add other isp_async cases- ISPASYNC_LOOP_DOWN and ISPASYNC_LOOP_UP. DOWN will cause internal queuing until UP, whereupon a timeout will fire up any pending xfers. It doesn't really keep commands from getting destroyed by loop down events, but at least minimizes the damage. This was much easier to implement with CAM.
1999-02-09 03:42:22 +03:00
}
Turn ISP_LOCK/ISP_UNLOCK into ISP_ILOCK/ISP_IUNLOCK macros for routines that are driven from a timeout (oops, they're on the interrupt stack).
2000-12-29 00:37:04 +03:00
ISP_IUNLOCK(isp);
Do SCSI Bus resets in this layer (now). Don't do it for Fibre Channel yet (we get LOOP DOWN events, and we'll hang on that at this time). Add other isp_async cases- ISPASYNC_LOOP_DOWN and ISPASYNC_LOOP_UP. DOWN will cause internal queuing until UP, whereupon a timeout will fire up any pending xfers. It doesn't really keep commands from getting destroyed by loop down events, but at least minimizes the damage. This was much easier to implement with CAM.
1999-02-09 03:42:22 +03:00
}
some comments added
1999-10-17 05:23:21 +04:00
Implement and use Fast Posting for both parallel && fibre. Redo a bit of the startup code. Implement a call to outer framework function so that asynchronous events can be handled (e.g., speed negotiation, target mode). Roll internal release tags.
1999-01-30 10:31:50 +03:00
int
isp_async(isp, cmd, arg)
struct ispsoftc *isp;
ispasync_t cmd;
void *arg;
{
Do a fairly large internal restructuring to accomodate dual-bus host adapters (e.g., the 1240). Include the new 1080/1240 NVRAM layout reading code. Some moderately significant mailbox changes were necessary also to accomodate a second channel.
1999-05-12 22:59:23 +04:00
int bus, tgt;
Do SCSI Bus resets in this layer (now). Don't do it for Fibre Channel yet (we get LOOP DOWN events, and we'll hang on that at this time). Add other isp_async cases- ISPASYNC_LOOP_DOWN and ISPASYNC_LOOP_UP. DOWN will cause internal queuing until UP, whereupon a timeout will fire up any pending xfers. It doesn't really keep commands from getting destroyed by loop down events, but at least minimizes the damage. This was much easier to implement with CAM.
1999-02-09 03:42:22 +03:00
int s = splbio();
Implement and use Fast Posting for both parallel && fibre. Redo a bit of the startup code. Implement a call to outer framework function so that asynchronous events can be handled (e.g., speed negotiation, target mode). Roll internal release tags.
1999-01-30 10:31:50 +03:00
switch (cmd) {
case ISPASYNC_NEW_TGT_PARAMS:
Add in an ioctl entry point so scsictl mediated bus resets will work. Redo how we start commands- do a 'slow' start function which then looks to see when we're done the configuration process at which point it *then* enables sync/wide mode. Set the max openings amount to the true max openings- not a synthetic. Add a timeout driven command requeue function so that Loop Down events well freeze things until a later point in time where they might be restarted.
1999-10-14 06:31:11 +04:00
if (IS_SCSI(isp) && isp->isp_dblev) {
sdparam *sdp = isp->isp_param;
char *wt;
int mhz, flags, period;
tgt = *((int *) arg);
bus = (tgt >> 16) & 0xffff;
tgt &= 0xffff;
Dual LVD (1280) support.
1999-12-16 08:35:42 +03:00
sdp += bus;
Add in an ioctl entry point so scsictl mediated bus resets will work. Redo how we start commands- do a 'slow' start function which then looks to see when we're done the configuration process at which point it *then* enables sync/wide mode. Set the max openings amount to the true max openings- not a synthetic. Add a timeout driven command requeue function so that Loop Down events well freeze things until a later point in time where they might be restarted.
1999-10-14 06:31:11 +04:00
flags = sdp->isp_devparam[tgt].cur_dflags;
period = sdp->isp_devparam[tgt].cur_period;
Dual LVD (1280) support.
1999-12-16 08:35:42 +03:00
Add in an ioctl entry point so scsictl mediated bus resets will work. Redo how we start commands- do a 'slow' start function which then looks to see when we're done the configuration process at which point it *then* enables sync/wide mode. Set the max openings amount to the true max openings- not a synthetic. Add a timeout driven command requeue function so that Loop Down events well freeze things until a later point in time where they might be restarted.
1999-10-14 06:31:11 +04:00
if ((flags & DPARM_SYNC) && period &&
(sdp->isp_devparam[tgt].cur_offset) != 0) {
Dual LVD (1280) support.
1999-12-16 08:35:42 +03:00
/*
* There's some ambiguity about our negotiated speed
* if we haven't detected LVD mode correctly (which
* seems to happen, unfortunately). If we're in LVD
* mode, then different rules apply about speed.
*/
if (sdp->isp_lvdmode || period < 0xc) {
Add in an ioctl entry point so scsictl mediated bus resets will work. Redo how we start commands- do a 'slow' start function which then looks to see when we're done the configuration process at which point it *then* enables sync/wide mode. Set the max openings amount to the true max openings- not a synthetic. Add a timeout driven command requeue function so that Loop Down events well freeze things until a later point in time where they might be restarted.
1999-10-14 06:31:11 +04:00
switch (period) {
add 80 MHz case
2000-02-12 05:25:28 +03:00
case 0x9:
mhz = 80;
break;
Add in an ioctl entry point so scsictl mediated bus resets will work. Redo how we start commands- do a 'slow' start function which then looks to see when we're done the configuration process at which point it *then* enables sync/wide mode. Set the max openings amount to the true max openings- not a synthetic. Add a timeout driven command requeue function so that Loop Down events well freeze things until a later point in time where they might be restarted.
1999-10-14 06:31:11 +04:00
case 0xa:
mhz = 40;
break;
case 0xb:
mhz = 33;
break;
case 0xc:
mhz = 25;
break;
default:
add isp1080 support and some basic PDB change stuff
1999-03-27 01:39:44 +03:00
mhz = 1000 / (period * 4);
Add in an ioctl entry point so scsictl mediated bus resets will work. Redo how we start commands- do a 'slow' start function which then looks to see when we're done the configuration process at which point it *then* enables sync/wide mode. Set the max openings amount to the true max openings- not a synthetic. Add a timeout driven command requeue function so that Loop Down events well freeze things until a later point in time where they might be restarted.
1999-10-14 06:31:11 +04:00
break;
add isp1080 support and some basic PDB change stuff
1999-03-27 01:39:44 +03:00
}
Implement and use Fast Posting for both parallel && fibre. Redo a bit of the startup code. Implement a call to outer framework function so that asynchronous events can be handled (e.g., speed negotiation, target mode). Roll internal release tags.
1999-01-30 10:31:50 +03:00
} else {
Add in an ioctl entry point so scsictl mediated bus resets will work. Redo how we start commands- do a 'slow' start function which then looks to see when we're done the configuration process at which point it *then* enables sync/wide mode. Set the max openings amount to the true max openings- not a synthetic. Add a timeout driven command requeue function so that Loop Down events well freeze things until a later point in time where they might be restarted.
1999-10-14 06:31:11 +04:00
mhz = 1000 / (period * 4);
Implement and use Fast Posting for both parallel && fibre. Redo a bit of the startup code. Implement a call to outer framework function so that asynchronous events can be handled (e.g., speed negotiation, target mode). Roll internal release tags.
1999-01-30 10:31:50 +03:00
}
Add in an ioctl entry point so scsictl mediated bus resets will work. Redo how we start commands- do a 'slow' start function which then looks to see when we're done the configuration process at which point it *then* enables sync/wide mode. Set the max openings amount to the true max openings- not a synthetic. Add a timeout driven command requeue function so that Loop Down events well freeze things until a later point in time where they might be restarted.
1999-10-14 06:31:11 +04:00
} else {
mhz = 0;
}
switch (flags & (DPARM_WIDE|DPARM_TQING)) {
case DPARM_WIDE:
Core version 2.0 (platform version 1.0) rewrite of ISP driver. Some interace cleanups, some new common functions. The major impact that will be noticeable right away is that if you boot with not Fibre connected to the FC cards, you no longer hang indefinitely.
2000-08-02 03:55:09 +04:00
wt = ", 16 bit wide";
Add in an ioctl entry point so scsictl mediated bus resets will work. Redo how we start commands- do a 'slow' start function which then looks to see when we're done the configuration process at which point it *then* enables sync/wide mode. Set the max openings amount to the true max openings- not a synthetic. Add a timeout driven command requeue function so that Loop Down events well freeze things until a later point in time where they might be restarted.
1999-10-14 06:31:11 +04:00
break;
case DPARM_TQING:
Core version 2.0 (platform version 1.0) rewrite of ISP driver. Some interace cleanups, some new common functions. The major impact that will be noticeable right away is that if you boot with not Fibre connected to the FC cards, you no longer hang indefinitely.
2000-08-02 03:55:09 +04:00
wt = ", Tagged Queueing Enabled";
Add in an ioctl entry point so scsictl mediated bus resets will work. Redo how we start commands- do a 'slow' start function which then looks to see when we're done the configuration process at which point it *then* enables sync/wide mode. Set the max openings amount to the true max openings- not a synthetic. Add a timeout driven command requeue function so that Loop Down events well freeze things until a later point in time where they might be restarted.
1999-10-14 06:31:11 +04:00
break;
case DPARM_WIDE|DPARM_TQING:
Core version 2.0 (platform version 1.0) rewrite of ISP driver. Some interace cleanups, some new common functions. The major impact that will be noticeable right away is that if you boot with not Fibre connected to the FC cards, you no longer hang indefinitely.
2000-08-02 03:55:09 +04:00
wt = ", 16 bit wide, Tagged Queueing Enabled";
Add in an ioctl entry point so scsictl mediated bus resets will work. Redo how we start commands- do a 'slow' start function which then looks to see when we're done the configuration process at which point it *then* enables sync/wide mode. Set the max openings amount to the true max openings- not a synthetic. Add a timeout driven command requeue function so that Loop Down events well freeze things until a later point in time where they might be restarted.
1999-10-14 06:31:11 +04:00
break;
default:
Core version 2.0 (platform version 1.0) rewrite of ISP driver. Some interace cleanups, some new common functions. The major impact that will be noticeable right away is that if you boot with not Fibre connected to the FC cards, you no longer hang indefinitely.
2000-08-02 03:55:09 +04:00
wt = " ";
Add in an ioctl entry point so scsictl mediated bus resets will work. Redo how we start commands- do a 'slow' start function which then looks to see when we're done the configuration process at which point it *then* enables sync/wide mode. Set the max openings amount to the true max openings- not a synthetic. Add a timeout driven command requeue function so that Loop Down events well freeze things until a later point in time where they might be restarted.
1999-10-14 06:31:11 +04:00
break;
}
if (mhz) {
Core version 2.0 (platform version 1.0) rewrite of ISP driver. Some interace cleanups, some new common functions. The major impact that will be noticeable right away is that if you boot with not Fibre connected to the FC cards, you no longer hang indefinitely.
2000-08-02 03:55:09 +04:00
isp_prt(isp, ISP_LOGINFO,
"Bus %d Target %d at %dMHz Max Offset %d%s",
bus, tgt, mhz, sdp->isp_devparam[tgt].cur_offset,
wt);
Add in an ioctl entry point so scsictl mediated bus resets will work. Redo how we start commands- do a 'slow' start function which then looks to see when we're done the configuration process at which point it *then* enables sync/wide mode. Set the max openings amount to the true max openings- not a synthetic. Add a timeout driven command requeue function so that Loop Down events well freeze things until a later point in time where they might be restarted.
1999-10-14 06:31:11 +04:00
} else {
Core version 2.0 (platform version 1.0) rewrite of ISP driver. Some interace cleanups, some new common functions. The major impact that will be noticeable right away is that if you boot with not Fibre connected to the FC cards, you no longer hang indefinitely.
2000-08-02 03:55:09 +04:00
isp_prt(isp, ISP_LOGINFO,
"Bus %d Target %d Async Mode%s", bus, tgt, wt);
Implement and use Fast Posting for both parallel && fibre. Redo a bit of the startup code. Implement a call to outer framework function so that asynchronous events can be handled (e.g., speed negotiation, target mode). Roll internal release tags.
1999-01-30 10:31:50 +03:00
}
break;
Add in an ioctl entry point so scsictl mediated bus resets will work. Redo how we start commands- do a 'slow' start function which then looks to see when we're done the configuration process at which point it *then* enables sync/wide mode. Set the max openings amount to the true max openings- not a synthetic. Add a timeout driven command requeue function so that Loop Down events well freeze things until a later point in time where they might be restarted.
1999-10-14 06:31:11 +04:00
}
Update the driver with some infrastructure for the 1080. Fix an embarrassing clock botch bug. Additional infrastructure for PDB change stuff.
1999-03-17 09:15:47 +03:00
case ISPASYNC_BUS_RESET:
Do a fairly large internal restructuring to accomodate dual-bus host adapters (e.g., the 1240). Include the new 1080/1240 NVRAM layout reading code. Some moderately significant mailbox changes were necessary also to accomodate a second channel.
1999-05-12 22:59:23 +04:00
if (arg)
bus = *((int *) arg);
else
bus = 0;
Core version 2.0 (platform version 1.0) rewrite of ISP driver. Some interace cleanups, some new common functions. The major impact that will be noticeable right away is that if you boot with not Fibre connected to the FC cards, you no longer hang indefinitely.
2000-08-02 03:55:09 +04:00
isp_prt(isp, ISP_LOGINFO, "SCSI bus %d reset detected", bus);
Update the driver with some infrastructure for the 1080. Fix an embarrassing clock botch bug. Additional infrastructure for PDB change stuff.
1999-03-17 09:15:47 +03:00
break;
case ISPASYNC_LOOP_DOWN:
/*
* Hopefully we get here in time to minimize the number
* of commands we are firing off that are sure to die.
*/
isp->isp_osinfo.blocked = 1;
Core version 2.0 (platform version 1.0) rewrite of ISP driver. Some interace cleanups, some new common functions. The major impact that will be noticeable right away is that if you boot with not Fibre connected to the FC cards, you no longer hang indefinitely.
2000-08-02 03:55:09 +04:00
isp_prt(isp, ISP_LOGINFO, "Loop DOWN");
Update the driver with some infrastructure for the 1080. Fix an embarrassing clock botch bug. Additional infrastructure for PDB change stuff.
1999-03-17 09:15:47 +03:00
break;
case ISPASYNC_LOOP_UP:
isp->isp_osinfo.blocked = 0;
New callout mechanism with two major improvements over the old timeout()/untimeout() API: - Clients supply callout handle storage, thus eliminating problems of resource allocation. - Insertion and removal of callouts is constant time, important as this facility is used quite a lot in the kernel. The old timeout()/untimeout() API has been removed from the kernel.
2000-03-23 10:01:25 +03:00
callout_reset(&isp->isp_osinfo._restart, 1,
isp_internal_restart, isp);
Core version 2.0 (platform version 1.0) rewrite of ISP driver. Some interace cleanups, some new common functions. The major impact that will be noticeable right away is that if you boot with not Fibre connected to the FC cards, you no longer hang indefinitely.
2000-08-02 03:55:09 +04:00
isp_prt(isp, ISP_LOGINFO, "Loop UP");
Update the driver with some infrastructure for the 1080. Fix an embarrassing clock botch bug. Additional infrastructure for PDB change stuff.
1999-03-17 09:15:47 +03:00
break;
Change to approved NASA/Ames copyright. Add Fabric support. Fix SCCLUN support. Add code that tries to track LoopID shifting.
1999-07-06 00:31:35 +04:00
case ISPASYNC_PDB_CHANGED:
Add in an ioctl entry point so scsictl mediated bus resets will work. Redo how we start commands- do a 'slow' start function which then looks to see when we're done the configuration process at which point it *then* enables sync/wide mode. Set the max openings amount to the true max openings- not a synthetic. Add a timeout driven command requeue function so that Loop Down events well freeze things until a later point in time where they might be restarted.
1999-10-14 06:31:11 +04:00
if (IS_FC(isp) && isp->isp_dblev) {
Apply nearly all of Bill Sommerfeld's -Wformat patches. The one change I didn't take is the %llu format- I can't have a common across multiple platform module assume a %ll argument capability- which really pointed out that I shouldn't be trying to *print* something which could long long.
2000-12-29 01:27:46 +03:00
const char fmt[] = "Target %d (Loop 0x%x) Port ID 0x%x "
Core version 2.0 (platform version 1.0) rewrite of ISP driver. Some interace cleanups, some new common functions. The major impact that will be noticeable right away is that if you boot with not Fibre connected to the FC cards, you no longer hang indefinitely.
2000-08-02 03:55:09 +04:00
"role %s %s\n Port WWN 0x%08x%08x\n Node WWN 0x%08x%08x";
Change to approved NASA/Ames copyright. Add Fabric support. Fix SCCLUN support. Add code that tries to track LoopID shifting.
1999-07-06 00:31:35 +04:00
const static char *roles[4] = {
Update the driver with some infrastructure for the 1080. Fix an embarrassing clock botch bug. Additional infrastructure for PDB change stuff.
1999-03-17 09:15:47 +03:00
"No", "Target", "Initiator", "Target/Initiator"
};
Change to approved NASA/Ames copyright. Add Fabric support. Fix SCCLUN support. Add code that tries to track LoopID shifting.
1999-07-06 00:31:35 +04:00
char *ptr;
fcparam *fcp = isp->isp_param;
int tgt = *((int *) arg);
struct lportdb *lp = &fcp->portdb[tgt];
if (lp->valid) {
ptr = "arrived";
} else {
ptr = "disappeared";
Update the driver with some infrastructure for the 1080. Fix an embarrassing clock botch bug. Additional infrastructure for PDB change stuff.
1999-03-17 09:15:47 +03:00
}
Core version 2.0 (platform version 1.0) rewrite of ISP driver. Some interace cleanups, some new common functions. The major impact that will be noticeable right away is that if you boot with not Fibre connected to the FC cards, you no longer hang indefinitely.
2000-08-02 03:55:09 +04:00
isp_prt(isp, ISP_LOGINFO, fmt, tgt, lp->loopid, lp->portid,
Change to approved NASA/Ames copyright. Add Fabric support. Fix SCCLUN support. Add code that tries to track LoopID shifting.
1999-07-06 00:31:35 +04:00
roles[lp->roles & 0x3], ptr,
(u_int32_t) (lp->port_wwn >> 32),
(u_int32_t) (lp->port_wwn & 0xffffffffLL),
(u_int32_t) (lp->node_wwn >> 32),
(u_int32_t) (lp->node_wwn & 0xffffffffLL));
Update the driver with some infrastructure for the 1080. Fix an embarrassing clock botch bug. Additional infrastructure for PDB change stuff.
1999-03-17 09:15:47 +03:00
break;
}
Change to approved NASA/Ames copyright. Add Fabric support. Fix SCCLUN support. Add code that tries to track LoopID shifting.
1999-07-06 00:31:35 +04:00
#ifdef ISP2100_FABRIC
Update the driver with some infrastructure for the 1080. Fix an embarrassing clock botch bug. Additional infrastructure for PDB change stuff.
1999-03-17 09:15:47 +03:00
case ISPASYNC_CHANGE_NOTIFY:
Core version 2.0 (platform version 1.0) rewrite of ISP driver. Some interace cleanups, some new common functions. The major impact that will be noticeable right away is that if you boot with not Fibre connected to the FC cards, you no longer hang indefinitely.
2000-08-02 03:55:09 +04:00
isp_prt(isp, ISP_LOGINFO, "Name Server Database Changed");
Update the driver with some infrastructure for the 1080. Fix an embarrassing clock botch bug. Additional infrastructure for PDB change stuff.
1999-03-17 09:15:47 +03:00
break;
Change to approved NASA/Ames copyright. Add Fabric support. Fix SCCLUN support. Add code that tries to track LoopID shifting.
1999-07-06 00:31:35 +04:00
case ISPASYNC_FABRIC_DEV:
{
int target;
struct lportdb *lp;
sns_scrsp_t *resp = (sns_scrsp_t *) arg;
u_int32_t portid;
u_int64_t wwn;
fcparam *fcp = isp->isp_param;
portid =
(((u_int32_t) resp->snscb_port_id[0]) << 16) |
(((u_int32_t) resp->snscb_port_id[1]) << 8) |
(((u_int32_t) resp->snscb_port_id[2]));
wwn =
(((u_int64_t)resp->snscb_portname[0]) << 56) |
(((u_int64_t)resp->snscb_portname[1]) << 48) |
(((u_int64_t)resp->snscb_portname[2]) << 40) |
(((u_int64_t)resp->snscb_portname[3]) << 32) |
(((u_int64_t)resp->snscb_portname[4]) << 24) |
(((u_int64_t)resp->snscb_portname[5]) << 16) |
(((u_int64_t)resp->snscb_portname[6]) << 8) |
(((u_int64_t)resp->snscb_portname[7]));
Add in commented-out hiwater measurement. Remove egregious older bug which had us refusing to log into fabric devices unless they were NL ports. Whuff.
2000-10-16 09:15:55 +04:00
Core version 2.0 (platform version 1.0) rewrite of ISP driver. Some interace cleanups, some new common functions. The major impact that will be noticeable right away is that if you boot with not Fibre connected to the FC cards, you no longer hang indefinitely.
2000-08-02 03:55:09 +04:00
isp_prt(isp, ISP_LOGINFO,
"Fabric Device (Type 0x%x)@PortID 0x%x WWN 0x%08x%08x",
resp->snscb_port_type, portid, ((u_int32_t)(wwn >> 32)),
Change to approved NASA/Ames copyright. Add Fabric support. Fix SCCLUN support. Add code that tries to track LoopID shifting.
1999-07-06 00:31:35 +04:00
((u_int32_t)(wwn & 0xffffffff)));
Add in commented-out hiwater measurement. Remove egregious older bug which had us refusing to log into fabric devices unless they were NL ports. Whuff.
2000-10-16 09:15:55 +04:00
Change to approved NASA/Ames copyright. Add Fabric support. Fix SCCLUN support. Add code that tries to track LoopID shifting.
1999-07-06 00:31:35 +04:00
for (target = FC_SNS_ID+1; target < MAX_FC_TARG; target++) {
lp = &fcp->portdb[target];
if (lp->port_wwn == wwn)
break;
}
if (target < MAX_FC_TARG) {
break;
}
for (target = FC_SNS_ID+1; target < MAX_FC_TARG; target++) {
lp = &fcp->portdb[target];
if (lp->port_wwn == 0)
break;
}
if (target == MAX_FC_TARG) {
Core version 2.0 (platform version 1.0) rewrite of ISP driver. Some interace cleanups, some new common functions. The major impact that will be noticeable right away is that if you boot with not Fibre connected to the FC cards, you no longer hang indefinitely.
2000-08-02 03:55:09 +04:00
isp_prt(isp, ISP_LOGWARN,
"no more space for fabric devices");
Change to approved NASA/Ames copyright. Add Fabric support. Fix SCCLUN support. Add code that tries to track LoopID shifting.
1999-07-06 00:31:35 +04:00
return (-1);
}
lp->port_wwn = lp->node_wwn = wwn;
lp->portid = portid;
break;
}
#endif
Implement and use Fast Posting for both parallel && fibre. Redo a bit of the startup code. Implement a call to outer framework function so that asynchronous events can be handled (e.g., speed negotiation, target mode). Roll internal release tags.
1999-01-30 10:31:50 +03:00
default:
break;
}
Do SCSI Bus resets in this layer (now). Don't do it for Fibre Channel yet (we get LOOP DOWN events, and we'll hang on that at this time). Add other isp_async cases- ISPASYNC_LOOP_DOWN and ISPASYNC_LOOP_UP. DOWN will cause internal queuing until UP, whereupon a timeout will fire up any pending xfers. It doesn't really keep commands from getting destroyed by loop down events, but at least minimizes the damage. This was much easier to implement with CAM.
1999-02-09 03:42:22 +03:00
(void) splx(s);
Implement and use Fast Posting for both parallel && fibre. Redo a bit of the startup code. Implement a call to outer framework function so that asynchronous events can be handled (e.g., speed negotiation, target mode). Roll internal release tags.
1999-01-30 10:31:50 +03:00
return (0);
}
Core version 2.0 (platform version 1.0) rewrite of ISP driver. Some interace cleanups, some new common functions. The major impact that will be noticeable right away is that if you boot with not Fibre connected to the FC cards, you no longer hang indefinitely.
2000-08-02 03:55:09 +04:00
#include <machine/stdarg.h>
void
#ifdef __STDC__
isp_prt(struct ispsoftc *isp, int level, const char *fmt, ...)
#else
isp_prt(isp, fmt, va_alist)
struct ispsoftc *isp;
char *fmt;
va_dcl;
#endif
{
va_list ap;
if (level != ISP_LOGALL && (level & isp->isp_dblev) == 0) {
return;
}
printf("%s: ", isp->isp_name);
va_start(ap, fmt);
vprintf(fmt, ap);
va_end(ap);
printf("\n");
}