d/freebsd-nq
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
0b220aa2e8
freebsd-nq/sys/dev/asr/asr.c

3826 lines
109 KiB
C
Raw Normal View History

Use __FBSDID(). Also some minor style cleanups.
2003-08-24 17:55:58 +00:00
/*-
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
* Copyright (c) 1996-2000 Distributed Processing Technology Corporation
Fresh code drop from the vendor. This fixes some critical bugs and adds support for a new class of controllers. Also adjust MAINTAINER at the approval of msmith. Obtained from: mark_salyzyn@adaptec.com
2001-08-21 19:58:16 +00:00
* Copyright (c) 2000-2001 Adaptec Corporation
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
* All rights reserved.
*
* TERMS AND CONDITIONS OF USE
*
* Redistribution and use in source form, with or without modification, are
* permitted provided that redistributions of source code must retain the
* above copyright notice, this list of conditions and the following disclaimer.
*
* This software is provided `as is' by Adaptec 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 Adaptec 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 interruptions) 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 driver software, even
* if advised of the possibility of such damage.
*
* SCSI I2O host adapter driver
*
Update the version history in the driver to note everything that has happened over the last three weeks.
2004-05-05 14:19:43 +00:00
* V1.10 2004/05/05 scottl@freebsd.org
* - Massive cleanup of the driver to remove dead code and
* non-conformant style.
* - Removed most i386-specific code to make it more portable.
* - Converted to the bus_space API.
unexpand -a everything
2002-05-14 01:09:03 +00:00
* V1.08 2001/08/21 Mark_Salyzyn@adaptec.com
* - The 2000S and 2005S do not initialize on some machines,
Fresh code drop from the vendor. This fixes some critical bugs and adds support for a new class of controllers. Also adjust MAINTAINER at the approval of msmith. Obtained from: mark_salyzyn@adaptec.com
2001-08-21 19:58:16 +00:00
* increased timeout to 255ms from 50ms for the StatusGet
* command.
unexpand -a everything
2002-05-14 01:09:03 +00:00
* V1.07 2001/05/22 Mark_Salyzyn@adaptec.com
* - I knew this one was too good to be true. The error return
* on ioctl commands needs to be compared to CAM_REQ_CMP, not
* to the bit masked status.
* V1.06 2001/05/08 Mark_Salyzyn@adaptec.com
* - The 2005S that was supported is affectionately called the
* Conjoined BAR Firmware. In order to support RAID-5 in a
* 16MB low-cost configuration, Firmware was forced to go
* to a Split BAR Firmware. This requires a separate IOP and
* Messaging base address.
* V1.05 2001/04/25 Mark_Salyzyn@adaptec.com
* - Handle support for 2005S Zero Channel RAID solution.
* - System locked up if the Adapter locked up. Do not try
* to send other commands if the resetIOP command fails. The
* fail outstanding command discovery loop was flawed as the
* removal of the command from the list prevented discovering
* all the commands.
* - Comment changes to clarify driver.
* - SysInfo searched for an EATA SmartROM, not an I2O SmartROM.
* - We do not use the AC_FOUND_DEV event because of I2O.
* Removed asr_async.
* V1.04 2000/09/22 Mark_Salyzyn@adaptec.com, msmith@freebsd.org,
* lampa@fee.vutbr.cz and Scott_Long@adaptec.com.
* - Removed support for PM1554, PM2554 and PM2654 in Mode-0
* mode as this is confused with competitor adapters in run
* mode.
* - critical locking needed in ASR_ccbAdd and ASR_ccbRemove
* to prevent operating system panic.
* - moved default major number to 154 from 97.
* V1.03 2000/07/12 Mark_Salyzyn@adaptec.com
* - The controller is not actually an ASR (Adaptec SCSI RAID)
* series that is visible, it's more of an internal code name.
* remove any visible references within reason for now.
* - bus_ptr->LUN was not correctly zeroed when initially
* allocated causing a possible panic of the operating system
* during boot.
* V1.02 2000/06/26 Mark_Salyzyn@adaptec.com
* - Code always fails for ASR_getTid affecting performance.
* - initiated a set of changes that resulted from a formal
* code inspection by Mark_Salyzyn@adaptec.com,
* George_Dake@adaptec.com, Jeff_Zeak@adaptec.com,
* Martin_Wilson@adaptec.com and Vincent_Trandoan@adaptec.com.
* Their findings were focussed on the LCT & TID handler, and
* all resulting changes were to improve code readability,
* consistency or have a positive effect on performance.
* V1.01 2000/06/14 Mark_Salyzyn@adaptec.com
* - Passthrough returned an incorrect error.
* - Passthrough did not migrate the intrinsic scsi layer wakeup
* on command completion.
* - generate control device nodes using make_dev and delete_dev.
* - Performance affected by TID caching reallocing.
* - Made suggested changes by Justin_Gibbs@adaptec.com
* - use splcam instead of splbio.
* - use cam_imask instead of bio_imask.
* - use u_int8_t instead of u_char.
* - use u_int16_t instead of u_short.
* - use u_int32_t instead of u_long where appropriate.
* - use 64 bit context handler instead of 32 bit.
* - create_ccb should only allocate the worst case
* requirements for the driver since CAM may evolve
* making union ccb much larger than needed here.
* renamed create_ccb to asr_alloc_ccb.
* - go nutz justifying all debug prints as macros
* defined at the top and remove unsightly ifdefs.
* - INLINE STATIC viewed as confusing. Historically
* utilized to affect code performance and debug
* issues in OS, Compiler or OEM specific situations.
* V1.00 2000/05/31 Mark_Salyzyn@adaptec.com
* - Ported from FreeBSD 2.2.X DPT I2O driver.
* changed struct scsi_xfer to union ccb/struct ccb_hdr
* changed variable name xs to ccb
* changed struct scsi_link to struct cam_path
* changed struct scsibus_data to struct cam_sim
* stopped using fordriver for holding on to the TID
* use proprietary packet creation instead of scsi_inquire
* CAM layer sends synchronize commands.
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
*/
Use __FBSDID(). Also some minor style cleanups.
2003-08-24 17:55:58 +00:00
#include <sys/cdefs.h>
Many more style cleanups. Switch complex macros to being inline functions. Put @includes in a better spot. Fix many cases of 2 space indents and spaces between a function name and the parens. Use KASSERT instead of a home-rolled ASSERT. Remove some undeeded caddr casts.
2004-05-01 02:27:06 +00:00
#include <sys/param.h> /* TRUE=1 and FALSE=0 defined here */
#include <sys/kernel.h>
Add missing <sys/module.h> includes
2004-05-30 20:08:47 +00:00
#include <sys/module.h>
Many more style cleanups. Switch complex macros to being inline functions. Put @includes in a better spot. Fix many cases of 2 space indents and spaces between a function name and the parens. Use KASSERT instead of a home-rolled ASSERT. Remove some undeeded caddr casts.
2004-05-01 02:27:06 +00:00
#include <sys/systm.h>
#include <sys/malloc.h>
#include <sys/conf.h>
#include <sys/ioccom.h>
Sweep kernel replacing suser(9) calls with priv(9) calls, assigning specific privilege names to a broad range of privileges. These may require some future tweaking. Sponsored by: nCircle Network Security, Inc. Obtained from: TrustedBSD Project Discussed on: arch@ Reviewed (at least in part) by: mlaier, jmg, pjd, bde, ceri, Alex Lyashkov <umka at sevcity dot net>, Skip Ford <skip dot ford at verizon dot net>, Antoine Brodin <antoine dot brodin at laposte dot net>
2006-11-06 13:42:10 +00:00
#include <sys/priv.h>
Many more style cleanups. Switch complex macros to being inline functions. Put @includes in a better spot. Fix many cases of 2 space indents and spaces between a function name and the parens. Use KASSERT instead of a home-rolled ASSERT. Remove some undeeded caddr casts.
2004-05-01 02:27:06 +00:00
#include <sys/proc.h>
#include <sys/bus.h>
#include <machine/resource.h>
#include <machine/bus.h>
#include <sys/rman.h>
#include <sys/stat.h>
Add the start of busdma infrastructure to this driver. Convert the ASR_resetIOPR function to use it.
2006-02-04 08:45:19 +00:00
#include <sys/bus_dma.h>
Many more style cleanups. Switch complex macros to being inline functions. Put @includes in a better spot. Fix many cases of 2 space indents and spaces between a function name and the parens. Use KASSERT instead of a home-rolled ASSERT. Remove some undeeded caddr casts.
2004-05-01 02:27:06 +00:00
#include <cam/cam.h>
#include <cam/cam_ccb.h>
#include <cam/cam_sim.h>
#include <cam/cam_xpt_sim.h>
#include <cam/cam_xpt_periph.h>
#include <cam/scsi/scsi_all.h>
#include <cam/scsi/scsi_message.h>
#include <vm/vm.h>
#include <vm/pmap.h>
#if defined(__i386__)
Bring back some of the ioctl junk that was removed in rev 1.59 as a i386-only kernel option, ASR_COMPAT, and under BURN_BRIDGES. It is really ugly, but raidutils depends on it. Discussed with: scottl
2005-03-17 01:20:49 +00:00
#include "opt_asr.h"
Many more style cleanups. Switch complex macros to being inline functions. Put @includes in a better spot. Fix many cases of 2 space indents and spaces between a function name and the parens. Use KASSERT instead of a home-rolled ASSERT. Remove some undeeded caddr casts.
2004-05-01 02:27:06 +00:00
#include <i386/include/cputypes.h>
Bring back some of the ioctl junk that was removed in rev 1.59 as a i386-only kernel option, ASR_COMPAT, and under BURN_BRIDGES. It is really ugly, but raidutils depends on it. Discussed with: scottl
2005-03-17 01:20:49 +00:00
#if defined(ASR_COMPAT)
#define ASR_IOCTL_COMPAT
#endif /* ASR_COMPAT */
Many more style cleanups. Switch complex macros to being inline functions. Put @includes in a better spot. Fix many cases of 2 space indents and spaces between a function name and the parens. Use KASSERT instead of a home-rolled ASSERT. Remove some undeeded caddr casts.
2004-05-01 02:27:06 +00:00
#endif
#include <machine/vmparam.h>
#include <dev/pci/pcivar.h>
#include <dev/pci/pcireg.h>
#define osdSwap4(x) ((u_long)ntohl((u_long)(x)))
#define KVTOPHYS(x) vtophys(x)
Add the start of busdma infrastructure to this driver. Convert the ASR_resetIOPR function to use it.
2006-02-04 08:45:19 +00:00
#include <dev/asr/dptalign.h>
#include <dev/asr/i2oexec.h>
#include <dev/asr/i2obscsi.h>
#include <dev/asr/i2odpt.h>
#include <dev/asr/i2oadptr.h>
Many more style cleanups. Switch complex macros to being inline functions. Put @includes in a better spot. Fix many cases of 2 space indents and spaces between a function name and the parens. Use KASSERT instead of a home-rolled ASSERT. Remove some undeeded caddr casts.
2004-05-01 02:27:06 +00:00
Add the start of busdma infrastructure to this driver. Convert the ASR_resetIOPR function to use it.
2006-02-04 08:45:19 +00:00
#include <dev/asr/sys_info.h>
Many more style cleanups. Switch complex macros to being inline functions. Put @includes in a better spot. Fix many cases of 2 space indents and spaces between a function name and the parens. Use KASSERT instead of a home-rolled ASSERT. Remove some undeeded caddr casts.
2004-05-01 02:27:06 +00:00
Use __FBSDID(). Also some minor style cleanups.
2003-08-24 17:55:58 +00:00
__FBSDID("$FreeBSD$");
`unexpand -a' should be run _before_ sed 's/^#define /#define^I/g'.
2002-05-14 21:59:10 +00:00
#define ASR_VERSION 1
Update the version history in the driver to note everything that has happened over the last three weeks.
2004-05-05 14:19:43 +00:00
#define ASR_REVISION '1'
#define ASR_SUBREVISION '0'
#define ASR_MONTH 5
#define ASR_DAY 5
#define ASR_YEAR (2004 - 1980)
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
/*
unexpand -a everything
2002-05-14 01:09:03 +00:00
* Debug macros to reduce the unsightly ifdefs
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
*/
#if (defined(DEBUG_ASR) || defined(DEBUG_ASR_USR_CMD) || defined(DEBUG_ASR_CMD))
Many more style cleanups. Switch complex macros to being inline functions. Put @includes in a better spot. Fix many cases of 2 space indents and spaces between a function name and the parens. Use KASSERT instead of a home-rolled ASSERT. Remove some undeeded caddr casts.
2004-05-01 02:27:06 +00:00
static __inline void
debug_asr_message(PI2O_MESSAGE_FRAME message)
{
u_int32_t * pointer = (u_int32_t *)message;
u_int32_t length = I2O_MESSAGE_FRAME_getMessageSize(message);
u_int32_t counter = 0;
while (length--) {
printf("%08lx%c", (u_long)*(pointer++),
(((++counter & 7) == 0) || (length == 0)) ? '\n' : ' ');
unexpand -a everything
2002-05-14 01:09:03 +00:00
}
Many more style cleanups. Switch complex macros to being inline functions. Put @includes in a better spot. Fix many cases of 2 space indents and spaces between a function name and the parens. Use KASSERT instead of a home-rolled ASSERT. Remove some undeeded caddr casts.
2004-05-01 02:27:06 +00:00
}
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
#endif /* DEBUG_ASR || DEBUG_ASR_USR_CMD || DEBUG_ASR_CMD */
Many more style cleanups. Switch complex macros to being inline functions. Put @includes in a better spot. Fix many cases of 2 space indents and spaces between a function name and the parens. Use KASSERT instead of a home-rolled ASSERT. Remove some undeeded caddr casts.
2004-05-01 02:27:06 +00:00
#ifdef DEBUG_ASR
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
/* Breaks on none STDC based compilers :-( */
Many more style cleanups. Switch complex macros to being inline functions. Put @includes in a better spot. Fix many cases of 2 space indents and spaces between a function name and the parens. Use KASSERT instead of a home-rolled ASSERT. Remove some undeeded caddr casts.
2004-05-01 02:27:06 +00:00
#define debug_asr_printf(fmt,args...) printf(fmt, ##args)
#define debug_asr_dump_message(message) debug_asr_message(message)
#define debug_asr_print_path(ccb) xpt_print_path(ccb->ccb_h.path);
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
#else /* DEBUG_ASR */
Don't indent preprocessor tokens.
2004-04-21 20:56:08 +00:00
#define debug_asr_printf(fmt,args...)
#define debug_asr_dump_message(message)
#define debug_asr_print_path(ccb)
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
#endif /* DEBUG_ASR */
/*
unexpand -a everything
2002-05-14 01:09:03 +00:00
* If DEBUG_ASR_CMD is defined:
* 0 - Display incoming SCSI commands
* 1 - add in a quick character before queueing.
* 2 - add in outgoing message frames.
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
*/
#if (defined(DEBUG_ASR_CMD))
Don't indent preprocessor tokens.
2004-04-21 20:56:08 +00:00
#define debug_asr_cmd_printf(fmt,args...) printf(fmt,##args)
Many more style cleanups. Switch complex macros to being inline functions. Put @includes in a better spot. Fix many cases of 2 space indents and spaces between a function name and the parens. Use KASSERT instead of a home-rolled ASSERT. Remove some undeeded caddr casts.
2004-05-01 02:27:06 +00:00
static __inline void
debug_asr_dump_ccb(union ccb *ccb)
{
u_int8_t *cp = (unsigned char *)&(ccb->csio.cdb_io);
int len = ccb->csio.cdb_len;
while (len) {
debug_asr_cmd_printf (" %02x", *(cp++));
--len;
unexpand -a everything
2002-05-14 01:09:03 +00:00
}
Many more style cleanups. Switch complex macros to being inline functions. Put @includes in a better spot. Fix many cases of 2 space indents and spaces between a function name and the parens. Use KASSERT instead of a home-rolled ASSERT. Remove some undeeded caddr casts.
2004-05-01 02:27:06 +00:00
}
Don't indent preprocessor tokens.
2004-04-21 20:56:08 +00:00
#if (DEBUG_ASR_CMD > 0)
#define debug_asr_cmd1_printf debug_asr_cmd_printf
#else
#define debug_asr_cmd1_printf(fmt,args...)
#endif
#if (DEBUG_ASR_CMD > 1)
Many more style cleanups. Switch complex macros to being inline functions. Put @includes in a better spot. Fix many cases of 2 space indents and spaces between a function name and the parens. Use KASSERT instead of a home-rolled ASSERT. Remove some undeeded caddr casts.
2004-05-01 02:27:06 +00:00
#define debug_asr_cmd2_printf debug_asr_cmd_printf
#define debug_asr_cmd2_dump_message(message) debug_asr_message(message)
Don't indent preprocessor tokens.
2004-04-21 20:56:08 +00:00
#else
#define debug_asr_cmd2_printf(fmt,args...)
#define debug_asr_cmd2_dump_message(message)
#endif
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
#else /* DEBUG_ASR_CMD */
Don't indent preprocessor tokens.
2004-04-21 20:56:08 +00:00
#define debug_asr_cmd_printf(fmt,args...)
Many more style cleanups. Switch complex macros to being inline functions. Put @includes in a better spot. Fix many cases of 2 space indents and spaces between a function name and the parens. Use KASSERT instead of a home-rolled ASSERT. Remove some undeeded caddr casts.
2004-05-01 02:27:06 +00:00
#define debug_asr_dump_ccb(ccb)
Don't indent preprocessor tokens.
2004-04-21 20:56:08 +00:00
#define debug_asr_cmd1_printf(fmt,args...)
#define debug_asr_cmd2_printf(fmt,args...)
#define debug_asr_cmd2_dump_message(message)
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
#endif /* DEBUG_ASR_CMD */
#if (defined(DEBUG_ASR_USR_CMD))
Don't indent preprocessor tokens.
2004-04-21 20:56:08 +00:00
#define debug_usr_cmd_printf(fmt,args...) printf(fmt,##args)
#define debug_usr_cmd_dump_message(message) debug_usr_message(message)
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
#else /* DEBUG_ASR_USR_CMD */
Don't indent preprocessor tokens.
2004-04-21 20:56:08 +00:00
#define debug_usr_cmd_printf(fmt,args...)
#define debug_usr_cmd_dump_message(message)
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
#endif /* DEBUG_ASR_USR_CMD */
Bring back some of the ioctl junk that was removed in rev 1.59 as a i386-only kernel option, ASR_COMPAT, and under BURN_BRIDGES. It is really ugly, but raidutils depends on it. Discussed with: scottl
2005-03-17 01:20:49 +00:00
#ifdef ASR_IOCTL_COMPAT
#define dsDescription_size 46 /* Snug as a bug in a rug */
#endif /* ASR_IOCTL_COMPAT */
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
#include "dev/asr/dptsig.h"
static dpt_sig_S ASR_sig = {
unexpand -a everything
2002-05-14 01:09:03 +00:00
{ 'd', 'P', 't', 'S', 'i', 'G'}, SIG_VERSION, PROC_INTEL,
PROC_386 | PROC_486 | PROC_PENTIUM | PROC_SEXIUM, FT_HBADRVR, 0,
Roto-till the ioctl code. Remove all the groping around in the x86 BIOS segment, remove the groping around in the Option ROM segments, remove the bogus tests for bcopy vs. copyout. There really is no reason for a management app to know these things other than to create l33t info tables for the user.
2004-05-05 05:29:19 +00:00
OEM_DPT, OS_FREE_BSD, CAP_ABOVE16MB, DEV_ALL, ADF_ALL_SC5,
unexpand -a everything
2002-05-14 01:09:03 +00:00
0, 0, ASR_VERSION, ASR_REVISION, ASR_SUBREVISION,
ASR_MONTH, ASR_DAY, ASR_YEAR,
/* 01234567890123456789012345678901234567890123456789 < 50 chars */
"Adaptec FreeBSD 4.0.0 Unix SCSI I2O HBA Driver"
/* ^^^^^ asr_attach alters these to match OS */
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
};
/* Configuration Definitions */
`unexpand -a' should be run _before_ sed 's/^#define /#define^I/g'.
2002-05-14 21:59:10 +00:00
#define SG_SIZE 58 /* Scatter Gather list Size */
#define MAX_TARGET_ID 126 /* Maximum Target ID supported */
#define MAX_LUN 255 /* Maximum LUN Supported */
#define MAX_CHANNEL 7 /* Maximum Channel # Supported by driver */
#define MAX_INBOUND 2000 /* Max CCBs, Also Max Queue Size */
#define MAX_OUTBOUND 256 /* Maximum outbound frames/adapter */
#define MAX_INBOUND_SIZE 512 /* Maximum inbound frame size */
#define MAX_MAP 4194304L /* Maximum mapping size of IOP */
unexpand -a everything
2002-05-14 01:09:03 +00:00
/* Also serves as the minimum map for */
/* the 2005S zero channel RAID product */
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
/* I2O register set */
Convert the asr driver to use the bus_space API. This does not represent a significant functional change, but it further cleans up the code and brings it closer to being portable. Thanks to Don Bowman for helping to test this.
2004-05-05 03:42:33 +00:00
#define I2O_REG_STATUS 0x30
#define I2O_REG_MASK 0x34
#define I2O_REG_TOFIFO 0x40
#define I2O_REG_FROMFIFO 0x44
#define Mask_InterruptsDisabled 0x08
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
/*
* A MIX of performance and space considerations for TID lookups
*/
typedef u_int16_t tid_t;
typedef struct {
unexpand -a everything
2002-05-14 01:09:03 +00:00
u_int32_t size; /* up to MAX_LUN */
tid_t TID[1];
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
} lun2tid_t;
typedef struct {
unexpand -a everything
2002-05-14 01:09:03 +00:00
u_int32_t size; /* up to MAX_TARGET */
lun2tid_t * LUN[1];
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
} target2lun_t;
/*
unexpand -a everything
2002-05-14 01:09:03 +00:00
* To ensure that we only allocate and use the worst case ccb here, lets
* make our own local ccb union. If asr_alloc_ccb is utilized for another
* ccb type, ensure that you add the additional structures into our local
* ccb union. To ensure strict type checking, we will utilize the local
* ccb definition wherever possible.
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
*/
union asr_ccb {
unexpand -a everything
2002-05-14 01:09:03 +00:00
struct ccb_hdr ccb_h; /* For convenience */
struct ccb_scsiio csio;
struct ccb_setasync csa;
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
};
Squash another use of vtophys. Instead of creating separate busdma objects for doing static memory transfers, start collecting them into a single object.
2006-02-04 17:56:17 +00:00
struct Asr_status_mem {
I2O_EXEC_STATUS_GET_REPLY status;
U32 rstatus;
};
Convert the asr driver to use the bus_space API. This does not represent a significant functional change, but it further cleans up the code and brings it closer to being portable. Thanks to Don Bowman for helping to test this.
2004-05-05 03:42:33 +00:00
/**************************************************************************
** ASR Host Adapter structure - One Structure For Each Host Adapter That **
** Is Configured Into The System. The Structure Supplies Configuration **
** Information, Status Info, Queue Info And An Active CCB List Pointer. **
***************************************************************************/
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
typedef struct Asr_softc {
Add the start of busdma infrastructure to this driver. Convert the ASR_resetIOPR function to use it.
2006-02-04 08:45:19 +00:00
device_t ha_dev;
unexpand -a everything
2002-05-14 01:09:03 +00:00
u_int16_t ha_irq;
Convert the asr driver to use the bus_space API. This does not represent a significant functional change, but it further cleans up the code and brings it closer to being portable. Thanks to Don Bowman for helping to test this.
2004-05-05 03:42:33 +00:00
u_long ha_Base; /* base port for each board */
bus_size_t ha_blinkLED;
bus_space_handle_t ha_i2o_bhandle;
bus_space_tag_t ha_i2o_btag;
bus_space_handle_t ha_frame_bhandle;
bus_space_tag_t ha_frame_btag;
unexpand -a everything
2002-05-14 01:09:03 +00:00
I2O_IOP_ENTRY ha_SystemTable;
LIST_HEAD(,ccb_hdr) ha_ccb; /* ccbs in use */
Add the start of busdma infrastructure to this driver. Convert the ASR_resetIOPR function to use it.
2006-02-04 08:45:19 +00:00
bus_dma_tag_t ha_parent_dmat;
Squash another use of vtophys. Instead of creating separate busdma objects for doing static memory transfers, start collecting them into a single object.
2006-02-04 17:56:17 +00:00
bus_dma_tag_t ha_statusmem_dmat;
bus_dmamap_t ha_statusmem_dmamap;
struct Asr_status_mem * ha_statusmem;
u_int32_t ha_rstatus_phys;
Add the start of busdma infrastructure to this driver. Convert the ASR_resetIOPR function to use it.
2006-02-04 08:45:19 +00:00
u_int32_t ha_status_phys;
unexpand -a everything
2002-05-14 01:09:03 +00:00
struct cam_path * ha_path[MAX_CHANNEL+1];
struct cam_sim * ha_sim[MAX_CHANNEL+1];
struct resource * ha_mem_res;
struct resource * ha_mes_res;
struct resource * ha_irq_res;
void * ha_intr;
PI2O_LCT ha_LCT; /* Complete list of devices */
Don't indent preprocessor tokens.
2004-04-21 20:56:08 +00:00
#define le_type IdentityTag[0]
#define I2O_BSA 0x20
#define I2O_FCA 0x40
#define I2O_SCSI 0x00
#define I2O_PORT 0x80
#define I2O_UNKNOWN 0x7F
#define le_bus IdentityTag[1]
#define le_target IdentityTag[2]
#define le_lun IdentityTag[3]
unexpand -a everything
2002-05-14 01:09:03 +00:00
target2lun_t * ha_targets[MAX_CHANNEL+1];
PI2O_SCSI_ERROR_REPLY_MESSAGE_FRAME ha_Msgs;
u_long ha_Msgs_Phys;
u_int8_t ha_in_reset;
Don't indent preprocessor tokens.
2004-04-21 20:56:08 +00:00
#define HA_OPERATIONAL 0
#define HA_IN_RESET 1
#define HA_OFF_LINE 2
#define HA_OFF_LINE_RECOVERY 3
unexpand -a everything
2002-05-14 01:09:03 +00:00
/* Configuration information */
/* The target id maximums we take */
u_int8_t ha_MaxBus; /* Maximum bus */
u_int8_t ha_MaxId; /* Maximum target ID */
u_int8_t ha_MaxLun; /* Maximum target LUN */
u_int8_t ha_SgSize; /* Max SG elements */
u_int8_t ha_pciBusNum;
u_int8_t ha_pciDeviceNum;
u_int8_t ha_adapter_target[MAX_CHANNEL+1];
u_int16_t ha_QueueSize; /* Max outstanding commands */
u_int16_t ha_Msgs_Count;
/* Links into other parents and HBAs */
struct Asr_softc * ha_next; /* HBA list */
Do the dreaded s/dev_t/struct cdev */ Bump __FreeBSD_version accordingly.
2004-06-16 09:47:26 +00:00
struct cdev *ha_devt;
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
} Asr_softc_t;
Give the asr softc list global a proper name.
2006-02-12 06:57:41 +00:00
static Asr_softc_t *Asr_softc_list;
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
/*
unexpand -a everything
2002-05-14 01:09:03 +00:00
* Prototypes of the routines we have in this object.
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
*/
/* I2O HDM interface */
Rename driver_t variables. Use device_printf() instead of homerolled printf. Use the provided softc instead of rolling our own.
2006-01-14 15:51:42 +00:00
static int asr_probe(device_t dev);
static int asr_attach(device_t dev);
Get rid of a bunch of useless macros. STATIC becomes static, INLINE becomes __inline where appropriate and gets nuked elsewhere, IN/OUT/INOUT go away. Reformat code affected by this.
2004-05-01 01:25:05 +00:00
Do the dreaded s/dev_t/struct cdev */ Bump __FreeBSD_version accordingly.
2004-06-16 09:47:26 +00:00
static int asr_ioctl(struct cdev *dev, u_long cmd, caddr_t data, int flag,
Get rid of a bunch of useless macros. STATIC becomes static, INLINE becomes __inline where appropriate and gets nuked elsewhere, IN/OUT/INOUT go away. Reformat code affected by this.
2004-05-01 01:25:05 +00:00
struct thread *td);
Do the dreaded s/dev_t/struct cdev */ Bump __FreeBSD_version accordingly.
2004-06-16 09:47:26 +00:00
static int asr_open(struct cdev *dev, int32_t flags, int32_t ifmt,
Get rid of a bunch of useless macros. STATIC becomes static, INLINE becomes __inline where appropriate and gets nuked elsewhere, IN/OUT/INOUT go away. Reformat code affected by this.
2004-05-01 01:25:05 +00:00
struct thread *td);
Do the dreaded s/dev_t/struct cdev */ Bump __FreeBSD_version accordingly.
2004-06-16 09:47:26 +00:00
static int asr_close(struct cdev *dev, int flags, int ifmt, struct thread *td);
Get rid of a bunch of useless macros. STATIC becomes static, INLINE becomes __inline where appropriate and gets nuked elsewhere, IN/OUT/INOUT go away. Reformat code affected by this.
2004-05-01 01:25:05 +00:00
static int asr_intr(Asr_softc_t *sc);
static void asr_timeout(void *arg);
static int ASR_init(Asr_softc_t *sc);
static int ASR_acquireLct(Asr_softc_t *sc);
static int ASR_acquireHrt(Asr_softc_t *sc);
static void asr_action(struct cam_sim *sim, union ccb *ccb);
static void asr_poll(struct cam_sim *sim);
static int ASR_queue(Asr_softc_t *sc, PI2O_MESSAGE_FRAME Message);
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
/*
unexpand -a everything
2002-05-14 01:09:03 +00:00
* Here is the auto-probe structure used to nest our tests appropriately
* during the startup phase of the operating system.
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
*/
Get rid of a bunch of useless macros. STATIC becomes static, INLINE becomes __inline where appropriate and gets nuked elsewhere, IN/OUT/INOUT go away. Reformat code affected by this.
2004-05-01 01:25:05 +00:00
static device_method_t asr_methods[] = {
unexpand -a everything
2002-05-14 01:09:03 +00:00
DEVMETHOD(device_probe, asr_probe),
DEVMETHOD(device_attach, asr_attach),
{ 0, 0 }
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
};
Get rid of a bunch of useless macros. STATIC becomes static, INLINE becomes __inline where appropriate and gets nuked elsewhere, IN/OUT/INOUT go away. Reformat code affected by this.
2004-05-01 01:25:05 +00:00
static driver_t asr_driver = {
unexpand -a everything
2002-05-14 01:09:03 +00:00
"asr",
asr_methods,
sizeof(Asr_softc_t)
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
};
Get rid of a bunch of useless macros. STATIC becomes static, INLINE becomes __inline where appropriate and gets nuked elsewhere, IN/OUT/INOUT go away. Reformat code affected by this.
2004-05-01 01:25:05 +00:00
static devclass_t asr_devclass;
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
DRIVER_MODULE(asr, pci, asr_driver, asr_devclass, 0, 0);
Add MODULE_DEPENDS for cam, pci, mca, eisa and isa where needed. PR: 106543 MFC after: 3 days
2006-12-11 18:28:31 +00:00
MODULE_DEPEND(asr, pci, 1, 1, 1);
MODULE_DEPEND(asr, cam, 1, 1, 1);
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
/*
* devsw for asr hba driver
*
* only ioctl is used. the sd driver provides all other access.
*/
Get rid of a bunch of useless macros. STATIC becomes static, INLINE becomes __inline where appropriate and gets nuked elsewhere, IN/OUT/INOUT go away. Reformat code affected by this.
2004-05-01 01:25:05 +00:00
static struct cdevsw asr_cdevsw = {
Device megapatch 4/6: Introduce d_version field in struct cdevsw, this must always be initialized to D_VERSION. Flip sense of D_NOGIANT flag to D_NEEDGIANT, this involves removing four D_NOGIANT flags and adding 145 D_NEEDGIANT flags.
2004-02-21 21:10:55 +00:00
.d_version = D_VERSION,
.d_flags = D_NEEDGIANT,
Gigacommit to improve device-driver source compatibility between branches: Initialize struct cdevsw using C99 sparse initializtion and remove all initializations to default values. This patch is automatically generated and has been tested by compiling LINT with all the fields in struct cdevsw in reverse order on alpha, sparc64 and i386. Approved by: re(scottl)
2003-03-03 12:15:54 +00:00
.d_open = asr_open,
.d_close = asr_close,
.d_ioctl = asr_ioctl,
.d_name = "asr",
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
};
/* I2O support routines */
Convert the asr driver to use the bus_space API. This does not represent a significant functional change, but it further cleans up the code and brings it closer to being portable. Thanks to Don Bowman for helping to test this.
2004-05-05 03:42:33 +00:00
static __inline u_int32_t
asr_get_FromFIFO(Asr_softc_t *sc)
{
return (bus_space_read_4(sc->ha_i2o_btag, sc->ha_i2o_bhandle,
I2O_REG_FROMFIFO));
}
static __inline u_int32_t
asr_get_ToFIFO(Asr_softc_t *sc)
{
return (bus_space_read_4(sc->ha_i2o_btag, sc->ha_i2o_bhandle,
I2O_REG_TOFIFO));
}
static __inline u_int32_t
asr_get_intr(Asr_softc_t *sc)
{
return (bus_space_read_4(sc->ha_i2o_btag, sc->ha_i2o_bhandle,
I2O_REG_MASK));
}
static __inline u_int32_t
asr_get_status(Asr_softc_t *sc)
{
return (bus_space_read_4(sc->ha_i2o_btag, sc->ha_i2o_bhandle,
I2O_REG_STATUS));
}
static __inline void
asr_set_FromFIFO(Asr_softc_t *sc, u_int32_t val)
{
bus_space_write_4(sc->ha_i2o_btag, sc->ha_i2o_bhandle, I2O_REG_FROMFIFO,
val);
}
static __inline void
asr_set_ToFIFO(Asr_softc_t *sc, u_int32_t val)
{
bus_space_write_4(sc->ha_i2o_btag, sc->ha_i2o_bhandle, I2O_REG_TOFIFO,
val);
}
static __inline void
asr_set_intr(Asr_softc_t *sc, u_int32_t val)
{
bus_space_write_4(sc->ha_i2o_btag, sc->ha_i2o_bhandle, I2O_REG_MASK,
val);
}
static __inline void
asr_set_frame(Asr_softc_t *sc, void *frame, u_int32_t offset, int len)
{
bus_space_write_region_4(sc->ha_frame_btag, sc->ha_frame_bhandle,
offset, (u_int32_t *)frame, len);
}
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
/*
unexpand -a everything
2002-05-14 01:09:03 +00:00
* Fill message with default.
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
*/
Get rid of a bunch of useless macros. STATIC becomes static, INLINE becomes __inline where appropriate and gets nuked elsewhere, IN/OUT/INOUT go away. Reformat code affected by this.
2004-05-01 01:25:05 +00:00
static PI2O_MESSAGE_FRAME
Remove the defAlignLong and getAlignLong macros. I guess that the original intent was to make sure that message structs allocated off of the stack were 4-byte aligned. However, the macros as defined did absolutely nothing. And since I2O forces you to manually copy messages down to the hardware, there really is no point of enforced alignment anyways.
2004-05-02 03:33:18 +00:00
ASR_fillMessage(void *Message, u_int16_t size)
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
{
Get rid of a bunch of useless macros. STATIC becomes static, INLINE becomes __inline where appropriate and gets nuked elsewhere, IN/OUT/INOUT go away. Reformat code affected by this.
2004-05-01 01:25:05 +00:00
PI2O_MESSAGE_FRAME Message_Ptr;
unexpand -a everything
2002-05-14 01:09:03 +00:00
Remove the defAlignLong and getAlignLong macros. I guess that the original intent was to make sure that message structs allocated off of the stack were 4-byte aligned. However, the macros as defined did absolutely nothing. And since I2O forces you to manually copy messages down to the hardware, there really is no point of enforced alignment anyways.
2004-05-02 03:33:18 +00:00
Message_Ptr = (I2O_MESSAGE_FRAME *)Message;
More whitespace style cleanups, also remove unneeded (void *) casts for bzero().
2004-05-01 03:06:54 +00:00
bzero(Message_Ptr, size);
unexpand -a everything
2002-05-14 01:09:03 +00:00
I2O_MESSAGE_FRAME_setVersionOffset(Message_Ptr, I2O_VERSION_11);
I2O_MESSAGE_FRAME_setMessageSize(Message_Ptr,
(size + sizeof(U32) - 1) >> 2);
I2O_MESSAGE_FRAME_setInitiatorAddress (Message_Ptr, 1);
Fix a typo from rev 1.56 that somehow escaped my testing.
2004-05-04 03:16:22 +00:00
KASSERT(Message_Ptr != NULL, ("Message_Ptr == NULL"));
unexpand -a everything
2002-05-14 01:09:03 +00:00
return (Message_Ptr);
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
} /* ASR_fillMessage */
Now that the U32 type is a really 32-bits wide, eliminate a bunch of other bad assumptions and long values.
2006-02-04 08:20:23 +00:00
#define EMPTY_QUEUE (0xffffffff)
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
Get rid of a bunch of useless macros. STATIC becomes static, INLINE becomes __inline where appropriate and gets nuked elsewhere, IN/OUT/INOUT go away. Reformat code affected by this.
2004-05-01 01:25:05 +00:00
static __inline U32
Convert the asr driver to use the bus_space API. This does not represent a significant functional change, but it further cleans up the code and brings it closer to being portable. Thanks to Don Bowman for helping to test this.
2004-05-05 03:42:33 +00:00
ASR_getMessage(Asr_softc_t *sc)
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
{
Get rid of a bunch of useless macros. STATIC becomes static, INLINE becomes __inline where appropriate and gets nuked elsewhere, IN/OUT/INOUT go away. Reformat code affected by this.
2004-05-01 01:25:05 +00:00
U32 MessageOffset;
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
Convert the asr driver to use the bus_space API. This does not represent a significant functional change, but it further cleans up the code and brings it closer to being portable. Thanks to Don Bowman for helping to test this.
2004-05-05 03:42:33 +00:00
MessageOffset = asr_get_ToFIFO(sc);
if (MessageOffset == EMPTY_QUEUE)
MessageOffset = asr_get_ToFIFO(sc);
unexpand -a everything
2002-05-14 01:09:03 +00:00
return (MessageOffset);
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
} /* ASR_getMessage */
/* Issue a polled command */
Get rid of a bunch of useless macros. STATIC becomes static, INLINE becomes __inline where appropriate and gets nuked elsewhere, IN/OUT/INOUT go away. Reformat code affected by this.
2004-05-01 01:25:05 +00:00
static U32
Convert the asr driver to use the bus_space API. This does not represent a significant functional change, but it further cleans up the code and brings it closer to being portable. Thanks to Don Bowman for helping to test this.
2004-05-05 03:42:33 +00:00
ASR_initiateCp(Asr_softc_t *sc, PI2O_MESSAGE_FRAME Message)
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
{
Squash a couple more invalid long casts.
2006-02-05 05:05:27 +00:00
U32 Mask = 0xffffffff;
Get rid of a bunch of useless macros. STATIC becomes static, INLINE becomes __inline where appropriate and gets nuked elsewhere, IN/OUT/INOUT go away. Reformat code affected by this.
2004-05-01 01:25:05 +00:00
U32 MessageOffset;
u_int Delay = 1500;
unexpand -a everything
2002-05-14 01:09:03 +00:00
/*
* ASR_initiateCp is only used for synchronous commands and will
* be made more resiliant to adapter delays since commands like
* resetIOP can cause the adapter to be deaf for a little time.
*/
Convert the asr driver to use the bus_space API. This does not represent a significant functional change, but it further cleans up the code and brings it closer to being portable. Thanks to Don Bowman for helping to test this.
2004-05-05 03:42:33 +00:00
while (((MessageOffset = ASR_getMessage(sc)) == EMPTY_QUEUE)
unexpand -a everything
2002-05-14 01:09:03 +00:00
&& (--Delay != 0)) {
DELAY (10000);
}
if (MessageOffset != EMPTY_QUEUE) {
Convert the asr driver to use the bus_space API. This does not represent a significant functional change, but it further cleans up the code and brings it closer to being portable. Thanks to Don Bowman for helping to test this.
2004-05-05 03:42:33 +00:00
asr_set_frame(sc, Message, MessageOffset,
I2O_MESSAGE_FRAME_getMessageSize(Message));
unexpand -a everything
2002-05-14 01:09:03 +00:00
/*
* Disable the Interrupts
*/
Convert the asr driver to use the bus_space API. This does not represent a significant functional change, but it further cleans up the code and brings it closer to being portable. Thanks to Don Bowman for helping to test this.
2004-05-05 03:42:33 +00:00
Mask = asr_get_intr(sc);
asr_set_intr(sc, Mask | Mask_InterruptsDisabled);
asr_set_ToFIFO(sc, MessageOffset);
unexpand -a everything
2002-05-14 01:09:03 +00:00
}
return (Mask);
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
} /* ASR_initiateCp */
/*
unexpand -a everything
2002-05-14 01:09:03 +00:00
* Reset the adapter.
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
*/
Get rid of a bunch of useless macros. STATIC becomes static, INLINE becomes __inline where appropriate and gets nuked elsewhere, IN/OUT/INOUT go away. Reformat code affected by this.
2004-05-01 01:25:05 +00:00
static U32
Convert the asr driver to use the bus_space API. This does not represent a significant functional change, but it further cleans up the code and brings it closer to being portable. Thanks to Don Bowman for helping to test this.
2004-05-05 03:42:33 +00:00
ASR_resetIOP(Asr_softc_t *sc)
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
{
Squash another use of vtophys. Instead of creating separate busdma objects for doing static memory transfers, start collecting them into a single object.
2006-02-04 17:56:17 +00:00
I2O_EXEC_IOP_RESET_MESSAGE Message;
unexpand -a everything
2002-05-14 01:09:03 +00:00
PI2O_EXEC_IOP_RESET_MESSAGE Message_Ptr;
Add the start of busdma infrastructure to this driver. Convert the ASR_resetIOPR function to use it.
2006-02-04 08:45:19 +00:00
U32 * Reply_Ptr;
unexpand -a everything
2002-05-14 01:09:03 +00:00
U32 Old;
/*
* Build up our copy of the Message.
*/
Remove the defAlignLong and getAlignLong macros. I guess that the original intent was to make sure that message structs allocated off of the stack were 4-byte aligned. However, the macros as defined did absolutely nothing. And since I2O forces you to manually copy messages down to the hardware, there really is no point of enforced alignment anyways.
2004-05-02 03:33:18 +00:00
Message_Ptr = (PI2O_EXEC_IOP_RESET_MESSAGE)ASR_fillMessage(&Message,
unexpand -a everything
2002-05-14 01:09:03 +00:00
sizeof(I2O_EXEC_IOP_RESET_MESSAGE));
I2O_EXEC_IOP_RESET_MESSAGE_setFunction(Message_Ptr, I2O_EXEC_IOP_RESET);
/*
* Reset the Reply Status
*/
Squash another use of vtophys. Instead of creating separate busdma objects for doing static memory transfers, start collecting them into a single object.
2006-02-04 17:56:17 +00:00
Reply_Ptr = &sc->ha_statusmem->rstatus;
Add the start of busdma infrastructure to this driver. Convert the ASR_resetIOPR function to use it.
2006-02-04 08:45:19 +00:00
*Reply_Ptr = 0;
unexpand -a everything
2002-05-14 01:09:03 +00:00
I2O_EXEC_IOP_RESET_MESSAGE_setStatusWordLowAddress(Message_Ptr,
Squash another use of vtophys. Instead of creating separate busdma objects for doing static memory transfers, start collecting them into a single object.
2006-02-04 17:56:17 +00:00
sc->ha_rstatus_phys);
unexpand -a everything
2002-05-14 01:09:03 +00:00
/*
* Send the Message out
*/
Now that the U32 type is a really 32-bits wide, eliminate a bunch of other bad assumptions and long values.
2006-02-04 08:20:23 +00:00
if ((Old = ASR_initiateCp(sc, (PI2O_MESSAGE_FRAME)Message_Ptr)) !=
0xffffffff) {
unexpand -a everything
2002-05-14 01:09:03 +00:00
/*
Many more style cleanups. Switch complex macros to being inline functions. Put @includes in a better spot. Fix many cases of 2 space indents and spaces between a function name and the parens. Use KASSERT instead of a home-rolled ASSERT. Remove some undeeded caddr casts.
2004-05-01 02:27:06 +00:00
* Wait for a response (Poll), timeouts are dangerous if
unexpand -a everything
2002-05-14 01:09:03 +00:00
* the card is truly responsive. We assume response in 2s.
*/
u_int8_t Delay = 200;
while ((*Reply_Ptr == 0) && (--Delay != 0)) {
DELAY (10000);
}
/*
* Re-enable the interrupts.
*/
Convert the asr driver to use the bus_space API. This does not represent a significant functional change, but it further cleans up the code and brings it closer to being portable. Thanks to Don Bowman for helping to test this.
2004-05-05 03:42:33 +00:00
asr_set_intr(sc, Old);
Correctly test *Reply_Ptr in ASR_resetIOP(). Thanks to dhartmei for pointing this out.
2004-05-01 06:32:01 +00:00
KASSERT(*Reply_Ptr != 0, ("*Reply_Ptr == 0"));
Many more style cleanups. Switch complex macros to being inline functions. Put @includes in a better spot. Fix many cases of 2 space indents and spaces between a function name and the parens. Use KASSERT instead of a home-rolled ASSERT. Remove some undeeded caddr casts.
2004-05-01 02:27:06 +00:00
return(*Reply_Ptr);
unexpand -a everything
2002-05-14 01:09:03 +00:00
}
Squash a couple more invalid long casts.
2006-02-05 05:05:27 +00:00
KASSERT(Old != 0xffffffff, ("Old == -1"));
unexpand -a everything
2002-05-14 01:09:03 +00:00
return (0);
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
} /* ASR_resetIOP */
/*
unexpand -a everything
2002-05-14 01:09:03 +00:00
* Get the curent state of the adapter
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
*/
Get rid of a bunch of useless macros. STATIC becomes static, INLINE becomes __inline where appropriate and gets nuked elsewhere, IN/OUT/INOUT go away. Reformat code affected by this.
2004-05-01 01:25:05 +00:00
static PI2O_EXEC_STATUS_GET_REPLY
Squash another use of vtophys. Instead of creating separate busdma objects for doing static memory transfers, start collecting them into a single object.
2006-02-04 17:56:17 +00:00
ASR_getStatus(Asr_softc_t *sc)
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
{
Remove the defAlignLong and getAlignLong macros. I guess that the original intent was to make sure that message structs allocated off of the stack were 4-byte aligned. However, the macros as defined did absolutely nothing. And since I2O forces you to manually copy messages down to the hardware, there really is no point of enforced alignment anyways.
2004-05-02 03:33:18 +00:00
I2O_EXEC_STATUS_GET_MESSAGE Message;
PI2O_EXEC_STATUS_GET_MESSAGE Message_Ptr;
Squash another use of vtophys. Instead of creating separate busdma objects for doing static memory transfers, start collecting them into a single object.
2006-02-04 17:56:17 +00:00
PI2O_EXEC_STATUS_GET_REPLY buffer;
Remove the defAlignLong and getAlignLong macros. I guess that the original intent was to make sure that message structs allocated off of the stack were 4-byte aligned. However, the macros as defined did absolutely nothing. And since I2O forces you to manually copy messages down to the hardware, there really is no point of enforced alignment anyways.
2004-05-02 03:33:18 +00:00
U32 Old;
unexpand -a everything
2002-05-14 01:09:03 +00:00
/*
* Build up our copy of the Message.
*/
Remove the defAlignLong and getAlignLong macros. I guess that the original intent was to make sure that message structs allocated off of the stack were 4-byte aligned. However, the macros as defined did absolutely nothing. And since I2O forces you to manually copy messages down to the hardware, there really is no point of enforced alignment anyways.
2004-05-02 03:33:18 +00:00
Message_Ptr = (PI2O_EXEC_STATUS_GET_MESSAGE)ASR_fillMessage(&Message,
Many more style cleanups. Switch complex macros to being inline functions. Put @includes in a better spot. Fix many cases of 2 space indents and spaces between a function name and the parens. Use KASSERT instead of a home-rolled ASSERT. Remove some undeeded caddr casts.
2004-05-01 02:27:06 +00:00
sizeof(I2O_EXEC_STATUS_GET_MESSAGE));
unexpand -a everything
2002-05-14 01:09:03 +00:00
I2O_EXEC_STATUS_GET_MESSAGE_setFunction(Message_Ptr,
Many more style cleanups. Switch complex macros to being inline functions. Put @includes in a better spot. Fix many cases of 2 space indents and spaces between a function name and the parens. Use KASSERT instead of a home-rolled ASSERT. Remove some undeeded caddr casts.
2004-05-01 02:27:06 +00:00
I2O_EXEC_STATUS_GET);
unexpand -a everything
2002-05-14 01:09:03 +00:00
I2O_EXEC_STATUS_GET_MESSAGE_setReplyBufferAddressLow(Message_Ptr,
Squash another use of vtophys. Instead of creating separate busdma objects for doing static memory transfers, start collecting them into a single object.
2006-02-04 17:56:17 +00:00
sc->ha_status_phys);
unexpand -a everything
2002-05-14 01:09:03 +00:00
/* This one is a Byte Count */
I2O_EXEC_STATUS_GET_MESSAGE_setReplyBufferLength(Message_Ptr,
Many more style cleanups. Switch complex macros to being inline functions. Put @includes in a better spot. Fix many cases of 2 space indents and spaces between a function name and the parens. Use KASSERT instead of a home-rolled ASSERT. Remove some undeeded caddr casts.
2004-05-01 02:27:06 +00:00
sizeof(I2O_EXEC_STATUS_GET_REPLY));
unexpand -a everything
2002-05-14 01:09:03 +00:00
/*
* Reset the Reply Status
*/
Squash another use of vtophys. Instead of creating separate busdma objects for doing static memory transfers, start collecting them into a single object.
2006-02-04 17:56:17 +00:00
buffer = &sc->ha_statusmem->status;
More whitespace style cleanups, also remove unneeded (void *) casts for bzero().
2004-05-01 03:06:54 +00:00
bzero(buffer, sizeof(I2O_EXEC_STATUS_GET_REPLY));
unexpand -a everything
2002-05-14 01:09:03 +00:00
/*
* Send the Message out
*/
Now that the U32 type is a really 32-bits wide, eliminate a bunch of other bad assumptions and long values.
2006-02-04 08:20:23 +00:00
if ((Old = ASR_initiateCp(sc, (PI2O_MESSAGE_FRAME)Message_Ptr)) !=
0xffffffff) {
unexpand -a everything
2002-05-14 01:09:03 +00:00
/*
* Wait for a response (Poll), timeouts are dangerous if
* the card is truly responsive. We assume response in 50ms.
*/
u_int8_t Delay = 255;
while (*((U8 * volatile)&(buffer->SyncByte)) == 0) {
if (--Delay == 0) {
Begin the slow, painful process of cleaning up asr. This eliminates many of the bogus casts of NULL. It also distracts me while I procrastinate on useful work.
2004-04-21 03:51:39 +00:00
buffer = NULL;
unexpand -a everything
2002-05-14 01:09:03 +00:00
break;
}
DELAY (1000);
}
/*
* Re-enable the interrupts.
*/
Convert the asr driver to use the bus_space API. This does not represent a significant functional change, but it further cleans up the code and brings it closer to being portable. Thanks to Don Bowman for helping to test this.
2004-05-05 03:42:33 +00:00
asr_set_intr(sc, Old);
unexpand -a everything
2002-05-14 01:09:03 +00:00
return (buffer);
}
Begin the slow, painful process of cleaning up asr. This eliminates many of the bogus casts of NULL. It also distracts me while I procrastinate on useful work.
2004-04-21 03:51:39 +00:00
return (NULL);
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
} /* ASR_getStatus */
/*
unexpand -a everything
2002-05-14 01:09:03 +00:00
* Check if the device is a SCSI I2O HBA, and add it to the list.
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
*/
/*
* Probe for ASR controller. If we find it, we will use it.
* virtual adapters.
*/
Remove the DOMINO and MODE0 device attachments. They never did anything. Remove a bunch of obfuscating macros.
2004-05-01 05:19:39 +00:00
static int
Rename driver_t variables. Use device_printf() instead of homerolled printf. Use the provided softc instead of rolling our own.
2006-01-14 15:51:42 +00:00
asr_probe(device_t dev)
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
{
Remove the DOMINO and MODE0 device attachments. They never did anything. Remove a bunch of obfuscating macros.
2004-05-01 05:19:39 +00:00
u_int32_t id;
Rename driver_t variables. Use device_printf() instead of homerolled printf. Use the provided softc instead of rolling our own.
2006-01-14 15:51:42 +00:00
id = (pci_get_device(dev) << 16) | pci_get_vendor(dev);
unexpand -a everything
2002-05-14 01:09:03 +00:00
if ((id == 0xA5011044) || (id == 0xA5111044)) {
Rename driver_t variables. Use device_printf() instead of homerolled printf. Use the provided softc instead of rolling our own.
2006-01-14 15:51:42 +00:00
device_set_desc(dev, "Adaptec Caching SCSI RAID");
Use BUS_PROBE_DEFAULT for pci probe return value
2005-03-05 19:06:12 +00:00
return (BUS_PROBE_DEFAULT);
unexpand -a everything
2002-05-14 01:09:03 +00:00
}
Remove the DOMINO and MODE0 device attachments. They never did anything. Remove a bunch of obfuscating macros.
2004-05-01 05:19:39 +00:00
return (ENXIO);
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
} /* asr_probe */
Get rid of a bunch of useless macros. STATIC becomes static, INLINE becomes __inline where appropriate and gets nuked elsewhere, IN/OUT/INOUT go away. Reformat code affected by this.
2004-05-01 01:25:05 +00:00
static __inline union asr_ccb *
asr_alloc_ccb(Asr_softc_t *sc)
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
{
Get rid of a bunch of useless macros. STATIC becomes static, INLINE becomes __inline where appropriate and gets nuked elsewhere, IN/OUT/INOUT go away. Reformat code affected by this.
2004-05-01 01:25:05 +00:00
union asr_ccb *new_ccb;
unexpand -a everything
2002-05-14 01:09:03 +00:00
if ((new_ccb = (union asr_ccb *)malloc(sizeof(*new_ccb),
Begin the slow, painful process of cleaning up asr. This eliminates many of the bogus casts of NULL. It also distracts me while I procrastinate on useful work.
2004-04-21 03:51:39 +00:00
M_DEVBUF, M_WAITOK | M_ZERO)) != NULL) {
unexpand -a everything
2002-05-14 01:09:03 +00:00
new_ccb->ccb_h.pinfo.priority = 1;
new_ccb->ccb_h.pinfo.index = CAM_UNQUEUED_INDEX;
new_ccb->ccb_h.spriv_ptr0 = sc;
}
return (new_ccb);
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
} /* asr_alloc_ccb */
Get rid of a bunch of useless macros. STATIC becomes static, INLINE becomes __inline where appropriate and gets nuked elsewhere, IN/OUT/INOUT go away. Reformat code affected by this.
2004-05-01 01:25:05 +00:00
static __inline void
asr_free_ccb(union asr_ccb *free_ccb)
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
{
unexpand -a everything
2002-05-14 01:09:03 +00:00
free(free_ccb, M_DEVBUF);
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
} /* asr_free_ccb */
/*
unexpand -a everything
2002-05-14 01:09:03 +00:00
* Print inquiry data `carefully'
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
*/
Get rid of a bunch of useless macros. STATIC becomes static, INLINE becomes __inline where appropriate and gets nuked elsewhere, IN/OUT/INOUT go away. Reformat code affected by this.
2004-05-01 01:25:05 +00:00
static void
ASR_prstring(u_int8_t *s, int len)
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
{
unexpand -a everything
2002-05-14 01:09:03 +00:00
while ((--len >= 0) && (*s) && (*s != ' ') && (*s != '-')) {
printf ("%c", *(s++));
}
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
} /* ASR_prstring */
/*
unexpand -a everything
2002-05-14 01:09:03 +00:00
* Send a message synchronously and without Interrupt to a ccb.
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
*/
Get rid of a bunch of useless macros. STATIC becomes static, INLINE becomes __inline where appropriate and gets nuked elsewhere, IN/OUT/INOUT go away. Reformat code affected by this.
2004-05-01 01:25:05 +00:00
static int
ASR_queue_s(union asr_ccb *ccb, PI2O_MESSAGE_FRAME Message)
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
{
Get rid of a bunch of useless macros. STATIC becomes static, INLINE becomes __inline where appropriate and gets nuked elsewhere, IN/OUT/INOUT go away. Reformat code affected by this.
2004-05-01 01:25:05 +00:00
int s;
U32 Mask;
Asr_softc_t *sc = (Asr_softc_t *)(ccb->ccb_h.spriv_ptr0);
unexpand -a everything
2002-05-14 01:09:03 +00:00
/*
* We do not need any (optional byteswapping) method access to
* the Initiator context field.
*/
I2O_MESSAGE_FRAME_setInitiatorContext64(Message, (long)ccb);
/* Prevent interrupt service */
s = splcam ();
Convert the asr driver to use the bus_space API. This does not represent a significant functional change, but it further cleans up the code and brings it closer to being portable. Thanks to Don Bowman for helping to test this.
2004-05-05 03:42:33 +00:00
Mask = asr_get_intr(sc);
asr_set_intr(sc, Mask | Mask_InterruptsDisabled);
unexpand -a everything
2002-05-14 01:09:03 +00:00
Convert the asr driver to use the bus_space API. This does not represent a significant functional change, but it further cleans up the code and brings it closer to being portable. Thanks to Don Bowman for helping to test this.
2004-05-05 03:42:33 +00:00
if (ASR_queue(sc, Message) == EMPTY_QUEUE) {
unexpand -a everything
2002-05-14 01:09:03 +00:00
ccb->ccb_h.status &= ~CAM_STATUS_MASK;
ccb->ccb_h.status |= CAM_REQUEUE_REQ;
}
/*
* Wait for this board to report a finished instruction.
*/
while ((ccb->ccb_h.status & CAM_STATUS_MASK) == CAM_REQ_INPROG) {
(void)asr_intr (sc);
}
/* Re-enable Interrupts */
Convert the asr driver to use the bus_space API. This does not represent a significant functional change, but it further cleans up the code and brings it closer to being portable. Thanks to Don Bowman for helping to test this.
2004-05-05 03:42:33 +00:00
asr_set_intr(sc, Mask);
unexpand -a everything
2002-05-14 01:09:03 +00:00
splx(s);
return (ccb->ccb_h.status);
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
} /* ASR_queue_s */
/*
Correct typos, mostly s/ a / an / where appropriate. Some whitespace cleanup, especially in troff files.
2003-01-01 18:49:04 +00:00
* Send a message synchronously to an Asr_softc_t.
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
*/
Get rid of a bunch of useless macros. STATIC becomes static, INLINE becomes __inline where appropriate and gets nuked elsewhere, IN/OUT/INOUT go away. Reformat code affected by this.
2004-05-01 01:25:05 +00:00
static int
ASR_queue_c(Asr_softc_t *sc, PI2O_MESSAGE_FRAME Message)
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
{
Get rid of a bunch of useless macros. STATIC becomes static, INLINE becomes __inline where appropriate and gets nuked elsewhere, IN/OUT/INOUT go away. Reformat code affected by this.
2004-05-01 01:25:05 +00:00
union asr_ccb *ccb;
int status;
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
Begin the slow, painful process of cleaning up asr. This eliminates many of the bogus casts of NULL. It also distracts me while I procrastinate on useful work.
2004-04-21 03:51:39 +00:00
if ((ccb = asr_alloc_ccb (sc)) == NULL) {
unexpand -a everything
2002-05-14 01:09:03 +00:00
return (CAM_REQUEUE_REQ);
}
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
unexpand -a everything
2002-05-14 01:09:03 +00:00
status = ASR_queue_s (ccb, Message);
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
unexpand -a everything
2002-05-14 01:09:03 +00:00
asr_free_ccb(ccb);
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
unexpand -a everything
2002-05-14 01:09:03 +00:00
return (status);
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
} /* ASR_queue_c */
/*
unexpand -a everything
2002-05-14 01:09:03 +00:00
* Add the specified ccb to the active queue
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
*/
Get rid of a bunch of useless macros. STATIC becomes static, INLINE becomes __inline where appropriate and gets nuked elsewhere, IN/OUT/INOUT go away. Reformat code affected by this.
2004-05-01 01:25:05 +00:00
static __inline void
ASR_ccbAdd(Asr_softc_t *sc, union asr_ccb *ccb)
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
{
unexpand -a everything
2002-05-14 01:09:03 +00:00
int s;
s = splcam();
LIST_INSERT_HEAD(&(sc->ha_ccb), &(ccb->ccb_h), sim_links.le);
if (ccb->ccb_h.timeout != CAM_TIME_INFINITY) {
if (ccb->ccb_h.timeout == CAM_TIME_DEFAULT) {
/*
* RAID systems can take considerable time to
* complete some commands given the large cache
* flashes switching from write back to write thru.
*/
ccb->ccb_h.timeout = 6 * 60 * 1000;
}
ccb->ccb_h.timeout_ch = timeout(asr_timeout, (caddr_t)ccb,
(ccb->ccb_h.timeout * hz) / 1000);
}
splx(s);
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
} /* ASR_ccbAdd */
/*
unexpand -a everything
2002-05-14 01:09:03 +00:00
* Remove the specified ccb from the active queue.
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
*/
Get rid of a bunch of useless macros. STATIC becomes static, INLINE becomes __inline where appropriate and gets nuked elsewhere, IN/OUT/INOUT go away. Reformat code affected by this.
2004-05-01 01:25:05 +00:00
static __inline void
ASR_ccbRemove(Asr_softc_t *sc, union asr_ccb *ccb)
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
{
unexpand -a everything
2002-05-14 01:09:03 +00:00
int s;
Fix a race in ASR_ccbAdd and ASR_ccbRemove, which were both manipulating the ccb queue without masking interrupts. Submitted by: Petr Lampa <lampa@fee.vutbr.cz>
2000-09-21 20:21:30 +00:00
unexpand -a everything
2002-05-14 01:09:03 +00:00
s = splcam();
untimeout(asr_timeout, (caddr_t)ccb, ccb->ccb_h.timeout_ch);
LIST_REMOVE(&(ccb->ccb_h), sim_links.le);
splx(s);
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
} /* ASR_ccbRemove */
/*
unexpand -a everything
2002-05-14 01:09:03 +00:00
* Fail all the active commands, so they get re-issued by the operating
* system.
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
*/
Get rid of a bunch of useless macros. STATIC becomes static, INLINE becomes __inline where appropriate and gets nuked elsewhere, IN/OUT/INOUT go away. Reformat code affected by this.
2004-05-01 01:25:05 +00:00
static void
ASR_failActiveCommands(Asr_softc_t *sc)
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
{
Get rid of a bunch of useless macros. STATIC becomes static, INLINE becomes __inline where appropriate and gets nuked elsewhere, IN/OUT/INOUT go away. Reformat code affected by this.
2004-05-01 01:25:05 +00:00
struct ccb_hdr *ccb;
int s;
Fresh code drop from the vendor. This fixes some critical bugs and adds support for a new class of controllers. Also adjust MAINTAINER at the approval of msmith. Obtained from: mark_salyzyn@adaptec.com
2001-08-21 19:58:16 +00:00
unexpand -a everything
2002-05-14 01:09:03 +00:00
s = splcam();
/*
* We do not need to inform the CAM layer that we had a bus
* reset since we manage it on our own, this also prevents the
* SCSI_DELAY settling that would be required on other systems.
* The `SCSI_DELAY' has already been handled by the card via the
* acquisition of the LCT table while we are at CAM priority level.
* for (int bus = 0; bus <= sc->ha_MaxBus; ++bus) {
* xpt_async (AC_BUS_RESET, sc->ha_path[bus], NULL);
* }
*/
Begin the slow, painful process of cleaning up asr. This eliminates many of the bogus casts of NULL. It also distracts me while I procrastinate on useful work.
2004-04-21 03:51:39 +00:00
while ((ccb = LIST_FIRST(&(sc->ha_ccb))) != NULL) {
unexpand -a everything
2002-05-14 01:09:03 +00:00
ASR_ccbRemove (sc, (union asr_ccb *)ccb);
ccb->status &= ~CAM_STATUS_MASK;
ccb->status |= CAM_REQUEUE_REQ;
/* Nothing Transfered */
((struct ccb_scsiio *)ccb)->resid
= ((struct ccb_scsiio *)ccb)->dxfer_len;
if (ccb->path) {
xpt_done ((union ccb *)ccb);
} else {
More low-hanging fruit: kill caddr_t in calls to wakeup(9) / [mt]sleep(9).
2003-03-02 16:54:40 +00:00
wakeup (ccb);
unexpand -a everything
2002-05-14 01:09:03 +00:00
}
}
splx(s);
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
} /* ASR_failActiveCommands */
/*
unexpand -a everything
2002-05-14 01:09:03 +00:00
* The following command causes the HBA to reset the specific bus
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
*/
Get rid of a bunch of useless macros. STATIC becomes static, INLINE becomes __inline where appropriate and gets nuked elsewhere, IN/OUT/INOUT go away. Reformat code affected by this.
2004-05-01 01:25:05 +00:00
static void
ASR_resetBus(Asr_softc_t *sc, int bus)
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
{
Remove the defAlignLong and getAlignLong macros. I guess that the original intent was to make sure that message structs allocated off of the stack were 4-byte aligned. However, the macros as defined did absolutely nothing. And since I2O forces you to manually copy messages down to the hardware, there really is no point of enforced alignment anyways.
2004-05-02 03:33:18 +00:00
I2O_HBA_BUS_RESET_MESSAGE Message;
I2O_HBA_BUS_RESET_MESSAGE *Message_Ptr;
PI2O_LCT_ENTRY Device;
unexpand -a everything
2002-05-14 01:09:03 +00:00
Remove the defAlignLong and getAlignLong macros. I guess that the original intent was to make sure that message structs allocated off of the stack were 4-byte aligned. However, the macros as defined did absolutely nothing. And since I2O forces you to manually copy messages down to the hardware, there really is no point of enforced alignment anyways.
2004-05-02 03:33:18 +00:00
Message_Ptr = (I2O_HBA_BUS_RESET_MESSAGE *)ASR_fillMessage(&Message,
unexpand -a everything
2002-05-14 01:09:03 +00:00
sizeof(I2O_HBA_BUS_RESET_MESSAGE));
I2O_MESSAGE_FRAME_setFunction(&Message_Ptr->StdMessageFrame,
I2O_HBA_BUS_RESET);
for (Device = sc->ha_LCT->LCTEntry; Device < (PI2O_LCT_ENTRY)
(((U32 *)sc->ha_LCT)+I2O_LCT_getTableSize(sc->ha_LCT));
++Device) {
if (((Device->le_type & I2O_PORT) != 0)
&& (Device->le_bus == bus)) {
I2O_MESSAGE_FRAME_setTargetAddress(
&Message_Ptr->StdMessageFrame,
I2O_LCT_ENTRY_getLocalTID(Device));
/* Asynchronous command, with no expectations */
(void)ASR_queue(sc, (PI2O_MESSAGE_FRAME)Message_Ptr);
break;
}
}
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
} /* ASR_resetBus */
Get rid of a bunch of useless macros. STATIC becomes static, INLINE becomes __inline where appropriate and gets nuked elsewhere, IN/OUT/INOUT go away. Reformat code affected by this.
2004-05-01 01:25:05 +00:00
static __inline int
ASR_getBlinkLedCode(Asr_softc_t *sc)
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
{
Convert the asr driver to use the bus_space API. This does not represent a significant functional change, but it further cleans up the code and brings it closer to being portable. Thanks to Don Bowman for helping to test this.
2004-05-05 03:42:33 +00:00
U8 blink;
if (sc == NULL)
return (0);
blink = bus_space_read_1(sc->ha_frame_btag,
sc->ha_frame_bhandle, sc->ha_blinkLED + 1);
if (blink != 0xBC)
return (0);
blink = bus_space_read_1(sc->ha_frame_btag,
sc->ha_frame_bhandle, sc->ha_blinkLED);
return (blink);
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
} /* ASR_getBlinkCode */
/*
unexpand -a everything
2002-05-14 01:09:03 +00:00
* Determine the address of an TID lookup. Must be done at high priority
* since the address can be changed by other threads of execution.
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
*
unexpand -a everything
2002-05-14 01:09:03 +00:00
* Returns NULL pointer if not indexible (but will attempt to generate
* an index if `new_entry' flag is set to TRUE).
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
*
unexpand -a everything
2002-05-14 01:09:03 +00:00
* All addressible entries are to be guaranteed zero if never initialized.
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
*/
Get rid of a bunch of useless macros. STATIC becomes static, INLINE becomes __inline where appropriate and gets nuked elsewhere, IN/OUT/INOUT go away. Reformat code affected by this.
2004-05-01 01:25:05 +00:00
static tid_t *
ASR_getTidAddress(Asr_softc_t *sc, int bus, int target, int lun, int new_entry)
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
{
Get rid of a bunch of useless macros. STATIC becomes static, INLINE becomes __inline where appropriate and gets nuked elsewhere, IN/OUT/INOUT go away. Reformat code affected by this.
2004-05-01 01:25:05 +00:00
target2lun_t *bus_ptr;
lun2tid_t *target_ptr;
unsigned new_size;
unexpand -a everything
2002-05-14 01:09:03 +00:00
/*
* Validity checking of incoming parameters. More of a bound
* expansion limit than an issue with the code dealing with the
* values.
*
* sc must be valid before it gets here, so that check could be
* dropped if speed a critical issue.
*/
Begin the slow, painful process of cleaning up asr. This eliminates many of the bogus casts of NULL. It also distracts me while I procrastinate on useful work.
2004-04-21 03:51:39 +00:00
if ((sc == NULL)
unexpand -a everything
2002-05-14 01:09:03 +00:00
|| (bus > MAX_CHANNEL)
|| (target > sc->ha_MaxId)
|| (lun > sc->ha_MaxLun)) {
debug_asr_printf("(%lx,%d,%d,%d) target out of range\n",
(u_long)sc, bus, target, lun);
Begin the slow, painful process of cleaning up asr. This eliminates many of the bogus casts of NULL. It also distracts me while I procrastinate on useful work.
2004-04-21 03:51:39 +00:00
return (NULL);
unexpand -a everything
2002-05-14 01:09:03 +00:00
}
/*
* See if there is an associated bus list.
*
* for performance, allocate in size of BUS_CHUNK chunks.
* BUS_CHUNK must be a power of two. This is to reduce
* fragmentation effects on the allocations.
*/
Don't indent preprocessor tokens.
2004-04-21 20:56:08 +00:00
#define BUS_CHUNK 8
unexpand -a everything
2002-05-14 01:09:03 +00:00
new_size = ((target + BUS_CHUNK - 1) & ~(BUS_CHUNK - 1));
Begin the slow, painful process of cleaning up asr. This eliminates many of the bogus casts of NULL. It also distracts me while I procrastinate on useful work.
2004-04-21 03:51:39 +00:00
if ((bus_ptr = sc->ha_targets[bus]) == NULL) {
unexpand -a everything
2002-05-14 01:09:03 +00:00
/*
* Allocate a new structure?
* Since one element in structure, the +1
* needed for size has been abstracted.
*/
if ((new_entry == FALSE)
|| ((sc->ha_targets[bus] = bus_ptr = (target2lun_t *)malloc (
sizeof(*bus_ptr) + (sizeof(bus_ptr->LUN) * new_size),
Back out M_* changes, per decision of the TRB. Approved by: trb
2003-02-19 05:47:46 +00:00
M_TEMP, M_WAITOK | M_ZERO))
Begin the slow, painful process of cleaning up asr. This eliminates many of the bogus casts of NULL. It also distracts me while I procrastinate on useful work.
2004-04-21 03:51:39 +00:00
== NULL)) {
unexpand -a everything
2002-05-14 01:09:03 +00:00
debug_asr_printf("failed to allocate bus list\n");
Begin the slow, painful process of cleaning up asr. This eliminates many of the bogus casts of NULL. It also distracts me while I procrastinate on useful work.
2004-04-21 03:51:39 +00:00
return (NULL);
unexpand -a everything
2002-05-14 01:09:03 +00:00
}
bus_ptr->size = new_size + 1;
} else if (bus_ptr->size <= new_size) {
target2lun_t * new_bus_ptr;
/*
* Reallocate a new structure?
* Since one element in structure, the +1
* needed for size has been abstracted.
*/
if ((new_entry == FALSE)
|| ((new_bus_ptr = (target2lun_t *)malloc (
sizeof(*bus_ptr) + (sizeof(bus_ptr->LUN) * new_size),
Begin the slow, painful process of cleaning up asr. This eliminates many of the bogus casts of NULL. It also distracts me while I procrastinate on useful work.
2004-04-21 03:51:39 +00:00
M_TEMP, M_WAITOK | M_ZERO)) == NULL)) {
unexpand -a everything
2002-05-14 01:09:03 +00:00
debug_asr_printf("failed to reallocate bus list\n");
Begin the slow, painful process of cleaning up asr. This eliminates many of the bogus casts of NULL. It also distracts me while I procrastinate on useful work.
2004-04-21 03:51:39 +00:00
return (NULL);
unexpand -a everything
2002-05-14 01:09:03 +00:00
}
/*
* Copy the whole thing, safer, simpler coding
* and not really performance critical at this point.
*/
Many more style cleanups. Switch complex macros to being inline functions. Put @includes in a better spot. Fix many cases of 2 space indents and spaces between a function name and the parens. Use KASSERT instead of a home-rolled ASSERT. Remove some undeeded caddr casts.
2004-05-01 02:27:06 +00:00
bcopy(bus_ptr, new_bus_ptr, sizeof(*bus_ptr)
+ (sizeof(bus_ptr->LUN) * (bus_ptr->size - 1)));
unexpand -a everything
2002-05-14 01:09:03 +00:00
sc->ha_targets[bus] = new_bus_ptr;
Remove the defAlignLong and getAlignLong macros. I guess that the original intent was to make sure that message structs allocated off of the stack were 4-byte aligned. However, the macros as defined did absolutely nothing. And since I2O forces you to manually copy messages down to the hardware, there really is no point of enforced alignment anyways.
2004-05-02 03:33:18 +00:00
free(bus_ptr, M_TEMP);
unexpand -a everything
2002-05-14 01:09:03 +00:00
bus_ptr = new_bus_ptr;
bus_ptr->size = new_size + 1;
}
/*
* We now have the bus list, lets get to the target list.
* Since most systems have only *one* lun, we do not allocate
* in chunks as above, here we allow one, then in chunk sizes.
* TARGET_CHUNK must be a power of two. This is to reduce
* fragmentation effects on the allocations.
*/
Don't indent preprocessor tokens.
2004-04-21 20:56:08 +00:00
#define TARGET_CHUNK 8
unexpand -a everything
2002-05-14 01:09:03 +00:00
if ((new_size = lun) != 0) {
new_size = ((lun + TARGET_CHUNK - 1) & ~(TARGET_CHUNK - 1));
}
Begin the slow, painful process of cleaning up asr. This eliminates many of the bogus casts of NULL. It also distracts me while I procrastinate on useful work.
2004-04-21 03:51:39 +00:00
if ((target_ptr = bus_ptr->LUN[target]) == NULL) {
unexpand -a everything
2002-05-14 01:09:03 +00:00
/*
* Allocate a new structure?
* Since one element in structure, the +1
* needed for size has been abstracted.
*/
if ((new_entry == FALSE)
|| ((bus_ptr->LUN[target] = target_ptr = (lun2tid_t *)malloc (
sizeof(*target_ptr) + (sizeof(target_ptr->TID) * new_size),
Begin the slow, painful process of cleaning up asr. This eliminates many of the bogus casts of NULL. It also distracts me while I procrastinate on useful work.
2004-04-21 03:51:39 +00:00
M_TEMP, M_WAITOK | M_ZERO)) == NULL)) {
unexpand -a everything
2002-05-14 01:09:03 +00:00
debug_asr_printf("failed to allocate target list\n");
Begin the slow, painful process of cleaning up asr. This eliminates many of the bogus casts of NULL. It also distracts me while I procrastinate on useful work.
2004-04-21 03:51:39 +00:00
return (NULL);
unexpand -a everything
2002-05-14 01:09:03 +00:00
}
target_ptr->size = new_size + 1;
} else if (target_ptr->size <= new_size) {
lun2tid_t * new_target_ptr;
/*
* Reallocate a new structure?
* Since one element in structure, the +1
* needed for size has been abstracted.
*/
if ((new_entry == FALSE)
|| ((new_target_ptr = (lun2tid_t *)malloc (
sizeof(*target_ptr) + (sizeof(target_ptr->TID) * new_size),
Begin the slow, painful process of cleaning up asr. This eliminates many of the bogus casts of NULL. It also distracts me while I procrastinate on useful work.
2004-04-21 03:51:39 +00:00
M_TEMP, M_WAITOK | M_ZERO)) == NULL)) {
unexpand -a everything
2002-05-14 01:09:03 +00:00
debug_asr_printf("failed to reallocate target list\n");
Begin the slow, painful process of cleaning up asr. This eliminates many of the bogus casts of NULL. It also distracts me while I procrastinate on useful work.
2004-04-21 03:51:39 +00:00
return (NULL);
unexpand -a everything
2002-05-14 01:09:03 +00:00
}
/*
* Copy the whole thing, safer, simpler coding
* and not really performance critical at this point.
*/
Many more style cleanups. Switch complex macros to being inline functions. Put @includes in a better spot. Fix many cases of 2 space indents and spaces between a function name and the parens. Use KASSERT instead of a home-rolled ASSERT. Remove some undeeded caddr casts.
2004-05-01 02:27:06 +00:00
bcopy(target_ptr, new_target_ptr, sizeof(*target_ptr)
+ (sizeof(target_ptr->TID) * (target_ptr->size - 1)));
unexpand -a everything
2002-05-14 01:09:03 +00:00
bus_ptr->LUN[target] = new_target_ptr;
Remove the defAlignLong and getAlignLong macros. I guess that the original intent was to make sure that message structs allocated off of the stack were 4-byte aligned. However, the macros as defined did absolutely nothing. And since I2O forces you to manually copy messages down to the hardware, there really is no point of enforced alignment anyways.
2004-05-02 03:33:18 +00:00
free(target_ptr, M_TEMP);
unexpand -a everything
2002-05-14 01:09:03 +00:00
target_ptr = new_target_ptr;
target_ptr->size = new_size + 1;
}
/*
* Now, acquire the TID address from the LUN indexed list.
*/
return (&(target_ptr->TID[lun]));
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
} /* ASR_getTidAddress */
/*
unexpand -a everything
2002-05-14 01:09:03 +00:00
* Get a pre-existing TID relationship.
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
*
unexpand -a everything
2002-05-14 01:09:03 +00:00
* If the TID was never set, return (tid_t)-1.
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
*
unexpand -a everything
2002-05-14 01:09:03 +00:00
* should use mutex rather than spl.
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
*/
Get rid of a bunch of useless macros. STATIC becomes static, INLINE becomes __inline where appropriate and gets nuked elsewhere, IN/OUT/INOUT go away. Reformat code affected by this.
2004-05-01 01:25:05 +00:00
static __inline tid_t
ASR_getTid(Asr_softc_t *sc, int bus, int target, int lun)
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
{
Get rid of a bunch of useless macros. STATIC becomes static, INLINE becomes __inline where appropriate and gets nuked elsewhere, IN/OUT/INOUT go away. Reformat code affected by this.
2004-05-01 01:25:05 +00:00
tid_t *tid_ptr;
int s;
tid_t retval;
unexpand -a everything
2002-05-14 01:09:03 +00:00
s = splcam();
Begin the slow, painful process of cleaning up asr. This eliminates many of the bogus casts of NULL. It also distracts me while I procrastinate on useful work.
2004-04-21 03:51:39 +00:00
if (((tid_ptr = ASR_getTidAddress(sc, bus, target, lun, FALSE)) == NULL)
unexpand -a everything
2002-05-14 01:09:03 +00:00
/* (tid_t)0 or (tid_t)-1 indicate no TID */
|| (*tid_ptr == (tid_t)0)) {
splx(s);
return ((tid_t)-1);
}
retval = *tid_ptr;
splx(s);
return (retval);
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
} /* ASR_getTid */
/*
unexpand -a everything
2002-05-14 01:09:03 +00:00
* Set a TID relationship.
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
*
unexpand -a everything
2002-05-14 01:09:03 +00:00
* If the TID was not set, return (tid_t)-1.
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
*
unexpand -a everything
2002-05-14 01:09:03 +00:00
* should use mutex rather than spl.
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
*/
Get rid of a bunch of useless macros. STATIC becomes static, INLINE becomes __inline where appropriate and gets nuked elsewhere, IN/OUT/INOUT go away. Reformat code affected by this.
2004-05-01 01:25:05 +00:00
static __inline tid_t
ASR_setTid(Asr_softc_t *sc, int bus, int target, int lun, tid_t TID)
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
{
Get rid of a bunch of useless macros. STATIC becomes static, INLINE becomes __inline where appropriate and gets nuked elsewhere, IN/OUT/INOUT go away. Reformat code affected by this.
2004-05-01 01:25:05 +00:00
tid_t *tid_ptr;
int s;
unexpand -a everything
2002-05-14 01:09:03 +00:00
if (TID != (tid_t)-1) {
if (TID == 0) {
return ((tid_t)-1);
}
s = splcam();
Get rid of a bunch of useless macros. STATIC becomes static, INLINE becomes __inline where appropriate and gets nuked elsewhere, IN/OUT/INOUT go away. Reformat code affected by this.
2004-05-01 01:25:05 +00:00
if ((tid_ptr = ASR_getTidAddress(sc, bus, target, lun, TRUE))
Begin the slow, painful process of cleaning up asr. This eliminates many of the bogus casts of NULL. It also distracts me while I procrastinate on useful work.
2004-04-21 03:51:39 +00:00
== NULL) {
unexpand -a everything
2002-05-14 01:09:03 +00:00
splx(s);
return ((tid_t)-1);
}
*tid_ptr = TID;
splx(s);
}
return (TID);
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
} /* ASR_setTid */
/*-------------------------------------------------------------------------*/
unexpand -a everything
2002-05-14 01:09:03 +00:00
/* Function ASR_rescan */
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
/*-------------------------------------------------------------------------*/
unexpand -a everything
2002-05-14 01:09:03 +00:00
/* The Parameters Passed To This Function Are : */
/* Asr_softc_t * : HBA miniport driver's adapter data storage. */
/* */
/* This Function Will rescan the adapter and resynchronize any data */
/* */
/* Return : 0 For OK, Error Code Otherwise */
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
/*-------------------------------------------------------------------------*/
Get rid of a bunch of useless macros. STATIC becomes static, INLINE becomes __inline where appropriate and gets nuked elsewhere, IN/OUT/INOUT go away. Reformat code affected by this.
2004-05-01 01:25:05 +00:00
static int
ASR_rescan(Asr_softc_t *sc)
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
{
Get rid of a bunch of useless macros. STATIC becomes static, INLINE becomes __inline where appropriate and gets nuked elsewhere, IN/OUT/INOUT go away. Reformat code affected by this.
2004-05-01 01:25:05 +00:00
int bus;
int error;
unexpand -a everything
2002-05-14 01:09:03 +00:00
/*
* Re-acquire the LCT table and synchronize us to the adapter.
*/
if ((error = ASR_acquireLct(sc)) == 0) {
error = ASR_acquireHrt(sc);
}
if (error != 0) {
return error;
}
bus = sc->ha_MaxBus;
/* Reset all existing cached TID lookups */
do {
int target, event = 0;
/*
* Scan for all targets on this bus to see if they
* got affected by the rescan.
*/
for (target = 0; target <= sc->ha_MaxId; ++target) {
int lun;
/* Stay away from the controller ID */
if (target == sc->ha_adapter_target[bus]) {
continue;
}
for (lun = 0; lun <= sc->ha_MaxLun; ++lun) {
PI2O_LCT_ENTRY Device;
tid_t TID = (tid_t)-1;
tid_t LastTID;
/*
* See if the cached TID changed. Search for
* the device in our new LCT.
*/
for (Device = sc->ha_LCT->LCTEntry;
Device < (PI2O_LCT_ENTRY)(((U32 *)sc->ha_LCT)
+ I2O_LCT_getTableSize(sc->ha_LCT));
++Device) {
if ((Device->le_type != I2O_UNKNOWN)
&& (Device->le_bus == bus)
&& (Device->le_target == target)
&& (Device->le_lun == lun)
&& (I2O_LCT_ENTRY_getUserTID(Device)
== 0xFFF)) {
TID = I2O_LCT_ENTRY_getLocalTID(
Device);
break;
}
}
/*
* Indicate to the OS that the label needs
* to be recalculated, or that the specific
* open device is no longer valid (Merde)
* because the cached TID changed.
*/
LastTID = ASR_getTid (sc, bus, target, lun);
if (LastTID != TID) {
struct cam_path * path;
if (xpt_create_path(&path,
/*periph*/NULL,
cam_sim_path(sc->ha_sim[bus]),
target, lun) != CAM_REQ_CMP) {
if (TID == (tid_t)-1) {
event |= AC_LOST_DEVICE;
} else {
event |= AC_INQ_CHANGED
| AC_GETDEV_CHANGED;
}
} else {
if (TID == (tid_t)-1) {
xpt_async(
AC_LOST_DEVICE,
path, NULL);
} else if (LastTID == (tid_t)-1) {
struct ccb_getdev ccb;
xpt_setup_ccb(
&(ccb.ccb_h),
path, /*priority*/5);
xpt_async(
AC_FOUND_DEVICE,
path,
&ccb);
} else {
xpt_async(
AC_INQ_CHANGED,
path, NULL);
xpt_async(
AC_GETDEV_CHANGED,
path, NULL);
}
}
}
/*
* We have the option of clearing the
* cached TID for it to be rescanned, or to
* set it now even if the device never got
* accessed. We chose the later since we
* currently do not use the condition that
* the TID ever got cached.
*/
ASR_setTid (sc, bus, target, lun, TID);
}
}
/*
* The xpt layer can not handle multiple events at the
* same call.
*/
if (event & AC_LOST_DEVICE) {
xpt_async(AC_LOST_DEVICE, sc->ha_path[bus], NULL);
}
if (event & AC_INQ_CHANGED) {
xpt_async(AC_INQ_CHANGED, sc->ha_path[bus], NULL);
}
if (event & AC_GETDEV_CHANGED) {
xpt_async(AC_GETDEV_CHANGED, sc->ha_path[bus], NULL);
}
} while (--bus >= 0);
return (error);
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
} /* ASR_rescan */
/*-------------------------------------------------------------------------*/
unexpand -a everything
2002-05-14 01:09:03 +00:00
/* Function ASR_reset */
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
/*-------------------------------------------------------------------------*/
unexpand -a everything
2002-05-14 01:09:03 +00:00
/* The Parameters Passed To This Function Are : */
/* Asr_softc_t * : HBA miniport driver's adapter data storage. */
/* */
/* This Function Will reset the adapter and resynchronize any data */
/* */
/* Return : None */
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
/*-------------------------------------------------------------------------*/
Get rid of a bunch of useless macros. STATIC becomes static, INLINE becomes __inline where appropriate and gets nuked elsewhere, IN/OUT/INOUT go away. Reformat code affected by this.
2004-05-01 01:25:05 +00:00
static int
ASR_reset(Asr_softc_t *sc)
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
{
Get rid of a bunch of useless macros. STATIC becomes static, INLINE becomes __inline where appropriate and gets nuked elsewhere, IN/OUT/INOUT go away. Reformat code affected by this.
2004-05-01 01:25:05 +00:00
int s, retVal;
unexpand -a everything
2002-05-14 01:09:03 +00:00
s = splcam();
if ((sc->ha_in_reset == HA_IN_RESET)
|| (sc->ha_in_reset == HA_OFF_LINE_RECOVERY)) {
splx (s);
return (EBUSY);
}
/*
* Promotes HA_OPERATIONAL to HA_IN_RESET,
* or HA_OFF_LINE to HA_OFF_LINE_RECOVERY.
*/
++(sc->ha_in_reset);
Convert the asr driver to use the bus_space API. This does not represent a significant functional change, but it further cleans up the code and brings it closer to being portable. Thanks to Don Bowman for helping to test this.
2004-05-05 03:42:33 +00:00
if (ASR_resetIOP(sc) == 0) {
unexpand -a everything
2002-05-14 01:09:03 +00:00
debug_asr_printf ("ASR_resetIOP failed\n");
/*
* We really need to take this card off-line, easier said
* than make sense. Better to keep retrying for now since if a
* UART cable is connected the blinkLEDs the adapter is now in
* a hard state requiring action from the monitor commands to
* the HBA to continue. For debugging waiting forever is a
* good thing. In a production system, however, one may wish
* to instead take the card off-line ...
*/
Remove more sead code.
2004-04-21 20:58:48 +00:00
/* Wait Forever */
Convert the asr driver to use the bus_space API. This does not represent a significant functional change, but it further cleans up the code and brings it closer to being portable. Thanks to Don Bowman for helping to test this.
2004-05-05 03:42:33 +00:00
while (ASR_resetIOP(sc) == 0);
unexpand -a everything
2002-05-14 01:09:03 +00:00
}
retVal = ASR_init (sc);
splx (s);
if (retVal != 0) {
debug_asr_printf ("ASR_init failed\n");
sc->ha_in_reset = HA_OFF_LINE;
return (ENXIO);
}
if (ASR_rescan (sc) != 0) {
debug_asr_printf ("ASR_rescan failed\n");
}
ASR_failActiveCommands (sc);
if (sc->ha_in_reset == HA_OFF_LINE_RECOVERY) {
printf ("asr%d: Brining adapter back on-line\n",
sc->ha_path[0]
? cam_sim_unit(xpt_path_sim(sc->ha_path[0]))
: 0);
}
sc->ha_in_reset = HA_OPERATIONAL;
return (0);
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
} /* ASR_reset */
/*
unexpand -a everything
2002-05-14 01:09:03 +00:00
* Device timeout handler.
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
*/
Get rid of a bunch of useless macros. STATIC becomes static, INLINE becomes __inline where appropriate and gets nuked elsewhere, IN/OUT/INOUT go away. Reformat code affected by this.
2004-05-01 01:25:05 +00:00
static void
asr_timeout(void *arg)
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
{
Get rid of a bunch of useless macros. STATIC becomes static, INLINE becomes __inline where appropriate and gets nuked elsewhere, IN/OUT/INOUT go away. Reformat code affected by this.
2004-05-01 01:25:05 +00:00
union asr_ccb *ccb = (union asr_ccb *)arg;
Asr_softc_t *sc = (Asr_softc_t *)(ccb->ccb_h.spriv_ptr0);
unexpand -a everything
2002-05-14 01:09:03 +00:00
int s;
debug_asr_print_path(ccb);
debug_asr_printf("timed out");
/*
* Check if the adapter has locked up?
*/
if ((s = ASR_getBlinkLedCode(sc)) != 0) {
/* Reset Adapter */
printf ("asr%d: Blink LED 0x%x resetting adapter\n",
cam_sim_unit(xpt_path_sim(ccb->ccb_h.path)), s);
if (ASR_reset (sc) == ENXIO) {
/* Try again later */
ccb->ccb_h.timeout_ch = timeout(asr_timeout,
(caddr_t)ccb,
(ccb->ccb_h.timeout * hz) / 1000);
}
return;
}
/*
* Abort does not function on the ASR card!!! Walking away from
* the SCSI command is also *very* dangerous. A SCSI BUS reset is
* our best bet, followed by a complete adapter reset if that fails.
*/
s = splcam();
/* Check if we already timed out once to raise the issue */
if ((ccb->ccb_h.status & CAM_STATUS_MASK) == CAM_CMD_TIMEOUT) {
debug_asr_printf (" AGAIN\nreinitializing adapter\n");
if (ASR_reset (sc) == ENXIO) {
ccb->ccb_h.timeout_ch = timeout(asr_timeout,
(caddr_t)ccb,
(ccb->ccb_h.timeout * hz) / 1000);
}
splx(s);
return;
}
debug_asr_printf ("\nresetting bus\n");
/* If the BUS reset does not take, then an adapter reset is next! */
ccb->ccb_h.status &= ~CAM_STATUS_MASK;
ccb->ccb_h.status |= CAM_CMD_TIMEOUT;
ccb->ccb_h.timeout_ch = timeout(asr_timeout, (caddr_t)ccb,
(ccb->ccb_h.timeout * hz) / 1000);
ASR_resetBus (sc, cam_sim_bus(xpt_path_sim(ccb->ccb_h.path)));
xpt_async (AC_BUS_RESET, ccb->ccb_h.path, NULL);
splx(s);
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
} /* asr_timeout */
/*
* send a message asynchronously
*/
Get rid of a bunch of useless macros. STATIC becomes static, INLINE becomes __inline where appropriate and gets nuked elsewhere, IN/OUT/INOUT go away. Reformat code affected by this.
2004-05-01 01:25:05 +00:00
static int
ASR_queue(Asr_softc_t *sc, PI2O_MESSAGE_FRAME Message)
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
{
Get rid of a bunch of useless macros. STATIC becomes static, INLINE becomes __inline where appropriate and gets nuked elsewhere, IN/OUT/INOUT go away. Reformat code affected by this.
2004-05-01 01:25:05 +00:00
U32 MessageOffset;
union asr_ccb *ccb;
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
Many more style cleanups. Switch complex macros to being inline functions. Put @includes in a better spot. Fix many cases of 2 space indents and spaces between a function name and the parens. Use KASSERT instead of a home-rolled ASSERT. Remove some undeeded caddr casts.
2004-05-01 02:27:06 +00:00
debug_asr_printf("Host Command Dump:\n");
debug_asr_dump_message(Message);
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
unexpand -a everything
2002-05-14 01:09:03 +00:00
ccb = (union asr_ccb *)(long)
I2O_MESSAGE_FRAME_getInitiatorContext64(Message);
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
Convert the asr driver to use the bus_space API. This does not represent a significant functional change, but it further cleans up the code and brings it closer to being portable. Thanks to Don Bowman for helping to test this.
2004-05-05 03:42:33 +00:00
if ((MessageOffset = ASR_getMessage(sc)) != EMPTY_QUEUE) {
asr_set_frame(sc, Message, MessageOffset,
I2O_MESSAGE_FRAME_getMessageSize(Message));
unexpand -a everything
2002-05-14 01:09:03 +00:00
if (ccb) {
ASR_ccbAdd (sc, ccb);
}
/* Post the command */
Convert the asr driver to use the bus_space API. This does not represent a significant functional change, but it further cleans up the code and brings it closer to being portable. Thanks to Don Bowman for helping to test this.
2004-05-05 03:42:33 +00:00
asr_set_ToFIFO(sc, MessageOffset);
unexpand -a everything
2002-05-14 01:09:03 +00:00
} else {
if (ASR_getBlinkLedCode(sc)) {
/*
* Unlikely we can do anything if we can't grab a
* message frame :-(, but lets give it a try.
*/
Get rid of a bunch of useless macros. STATIC becomes static, INLINE becomes __inline where appropriate and gets nuked elsewhere, IN/OUT/INOUT go away. Reformat code affected by this.
2004-05-01 01:25:05 +00:00
(void)ASR_reset(sc);
unexpand -a everything
2002-05-14 01:09:03 +00:00
}
}
return (MessageOffset);
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
} /* ASR_queue */
/* Simple Scatter Gather elements */
`unexpand -a' should be run _before_ sed 's/^#define /#define^I/g'.
2002-05-14 21:59:10 +00:00
#define SG(SGL,Index,Flags,Buffer,Size) \
unexpand -a everything
2002-05-14 01:09:03 +00:00
I2O_FLAGS_COUNT_setCount( \
&(((PI2O_SG_ELEMENT)(SGL))->u.Simple[Index].FlagsCount), \
Size); \
I2O_FLAGS_COUNT_setFlags( \
&(((PI2O_SG_ELEMENT)(SGL))->u.Simple[Index].FlagsCount), \
I2O_SGL_FLAGS_SIMPLE_ADDRESS_ELEMENT | (Flags)); \
I2O_SGE_SIMPLE_ELEMENT_setPhysicalAddress( \
&(((PI2O_SG_ELEMENT)(SGL))->u.Simple[Index]), \
Replace (ab)uses of "NULL" where "0" is really meant.
2002-08-22 21:24:01 +00:00
(Buffer == NULL) ? 0 : KVTOPHYS(Buffer))
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
/*
unexpand -a everything
2002-05-14 01:09:03 +00:00
* Retrieve Parameter Group.
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
*/
Get rid of a bunch of useless macros. STATIC becomes static, INLINE becomes __inline where appropriate and gets nuked elsewhere, IN/OUT/INOUT go away. Reformat code affected by this.
2004-05-01 01:25:05 +00:00
static void *
ASR_getParams(Asr_softc_t *sc, tid_t TID, int Group, void *Buffer,
unsigned BufferSize)
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
{
unexpand -a everything
2002-05-14 01:09:03 +00:00
struct paramGetMessage {
I2O_UTIL_PARAMS_GET_MESSAGE M;
Get rid of a bunch of useless macros. STATIC becomes static, INLINE becomes __inline where appropriate and gets nuked elsewhere, IN/OUT/INOUT go away. Reformat code affected by this.
2004-05-01 01:25:05 +00:00
char
F[sizeof(I2O_SGE_SIMPLE_ELEMENT)*2 - sizeof(I2O_SG_ELEMENT)];
unexpand -a everything
2002-05-14 01:09:03 +00:00
struct Operations {
I2O_PARAM_OPERATIONS_LIST_HEADER Header;
I2O_PARAM_OPERATION_ALL_TEMPLATE Template[1];
} O;
Remove the defAlignLong and getAlignLong macros. I guess that the original intent was to make sure that message structs allocated off of the stack were 4-byte aligned. However, the macros as defined did absolutely nothing. And since I2O forces you to manually copy messages down to the hardware, there really is no point of enforced alignment anyways.
2004-05-02 03:33:18 +00:00
} Message;
struct Operations *Operations_Ptr;
I2O_UTIL_PARAMS_GET_MESSAGE *Message_Ptr;
unexpand -a everything
2002-05-14 01:09:03 +00:00
struct ParamBuffer {
I2O_PARAM_RESULTS_LIST_HEADER Header;
I2O_PARAM_READ_OPERATION_RESULT Read;
char Info[1];
Remove the defAlignLong and getAlignLong macros. I guess that the original intent was to make sure that message structs allocated off of the stack were 4-byte aligned. However, the macros as defined did absolutely nothing. And since I2O forces you to manually copy messages down to the hardware, there really is no point of enforced alignment anyways.
2004-05-02 03:33:18 +00:00
} *Buffer_Ptr;
unexpand -a everything
2002-05-14 01:09:03 +00:00
Remove the defAlignLong and getAlignLong macros. I guess that the original intent was to make sure that message structs allocated off of the stack were 4-byte aligned. However, the macros as defined did absolutely nothing. And since I2O forces you to manually copy messages down to the hardware, there really is no point of enforced alignment anyways.
2004-05-02 03:33:18 +00:00
Message_Ptr = (I2O_UTIL_PARAMS_GET_MESSAGE *)ASR_fillMessage(&Message,
unexpand -a everything
2002-05-14 01:09:03 +00:00
sizeof(I2O_UTIL_PARAMS_GET_MESSAGE)
+ sizeof(I2O_SGE_SIMPLE_ELEMENT)*2 - sizeof(I2O_SG_ELEMENT));
Operations_Ptr = (struct Operations *)((char *)Message_Ptr
+ sizeof(I2O_UTIL_PARAMS_GET_MESSAGE)
+ sizeof(I2O_SGE_SIMPLE_ELEMENT)*2 - sizeof(I2O_SG_ELEMENT));
More whitespace style cleanups, also remove unneeded (void *) casts for bzero().
2004-05-01 03:06:54 +00:00
bzero(Operations_Ptr, sizeof(struct Operations));
unexpand -a everything
2002-05-14 01:09:03 +00:00
I2O_PARAM_OPERATIONS_LIST_HEADER_setOperationCount(
&(Operations_Ptr->Header), 1);
I2O_PARAM_OPERATION_ALL_TEMPLATE_setOperation(
&(Operations_Ptr->Template[0]), I2O_PARAMS_OPERATION_FIELD_GET);
I2O_PARAM_OPERATION_ALL_TEMPLATE_setFieldCount(
&(Operations_Ptr->Template[0]), 0xFFFF);
I2O_PARAM_OPERATION_ALL_TEMPLATE_setGroupNumber(
&(Operations_Ptr->Template[0]), Group);
Remove the defAlignLong and getAlignLong macros. I guess that the original intent was to make sure that message structs allocated off of the stack were 4-byte aligned. However, the macros as defined did absolutely nothing. And since I2O forces you to manually copy messages down to the hardware, there really is no point of enforced alignment anyways.
2004-05-02 03:33:18 +00:00
Buffer_Ptr = (struct ParamBuffer *)Buffer;
More whitespace style cleanups, also remove unneeded (void *) casts for bzero().
2004-05-01 03:06:54 +00:00
bzero(Buffer_Ptr, BufferSize);
unexpand -a everything
2002-05-14 01:09:03 +00:00
I2O_MESSAGE_FRAME_setVersionOffset(&(Message_Ptr->StdMessageFrame),
I2O_VERSION_11
+ (((sizeof(I2O_UTIL_PARAMS_GET_MESSAGE) - sizeof(I2O_SG_ELEMENT))
/ sizeof(U32)) << 4));
I2O_MESSAGE_FRAME_setTargetAddress (&(Message_Ptr->StdMessageFrame),
TID);
I2O_MESSAGE_FRAME_setFunction (&(Message_Ptr->StdMessageFrame),
I2O_UTIL_PARAMS_GET);
/*
* Set up the buffers as scatter gather elements.
*/
SG(&(Message_Ptr->SGL), 0,
I2O_SGL_FLAGS_DIR | I2O_SGL_FLAGS_END_OF_BUFFER,
Operations_Ptr, sizeof(struct Operations));
SG(&(Message_Ptr->SGL), 1,
I2O_SGL_FLAGS_LAST_ELEMENT | I2O_SGL_FLAGS_END_OF_BUFFER,
Buffer_Ptr, BufferSize);
if ((ASR_queue_c(sc, (PI2O_MESSAGE_FRAME)Message_Ptr) == CAM_REQ_CMP)
&& (Buffer_Ptr->Header.ResultCount)) {
return ((void *)(Buffer_Ptr->Info));
}
Begin the slow, painful process of cleaning up asr. This eliminates many of the bogus casts of NULL. It also distracts me while I procrastinate on useful work.
2004-04-21 03:51:39 +00:00
return (NULL);
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
} /* ASR_getParams */
/*
unexpand -a everything
2002-05-14 01:09:03 +00:00
* Acquire the LCT information.
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
*/
Get rid of a bunch of useless macros. STATIC becomes static, INLINE becomes __inline where appropriate and gets nuked elsewhere, IN/OUT/INOUT go away. Reformat code affected by this.
2004-05-01 01:25:05 +00:00
static int
ASR_acquireLct(Asr_softc_t *sc)
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
{
Get rid of a bunch of useless macros. STATIC becomes static, INLINE becomes __inline where appropriate and gets nuked elsewhere, IN/OUT/INOUT go away. Reformat code affected by this.
2004-05-01 01:25:05 +00:00
PI2O_EXEC_LCT_NOTIFY_MESSAGE Message_Ptr;
PI2O_SGE_SIMPLE_ELEMENT sg;
int MessageSizeInBytes;
caddr_t v;
int len;
I2O_LCT Table;
PI2O_LCT_ENTRY Entry;
unexpand -a everything
2002-05-14 01:09:03 +00:00
/*
* sc value assumed valid
*/
Fix a typo from rev 1.56 that somehow escaped my testing.
2004-05-04 03:16:22 +00:00
MessageSizeInBytes = sizeof(I2O_EXEC_LCT_NOTIFY_MESSAGE) -
sizeof(I2O_SG_ELEMENT) + sizeof(I2O_SGE_SIMPLE_ELEMENT);
if ((Message_Ptr = (PI2O_EXEC_LCT_NOTIFY_MESSAGE)malloc(
MessageSizeInBytes, M_TEMP, M_WAITOK)) == NULL) {
unexpand -a everything
2002-05-14 01:09:03 +00:00
return (ENOMEM);
}
Remove the defAlignLong and getAlignLong macros. I guess that the original intent was to make sure that message structs allocated off of the stack were 4-byte aligned. However, the macros as defined did absolutely nothing. And since I2O forces you to manually copy messages down to the hardware, there really is no point of enforced alignment anyways.
2004-05-02 03:33:18 +00:00
(void)ASR_fillMessage((void *)Message_Ptr, MessageSizeInBytes);
unexpand -a everything
2002-05-14 01:09:03 +00:00
I2O_MESSAGE_FRAME_setVersionOffset(&(Message_Ptr->StdMessageFrame),
Fix a typo from rev 1.56 that somehow escaped my testing.
2004-05-04 03:16:22 +00:00
(I2O_VERSION_11 + (((sizeof(I2O_EXEC_LCT_NOTIFY_MESSAGE) -
sizeof(I2O_SG_ELEMENT)) / sizeof(U32)) << 4)));
unexpand -a everything
2002-05-14 01:09:03 +00:00
I2O_MESSAGE_FRAME_setFunction(&(Message_Ptr->StdMessageFrame),
Fix a typo from rev 1.56 that somehow escaped my testing.
2004-05-04 03:16:22 +00:00
I2O_EXEC_LCT_NOTIFY);
unexpand -a everything
2002-05-14 01:09:03 +00:00
I2O_EXEC_LCT_NOTIFY_MESSAGE_setClassIdentifier(Message_Ptr,
Fix a typo from rev 1.56 that somehow escaped my testing.
2004-05-04 03:16:22 +00:00
I2O_CLASS_MATCH_ANYCLASS);
unexpand -a everything
2002-05-14 01:09:03 +00:00
/*
* Call the LCT table to determine the number of device entries
* to reserve space for.
*/
SG(&(Message_Ptr->SGL), 0,
I2O_SGL_FLAGS_LAST_ELEMENT | I2O_SGL_FLAGS_END_OF_BUFFER, &Table,
sizeof(I2O_LCT));
/*
* since this code is reused in several systems, code efficiency
* is greater by using a shift operation rather than a divide by
* sizeof(u_int32_t).
*/
I2O_LCT_setTableSize(&Table,
(sizeof(I2O_LCT) - sizeof(I2O_LCT_ENTRY)) >> 2);
(void)ASR_queue_c(sc, (PI2O_MESSAGE_FRAME)Message_Ptr);
/*
* Determine the size of the LCT table.
*/
if (sc->ha_LCT) {
Remove the defAlignLong and getAlignLong macros. I guess that the original intent was to make sure that message structs allocated off of the stack were 4-byte aligned. However, the macros as defined did absolutely nothing. And since I2O forces you to manually copy messages down to the hardware, there really is no point of enforced alignment anyways.
2004-05-02 03:33:18 +00:00
free(sc->ha_LCT, M_TEMP);
unexpand -a everything
2002-05-14 01:09:03 +00:00
}
/*
* malloc only generates contiguous memory when less than a
* page is expected. We must break the request up into an SG list ...
*/
if (((len = (I2O_LCT_getTableSize(&Table) << 2)) <=
(sizeof(I2O_LCT) - sizeof(I2O_LCT_ENTRY)))
|| (len > (128 * 1024))) { /* Arbitrary */
Remove the defAlignLong and getAlignLong macros. I guess that the original intent was to make sure that message structs allocated off of the stack were 4-byte aligned. However, the macros as defined did absolutely nothing. And since I2O forces you to manually copy messages down to the hardware, there really is no point of enforced alignment anyways.
2004-05-02 03:33:18 +00:00
free(Message_Ptr, M_TEMP);
unexpand -a everything
2002-05-14 01:09:03 +00:00
return (EINVAL);
}
Begin the slow, painful process of cleaning up asr. This eliminates many of the bogus casts of NULL. It also distracts me while I procrastinate on useful work.
2004-04-21 03:51:39 +00:00
if ((sc->ha_LCT = (PI2O_LCT)malloc (len, M_TEMP, M_WAITOK)) == NULL) {
Remove the defAlignLong and getAlignLong macros. I guess that the original intent was to make sure that message structs allocated off of the stack were 4-byte aligned. However, the macros as defined did absolutely nothing. And since I2O forces you to manually copy messages down to the hardware, there really is no point of enforced alignment anyways.
2004-05-02 03:33:18 +00:00
free(Message_Ptr, M_TEMP);
unexpand -a everything
2002-05-14 01:09:03 +00:00
return (ENOMEM);
}
/*
* since this code is reused in several systems, code efficiency
* is greater by using a shift operation rather than a divide by
* sizeof(u_int32_t).
*/
I2O_LCT_setTableSize(sc->ha_LCT,
(sizeof(I2O_LCT) - sizeof(I2O_LCT_ENTRY)) >> 2);
/*
* Convert the access to the LCT table into a SG list.
*/
sg = Message_Ptr->SGL.u.Simple;
v = (caddr_t)(sc->ha_LCT);
for (;;) {
int next, base, span;
span = 0;
next = base = KVTOPHYS(v);
I2O_SGE_SIMPLE_ELEMENT_setPhysicalAddress(sg, base);
/* How far can we go contiguously */
while ((len > 0) && (base == next)) {
int size;
next = trunc_page(base) + PAGE_SIZE;
size = next - base;
if (size > len) {
size = len;
}
span += size;
v += size;
len -= size;
base = KVTOPHYS(v);
}
/* Construct the Flags */
I2O_FLAGS_COUNT_setCount(&(sg->FlagsCount), span);
{
int rw = I2O_SGL_FLAGS_SIMPLE_ADDRESS_ELEMENT;
if (len <= 0) {
rw = (I2O_SGL_FLAGS_SIMPLE_ADDRESS_ELEMENT
| I2O_SGL_FLAGS_LAST_ELEMENT
| I2O_SGL_FLAGS_END_OF_BUFFER);
}
I2O_FLAGS_COUNT_setFlags(&(sg->FlagsCount), rw);
}
if (len <= 0) {
break;
}
/*
* Incrementing requires resizing of the packet.
*/
++sg;
MessageSizeInBytes += sizeof(*sg);
I2O_MESSAGE_FRAME_setMessageSize(
&(Message_Ptr->StdMessageFrame),
I2O_MESSAGE_FRAME_getMessageSize(
&(Message_Ptr->StdMessageFrame))
+ (sizeof(*sg) / sizeof(U32)));
{
PI2O_EXEC_LCT_NOTIFY_MESSAGE NewMessage_Ptr;
if ((NewMessage_Ptr = (PI2O_EXEC_LCT_NOTIFY_MESSAGE)
Begin the slow, painful process of cleaning up asr. This eliminates many of the bogus casts of NULL. It also distracts me while I procrastinate on useful work.
2004-04-21 03:51:39 +00:00
malloc(MessageSizeInBytes, M_TEMP, M_WAITOK))
== NULL) {
Remove the defAlignLong and getAlignLong macros. I guess that the original intent was to make sure that message structs allocated off of the stack were 4-byte aligned. However, the macros as defined did absolutely nothing. And since I2O forces you to manually copy messages down to the hardware, there really is no point of enforced alignment anyways.
2004-05-02 03:33:18 +00:00
free(sc->ha_LCT, M_TEMP);
Begin the slow, painful process of cleaning up asr. This eliminates many of the bogus casts of NULL. It also distracts me while I procrastinate on useful work.
2004-04-21 03:51:39 +00:00
sc->ha_LCT = NULL;
Remove the defAlignLong and getAlignLong macros. I guess that the original intent was to make sure that message structs allocated off of the stack were 4-byte aligned. However, the macros as defined did absolutely nothing. And since I2O forces you to manually copy messages down to the hardware, there really is no point of enforced alignment anyways.
2004-05-02 03:33:18 +00:00
free(Message_Ptr, M_TEMP);
unexpand -a everything
2002-05-14 01:09:03 +00:00
return (ENOMEM);
}
span = ((caddr_t)sg) - (caddr_t)Message_Ptr;
Many more style cleanups. Switch complex macros to being inline functions. Put @includes in a better spot. Fix many cases of 2 space indents and spaces between a function name and the parens. Use KASSERT instead of a home-rolled ASSERT. Remove some undeeded caddr casts.
2004-05-01 02:27:06 +00:00
bcopy(Message_Ptr, NewMessage_Ptr, span);
free(Message_Ptr, M_TEMP);
unexpand -a everything
2002-05-14 01:09:03 +00:00
sg = (PI2O_SGE_SIMPLE_ELEMENT)
(((caddr_t)NewMessage_Ptr) + span);
Message_Ptr = NewMessage_Ptr;
}
}
{ int retval;
retval = ASR_queue_c(sc, (PI2O_MESSAGE_FRAME)Message_Ptr);
Remove the defAlignLong and getAlignLong macros. I guess that the original intent was to make sure that message structs allocated off of the stack were 4-byte aligned. However, the macros as defined did absolutely nothing. And since I2O forces you to manually copy messages down to the hardware, there really is no point of enforced alignment anyways.
2004-05-02 03:33:18 +00:00
free(Message_Ptr, M_TEMP);
unexpand -a everything
2002-05-14 01:09:03 +00:00
if (retval != CAM_REQ_CMP) {
return (ENODEV);
}
}
/* If the LCT table grew, lets truncate accesses */
if (I2O_LCT_getTableSize(&Table) < I2O_LCT_getTableSize(sc->ha_LCT)) {
I2O_LCT_setTableSize(sc->ha_LCT, I2O_LCT_getTableSize(&Table));
}
for (Entry = sc->ha_LCT->LCTEntry; Entry < (PI2O_LCT_ENTRY)
(((U32 *)sc->ha_LCT)+I2O_LCT_getTableSize(sc->ha_LCT));
++Entry) {
Entry->le_type = I2O_UNKNOWN;
switch (I2O_CLASS_ID_getClass(&(Entry->ClassID))) {
case I2O_CLASS_RANDOM_BLOCK_STORAGE:
Entry->le_type = I2O_BSA;
break;
case I2O_CLASS_SCSI_PERIPHERAL:
Entry->le_type = I2O_SCSI;
break;
case I2O_CLASS_FIBRE_CHANNEL_PERIPHERAL:
Entry->le_type = I2O_FCA;
break;
case I2O_CLASS_BUS_ADAPTER_PORT:
Entry->le_type = I2O_PORT | I2O_SCSI;
/* FALLTHRU */
case I2O_CLASS_FIBRE_CHANNEL_PORT:
if (I2O_CLASS_ID_getClass(&(Entry->ClassID)) ==
I2O_CLASS_FIBRE_CHANNEL_PORT) {
Entry->le_type = I2O_PORT | I2O_FCA;
}
{ struct ControllerInfo {
I2O_PARAM_RESULTS_LIST_HEADER Header;
I2O_PARAM_READ_OPERATION_RESULT Read;
I2O_HBA_SCSI_CONTROLLER_INFO_SCALAR Info;
Remove the defAlignLong and getAlignLong macros. I guess that the original intent was to make sure that message structs allocated off of the stack were 4-byte aligned. However, the macros as defined did absolutely nothing. And since I2O forces you to manually copy messages down to the hardware, there really is no point of enforced alignment anyways.
2004-05-02 03:33:18 +00:00
} Buffer;
unexpand -a everything
2002-05-14 01:09:03 +00:00
PI2O_HBA_SCSI_CONTROLLER_INFO_SCALAR Info;
Entry->le_bus = 0xff;
Entry->le_target = 0xff;
Entry->le_lun = 0xff;
if ((Info = (PI2O_HBA_SCSI_CONTROLLER_INFO_SCALAR)
ASR_getParams(sc,
I2O_LCT_ENTRY_getLocalTID(Entry),
I2O_HBA_SCSI_CONTROLLER_INFO_GROUP_NO,
Remove the defAlignLong and getAlignLong macros. I guess that the original intent was to make sure that message structs allocated off of the stack were 4-byte aligned. However, the macros as defined did absolutely nothing. And since I2O forces you to manually copy messages down to the hardware, there really is no point of enforced alignment anyways.
2004-05-02 03:33:18 +00:00
&Buffer, sizeof(struct ControllerInfo))) == NULL) {
unexpand -a everything
2002-05-14 01:09:03 +00:00
continue;
}
Entry->le_target
= I2O_HBA_SCSI_CONTROLLER_INFO_SCALAR_getInitiatorID(
Info);
Entry->le_lun = 0;
} /* FALLTHRU */
default:
continue;
}
{ struct DeviceInfo {
I2O_PARAM_RESULTS_LIST_HEADER Header;
I2O_PARAM_READ_OPERATION_RESULT Read;
I2O_DPT_DEVICE_INFO_SCALAR Info;
Remove the defAlignLong and getAlignLong macros. I guess that the original intent was to make sure that message structs allocated off of the stack were 4-byte aligned. However, the macros as defined did absolutely nothing. And since I2O forces you to manually copy messages down to the hardware, there really is no point of enforced alignment anyways.
2004-05-02 03:33:18 +00:00
} Buffer;
unexpand -a everything
2002-05-14 01:09:03 +00:00
PI2O_DPT_DEVICE_INFO_SCALAR Info;
Entry->le_bus = 0xff;
Entry->le_target = 0xff;
Entry->le_lun = 0xff;
if ((Info = (PI2O_DPT_DEVICE_INFO_SCALAR)
ASR_getParams(sc,
I2O_LCT_ENTRY_getLocalTID(Entry),
I2O_DPT_DEVICE_INFO_GROUP_NO,
Remove the defAlignLong and getAlignLong macros. I guess that the original intent was to make sure that message structs allocated off of the stack were 4-byte aligned. However, the macros as defined did absolutely nothing. And since I2O forces you to manually copy messages down to the hardware, there really is no point of enforced alignment anyways.
2004-05-02 03:33:18 +00:00
&Buffer, sizeof(struct DeviceInfo))) == NULL) {
unexpand -a everything
2002-05-14 01:09:03 +00:00
continue;
}
Entry->le_type
|= I2O_DPT_DEVICE_INFO_SCALAR_getDeviceType(Info);
Entry->le_bus
= I2O_DPT_DEVICE_INFO_SCALAR_getBus(Info);
if ((Entry->le_bus > sc->ha_MaxBus)
&& (Entry->le_bus <= MAX_CHANNEL)) {
sc->ha_MaxBus = Entry->le_bus;
}
Entry->le_target
= I2O_DPT_DEVICE_INFO_SCALAR_getIdentifier(Info);
Entry->le_lun
= I2O_DPT_DEVICE_INFO_SCALAR_getLunInfo(Info);
}
}
/*
* A zero return value indicates success.
*/
return (0);
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
} /* ASR_acquireLct */
/*
* Initialize a message frame.
* We assume that the CDB has already been set up, so all we do here is
* generate the Scatter Gather list.
*/
Get rid of a bunch of useless macros. STATIC becomes static, INLINE becomes __inline where appropriate and gets nuked elsewhere, IN/OUT/INOUT go away. Reformat code affected by this.
2004-05-01 01:25:05 +00:00
static PI2O_MESSAGE_FRAME
ASR_init_message(union asr_ccb *ccb, PI2O_MESSAGE_FRAME Message)
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
{
Get rid of a bunch of useless macros. STATIC becomes static, INLINE becomes __inline where appropriate and gets nuked elsewhere, IN/OUT/INOUT go away. Reformat code affected by this.
2004-05-01 01:25:05 +00:00
PI2O_MESSAGE_FRAME Message_Ptr;
unexpand -a everything
2002-05-14 01:09:03 +00:00
PI2O_SGE_SIMPLE_ELEMENT sg;
Get rid of a bunch of useless macros. STATIC becomes static, INLINE becomes __inline where appropriate and gets nuked elsewhere, IN/OUT/INOUT go away. Reformat code affected by this.
2004-05-01 01:25:05 +00:00
Asr_softc_t *sc = (Asr_softc_t *)(ccb->ccb_h.spriv_ptr0);
unexpand -a everything
2002-05-14 01:09:03 +00:00
vm_size_t size, len;
Get rid of a bunch of useless macros. STATIC becomes static, INLINE becomes __inline where appropriate and gets nuked elsewhere, IN/OUT/INOUT go away. Reformat code affected by this.
2004-05-01 01:25:05 +00:00
caddr_t v;
unexpand -a everything
2002-05-14 01:09:03 +00:00
U32 MessageSize;
Get rid of a bunch of useless macros. STATIC becomes static, INLINE becomes __inline where appropriate and gets nuked elsewhere, IN/OUT/INOUT go away. Reformat code affected by this.
2004-05-01 01:25:05 +00:00
int next, span, base, rw;
int target = ccb->ccb_h.target_id;
int lun = ccb->ccb_h.target_lun;
int bus =cam_sim_bus(xpt_path_sim(ccb->ccb_h.path));
tid_t TID;
unexpand -a everything
2002-05-14 01:09:03 +00:00
/* We only need to zero out the PRIVATE_SCSI_SCB_EXECUTE_MESSAGE */
Remove the defAlignLong and getAlignLong macros. I guess that the original intent was to make sure that message structs allocated off of the stack were 4-byte aligned. However, the macros as defined did absolutely nothing. And since I2O forces you to manually copy messages down to the hardware, there really is no point of enforced alignment anyways.
2004-05-02 03:33:18 +00:00
Message_Ptr = (I2O_MESSAGE_FRAME *)Message;
Get rid of a bunch of useless macros. STATIC becomes static, INLINE becomes __inline where appropriate and gets nuked elsewhere, IN/OUT/INOUT go away. Reformat code affected by this.
2004-05-01 01:25:05 +00:00
bzero(Message_Ptr, (sizeof(PRIVATE_SCSI_SCB_EXECUTE_MESSAGE) -
sizeof(I2O_SG_ELEMENT)));
unexpand -a everything
2002-05-14 01:09:03 +00:00
Get rid of a bunch of useless macros. STATIC becomes static, INLINE becomes __inline where appropriate and gets nuked elsewhere, IN/OUT/INOUT go away. Reformat code affected by this.
2004-05-01 01:25:05 +00:00
if ((TID = ASR_getTid (sc, bus, target, lun)) == (tid_t)-1) {
PI2O_LCT_ENTRY Device;
TID = 0;
for (Device = sc->ha_LCT->LCTEntry; Device < (PI2O_LCT_ENTRY)
(((U32 *)sc->ha_LCT) + I2O_LCT_getTableSize(sc->ha_LCT));
++Device) {
if ((Device->le_type != I2O_UNKNOWN)
&& (Device->le_bus == bus)
&& (Device->le_target == target)
&& (Device->le_lun == lun)
&& (I2O_LCT_ENTRY_getUserTID(Device) == 0xFFF)) {
TID = I2O_LCT_ENTRY_getLocalTID(Device);
ASR_setTid(sc, Device->le_bus,
Device->le_target, Device->le_lun,
TID);
break;
unexpand -a everything
2002-05-14 01:09:03 +00:00
}
}
}
Get rid of a bunch of useless macros. STATIC becomes static, INLINE becomes __inline where appropriate and gets nuked elsewhere, IN/OUT/INOUT go away. Reformat code affected by this.
2004-05-01 01:25:05 +00:00
if (TID == (tid_t)0) {
return (NULL);
}
I2O_MESSAGE_FRAME_setTargetAddress(Message_Ptr, TID);
PRIVATE_SCSI_SCB_EXECUTE_MESSAGE_setTID(
(PPRIVATE_SCSI_SCB_EXECUTE_MESSAGE)Message_Ptr, TID);
unexpand -a everything
2002-05-14 01:09:03 +00:00
I2O_MESSAGE_FRAME_setVersionOffset(Message_Ptr, I2O_VERSION_11 |
(((sizeof(PRIVATE_SCSI_SCB_EXECUTE_MESSAGE) - sizeof(I2O_SG_ELEMENT))
/ sizeof(U32)) << 4));
I2O_MESSAGE_FRAME_setMessageSize(Message_Ptr,
(sizeof(PRIVATE_SCSI_SCB_EXECUTE_MESSAGE)
- sizeof(I2O_SG_ELEMENT)) / sizeof(U32));
I2O_MESSAGE_FRAME_setInitiatorAddress (Message_Ptr, 1);
I2O_MESSAGE_FRAME_setFunction(Message_Ptr, I2O_PRIVATE_MESSAGE);
I2O_PRIVATE_MESSAGE_FRAME_setXFunctionCode (
(PI2O_PRIVATE_MESSAGE_FRAME)Message_Ptr, I2O_SCSI_SCB_EXEC);
PRIVATE_SCSI_SCB_EXECUTE_MESSAGE_setSCBFlags (
(PPRIVATE_SCSI_SCB_EXECUTE_MESSAGE)Message_Ptr,
I2O_SCB_FLAG_ENABLE_DISCONNECT
| I2O_SCB_FLAG_SIMPLE_QUEUE_TAG
| I2O_SCB_FLAG_SENSE_DATA_IN_BUFFER);
/*
* We do not need any (optional byteswapping) method access to
* the Initiator & Transaction context field.
*/
I2O_MESSAGE_FRAME_setInitiatorContext64(Message, (long)ccb);
I2O_PRIVATE_MESSAGE_FRAME_setOrganizationID(
(PI2O_PRIVATE_MESSAGE_FRAME)Message_Ptr, DPT_ORGANIZATION_ID);
/*
* copy the cdb over
*/
PRIVATE_SCSI_SCB_EXECUTE_MESSAGE_setCDBLength(
Many more style cleanups. Switch complex macros to being inline functions. Put @includes in a better spot. Fix many cases of 2 space indents and spaces between a function name and the parens. Use KASSERT instead of a home-rolled ASSERT. Remove some undeeded caddr casts.
2004-05-01 02:27:06 +00:00
(PPRIVATE_SCSI_SCB_EXECUTE_MESSAGE)Message_Ptr, ccb->csio.cdb_len);
bcopy(&(ccb->csio.cdb_io),
((PPRIVATE_SCSI_SCB_EXECUTE_MESSAGE)Message_Ptr)->CDB,
ccb->csio.cdb_len);
unexpand -a everything
2002-05-14 01:09:03 +00:00
/*
* Given a buffer describing a transfer, set up a scatter/gather map
* in a ccb to map that SCSI transfer.
*/
rw = (ccb->ccb_h.flags & CAM_DIR_IN) ? 0 : I2O_SGL_FLAGS_DIR;
PRIVATE_SCSI_SCB_EXECUTE_MESSAGE_setSCBFlags (
(PPRIVATE_SCSI_SCB_EXECUTE_MESSAGE)Message_Ptr,
(ccb->csio.dxfer_len)
? ((rw) ? (I2O_SCB_FLAG_XFER_TO_DEVICE
| I2O_SCB_FLAG_ENABLE_DISCONNECT
| I2O_SCB_FLAG_SIMPLE_QUEUE_TAG
| I2O_SCB_FLAG_SENSE_DATA_IN_BUFFER)
: (I2O_SCB_FLAG_XFER_FROM_DEVICE
| I2O_SCB_FLAG_ENABLE_DISCONNECT
| I2O_SCB_FLAG_SIMPLE_QUEUE_TAG
| I2O_SCB_FLAG_SENSE_DATA_IN_BUFFER))
: (I2O_SCB_FLAG_ENABLE_DISCONNECT
| I2O_SCB_FLAG_SIMPLE_QUEUE_TAG
| I2O_SCB_FLAG_SENSE_DATA_IN_BUFFER));
/*
* Given a transfer described by a `data', fill in the SG list.
*/
sg = &((PPRIVATE_SCSI_SCB_EXECUTE_MESSAGE)Message_Ptr)->SGL.u.Simple[0];
len = ccb->csio.dxfer_len;
v = ccb->csio.data_ptr;
Many more style cleanups. Switch complex macros to being inline functions. Put @includes in a better spot. Fix many cases of 2 space indents and spaces between a function name and the parens. Use KASSERT instead of a home-rolled ASSERT. Remove some undeeded caddr casts.
2004-05-01 02:27:06 +00:00
KASSERT(ccb->csio.dxfer_len >= 0, ("csio.dxfer_len < 0"));
unexpand -a everything
2002-05-14 01:09:03 +00:00
MessageSize = I2O_MESSAGE_FRAME_getMessageSize(Message_Ptr);
PRIVATE_SCSI_SCB_EXECUTE_MESSAGE_setByteCount(
(PPRIVATE_SCSI_SCB_EXECUTE_MESSAGE)Message_Ptr, len);
while ((len > 0) && (sg < &((PPRIVATE_SCSI_SCB_EXECUTE_MESSAGE)
Message_Ptr)->SGL.u.Simple[SG_SIZE])) {
span = 0;
next = base = KVTOPHYS(v);
I2O_SGE_SIMPLE_ELEMENT_setPhysicalAddress(sg, base);
/* How far can we go contiguously */
while ((len > 0) && (base == next)) {
next = trunc_page(base) + PAGE_SIZE;
size = next - base;
if (size > len) {
size = len;
}
span += size;
v += size;
len -= size;
base = KVTOPHYS(v);
}
I2O_FLAGS_COUNT_setCount(&(sg->FlagsCount), span);
if (len == 0) {
rw |= I2O_SGL_FLAGS_LAST_ELEMENT;
}
I2O_FLAGS_COUNT_setFlags(&(sg->FlagsCount),
I2O_SGL_FLAGS_SIMPLE_ADDRESS_ELEMENT | rw);
++sg;
MessageSize += sizeof(*sg) / sizeof(U32);
}
/* We always do the request sense ... */
if ((span = ccb->csio.sense_len) == 0) {
span = sizeof(ccb->csio.sense_data);
}
SG(sg, 0, I2O_SGL_FLAGS_LAST_ELEMENT | I2O_SGL_FLAGS_END_OF_BUFFER,
&(ccb->csio.sense_data), span);
I2O_MESSAGE_FRAME_setMessageSize(Message_Ptr,
MessageSize + (sizeof(*sg) / sizeof(U32)));
return (Message_Ptr);
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
} /* ASR_init_message */
/*
unexpand -a everything
2002-05-14 01:09:03 +00:00
* Reset the adapter.
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
*/
Get rid of a bunch of useless macros. STATIC becomes static, INLINE becomes __inline where appropriate and gets nuked elsewhere, IN/OUT/INOUT go away. Reformat code affected by this.
2004-05-01 01:25:05 +00:00
static U32
ASR_initOutBound(Asr_softc_t *sc)
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
{
unexpand -a everything
2002-05-14 01:09:03 +00:00
struct initOutBoundMessage {
I2O_EXEC_OUTBOUND_INIT_MESSAGE M;
U32 R;
Remove the defAlignLong and getAlignLong macros. I guess that the original intent was to make sure that message structs allocated off of the stack were 4-byte aligned. However, the macros as defined did absolutely nothing. And since I2O forces you to manually copy messages down to the hardware, there really is no point of enforced alignment anyways.
2004-05-02 03:33:18 +00:00
} Message;
PI2O_EXEC_OUTBOUND_INIT_MESSAGE Message_Ptr;
U32 *volatile Reply_Ptr;
U32 Old;
unexpand -a everything
2002-05-14 01:09:03 +00:00
/*
* Build up our copy of the Message.
*/
Remove the defAlignLong and getAlignLong macros. I guess that the original intent was to make sure that message structs allocated off of the stack were 4-byte aligned. However, the macros as defined did absolutely nothing. And since I2O forces you to manually copy messages down to the hardware, there really is no point of enforced alignment anyways.
2004-05-02 03:33:18 +00:00
Message_Ptr = (PI2O_EXEC_OUTBOUND_INIT_MESSAGE)ASR_fillMessage(&Message,
unexpand -a everything
2002-05-14 01:09:03 +00:00
sizeof(I2O_EXEC_OUTBOUND_INIT_MESSAGE));
I2O_MESSAGE_FRAME_setFunction(&(Message_Ptr->StdMessageFrame),
I2O_EXEC_OUTBOUND_INIT);
I2O_EXEC_OUTBOUND_INIT_MESSAGE_setHostPageFrameSize(Message_Ptr, PAGE_SIZE);
I2O_EXEC_OUTBOUND_INIT_MESSAGE_setOutboundMFrameSize(Message_Ptr,
sizeof(I2O_SCSI_ERROR_REPLY_MESSAGE_FRAME));
/*
* Reset the Reply Status
*/
*(Reply_Ptr = (U32 *)((char *)Message_Ptr
+ sizeof(I2O_EXEC_OUTBOUND_INIT_MESSAGE))) = 0;
SG (&(Message_Ptr->SGL), 0, I2O_SGL_FLAGS_LAST_ELEMENT, Reply_Ptr,
sizeof(U32));
/*
* Send the Message out
*/
Now that the U32 type is a really 32-bits wide, eliminate a bunch of other bad assumptions and long values.
2006-02-04 08:20:23 +00:00
if ((Old = ASR_initiateCp(sc, (PI2O_MESSAGE_FRAME)Message_Ptr)) !=
0xffffffff) {
unexpand -a everything
2002-05-14 01:09:03 +00:00
u_long size, addr;
/*
* Wait for a response (Poll).
*/
while (*Reply_Ptr < I2O_EXEC_OUTBOUND_INIT_REJECTED);
/*
* Re-enable the interrupts.
*/
Convert the asr driver to use the bus_space API. This does not represent a significant functional change, but it further cleans up the code and brings it closer to being portable. Thanks to Don Bowman for helping to test this.
2004-05-05 03:42:33 +00:00
asr_set_intr(sc, Old);
unexpand -a everything
2002-05-14 01:09:03 +00:00
/*
* Populate the outbound table.
*/
Begin the slow, painful process of cleaning up asr. This eliminates many of the bogus casts of NULL. It also distracts me while I procrastinate on useful work.
2004-04-21 03:51:39 +00:00
if (sc->ha_Msgs == NULL) {
unexpand -a everything
2002-05-14 01:09:03 +00:00
/* Allocate the reply frames */
size = sizeof(I2O_SCSI_ERROR_REPLY_MESSAGE_FRAME)
* sc->ha_Msgs_Count;
/*
* contigmalloc only works reliably at
* initialization time.
*/
if ((sc->ha_Msgs = (PI2O_SCSI_ERROR_REPLY_MESSAGE_FRAME)
Back out M_* changes, per decision of the TRB. Approved by: trb
2003-02-19 05:47:46 +00:00
contigmalloc (size, M_DEVBUF, M_WAITOK, 0ul,
Begin the slow, painful process of cleaning up asr. This eliminates many of the bogus casts of NULL. It also distracts me while I procrastinate on useful work.
2004-04-21 03:51:39 +00:00
0xFFFFFFFFul, (u_long)sizeof(U32), 0ul)) != NULL) {
More whitespace style cleanups, also remove unneeded (void *) casts for bzero().
2004-05-01 03:06:54 +00:00
bzero(sc->ha_Msgs, size);
unexpand -a everything
2002-05-14 01:09:03 +00:00
sc->ha_Msgs_Phys = KVTOPHYS(sc->ha_Msgs);
}
}
/* Initialize the outbound FIFO */
Begin the slow, painful process of cleaning up asr. This eliminates many of the bogus casts of NULL. It also distracts me while I procrastinate on useful work.
2004-04-21 03:51:39 +00:00
if (sc->ha_Msgs != NULL)
Convert the asr driver to use the bus_space API. This does not represent a significant functional change, but it further cleans up the code and brings it closer to being portable. Thanks to Don Bowman for helping to test this.
2004-05-05 03:42:33 +00:00
for(size = sc->ha_Msgs_Count, addr = sc->ha_Msgs_Phys;
size; --size) {
asr_set_FromFIFO(sc, addr);
unexpand -a everything
2002-05-14 01:09:03 +00:00
addr += sizeof(I2O_SCSI_ERROR_REPLY_MESSAGE_FRAME);
}
return (*Reply_Ptr);
}
return (0);
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
} /* ASR_initOutBound */
/*
unexpand -a everything
2002-05-14 01:09:03 +00:00
* Set the system table
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
*/
Get rid of a bunch of useless macros. STATIC becomes static, INLINE becomes __inline where appropriate and gets nuked elsewhere, IN/OUT/INOUT go away. Reformat code affected by this.
2004-05-01 01:25:05 +00:00
static int
ASR_setSysTab(Asr_softc_t *sc)
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
{
unexpand -a everything
2002-05-14 01:09:03 +00:00
PI2O_EXEC_SYS_TAB_SET_MESSAGE Message_Ptr;
PI2O_SET_SYSTAB_HEADER SystemTable;
Asr_softc_t * ha;
PI2O_SGE_SIMPLE_ELEMENT sg;
int retVal;
if ((SystemTable = (PI2O_SET_SYSTAB_HEADER)malloc (
Begin the slow, painful process of cleaning up asr. This eliminates many of the bogus casts of NULL. It also distracts me while I procrastinate on useful work.
2004-04-21 03:51:39 +00:00
sizeof(I2O_SET_SYSTAB_HEADER), M_TEMP, M_WAITOK | M_ZERO)) == NULL) {
unexpand -a everything
2002-05-14 01:09:03 +00:00
return (ENOMEM);
}
Give the asr softc list global a proper name.
2006-02-12 06:57:41 +00:00
for (ha = Asr_softc_list; ha; ha = ha->ha_next) {
unexpand -a everything
2002-05-14 01:09:03 +00:00
++SystemTable->NumberEntries;
}
if ((Message_Ptr = (PI2O_EXEC_SYS_TAB_SET_MESSAGE)malloc (
sizeof(I2O_EXEC_SYS_TAB_SET_MESSAGE) - sizeof(I2O_SG_ELEMENT)
+ ((3+SystemTable->NumberEntries) * sizeof(I2O_SGE_SIMPLE_ELEMENT)),
Begin the slow, painful process of cleaning up asr. This eliminates many of the bogus casts of NULL. It also distracts me while I procrastinate on useful work.
2004-04-21 03:51:39 +00:00
M_TEMP, M_WAITOK)) == NULL) {
Remove the defAlignLong and getAlignLong macros. I guess that the original intent was to make sure that message structs allocated off of the stack were 4-byte aligned. However, the macros as defined did absolutely nothing. And since I2O forces you to manually copy messages down to the hardware, there really is no point of enforced alignment anyways.
2004-05-02 03:33:18 +00:00
free(SystemTable, M_TEMP);
unexpand -a everything
2002-05-14 01:09:03 +00:00
return (ENOMEM);
}
Fix a typo from rev 1.56 that somehow escaped my testing.
2004-05-04 03:16:22 +00:00
(void)ASR_fillMessage((void *)Message_Ptr,
unexpand -a everything
2002-05-14 01:09:03 +00:00
sizeof(I2O_EXEC_SYS_TAB_SET_MESSAGE) - sizeof(I2O_SG_ELEMENT)
+ ((3+SystemTable->NumberEntries) * sizeof(I2O_SGE_SIMPLE_ELEMENT)));
I2O_MESSAGE_FRAME_setVersionOffset(&(Message_Ptr->StdMessageFrame),
(I2O_VERSION_11 +
(((sizeof(I2O_EXEC_SYS_TAB_SET_MESSAGE) - sizeof(I2O_SG_ELEMENT))
/ sizeof(U32)) << 4)));
I2O_MESSAGE_FRAME_setFunction(&(Message_Ptr->StdMessageFrame),
I2O_EXEC_SYS_TAB_SET);
/*
* Call the LCT table to determine the number of device entries
* to reserve space for.
* since this code is reused in several systems, code efficiency
* is greater by using a shift operation rather than a divide by
* sizeof(u_int32_t).
*/
sg = (PI2O_SGE_SIMPLE_ELEMENT)((char *)Message_Ptr
+ ((I2O_MESSAGE_FRAME_getVersionOffset(
&(Message_Ptr->StdMessageFrame)) & 0xF0) >> 2));
SG(sg, 0, I2O_SGL_FLAGS_DIR, SystemTable, sizeof(I2O_SET_SYSTAB_HEADER));
++sg;
Give the asr softc list global a proper name.
2006-02-12 06:57:41 +00:00
for (ha = Asr_softc_list; ha; ha = ha->ha_next) {
unexpand -a everything
2002-05-14 01:09:03 +00:00
SG(sg, 0,
((ha->ha_next)
? (I2O_SGL_FLAGS_DIR)
: (I2O_SGL_FLAGS_DIR | I2O_SGL_FLAGS_END_OF_BUFFER)),
&(ha->ha_SystemTable), sizeof(ha->ha_SystemTable));
++sg;
}
SG(sg, 0, I2O_SGL_FLAGS_DIR | I2O_SGL_FLAGS_END_OF_BUFFER, NULL, 0);
SG(sg, 1, I2O_SGL_FLAGS_DIR | I2O_SGL_FLAGS_LAST_ELEMENT
| I2O_SGL_FLAGS_END_OF_BUFFER, NULL, 0);
retVal = ASR_queue_c(sc, (PI2O_MESSAGE_FRAME)Message_Ptr);
Remove the defAlignLong and getAlignLong macros. I guess that the original intent was to make sure that message structs allocated off of the stack were 4-byte aligned. However, the macros as defined did absolutely nothing. And since I2O forces you to manually copy messages down to the hardware, there really is no point of enforced alignment anyways.
2004-05-02 03:33:18 +00:00
free(Message_Ptr, M_TEMP);
free(SystemTable, M_TEMP);
unexpand -a everything
2002-05-14 01:09:03 +00:00
return (retVal);
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
} /* ASR_setSysTab */
Get rid of a bunch of useless macros. STATIC becomes static, INLINE becomes __inline where appropriate and gets nuked elsewhere, IN/OUT/INOUT go away. Reformat code affected by this.
2004-05-01 01:25:05 +00:00
static int
ASR_acquireHrt(Asr_softc_t *sc)
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
{
Remove the defAlignLong and getAlignLong macros. I guess that the original intent was to make sure that message structs allocated off of the stack were 4-byte aligned. However, the macros as defined did absolutely nothing. And since I2O forces you to manually copy messages down to the hardware, there really is no point of enforced alignment anyways.
2004-05-02 03:33:18 +00:00
I2O_EXEC_HRT_GET_MESSAGE Message;
I2O_EXEC_HRT_GET_MESSAGE *Message_Ptr;
unexpand -a everything
2002-05-14 01:09:03 +00:00
struct {
I2O_HRT Header;
I2O_HRT_ENTRY Entry[MAX_CHANNEL];
Remove the defAlignLong and getAlignLong macros. I guess that the original intent was to make sure that message structs allocated off of the stack were 4-byte aligned. However, the macros as defined did absolutely nothing. And since I2O forces you to manually copy messages down to the hardware, there really is no point of enforced alignment anyways.
2004-05-02 03:33:18 +00:00
} Hrt;
u_int8_t NumberOfEntries;
PI2O_HRT_ENTRY Entry;
unexpand -a everything
2002-05-14 01:09:03 +00:00
More whitespace style cleanups, also remove unneeded (void *) casts for bzero().
2004-05-01 03:06:54 +00:00
bzero(&Hrt, sizeof (Hrt));
Remove the defAlignLong and getAlignLong macros. I guess that the original intent was to make sure that message structs allocated off of the stack were 4-byte aligned. However, the macros as defined did absolutely nothing. And since I2O forces you to manually copy messages down to the hardware, there really is no point of enforced alignment anyways.
2004-05-02 03:33:18 +00:00
Message_Ptr = (I2O_EXEC_HRT_GET_MESSAGE *)ASR_fillMessage(&Message,
unexpand -a everything
2002-05-14 01:09:03 +00:00
sizeof(I2O_EXEC_HRT_GET_MESSAGE) - sizeof(I2O_SG_ELEMENT)
+ sizeof(I2O_SGE_SIMPLE_ELEMENT));
I2O_MESSAGE_FRAME_setVersionOffset(&(Message_Ptr->StdMessageFrame),
(I2O_VERSION_11
+ (((sizeof(I2O_EXEC_HRT_GET_MESSAGE) - sizeof(I2O_SG_ELEMENT))
/ sizeof(U32)) << 4)));
I2O_MESSAGE_FRAME_setFunction (&(Message_Ptr->StdMessageFrame),
I2O_EXEC_HRT_GET);
/*
* Set up the buffers as scatter gather elements.
*/
SG(&(Message_Ptr->SGL), 0,
I2O_SGL_FLAGS_LAST_ELEMENT | I2O_SGL_FLAGS_END_OF_BUFFER,
&Hrt, sizeof(Hrt));
if (ASR_queue_c(sc, (PI2O_MESSAGE_FRAME)Message_Ptr) != CAM_REQ_CMP) {
return (ENODEV);
}
if ((NumberOfEntries = I2O_HRT_getNumberEntries(&Hrt.Header))
> (MAX_CHANNEL + 1)) {
NumberOfEntries = MAX_CHANNEL + 1;
}
for (Entry = Hrt.Header.HRTEntry;
NumberOfEntries != 0;
++Entry, --NumberOfEntries) {
PI2O_LCT_ENTRY Device;
for (Device = sc->ha_LCT->LCTEntry; Device < (PI2O_LCT_ENTRY)
(((U32 *)sc->ha_LCT)+I2O_LCT_getTableSize(sc->ha_LCT));
++Device) {
if (I2O_LCT_ENTRY_getLocalTID(Device)
== (I2O_HRT_ENTRY_getAdapterID(Entry) & 0xFFF)) {
Device->le_bus = I2O_HRT_ENTRY_getAdapterID(
Entry) >> 16;
if ((Device->le_bus > sc->ha_MaxBus)
&& (Device->le_bus <= MAX_CHANNEL)) {
sc->ha_MaxBus = Device->le_bus;
}
}
}
}
return (0);
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
} /* ASR_acquireHrt */
/*
unexpand -a everything
2002-05-14 01:09:03 +00:00
* Enable the adapter.
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
*/
Get rid of a bunch of useless macros. STATIC becomes static, INLINE becomes __inline where appropriate and gets nuked elsewhere, IN/OUT/INOUT go away. Reformat code affected by this.
2004-05-01 01:25:05 +00:00
static int
ASR_enableSys(Asr_softc_t *sc)
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
{
Remove the defAlignLong and getAlignLong macros. I guess that the original intent was to make sure that message structs allocated off of the stack were 4-byte aligned. However, the macros as defined did absolutely nothing. And since I2O forces you to manually copy messages down to the hardware, there really is no point of enforced alignment anyways.
2004-05-02 03:33:18 +00:00
I2O_EXEC_SYS_ENABLE_MESSAGE Message;
PI2O_EXEC_SYS_ENABLE_MESSAGE Message_Ptr;
unexpand -a everything
2002-05-14 01:09:03 +00:00
Remove the defAlignLong and getAlignLong macros. I guess that the original intent was to make sure that message structs allocated off of the stack were 4-byte aligned. However, the macros as defined did absolutely nothing. And since I2O forces you to manually copy messages down to the hardware, there really is no point of enforced alignment anyways.
2004-05-02 03:33:18 +00:00
Message_Ptr = (PI2O_EXEC_SYS_ENABLE_MESSAGE)ASR_fillMessage(&Message,
unexpand -a everything
2002-05-14 01:09:03 +00:00
sizeof(I2O_EXEC_SYS_ENABLE_MESSAGE));
I2O_MESSAGE_FRAME_setFunction(&(Message_Ptr->StdMessageFrame),
I2O_EXEC_SYS_ENABLE);
return (ASR_queue_c(sc, (PI2O_MESSAGE_FRAME)Message_Ptr) != 0);
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
} /* ASR_enableSys */
/*
unexpand -a everything
2002-05-14 01:09:03 +00:00
* Perform the stages necessary to initialize the adapter
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
*/
Get rid of a bunch of useless macros. STATIC becomes static, INLINE becomes __inline where appropriate and gets nuked elsewhere, IN/OUT/INOUT go away. Reformat code affected by this.
2004-05-01 01:25:05 +00:00
static int
ASR_init(Asr_softc_t *sc)
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
{
unexpand -a everything
2002-05-14 01:09:03 +00:00
return ((ASR_initOutBound(sc) == 0)
|| (ASR_setSysTab(sc) != CAM_REQ_CMP)
|| (ASR_enableSys(sc) != CAM_REQ_CMP));
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
} /* ASR_init */
/*
unexpand -a everything
2002-05-14 01:09:03 +00:00
* Send a Synchronize Cache command to the target device.
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
*/
Get rid of a bunch of useless macros. STATIC becomes static, INLINE becomes __inline where appropriate and gets nuked elsewhere, IN/OUT/INOUT go away. Reformat code affected by this.
2004-05-01 01:25:05 +00:00
static void
ASR_sync(Asr_softc_t *sc, int bus, int target, int lun)
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
{
Get rid of a bunch of useless macros. STATIC becomes static, INLINE becomes __inline where appropriate and gets nuked elsewhere, IN/OUT/INOUT go away. Reformat code affected by this.
2004-05-01 01:25:05 +00:00
tid_t TID;
unexpand -a everything
2002-05-14 01:09:03 +00:00
/*
* We will not synchronize the device when there are outstanding
* commands issued by the OS (this is due to a locked up device,
* as the OS normally would flush all outstanding commands before
* issuing a shutdown or an adapter reset).
*/
Begin the slow, painful process of cleaning up asr. This eliminates many of the bogus casts of NULL. It also distracts me while I procrastinate on useful work.
2004-04-21 03:51:39 +00:00
if ((sc != NULL)
&& (LIST_FIRST(&(sc->ha_ccb)) != NULL)
unexpand -a everything
2002-05-14 01:09:03 +00:00
&& ((TID = ASR_getTid (sc, bus, target, lun)) != (tid_t)-1)
&& (TID != (tid_t)0)) {
Remove the defAlignLong and getAlignLong macros. I guess that the original intent was to make sure that message structs allocated off of the stack were 4-byte aligned. However, the macros as defined did absolutely nothing. And since I2O forces you to manually copy messages down to the hardware, there really is no point of enforced alignment anyways.
2004-05-02 03:33:18 +00:00
PRIVATE_SCSI_SCB_EXECUTE_MESSAGE Message;
PPRIVATE_SCSI_SCB_EXECUTE_MESSAGE Message_Ptr;
unexpand -a everything
2002-05-14 01:09:03 +00:00
Remove the defAlignLong and getAlignLong macros. I guess that the original intent was to make sure that message structs allocated off of the stack were 4-byte aligned. However, the macros as defined did absolutely nothing. And since I2O forces you to manually copy messages down to the hardware, there really is no point of enforced alignment anyways.
2004-05-02 03:33:18 +00:00
Message_Ptr = (PRIVATE_SCSI_SCB_EXECUTE_MESSAGE *)&Message;
More whitespace style cleanups, also remove unneeded (void *) casts for bzero().
2004-05-01 03:06:54 +00:00
bzero(Message_Ptr, sizeof(PRIVATE_SCSI_SCB_EXECUTE_MESSAGE)
- sizeof(I2O_SG_ELEMENT) + sizeof(I2O_SGE_SIMPLE_ELEMENT));
unexpand -a everything
2002-05-14 01:09:03 +00:00
I2O_MESSAGE_FRAME_setVersionOffset(
(PI2O_MESSAGE_FRAME)Message_Ptr,
I2O_VERSION_11
| (((sizeof(PRIVATE_SCSI_SCB_EXECUTE_MESSAGE)
- sizeof(I2O_SG_ELEMENT))
/ sizeof(U32)) << 4));
I2O_MESSAGE_FRAME_setMessageSize(
(PI2O_MESSAGE_FRAME)Message_Ptr,
(sizeof(PRIVATE_SCSI_SCB_EXECUTE_MESSAGE)
- sizeof(I2O_SG_ELEMENT))
/ sizeof(U32));
I2O_MESSAGE_FRAME_setInitiatorAddress (
(PI2O_MESSAGE_FRAME)Message_Ptr, 1);
I2O_MESSAGE_FRAME_setFunction(
(PI2O_MESSAGE_FRAME)Message_Ptr, I2O_PRIVATE_MESSAGE);
I2O_MESSAGE_FRAME_setTargetAddress(
(PI2O_MESSAGE_FRAME)Message_Ptr, TID);
I2O_PRIVATE_MESSAGE_FRAME_setXFunctionCode (
(PI2O_PRIVATE_MESSAGE_FRAME)Message_Ptr,
I2O_SCSI_SCB_EXEC);
PRIVATE_SCSI_SCB_EXECUTE_MESSAGE_setTID(Message_Ptr, TID);
PRIVATE_SCSI_SCB_EXECUTE_MESSAGE_setSCBFlags (Message_Ptr,
I2O_SCB_FLAG_ENABLE_DISCONNECT
| I2O_SCB_FLAG_SIMPLE_QUEUE_TAG
| I2O_SCB_FLAG_SENSE_DATA_IN_BUFFER);
I2O_PRIVATE_MESSAGE_FRAME_setOrganizationID(
(PI2O_PRIVATE_MESSAGE_FRAME)Message_Ptr,
DPT_ORGANIZATION_ID);
PRIVATE_SCSI_SCB_EXECUTE_MESSAGE_setCDBLength(Message_Ptr, 6);
Message_Ptr->CDB[0] = SYNCHRONIZE_CACHE;
Message_Ptr->CDB[1] = (lun << 5);
PRIVATE_SCSI_SCB_EXECUTE_MESSAGE_setSCBFlags (Message_Ptr,
(I2O_SCB_FLAG_XFER_FROM_DEVICE
| I2O_SCB_FLAG_ENABLE_DISCONNECT
| I2O_SCB_FLAG_SIMPLE_QUEUE_TAG
| I2O_SCB_FLAG_SENSE_DATA_IN_BUFFER));
(void)ASR_queue_c(sc, (PI2O_MESSAGE_FRAME)Message_Ptr);
}
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
}
Get rid of a bunch of useless macros. STATIC becomes static, INLINE becomes __inline where appropriate and gets nuked elsewhere, IN/OUT/INOUT go away. Reformat code affected by this.
2004-05-01 01:25:05 +00:00
static void
ASR_synchronize(Asr_softc_t *sc)
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
{
Get rid of a bunch of useless macros. STATIC becomes static, INLINE becomes __inline where appropriate and gets nuked elsewhere, IN/OUT/INOUT go away. Reformat code affected by this.
2004-05-01 01:25:05 +00:00
int bus, target, lun;
unexpand -a everything
2002-05-14 01:09:03 +00:00
for (bus = 0; bus <= sc->ha_MaxBus; ++bus) {
for (target = 0; target <= sc->ha_MaxId; ++target) {
for (lun = 0; lun <= sc->ha_MaxLun; ++lun) {
ASR_sync(sc,bus,target,lun);
}
}
}
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
}
/*
unexpand -a everything
2002-05-14 01:09:03 +00:00
* Reset the HBA, targets and BUS.
* Currently this resets *all* the SCSI busses.
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
*/
Get rid of a bunch of useless macros. STATIC becomes static, INLINE becomes __inline where appropriate and gets nuked elsewhere, IN/OUT/INOUT go away. Reformat code affected by this.
2004-05-01 01:25:05 +00:00
static __inline void
asr_hbareset(Asr_softc_t *sc)
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
{
Get rid of a bunch of useless macros. STATIC becomes static, INLINE becomes __inline where appropriate and gets nuked elsewhere, IN/OUT/INOUT go away. Reformat code affected by this.
2004-05-01 01:25:05 +00:00
ASR_synchronize(sc);
(void)ASR_reset(sc);
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
} /* asr_hbareset */
/*
unexpand -a everything
2002-05-14 01:09:03 +00:00
* A reduced copy of the real pci_map_mem, incorporating the MAX_MAP
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
* limit and a reduction in error checking (in the pre 4.0 case).
*/
Get rid of a bunch of useless macros. STATIC becomes static, INLINE becomes __inline where appropriate and gets nuked elsewhere, IN/OUT/INOUT go away. Reformat code affected by this.
2004-05-01 01:25:05 +00:00
static int
Rename driver_t variables. Use device_printf() instead of homerolled printf. Use the provided softc instead of rolling our own.
2006-01-14 15:51:42 +00:00
asr_pci_map_mem(device_t dev, Asr_softc_t *sc)
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
{
Get rid of a bunch of useless macros. STATIC becomes static, INLINE becomes __inline where appropriate and gets nuked elsewhere, IN/OUT/INOUT go away. Reformat code affected by this.
2004-05-01 01:25:05 +00:00
int rid;
u_int32_t p, l, s;
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
unexpand -a everything
2002-05-14 01:09:03 +00:00
/*
* I2O specification says we must find first *memory* mapped BAR
*/
Use PCIR_BAR(x) instead of PCIR_MAPS. Glanced over by: imp, gibbs Tested by: i386 LINT
2003-09-02 17:30:40 +00:00
for (rid = 0; rid < 4; rid++) {
Rename driver_t variables. Use device_printf() instead of homerolled printf. Use the provided softc instead of rolling our own.
2006-01-14 15:51:42 +00:00
p = pci_read_config(dev, PCIR_BAR(rid), sizeof(p));
unexpand -a everything
2002-05-14 01:09:03 +00:00
if ((p & 1) == 0) {
break;
}
}
/*
* Give up?
*/
Use PCIR_BAR(x) instead of PCIR_MAPS. Glanced over by: imp, gibbs Tested by: i386 LINT
2003-09-02 17:30:40 +00:00
if (rid >= 4) {
rid = 0;
unexpand -a everything
2002-05-14 01:09:03 +00:00
}
Use PCIR_BAR(x) instead of PCIR_MAPS. Glanced over by: imp, gibbs Tested by: i386 LINT
2003-09-02 17:30:40 +00:00
rid = PCIR_BAR(rid);
Rename driver_t variables. Use device_printf() instead of homerolled printf. Use the provided softc instead of rolling our own.
2006-01-14 15:51:42 +00:00
p = pci_read_config(dev, rid, sizeof(p));
pci_write_config(dev, rid, -1, sizeof(p));
l = 0 - (pci_read_config(dev, rid, sizeof(l)) & ~15);
pci_write_config(dev, rid, p, sizeof(p));
unexpand -a everything
2002-05-14 01:09:03 +00:00
if (l > MAX_MAP) {
l = MAX_MAP;
}
/*
* The 2005S Zero Channel RAID solution is not a perfect PCI
* citizen. It asks for 4MB on BAR0, and 0MB on BAR1, once
* enabled it rewrites the size of BAR0 to 2MB, sets BAR1 to
* BAR0+2MB and sets it's size to 2MB. The IOP registers are
* accessible via BAR0, the messaging registers are accessible
* via BAR1. If the subdevice code is 50 to 59 decimal.
*/
Rename driver_t variables. Use device_printf() instead of homerolled printf. Use the provided softc instead of rolling our own.
2006-01-14 15:51:42 +00:00
s = pci_read_config(dev, PCIR_DEVVENDOR, sizeof(s));
unexpand -a everything
2002-05-14 01:09:03 +00:00
if (s != 0xA5111044) {
Rename driver_t variables. Use device_printf() instead of homerolled printf. Use the provided softc instead of rolling our own.
2006-01-14 15:51:42 +00:00
s = pci_read_config(dev, PCIR_SUBVEND_0, sizeof(s));
unexpand -a everything
2002-05-14 01:09:03 +00:00
if ((((ADPTDOMINATOR_SUB_ID_START ^ s) & 0xF000FFFF) == 0)
&& (ADPTDOMINATOR_SUB_ID_START <= s)
&& (s <= ADPTDOMINATOR_SUB_ID_END)) {
l = MAX_MAP; /* Conjoined BAR Raptor Daptor */
}
}
p &= ~15;
Rename driver_t variables. Use device_printf() instead of homerolled printf. Use the provided softc instead of rolling our own.
2006-01-14 15:51:42 +00:00
sc->ha_mem_res = bus_alloc_resource(dev, SYS_RES_MEMORY, &rid,
unexpand -a everything
2002-05-14 01:09:03 +00:00
p, p + l, l, RF_ACTIVE);
Begin the slow, painful process of cleaning up asr. This eliminates many of the bogus casts of NULL. It also distracts me while I procrastinate on useful work.
2004-04-21 03:51:39 +00:00
if (sc->ha_mem_res == NULL) {
unexpand -a everything
2002-05-14 01:09:03 +00:00
return (0);
}
Convert the asr driver to use the bus_space API. This does not represent a significant functional change, but it further cleans up the code and brings it closer to being portable. Thanks to Don Bowman for helping to test this.
2004-05-05 03:42:33 +00:00
sc->ha_Base = rman_get_start(sc->ha_mem_res);
sc->ha_i2o_bhandle = rman_get_bushandle(sc->ha_mem_res);
sc->ha_i2o_btag = rman_get_bustag(sc->ha_mem_res);
unexpand -a everything
2002-05-14 01:09:03 +00:00
if (s == 0xA5111044) { /* Split BAR Raptor Daptor */
Use PCIR_BAR(x) instead of PCIR_MAPS. Glanced over by: imp, gibbs Tested by: i386 LINT
2003-09-02 17:30:40 +00:00
if ((rid += sizeof(u_int32_t)) >= PCIR_BAR(4)) {
unexpand -a everything
2002-05-14 01:09:03 +00:00
return (0);
}
Rename driver_t variables. Use device_printf() instead of homerolled printf. Use the provided softc instead of rolling our own.
2006-01-14 15:51:42 +00:00
p = pci_read_config(dev, rid, sizeof(p));
pci_write_config(dev, rid, -1, sizeof(p));
l = 0 - (pci_read_config(dev, rid, sizeof(l)) & ~15);
pci_write_config(dev, rid, p, sizeof(p));
unexpand -a everything
2002-05-14 01:09:03 +00:00
if (l > MAX_MAP) {
l = MAX_MAP;
}
p &= ~15;
Rename driver_t variables. Use device_printf() instead of homerolled printf. Use the provided softc instead of rolling our own.
2006-01-14 15:51:42 +00:00
sc->ha_mes_res = bus_alloc_resource(dev, SYS_RES_MEMORY, &rid,
unexpand -a everything
2002-05-14 01:09:03 +00:00
p, p + l, l, RF_ACTIVE);
Begin the slow, painful process of cleaning up asr. This eliminates many of the bogus casts of NULL. It also distracts me while I procrastinate on useful work.
2004-04-21 03:51:39 +00:00
if (sc->ha_mes_res == NULL) {
unexpand -a everything
2002-05-14 01:09:03 +00:00
return (0);
}
Convert the asr driver to use the bus_space API. This does not represent a significant functional change, but it further cleans up the code and brings it closer to being portable. Thanks to Don Bowman for helping to test this.
2004-05-05 03:42:33 +00:00
sc->ha_frame_bhandle = rman_get_bushandle(sc->ha_mes_res);
sc->ha_frame_btag = rman_get_bustag(sc->ha_mes_res);
unexpand -a everything
2002-05-14 01:09:03 +00:00
} else {
Convert the asr driver to use the bus_space API. This does not represent a significant functional change, but it further cleans up the code and brings it closer to being portable. Thanks to Don Bowman for helping to test this.
2004-05-05 03:42:33 +00:00
sc->ha_frame_bhandle = sc->ha_i2o_bhandle;
sc->ha_frame_btag = sc->ha_i2o_btag;
unexpand -a everything
2002-05-14 01:09:03 +00:00
}
return (1);
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
} /* asr_pci_map_mem */
/*
unexpand -a everything
2002-05-14 01:09:03 +00:00
* A simplified copy of the real pci_map_int with additional
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
* registration requirements.
*/
Get rid of a bunch of useless macros. STATIC becomes static, INLINE becomes __inline where appropriate and gets nuked elsewhere, IN/OUT/INOUT go away. Reformat code affected by this.
2004-05-01 01:25:05 +00:00
static int
Rename driver_t variables. Use device_printf() instead of homerolled printf. Use the provided softc instead of rolling our own.
2006-01-14 15:51:42 +00:00
asr_pci_map_int(device_t dev, Asr_softc_t *sc)
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
{
Get rid of a bunch of useless macros. STATIC becomes static, INLINE becomes __inline where appropriate and gets nuked elsewhere, IN/OUT/INOUT go away. Reformat code affected by this.
2004-05-01 01:25:05 +00:00
int rid = 0;
unexpand -a everything
2002-05-14 01:09:03 +00:00
Rename driver_t variables. Use device_printf() instead of homerolled printf. Use the provided softc instead of rolling our own.
2006-01-14 15:51:42 +00:00
sc->ha_irq_res = bus_alloc_resource_any(dev, SYS_RES_IRQ, &rid,
Convert callers to the new bus_alloc_resource_any(9) API. Submitted by: Mark Santcroos <marks@ripe.net> Reviewed by: imp, dfr, bde
2004-03-17 17:50:55 +00:00
RF_ACTIVE | RF_SHAREABLE);
Begin the slow, painful process of cleaning up asr. This eliminates many of the bogus casts of NULL. It also distracts me while I procrastinate on useful work.
2004-04-21 03:51:39 +00:00
if (sc->ha_irq_res == NULL) {
unexpand -a everything
2002-05-14 01:09:03 +00:00
return (0);
}
Rename driver_t variables. Use device_printf() instead of homerolled printf. Use the provided softc instead of rolling our own.
2006-01-14 15:51:42 +00:00
if (bus_setup_intr(dev, sc->ha_irq_res, INTR_TYPE_CAM | INTR_ENTROPY,
o break newbus api: add a new argument of type driver_filter_t to bus_setup_intr() o add an int return code to all fast handlers o retire INTR_FAST/IH_FAST For more info: http://docs.freebsd.org/cgi/getmsg.cgi?fetch=465712+0+current/freebsd-current Reviewed by: many Approved by: re@
2007-02-23 12:19:07 +00:00
NULL, (driver_intr_t *)asr_intr, (void *)sc, &(sc->ha_intr))) {
unexpand -a everything
2002-05-14 01:09:03 +00:00
return (0);
}
Rename driver_t variables. Use device_printf() instead of homerolled printf. Use the provided softc instead of rolling our own.
2006-01-14 15:51:42 +00:00
sc->ha_irq = pci_read_config(dev, PCIR_INTLINE, sizeof(char));
unexpand -a everything
2002-05-14 01:09:03 +00:00
return (1);
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
} /* asr_pci_map_int */
Add the start of busdma infrastructure to this driver. Convert the ASR_resetIOPR function to use it.
2006-02-04 08:45:19 +00:00
static void
asr_status_cb(void *arg, bus_dma_segment_t *segs, int nseg, int error)
{
Asr_softc_t *sc;
if (error)
return;
sc = (Asr_softc_t *)arg;
/* XXX
* The status word can be at a 64-bit address, but the existing
* accessor macros simply cannot manipulate 64-bit addresses.
*/
Squash another use of vtophys. Instead of creating separate busdma objects for doing static memory transfers, start collecting them into a single object.
2006-02-04 17:56:17 +00:00
sc->ha_status_phys = (u_int32_t)segs[0].ds_addr +
offsetof(struct Asr_status_mem, status);
sc->ha_rstatus_phys = (u_int32_t)segs[0].ds_addr +
offsetof(struct Asr_status_mem, rstatus);
Add the start of busdma infrastructure to this driver. Convert the ASR_resetIOPR function to use it.
2006-02-04 08:45:19 +00:00
}
static int
asr_alloc_dma(Asr_softc_t *sc)
{
device_t dev;
dev = sc->ha_dev;
if (bus_dma_tag_create(NULL, /* parent */
1, 0, /* algnmnt, boundary */
BUS_SPACE_MAXADDR_32BIT, /* lowaddr */
BUS_SPACE_MAXADDR, /* highaddr */
NULL, NULL, /* filter, filterarg */
BUS_SPACE_MAXSIZE_32BIT, /* maxsize */
BUS_SPACE_UNRESTRICTED, /* nsegments */
BUS_SPACE_MAXSIZE_32BIT, /* maxsegsize */
0, /* flags */
NULL, NULL, /* lockfunc, lockarg */
&sc->ha_parent_dmat)) {
device_printf(dev, "Cannot allocate parent DMA tag\n");
return (ENOMEM);
}
if (bus_dma_tag_create(sc->ha_parent_dmat, /* parent */
1, 0, /* algnmnt, boundary */
BUS_SPACE_MAXADDR_32BIT, /* lowaddr */
BUS_SPACE_MAXADDR, /* highaddr */
NULL, NULL, /* filter, filterarg */
Squash another use of vtophys. Instead of creating separate busdma objects for doing static memory transfers, start collecting them into a single object.
2006-02-04 17:56:17 +00:00
sizeof(sc->ha_statusmem),/* maxsize */
Add the start of busdma infrastructure to this driver. Convert the ASR_resetIOPR function to use it.
2006-02-04 08:45:19 +00:00
1, /* nsegments */
Squash another use of vtophys. Instead of creating separate busdma objects for doing static memory transfers, start collecting them into a single object.
2006-02-04 17:56:17 +00:00
sizeof(sc->ha_statusmem),/* maxsegsize */
Add the start of busdma infrastructure to this driver. Convert the ASR_resetIOPR function to use it.
2006-02-04 08:45:19 +00:00
0, /* flags */
NULL, NULL, /* lockfunc, lockarg */
Squash another use of vtophys. Instead of creating separate busdma objects for doing static memory transfers, start collecting them into a single object.
2006-02-04 17:56:17 +00:00
&sc->ha_statusmem_dmat)) {
Add the start of busdma infrastructure to this driver. Convert the ASR_resetIOPR function to use it.
2006-02-04 08:45:19 +00:00
device_printf(dev, "Cannot allocate status DMA tag\n");
bus_dma_tag_destroy(sc->ha_parent_dmat);
return (ENOMEM);
}
Squash another use of vtophys. Instead of creating separate busdma objects for doing static memory transfers, start collecting them into a single object.
2006-02-04 17:56:17 +00:00
if (bus_dmamem_alloc(sc->ha_statusmem_dmat, (void **)&sc->ha_statusmem,
BUS_DMA_NOWAIT, &sc->ha_statusmem_dmamap)) {
Add the start of busdma infrastructure to this driver. Convert the ASR_resetIOPR function to use it.
2006-02-04 08:45:19 +00:00
device_printf(dev, "Cannot allocate status memory\n");
Squash another use of vtophys. Instead of creating separate busdma objects for doing static memory transfers, start collecting them into a single object.
2006-02-04 17:56:17 +00:00
bus_dma_tag_destroy(sc->ha_statusmem_dmat);
Add the start of busdma infrastructure to this driver. Convert the ASR_resetIOPR function to use it.
2006-02-04 08:45:19 +00:00
bus_dma_tag_destroy(sc->ha_parent_dmat);
return (ENOMEM);
}
Squash another use of vtophys. Instead of creating separate busdma objects for doing static memory transfers, start collecting them into a single object.
2006-02-04 17:56:17 +00:00
(void)bus_dmamap_load(sc->ha_statusmem_dmat, sc->ha_statusmem_dmamap,
sc->ha_statusmem, sizeof(sc->ha_statusmem), asr_status_cb, sc, 0);
Add the start of busdma infrastructure to this driver. Convert the ASR_resetIOPR function to use it.
2006-02-04 08:45:19 +00:00
return (0);
}
static void
asr_release_dma(Asr_softc_t *sc)
{
Squash another use of vtophys. Instead of creating separate busdma objects for doing static memory transfers, start collecting them into a single object.
2006-02-04 17:56:17 +00:00
if (sc->ha_rstatus_phys != 0)
bus_dmamap_unload(sc->ha_statusmem_dmat,
sc->ha_statusmem_dmamap);
if (sc->ha_statusmem != NULL)
bus_dmamem_free(sc->ha_statusmem_dmat, sc->ha_statusmem,
sc->ha_statusmem_dmamap);
if (sc->ha_statusmem_dmat != NULL)
bus_dma_tag_destroy(sc->ha_statusmem_dmat);
Add the start of busdma infrastructure to this driver. Convert the ASR_resetIOPR function to use it.
2006-02-04 08:45:19 +00:00
if (sc->ha_parent_dmat != NULL)
bus_dma_tag_destroy(sc->ha_parent_dmat);
}
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
/*
unexpand -a everything
2002-05-14 01:09:03 +00:00
* Attach the devices, and virtual devices to the driver list.
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
*/
Remove the DOMINO and MODE0 device attachments. They never did anything. Remove a bunch of obfuscating macros.
2004-05-01 05:19:39 +00:00
static int
Rename driver_t variables. Use device_printf() instead of homerolled printf. Use the provided softc instead of rolling our own.
2006-01-14 15:51:42 +00:00
asr_attach(device_t dev)
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
{
Re-indent some silly sub-blocks in asr_attach().
2004-05-01 06:12:58 +00:00
PI2O_EXEC_STATUS_GET_REPLY status;
PI2O_LCT_ENTRY Device;
Remove ASR_get_sc() and reference the softc in the dev_t. For some nefarious reason, the I2O protocol requires knowledge of all I2O devices in the system, so we can't get rid of the evil linked-list of softc's yet.
2004-05-01 05:56:57 +00:00
Asr_softc_t *sc, **ha;
Get rid of a bunch of useless macros. STATIC becomes static, INLINE becomes __inline where appropriate and gets nuked elsewhere, IN/OUT/INOUT go away. Reformat code affected by this.
2004-05-01 01:25:05 +00:00
struct scsi_inquiry_data *iq;
Rename driver_t variables. Use device_printf() instead of homerolled printf. Use the provided softc instead of rolling our own.
2006-01-14 15:51:42 +00:00
int bus, size, unit;
Add the start of busdma infrastructure to this driver. Convert the ASR_resetIOPR function to use it.
2006-02-04 08:45:19 +00:00
int error;
Rename driver_t variables. Use device_printf() instead of homerolled printf. Use the provided softc instead of rolling our own.
2006-01-14 15:51:42 +00:00
sc = device_get_softc(dev);
unit = device_get_unit(dev);
Add the start of busdma infrastructure to this driver. Convert the ASR_resetIOPR function to use it.
2006-02-04 08:45:19 +00:00
sc->ha_dev = dev;
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
Give the asr softc list global a proper name.
2006-02-12 06:57:41 +00:00
if (Asr_softc_list == NULL) {
unexpand -a everything
2002-05-14 01:09:03 +00:00
/*
* Fixup the OS revision as saved in the dptsig for the
* engine (dptioctl.h) to pick up.
*/
Many more style cleanups. Switch complex macros to being inline functions. Put @includes in a better spot. Fix many cases of 2 space indents and spaces between a function name and the parens. Use KASSERT instead of a home-rolled ASSERT. Remove some undeeded caddr casts.
2004-05-01 02:27:06 +00:00
bcopy(osrelease, &ASR_sig.dsDescription[16], 5);
unexpand -a everything
2002-05-14 01:09:03 +00:00
}
/*
* Initialize the software structure
*/
LIST_INIT(&(sc->ha_ccb));
/* Link us into the HA list */
Give the asr softc list global a proper name.
2006-02-12 06:57:41 +00:00
for (ha = &Asr_softc_list; *ha; ha = &((*ha)->ha_next));
unexpand -a everything
2002-05-14 01:09:03 +00:00
*(ha) = sc;
Re-indent some silly sub-blocks in asr_attach().
2004-05-01 06:12:58 +00:00
/*
* This is the real McCoy!
*/
Rename driver_t variables. Use device_printf() instead of homerolled printf. Use the provided softc instead of rolling our own.
2006-01-14 15:51:42 +00:00
if (!asr_pci_map_mem(dev, sc)) {
device_printf(dev, "could not map memory\n");
Re-indent some silly sub-blocks in asr_attach().
2004-05-01 06:12:58 +00:00
return(ENXIO);
}
/* Enable if not formerly enabled */
Rename driver_t variables. Use device_printf() instead of homerolled printf. Use the provided softc instead of rolling our own.
2006-01-14 15:51:42 +00:00
pci_write_config(dev, PCIR_COMMAND,
pci_read_config(dev, PCIR_COMMAND, sizeof(char)) |
Re-indent some silly sub-blocks in asr_attach().
2004-05-01 06:12:58 +00:00
PCIM_CMD_MEMEN | PCIM_CMD_BUSMASTEREN, sizeof(char));
Replace bogus code with real code.
2006-01-14 15:59:54 +00:00
sc->ha_pciBusNum = pci_get_bus(dev);
sc->ha_pciDeviceNum = (pci_get_slot(dev) << 3) | pci_get_function(dev);
Add the start of busdma infrastructure to this driver. Convert the ASR_resetIOPR function to use it.
2006-02-04 08:45:19 +00:00
if ((error = asr_alloc_dma(sc)) != 0)
return (error);
Re-indent some silly sub-blocks in asr_attach().
2004-05-01 06:12:58 +00:00
/* Check if the device is there? */
Fix a possible memory leak in asr_attach.
2006-02-04 06:08:19 +00:00
if (ASR_resetIOP(sc) == 0) {
device_printf(dev, "Cannot reset adapter\n");
Add the start of busdma infrastructure to this driver. Convert the ASR_resetIOPR function to use it.
2006-02-04 08:45:19 +00:00
asr_release_dma(sc);
Fix a possible memory leak in asr_attach.
2006-02-04 06:08:19 +00:00
return (EIO);
}
Squash another use of vtophys. Instead of creating separate busdma objects for doing static memory transfers, start collecting them into a single object.
2006-02-04 17:56:17 +00:00
status = &sc->ha_statusmem->status;
if (ASR_getStatus(sc) == NULL) {
Rename driver_t variables. Use device_printf() instead of homerolled printf. Use the provided softc instead of rolling our own.
2006-01-14 15:51:42 +00:00
device_printf(dev, "could not initialize hardware\n");
Add the start of busdma infrastructure to this driver. Convert the ASR_resetIOPR function to use it.
2006-02-04 08:45:19 +00:00
asr_release_dma(sc);
Fix a possible memory leak in asr_attach.
2006-02-04 06:08:19 +00:00
return(ENODEV);
Re-indent some silly sub-blocks in asr_attach().
2004-05-01 06:12:58 +00:00
}
sc->ha_SystemTable.OrganizationID = status->OrganizationID;
sc->ha_SystemTable.IOP_ID = status->IOP_ID;
sc->ha_SystemTable.I2oVersion = status->I2oVersion;
sc->ha_SystemTable.IopState = status->IopState;
sc->ha_SystemTable.MessengerType = status->MessengerType;
sc->ha_SystemTable.InboundMessageFrameSize = status->InboundMFrameSize;
sc->ha_SystemTable.MessengerInfo.InboundMessagePortAddressLow =
Convert the asr driver to use the bus_space API. This does not represent a significant functional change, but it further cleans up the code and brings it closer to being portable. Thanks to Don Bowman for helping to test this.
2004-05-05 03:42:33 +00:00
(U32)(sc->ha_Base + I2O_REG_TOFIFO); /* XXX 64-bit */
Re-indent some silly sub-blocks in asr_attach().
2004-05-01 06:12:58 +00:00
Rename driver_t variables. Use device_printf() instead of homerolled printf. Use the provided softc instead of rolling our own.
2006-01-14 15:51:42 +00:00
if (!asr_pci_map_int(dev, (void *)sc)) {
device_printf(dev, "could not map interrupt\n");
Add the start of busdma infrastructure to this driver. Convert the ASR_resetIOPR function to use it.
2006-02-04 08:45:19 +00:00
asr_release_dma(sc);
Re-indent some silly sub-blocks in asr_attach().
2004-05-01 06:12:58 +00:00
return(ENXIO);
}
unexpand -a everything
2002-05-14 01:09:03 +00:00
Re-indent some silly sub-blocks in asr_attach().
2004-05-01 06:12:58 +00:00
/* Adjust the maximim inbound count */
if (((sc->ha_QueueSize =
I2O_EXEC_STATUS_GET_REPLY_getMaxInboundMFrames(status)) >
MAX_INBOUND) || (sc->ha_QueueSize == 0)) {
sc->ha_QueueSize = MAX_INBOUND;
}
unexpand -a everything
2002-05-14 01:09:03 +00:00
Re-indent some silly sub-blocks in asr_attach().
2004-05-01 06:12:58 +00:00
/* Adjust the maximum outbound count */
if (((sc->ha_Msgs_Count =
I2O_EXEC_STATUS_GET_REPLY_getMaxOutboundMFrames(status)) >
MAX_OUTBOUND) || (sc->ha_Msgs_Count == 0)) {
sc->ha_Msgs_Count = MAX_OUTBOUND;
}
if (sc->ha_Msgs_Count > sc->ha_QueueSize) {
sc->ha_Msgs_Count = sc->ha_QueueSize;
}
unexpand -a everything
2002-05-14 01:09:03 +00:00
Re-indent some silly sub-blocks in asr_attach().
2004-05-01 06:12:58 +00:00
/* Adjust the maximum SG size to adapter */
if ((size = (I2O_EXEC_STATUS_GET_REPLY_getInboundMFrameSize(status) <<
2)) > MAX_INBOUND_SIZE) {
size = MAX_INBOUND_SIZE;
unexpand -a everything
2002-05-14 01:09:03 +00:00
}
Re-indent some silly sub-blocks in asr_attach().
2004-05-01 06:12:58 +00:00
sc->ha_SgSize = (size - sizeof(PRIVATE_SCSI_SCB_EXECUTE_MESSAGE)
+ sizeof(I2O_SG_ELEMENT)) / sizeof(I2O_SGE_SIMPLE_ELEMENT);
unexpand -a everything
2002-05-14 01:09:03 +00:00
/*
* Only do a bus/HBA reset on the first time through. On this
* first time through, we do not send a flush to the devices.
*/
if (ASR_init(sc) == 0) {
struct BufferInfo {
I2O_PARAM_RESULTS_LIST_HEADER Header;
I2O_PARAM_READ_OPERATION_RESULT Read;
I2O_DPT_EXEC_IOP_BUFFERS_SCALAR Info;
Remove the defAlignLong and getAlignLong macros. I guess that the original intent was to make sure that message structs allocated off of the stack were 4-byte aligned. However, the macros as defined did absolutely nothing. And since I2O forces you to manually copy messages down to the hardware, there really is no point of enforced alignment anyways.
2004-05-02 03:33:18 +00:00
} Buffer;
unexpand -a everything
2002-05-14 01:09:03 +00:00
PI2O_DPT_EXEC_IOP_BUFFERS_SCALAR Info;
Don't indent preprocessor tokens.
2004-04-21 20:56:08 +00:00
#define FW_DEBUG_BLED_OFFSET 8
unexpand -a everything
2002-05-14 01:09:03 +00:00
if ((Info = (PI2O_DPT_EXEC_IOP_BUFFERS_SCALAR)
Convert the asr driver to use the bus_space API. This does not represent a significant functional change, but it further cleans up the code and brings it closer to being portable. Thanks to Don Bowman for helping to test this.
2004-05-05 03:42:33 +00:00
ASR_getParams(sc, 0, I2O_DPT_EXEC_IOP_BUFFERS_GROUP_NO,
Remove the defAlignLong and getAlignLong macros. I guess that the original intent was to make sure that message structs allocated off of the stack were 4-byte aligned. However, the macros as defined did absolutely nothing. And since I2O forces you to manually copy messages down to the hardware, there really is no point of enforced alignment anyways.
2004-05-02 03:33:18 +00:00
&Buffer, sizeof(struct BufferInfo))) != NULL) {
Convert the asr driver to use the bus_space API. This does not represent a significant functional change, but it further cleans up the code and brings it closer to being portable. Thanks to Don Bowman for helping to test this.
2004-05-05 03:42:33 +00:00
sc->ha_blinkLED = FW_DEBUG_BLED_OFFSET +
I2O_DPT_EXEC_IOP_BUFFERS_SCALAR_getSerialOutputOffset(Info);
unexpand -a everything
2002-05-14 01:09:03 +00:00
}
if (ASR_acquireLct(sc) == 0) {
(void)ASR_acquireHrt(sc);
}
} else {
Rename driver_t variables. Use device_printf() instead of homerolled printf. Use the provided softc instead of rolling our own.
2006-01-14 15:51:42 +00:00
device_printf(dev, "failed to initialize\n");
Add the start of busdma infrastructure to this driver. Convert the ASR_resetIOPR function to use it.
2006-02-04 08:45:19 +00:00
asr_release_dma(sc);
Remove the DOMINO and MODE0 device attachments. They never did anything. Remove a bunch of obfuscating macros.
2004-05-01 05:19:39 +00:00
return(ENXIO);
unexpand -a everything
2002-05-14 01:09:03 +00:00
}
/*
* Add in additional probe responses for more channels. We
* are reusing the variable `target' for a channel loop counter.
* Done here because of we need both the acquireLct and
* acquireHrt data.
*/
Re-indent some silly sub-blocks in asr_attach().
2004-05-01 06:12:58 +00:00
for (Device = sc->ha_LCT->LCTEntry; Device < (PI2O_LCT_ENTRY)
(((U32 *)sc->ha_LCT)+I2O_LCT_getTableSize(sc->ha_LCT)); ++Device) {
if (Device->le_type == I2O_UNKNOWN) {
continue;
}
if (I2O_LCT_ENTRY_getUserTID(Device) == 0xFFF) {
if (Device->le_target > sc->ha_MaxId) {
sc->ha_MaxId = Device->le_target;
unexpand -a everything
2002-05-14 01:09:03 +00:00
}
Re-indent some silly sub-blocks in asr_attach().
2004-05-01 06:12:58 +00:00
if (Device->le_lun > sc->ha_MaxLun) {
sc->ha_MaxLun = Device->le_lun;
unexpand -a everything
2002-05-14 01:09:03 +00:00
}
}
Re-indent some silly sub-blocks in asr_attach().
2004-05-01 06:12:58 +00:00
if (((Device->le_type & I2O_PORT) != 0)
&& (Device->le_bus <= MAX_CHANNEL)) {
/* Do not increase MaxId for efficiency */
sc->ha_adapter_target[Device->le_bus] =
Device->le_target;
}
unexpand -a everything
2002-05-14 01:09:03 +00:00
}
/*
* Print the HBA model number as inquired from the card.
*/
Rename driver_t variables. Use device_printf() instead of homerolled printf. Use the provided softc instead of rolling our own.
2006-01-14 15:51:42 +00:00
device_printf(dev, " ");
unexpand -a everything
2002-05-14 01:09:03 +00:00
Re-indent some silly sub-blocks in asr_attach().
2004-05-01 06:12:58 +00:00
if ((iq = (struct scsi_inquiry_data *)malloc(
sizeof(struct scsi_inquiry_data), M_TEMP, M_WAITOK | M_ZERO)) !=
NULL) {
Remove the defAlignLong and getAlignLong macros. I guess that the original intent was to make sure that message structs allocated off of the stack were 4-byte aligned. However, the macros as defined did absolutely nothing. And since I2O forces you to manually copy messages down to the hardware, there really is no point of enforced alignment anyways.
2004-05-02 03:33:18 +00:00
PRIVATE_SCSI_SCB_EXECUTE_MESSAGE Message;
PPRIVATE_SCSI_SCB_EXECUTE_MESSAGE Message_Ptr;
int posted = 0;
unexpand -a everything
2002-05-14 01:09:03 +00:00
Remove the defAlignLong and getAlignLong macros. I guess that the original intent was to make sure that message structs allocated off of the stack were 4-byte aligned. However, the macros as defined did absolutely nothing. And since I2O forces you to manually copy messages down to the hardware, there really is no point of enforced alignment anyways.
2004-05-02 03:33:18 +00:00
Message_Ptr = (PRIVATE_SCSI_SCB_EXECUTE_MESSAGE *)&Message;
More whitespace style cleanups, also remove unneeded (void *) casts for bzero().
2004-05-01 03:06:54 +00:00
bzero(Message_Ptr, sizeof(PRIVATE_SCSI_SCB_EXECUTE_MESSAGE) -
sizeof(I2O_SG_ELEMENT) + sizeof(I2O_SGE_SIMPLE_ELEMENT));
unexpand -a everything
2002-05-14 01:09:03 +00:00
I2O_MESSAGE_FRAME_setVersionOffset(
More whitespace style cleanups, also remove unneeded (void *) casts for bzero().
2004-05-01 03:06:54 +00:00
(PI2O_MESSAGE_FRAME)Message_Ptr, I2O_VERSION_11 |
(((sizeof(PRIVATE_SCSI_SCB_EXECUTE_MESSAGE)
- sizeof(I2O_SG_ELEMENT)) / sizeof(U32)) << 4));
unexpand -a everything
2002-05-14 01:09:03 +00:00
I2O_MESSAGE_FRAME_setMessageSize(
More whitespace style cleanups, also remove unneeded (void *) casts for bzero().
2004-05-01 03:06:54 +00:00
(PI2O_MESSAGE_FRAME)Message_Ptr,
(sizeof(PRIVATE_SCSI_SCB_EXECUTE_MESSAGE) -
sizeof(I2O_SG_ELEMENT) + sizeof(I2O_SGE_SIMPLE_ELEMENT)) /
sizeof(U32));
I2O_MESSAGE_FRAME_setInitiatorAddress(
(PI2O_MESSAGE_FRAME)Message_Ptr, 1);
unexpand -a everything
2002-05-14 01:09:03 +00:00
I2O_MESSAGE_FRAME_setFunction(
More whitespace style cleanups, also remove unneeded (void *) casts for bzero().
2004-05-01 03:06:54 +00:00
(PI2O_MESSAGE_FRAME)Message_Ptr, I2O_PRIVATE_MESSAGE);
I2O_PRIVATE_MESSAGE_FRAME_setXFunctionCode(
(PI2O_PRIVATE_MESSAGE_FRAME)Message_Ptr, I2O_SCSI_SCB_EXEC);
unexpand -a everything
2002-05-14 01:09:03 +00:00
PRIVATE_SCSI_SCB_EXECUTE_MESSAGE_setSCBFlags (Message_Ptr,
I2O_SCB_FLAG_ENABLE_DISCONNECT
| I2O_SCB_FLAG_SIMPLE_QUEUE_TAG
| I2O_SCB_FLAG_SENSE_DATA_IN_BUFFER);
PRIVATE_SCSI_SCB_EXECUTE_MESSAGE_setInterpret(Message_Ptr, 1);
I2O_PRIVATE_MESSAGE_FRAME_setOrganizationID(
More whitespace style cleanups, also remove unneeded (void *) casts for bzero().
2004-05-01 03:06:54 +00:00
(PI2O_PRIVATE_MESSAGE_FRAME)Message_Ptr,
DPT_ORGANIZATION_ID);
unexpand -a everything
2002-05-14 01:09:03 +00:00
PRIVATE_SCSI_SCB_EXECUTE_MESSAGE_setCDBLength(Message_Ptr, 6);
Message_Ptr->CDB[0] = INQUIRY;
Re-indent some silly sub-blocks in asr_attach().
2004-05-01 06:12:58 +00:00
Message_Ptr->CDB[4] =
(unsigned char)sizeof(struct scsi_inquiry_data);
unexpand -a everything
2002-05-14 01:09:03 +00:00
if (Message_Ptr->CDB[4] == 0) {
Message_Ptr->CDB[4] = 255;
}
PRIVATE_SCSI_SCB_EXECUTE_MESSAGE_setSCBFlags (Message_Ptr,
(I2O_SCB_FLAG_XFER_FROM_DEVICE
| I2O_SCB_FLAG_ENABLE_DISCONNECT
| I2O_SCB_FLAG_SIMPLE_QUEUE_TAG
| I2O_SCB_FLAG_SENSE_DATA_IN_BUFFER));
PRIVATE_SCSI_SCB_EXECUTE_MESSAGE_setByteCount(
(PPRIVATE_SCSI_SCB_EXECUTE_MESSAGE)Message_Ptr,
sizeof(struct scsi_inquiry_data));
SG(&(Message_Ptr->SGL), 0,
I2O_SGL_FLAGS_LAST_ELEMENT | I2O_SGL_FLAGS_END_OF_BUFFER,
iq, sizeof(struct scsi_inquiry_data));
(void)ASR_queue_c(sc, (PI2O_MESSAGE_FRAME)Message_Ptr);
if (iq->vendor[0] && (iq->vendor[0] != ' ')) {
printf (" ");
ASR_prstring (iq->vendor, 8);
++posted;
}
if (iq->product[0] && (iq->product[0] != ' ')) {
printf (" ");
ASR_prstring (iq->product, 16);
++posted;
}
if (iq->revision[0] && (iq->revision[0] != ' ')) {
printf (" FW Rev. ");
ASR_prstring (iq->revision, 4);
++posted;
}
Remove the defAlignLong and getAlignLong macros. I guess that the original intent was to make sure that message structs allocated off of the stack were 4-byte aligned. However, the macros as defined did absolutely nothing. And since I2O forces you to manually copy messages down to the hardware, there really is no point of enforced alignment anyways.
2004-05-02 03:33:18 +00:00
free(iq, M_TEMP);
unexpand -a everything
2002-05-14 01:09:03 +00:00
if (posted) {
printf (",");
}
}
printf (" %d channel, %d CCBs, Protocol I2O\n", sc->ha_MaxBus + 1,
(sc->ha_QueueSize > MAX_INBOUND) ? MAX_INBOUND : sc->ha_QueueSize);
Re-indent some silly sub-blocks in asr_attach().
2004-05-01 06:12:58 +00:00
for (bus = 0; bus <= sc->ha_MaxBus; ++bus) {
struct cam_devq * devq;
int QueueSize = sc->ha_QueueSize;
unexpand -a everything
2002-05-14 01:09:03 +00:00
Re-indent some silly sub-blocks in asr_attach().
2004-05-01 06:12:58 +00:00
if (QueueSize > MAX_INBOUND) {
QueueSize = MAX_INBOUND;
unexpand -a everything
2002-05-14 01:09:03 +00:00
}
Re-indent some silly sub-blocks in asr_attach().
2004-05-01 06:12:58 +00:00
/*
* Create the device queue for our SIM(s).
*/
if ((devq = cam_simq_alloc(QueueSize)) == NULL) {
continue;
}
unexpand -a everything
2002-05-14 01:09:03 +00:00
Re-indent some silly sub-blocks in asr_attach().
2004-05-01 06:12:58 +00:00
/*
* Construct our first channel SIM entry
*/
sc->ha_sim[bus] = cam_sim_alloc(asr_action, asr_poll, "asr", sc,
Remove Giant from CAM. Drivers (SIMs) now register a mutex that CAM will use to synchornize and protect all data objects that are used for that SIM. Drivers that are not yet MPSAFE register Giant and operate as usual. RIght now, no drivers are MPSAFE, though a few will be changed in the coming week as this work settles down. The driver API has changed, so all CAM drivers will need to be recompiled. The userland API has not changed, so tools like camcontrol do not need to be recompiled.
2007-04-15 08:49:19 +00:00
unit, &Giant,
1, QueueSize, devq);
Re-indent some silly sub-blocks in asr_attach().
2004-05-01 06:12:58 +00:00
if (sc->ha_sim[bus] == NULL) {
continue;
}
unexpand -a everything
2002-05-14 01:09:03 +00:00
Prepare for future integration between CAM and newbus. xpt_bus_register now takes a device_t to be the parent of the bus that is being created. Most SIMs have been updated with a reasonable argument, but a few exceptions just pass NULL for now. This argument isn't used yet and the newbus integration likely won't be ready until after 7.0-RELEASE.
2007-06-17 05:55:54 +00:00
if (xpt_bus_register(sc->ha_sim[bus], dev, bus) != CAM_SUCCESS){
Re-indent some silly sub-blocks in asr_attach().
2004-05-01 06:12:58 +00:00
cam_sim_free(sc->ha_sim[bus],
/*free_devq*/TRUE);
sc->ha_sim[bus] = NULL;
continue;
}
unexpand -a everything
2002-05-14 01:09:03 +00:00
Re-indent some silly sub-blocks in asr_attach().
2004-05-01 06:12:58 +00:00
if (xpt_create_path(&(sc->ha_path[bus]), /*periph*/NULL,
cam_sim_path(sc->ha_sim[bus]), CAM_TARGET_WILDCARD,
CAM_LUN_WILDCARD) != CAM_REQ_CMP) {
xpt_bus_deregister( cam_sim_path(sc->ha_sim[bus]));
cam_sim_free(sc->ha_sim[bus], /*free_devq*/TRUE);
sc->ha_sim[bus] = NULL;
continue;
unexpand -a everything
2002-05-14 01:09:03 +00:00
}
}
Don't allocate an asr_ccb in asr_attach, it hasn't been needed for years.
2006-01-14 16:01:01 +00:00
unexpand -a everything
2002-05-14 01:09:03 +00:00
/*
* Generate the device node information
*/
Remove ASR_get_sc() and reference the softc in the dev_t. For some nefarious reason, the I2O protocol requires knowledge of all I2O devices in the system, so we can't get rid of the evil linked-list of softc's yet.
2004-05-01 05:56:57 +00:00
sc->ha_devt = make_dev(&asr_cdevsw, unit, UID_ROOT, GID_OPERATOR, 0640,
Remove the bogus printing of the asr control device major number. Also rename the control device from rasr%d to asr%d. This starts us down the path of divorcing ourselves from a very bogus design in the management apps. Since the apps are open source now, they will likely be updated and fixed before 5.3.
2004-05-02 00:27:54 +00:00
"asr%d", unit);
Only create the rdpti alias if the asr device creation succeeds.
2005-04-29 04:47:11 +00:00
if (sc->ha_devt != NULL)
(void)make_dev_alias(sc->ha_devt, "rdpti%d", unit);
Remove ASR_get_sc() and reference the softc in the dev_t. For some nefarious reason, the I2O protocol requires knowledge of all I2O devices in the system, so we can't get rid of the evil linked-list of softc's yet.
2004-05-01 05:56:57 +00:00
sc->ha_devt->si_drv1 = sc;
Remove the DOMINO and MODE0 device attachments. They never did anything. Remove a bunch of obfuscating macros.
2004-05-01 05:19:39 +00:00
return(0);
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
} /* asr_attach */
Get rid of a bunch of useless macros. STATIC becomes static, INLINE becomes __inline where appropriate and gets nuked elsewhere, IN/OUT/INOUT go away. Reformat code affected by this.
2004-05-01 01:25:05 +00:00
static void
asr_poll(struct cam_sim *sim)
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
{
unexpand -a everything
2002-05-14 01:09:03 +00:00
asr_intr(cam_sim_softc(sim));
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
} /* asr_poll */
Get rid of a bunch of useless macros. STATIC becomes static, INLINE becomes __inline where appropriate and gets nuked elsewhere, IN/OUT/INOUT go away. Reformat code affected by this.
2004-05-01 01:25:05 +00:00
static void
asr_action(struct cam_sim *sim, union ccb *ccb)
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
{
Get rid of a bunch of useless macros. STATIC becomes static, INLINE becomes __inline where appropriate and gets nuked elsewhere, IN/OUT/INOUT go away. Reformat code affected by this.
2004-05-01 01:25:05 +00:00
struct Asr_softc *sc;
unexpand -a everything
2002-05-14 01:09:03 +00:00
Many more style cleanups. Switch complex macros to being inline functions. Put @includes in a better spot. Fix many cases of 2 space indents and spaces between a function name and the parens. Use KASSERT instead of a home-rolled ASSERT. Remove some undeeded caddr casts.
2004-05-01 02:27:06 +00:00
debug_asr_printf("asr_action(%lx,%lx{%x})\n", (u_long)sim, (u_long)ccb,
ccb->ccb_h.func_code);
unexpand -a everything
2002-05-14 01:09:03 +00:00
CAM_DEBUG(ccb->ccb_h.path, CAM_DEBUG_TRACE, ("asr_action\n"));
ccb->ccb_h.spriv_ptr0 = sc = (struct Asr_softc *)cam_sim_softc(sim);
switch (ccb->ccb_h.func_code) {
/* Common cases first */
case XPT_SCSI_IO: /* Execute the requested I/O operation */
{
struct Message {
char M[MAX_INBOUND_SIZE];
Remove the defAlignLong and getAlignLong macros. I guess that the original intent was to make sure that message structs allocated off of the stack were 4-byte aligned. However, the macros as defined did absolutely nothing. And since I2O forces you to manually copy messages down to the hardware, there really is no point of enforced alignment anyways.
2004-05-02 03:33:18 +00:00
} Message;
unexpand -a everything
2002-05-14 01:09:03 +00:00
PI2O_MESSAGE_FRAME Message_Ptr;
/* Reject incoming commands while we are resetting the card */
if (sc->ha_in_reset != HA_OPERATIONAL) {
ccb->ccb_h.status &= ~CAM_STATUS_MASK;
if (sc->ha_in_reset >= HA_OFF_LINE) {
/* HBA is now off-line */
ccb->ccb_h.status |= CAM_UNREC_HBA_ERROR;
} else {
/* HBA currently resetting, try again later. */
ccb->ccb_h.status |= CAM_REQUEUE_REQ;
}
debug_asr_cmd_printf (" e\n");
xpt_done(ccb);
debug_asr_cmd_printf (" q\n");
break;
}
if ((ccb->ccb_h.status & CAM_STATUS_MASK) != CAM_REQ_INPROG) {
printf(
"asr%d WARNING: scsi_cmd(%x) already done on b%dt%du%d\n",
cam_sim_unit(xpt_path_sim(ccb->ccb_h.path)),
ccb->csio.cdb_io.cdb_bytes[0],
cam_sim_bus(sim),
ccb->ccb_h.target_id,
ccb->ccb_h.target_lun);
}
Many more style cleanups. Switch complex macros to being inline functions. Put @includes in a better spot. Fix many cases of 2 space indents and spaces between a function name and the parens. Use KASSERT instead of a home-rolled ASSERT. Remove some undeeded caddr casts.
2004-05-01 02:27:06 +00:00
debug_asr_cmd_printf("(%d,%d,%d,%d)", cam_sim_unit(sim),
cam_sim_bus(sim), ccb->ccb_h.target_id,
ccb->ccb_h.target_lun);
debug_asr_dump_ccb(ccb);
unexpand -a everything
2002-05-14 01:09:03 +00:00
Remove the defAlignLong and getAlignLong macros. I guess that the original intent was to make sure that message structs allocated off of the stack were 4-byte aligned. However, the macros as defined did absolutely nothing. And since I2O forces you to manually copy messages down to the hardware, there really is no point of enforced alignment anyways.
2004-05-02 03:33:18 +00:00
if ((Message_Ptr = ASR_init_message((union asr_ccb *)ccb,
(PI2O_MESSAGE_FRAME)&Message)) != NULL) {
unexpand -a everything
2002-05-14 01:09:03 +00:00
debug_asr_cmd2_printf ("TID=%x:\n",
PRIVATE_SCSI_SCB_EXECUTE_MESSAGE_getTID(
(PPRIVATE_SCSI_SCB_EXECUTE_MESSAGE)Message_Ptr));
debug_asr_cmd2_dump_message(Message_Ptr);
debug_asr_cmd1_printf (" q");
if (ASR_queue (sc, Message_Ptr) == EMPTY_QUEUE) {
ccb->ccb_h.status &= ~CAM_STATUS_MASK;
ccb->ccb_h.status |= CAM_REQUEUE_REQ;
debug_asr_cmd_printf (" E\n");
xpt_done(ccb);
}
Many more style cleanups. Switch complex macros to being inline functions. Put @includes in a better spot. Fix many cases of 2 space indents and spaces between a function name and the parens. Use KASSERT instead of a home-rolled ASSERT. Remove some undeeded caddr casts.
2004-05-01 02:27:06 +00:00
debug_asr_cmd_printf(" Q\n");
unexpand -a everything
2002-05-14 01:09:03 +00:00
break;
}
/*
* We will get here if there is no valid TID for the device
* referenced in the scsi command packet.
*/
ccb->ccb_h.status &= ~CAM_STATUS_MASK;
ccb->ccb_h.status |= CAM_SEL_TIMEOUT;
debug_asr_cmd_printf (" B\n");
xpt_done(ccb);
break;
}
case XPT_RESET_DEV: /* Bus Device Reset the specified SCSI device */
/* Rese HBA device ... */
asr_hbareset (sc);
ccb->ccb_h.status = CAM_REQ_CMP;
xpt_done(ccb);
break;
Don't indent preprocessor tokens.
2004-04-21 20:56:08 +00:00
#if (defined(REPORT_LUNS))
unexpand -a everything
2002-05-14 01:09:03 +00:00
case REPORT_LUNS:
Don't indent preprocessor tokens.
2004-04-21 20:56:08 +00:00
#endif
unexpand -a everything
2002-05-14 01:09:03 +00:00
case XPT_ABORT: /* Abort the specified CCB */
/* XXX Implement */
ccb->ccb_h.status = CAM_REQ_INVALID;
xpt_done(ccb);
break;
case XPT_SET_TRAN_SETTINGS:
/* XXX Implement */
ccb->ccb_h.status = CAM_FUNC_NOTAVAIL;
xpt_done(ccb);
break;
case XPT_GET_TRAN_SETTINGS:
/* Get default/user set transfer settings for the target */
{
The first of 3 major steps to move the CAM layer forward to using the CAM_NEW_TRAN_CODE that has been in the tree for some years now. This first step consists solely of adding to or correcting CAM_NEW_TRAN_CODE pieces in the kernel source tree such that a both a GENERIC (at least on i386) and a LINT build with CAM_NEW_TRAN_CODE as an option will compile correctly and run (at least with some the h/w I have). After a short settle time, the other pieces (making CAM_NEW_TRAN_CODE the default and updating libcam and camcontrol) will be brought in. This will be an incompatible change in that the size of structures related to XPT_PATH_INQ and XPT_{GET,SET}_TRAN_SETTINGS change in both size and content. However, basic system operation and basic system utilities work well enough with this change. Reviewed by: freebsd-scsi and specific stakeholders
2006-10-31 05:53:29 +00:00
struct ccb_trans_settings *cts = &(ccb->cts);
struct ccb_trans_settings_scsi *scsi =
&cts->proto_specific.scsi;
struct ccb_trans_settings_spi *spi =
&cts->xport_specific.spi;
if (cts->type == CTS_TYPE_USER_SETTINGS) {
cts->protocol = PROTO_SCSI;
cts->protocol_version = SCSI_REV_2;
cts->transport = XPORT_SPI;
cts->transport_version = 2;
scsi->flags = CTS_SCSI_FLAGS_TAG_ENB;
spi->flags = CTS_SPI_FLAGS_DISC_ENB;
spi->bus_width = MSG_EXT_WDTR_BUS_16_BIT;
spi->sync_period = 6; /* 40MHz */
spi->sync_offset = 15;
spi->valid = CTS_SPI_VALID_SYNC_RATE
| CTS_SPI_VALID_SYNC_OFFSET
| CTS_SPI_VALID_BUS_WIDTH
| CTS_SPI_VALID_DISC;
scsi->valid = CTS_SCSI_VALID_TQ;
unexpand -a everything
2002-05-14 01:09:03 +00:00
The first of 3 major steps to move the CAM layer forward to using the CAM_NEW_TRAN_CODE that has been in the tree for some years now. This first step consists solely of adding to or correcting CAM_NEW_TRAN_CODE pieces in the kernel source tree such that a both a GENERIC (at least on i386) and a LINT build with CAM_NEW_TRAN_CODE as an option will compile correctly and run (at least with some the h/w I have). After a short settle time, the other pieces (making CAM_NEW_TRAN_CODE the default and updating libcam and camcontrol) will be brought in. This will be an incompatible change in that the size of structures related to XPT_PATH_INQ and XPT_{GET,SET}_TRAN_SETTINGS change in both size and content. However, basic system operation and basic system utilities work well enough with this change. Reviewed by: freebsd-scsi and specific stakeholders
2006-10-31 05:53:29 +00:00
ccb->ccb_h.status = CAM_REQ_CMP;
} else {
ccb->ccb_h.status = CAM_FUNC_NOTAVAIL;
}
unexpand -a everything
2002-05-14 01:09:03 +00:00
xpt_done(ccb);
break;
}
case XPT_CALC_GEOMETRY:
{
struct ccb_calc_geometry *ccg;
u_int32_t size_mb;
u_int32_t secs_per_cylinder;
ccg = &(ccb->ccg);
size_mb = ccg->volume_size
/ ((1024L * 1024L) / ccg->block_size);
if (size_mb > 4096) {
ccg->heads = 255;
ccg->secs_per_track = 63;
} else if (size_mb > 2048) {
ccg->heads = 128;
ccg->secs_per_track = 63;
} else if (size_mb > 1024) {
ccg->heads = 65;
ccg->secs_per_track = 63;
} else {
ccg->heads = 64;
ccg->secs_per_track = 32;
}
secs_per_cylinder = ccg->heads * ccg->secs_per_track;
ccg->cylinders = ccg->volume_size / secs_per_cylinder;
ccb->ccb_h.status = CAM_REQ_CMP;
xpt_done(ccb);
break;
}
case XPT_RESET_BUS: /* Reset the specified SCSI bus */
ASR_resetBus (sc, cam_sim_bus(sim));
ccb->ccb_h.status = CAM_REQ_CMP;
xpt_done(ccb);
break;
case XPT_TERM_IO: /* Terminate the I/O process */
/* XXX Implement */
ccb->ccb_h.status = CAM_REQ_INVALID;
xpt_done(ccb);
break;
case XPT_PATH_INQ: /* Path routing inquiry */
{
struct ccb_pathinq *cpi = &(ccb->cpi);
cpi->version_num = 1; /* XXX??? */
cpi->hba_inquiry = PI_SDTR_ABLE|PI_TAG_ABLE|PI_WIDE_16;
cpi->target_sprt = 0;
/* Not necessary to reset bus, done by HDM initialization */
cpi->hba_misc = PIM_NOBUSRESET;
cpi->hba_eng_cnt = 0;
cpi->max_target = sc->ha_MaxId;
cpi->max_lun = sc->ha_MaxLun;
cpi->initiator_id = sc->ha_adapter_target[cam_sim_bus(sim)];
cpi->bus_id = cam_sim_bus(sim);
cpi->base_transfer_speed = 3300;
strncpy(cpi->sim_vid, "FreeBSD", SIM_IDLEN);
strncpy(cpi->hba_vid, "Adaptec", HBA_IDLEN);
strncpy(cpi->dev_name, cam_sim_name(sim), DEV_IDLEN);
cpi->unit_number = cam_sim_unit(sim);
cpi->ccb_h.status = CAM_REQ_CMP;
The first of 3 major steps to move the CAM layer forward to using the CAM_NEW_TRAN_CODE that has been in the tree for some years now. This first step consists solely of adding to or correcting CAM_NEW_TRAN_CODE pieces in the kernel source tree such that a both a GENERIC (at least on i386) and a LINT build with CAM_NEW_TRAN_CODE as an option will compile correctly and run (at least with some the h/w I have). After a short settle time, the other pieces (making CAM_NEW_TRAN_CODE the default and updating libcam and camcontrol) will be brought in. This will be an incompatible change in that the size of structures related to XPT_PATH_INQ and XPT_{GET,SET}_TRAN_SETTINGS change in both size and content. However, basic system operation and basic system utilities work well enough with this change. Reviewed by: freebsd-scsi and specific stakeholders
2006-10-31 05:53:29 +00:00
cpi->transport = XPORT_SPI;
cpi->transport_version = 2;
cpi->protocol = PROTO_SCSI;
cpi->protocol_version = SCSI_REV_2;
unexpand -a everything
2002-05-14 01:09:03 +00:00
xpt_done(ccb);
break;
}
default:
ccb->ccb_h.status = CAM_REQ_INVALID;
xpt_done(ccb);
break;
}
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
} /* asr_action */
/*
* Handle processing of current CCB as pointed to by the Status.
*/
Get rid of a bunch of useless macros. STATIC becomes static, INLINE becomes __inline where appropriate and gets nuked elsewhere, IN/OUT/INOUT go away. Reformat code affected by this.
2004-05-01 01:25:05 +00:00
static int
asr_intr(Asr_softc_t *sc)
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
{
Get rid of a bunch of useless macros. STATIC becomes static, INLINE becomes __inline where appropriate and gets nuked elsewhere, IN/OUT/INOUT go away. Reformat code affected by this.
2004-05-01 01:25:05 +00:00
int processed;
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
Convert the asr driver to use the bus_space API. This does not represent a significant functional change, but it further cleans up the code and brings it closer to being portable. Thanks to Don Bowman for helping to test this.
2004-05-05 03:42:33 +00:00
for(processed = 0; asr_get_status(sc) & Mask_InterruptsDisabled;
Get rid of a bunch of useless macros. STATIC becomes static, INLINE becomes __inline where appropriate and gets nuked elsewhere, IN/OUT/INOUT go away. Reformat code affected by this.
2004-05-01 01:25:05 +00:00
processed = 1) {
union asr_ccb *ccb;
- Set the CAM status to CAM_SCSI_STATUS_ERROR rather than CAM_REQ_CMP in case of a CHECK CONDITION. - Make this driver return SCSI status information. - While here, factor out the clearing of the CAM status from every element of the switch statement to only once before the switch. This fixes burning CDs with recent cdrecord 2.01 alpha versions and burners attached to asr(4) controllers but there could have been other applications and da(4) etc. also affected. Reviewed by: gibbs, scottl MFC after: 2 weeks
2004-08-02 23:32:23 +00:00
u_int dsc;
unexpand -a everything
2002-05-14 01:09:03 +00:00
U32 ReplyOffset;
PI2O_SCSI_ERROR_REPLY_MESSAGE_FRAME Reply;
Convert the asr driver to use the bus_space API. This does not represent a significant functional change, but it further cleans up the code and brings it closer to being portable. Thanks to Don Bowman for helping to test this.
2004-05-05 03:42:33 +00:00
if (((ReplyOffset = asr_get_FromFIFO(sc)) == EMPTY_QUEUE)
&& ((ReplyOffset = asr_get_FromFIFO(sc)) == EMPTY_QUEUE)) {
unexpand -a everything
2002-05-14 01:09:03 +00:00
break;
}
Reply = (PI2O_SCSI_ERROR_REPLY_MESSAGE_FRAME)(ReplyOffset
- sc->ha_Msgs_Phys + (char *)(sc->ha_Msgs));
/*
* We do not need any (optional byteswapping) method access to
* the Initiator context field.
*/
ccb = (union asr_ccb *)(long)
I2O_MESSAGE_FRAME_getInitiatorContext64(
&(Reply->StdReplyFrame.StdMessageFrame));
if (I2O_MESSAGE_FRAME_getMsgFlags(
&(Reply->StdReplyFrame.StdMessageFrame))
& I2O_MESSAGE_FLAGS_FAIL) {
Remove the defAlignLong and getAlignLong macros. I guess that the original intent was to make sure that message structs allocated off of the stack were 4-byte aligned. However, the macros as defined did absolutely nothing. And since I2O forces you to manually copy messages down to the hardware, there really is no point of enforced alignment anyways.
2004-05-02 03:33:18 +00:00
I2O_UTIL_NOP_MESSAGE Message;
PI2O_UTIL_NOP_MESSAGE Message_Ptr;
U32 MessageOffset;
unexpand -a everything
2002-05-14 01:09:03 +00:00
MessageOffset = (u_long)
I2O_FAILURE_REPLY_MESSAGE_FRAME_getPreservedMFA(
(PI2O_FAILURE_REPLY_MESSAGE_FRAME)Reply);
/*
* Get the Original Message Frame's address, and get
* it's Transaction Context into our space. (Currently
* unused at original authorship, but better to be
* safe than sorry). Straight copy means that we
* need not concern ourselves with the (optional
* byteswapping) method access.
*/
Convert the asr driver to use the bus_space API. This does not represent a significant functional change, but it further cleans up the code and brings it closer to being portable. Thanks to Don Bowman for helping to test this.
2004-05-05 03:42:33 +00:00
Reply->StdReplyFrame.TransactionContext =
bus_space_read_4(sc->ha_frame_btag,
sc->ha_frame_bhandle, MessageOffset +
offsetof(I2O_SINGLE_REPLY_MESSAGE_FRAME,
TransactionContext));
unexpand -a everything
2002-05-14 01:09:03 +00:00
/*
* For 64 bit machines, we need to reconstruct the
* 64 bit context.
*/
ccb = (union asr_ccb *)(long)
I2O_MESSAGE_FRAME_getInitiatorContext64(
&(Reply->StdReplyFrame.StdMessageFrame));
/*
* Unique error code for command failure.
*/
I2O_SINGLE_REPLY_MESSAGE_FRAME_setDetailedStatusCode(
&(Reply->StdReplyFrame), (u_int16_t)-2);
/*
* Modify the message frame to contain a NOP and
* re-issue it to the controller.
*/
Message_Ptr = (PI2O_UTIL_NOP_MESSAGE)ASR_fillMessage(
Remove the defAlignLong and getAlignLong macros. I guess that the original intent was to make sure that message structs allocated off of the stack were 4-byte aligned. However, the macros as defined did absolutely nothing. And since I2O forces you to manually copy messages down to the hardware, there really is no point of enforced alignment anyways.
2004-05-02 03:33:18 +00:00
&Message, sizeof(I2O_UTIL_NOP_MESSAGE));
Don't indent preprocessor tokens.
2004-04-21 20:56:08 +00:00
#if (I2O_UTIL_NOP != 0)
unexpand -a everything
2002-05-14 01:09:03 +00:00
I2O_MESSAGE_FRAME_setFunction (
&(Message_Ptr->StdMessageFrame),
I2O_UTIL_NOP);
Don't indent preprocessor tokens.
2004-04-21 20:56:08 +00:00
#endif
unexpand -a everything
2002-05-14 01:09:03 +00:00
/*
* Copy the packet out to the Original Message
*/
Convert the asr driver to use the bus_space API. This does not represent a significant functional change, but it further cleans up the code and brings it closer to being portable. Thanks to Don Bowman for helping to test this.
2004-05-05 03:42:33 +00:00
asr_set_frame(sc, Message_Ptr, MessageOffset,
sizeof(I2O_UTIL_NOP_MESSAGE));
unexpand -a everything
2002-05-14 01:09:03 +00:00
/*
* Issue the NOP
*/
Convert the asr driver to use the bus_space API. This does not represent a significant functional change, but it further cleans up the code and brings it closer to being portable. Thanks to Don Bowman for helping to test this.
2004-05-05 03:42:33 +00:00
asr_set_ToFIFO(sc, MessageOffset);
unexpand -a everything
2002-05-14 01:09:03 +00:00
}
/*
* Asynchronous command with no return requirements,
* and a generic handler for immunity against odd error
* returns from the adapter.
*/
Begin the slow, painful process of cleaning up asr. This eliminates many of the bogus casts of NULL. It also distracts me while I procrastinate on useful work.
2004-04-21 03:51:39 +00:00
if (ccb == NULL) {
unexpand -a everything
2002-05-14 01:09:03 +00:00
/*
* Return Reply so that it can be used for the
* next command
*/
Convert the asr driver to use the bus_space API. This does not represent a significant functional change, but it further cleans up the code and brings it closer to being portable. Thanks to Don Bowman for helping to test this.
2004-05-05 03:42:33 +00:00
asr_set_FromFIFO(sc, ReplyOffset);
unexpand -a everything
2002-05-14 01:09:03 +00:00
continue;
}
/* Welease Wadjah! (and stop timeouts) */
ASR_ccbRemove (sc, ccb);
- Set the CAM status to CAM_SCSI_STATUS_ERROR rather than CAM_REQ_CMP in case of a CHECK CONDITION. - Make this driver return SCSI status information. - While here, factor out the clearing of the CAM status from every element of the switch statement to only once before the switch. This fixes burning CDs with recent cdrecord 2.01 alpha versions and burners attached to asr(4) controllers but there could have been other applications and da(4) etc. also affected. Reviewed by: gibbs, scottl MFC after: 2 weeks
2004-08-02 23:32:23 +00:00
dsc = I2O_SINGLE_REPLY_MESSAGE_FRAME_getDetailedStatusCode(
&(Reply->StdReplyFrame));
ccb->csio.scsi_status = dsc & I2O_SCSI_DEVICE_DSC_MASK;
ccb->ccb_h.status &= ~CAM_STATUS_MASK;
switch (dsc) {
unexpand -a everything
2002-05-14 01:09:03 +00:00
case I2O_SCSI_DSC_SUCCESS:
ccb->ccb_h.status |= CAM_REQ_CMP;
break;
case I2O_SCSI_DSC_CHECK_CONDITION:
- Set the CAM status to CAM_SCSI_STATUS_ERROR rather than CAM_REQ_CMP in case of a CHECK CONDITION. - Make this driver return SCSI status information. - While here, factor out the clearing of the CAM status from every element of the switch statement to only once before the switch. This fixes burning CDs with recent cdrecord 2.01 alpha versions and burners attached to asr(4) controllers but there could have been other applications and da(4) etc. also affected. Reviewed by: gibbs, scottl MFC after: 2 weeks
2004-08-02 23:32:23 +00:00
ccb->ccb_h.status |= CAM_SCSI_STATUS_ERROR |
CAM_AUTOSNS_VALID;
unexpand -a everything
2002-05-14 01:09:03 +00:00
break;
case I2O_SCSI_DSC_BUSY:
/* FALLTHRU */
case I2O_SCSI_HBA_DSC_ADAPTER_BUSY:
/* FALLTHRU */
case I2O_SCSI_HBA_DSC_SCSI_BUS_RESET:
/* FALLTHRU */
case I2O_SCSI_HBA_DSC_BUS_BUSY:
ccb->ccb_h.status |= CAM_SCSI_BUSY;
break;
case I2O_SCSI_HBA_DSC_SELECTION_TIMEOUT:
ccb->ccb_h.status |= CAM_SEL_TIMEOUT;
break;
case I2O_SCSI_HBA_DSC_COMMAND_TIMEOUT:
/* FALLTHRU */
case I2O_SCSI_HBA_DSC_DEVICE_NOT_PRESENT:
/* FALLTHRU */
case I2O_SCSI_HBA_DSC_LUN_INVALID:
/* FALLTHRU */
case I2O_SCSI_HBA_DSC_SCSI_TID_INVALID:
ccb->ccb_h.status |= CAM_CMD_TIMEOUT;
break;
case I2O_SCSI_HBA_DSC_DATA_OVERRUN:
/* FALLTHRU */
case I2O_SCSI_HBA_DSC_REQUEST_LENGTH_ERROR:
ccb->ccb_h.status |= CAM_DATA_RUN_ERR;
break;
default:
ccb->ccb_h.status |= CAM_REQUEUE_REQ;
break;
}
if ((ccb->csio.resid = ccb->csio.dxfer_len) != 0) {
ccb->csio.resid -=
I2O_SCSI_ERROR_REPLY_MESSAGE_FRAME_getTransferCount(
Reply);
}
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
unexpand -a everything
2002-05-14 01:09:03 +00:00
/* Sense data in reply packet */
if (ccb->ccb_h.status & CAM_AUTOSNS_VALID) {
u_int16_t size = I2O_SCSI_ERROR_REPLY_MESSAGE_FRAME_getAutoSenseTransferCount(Reply);
if (size) {
if (size > sizeof(ccb->csio.sense_data)) {
size = sizeof(ccb->csio.sense_data);
}
if (size > I2O_SCSI_SENSE_DATA_SZ) {
size = I2O_SCSI_SENSE_DATA_SZ;
}
if ((ccb->csio.sense_len)
&& (size > ccb->csio.sense_len)) {
size = ccb->csio.sense_len;
}
Many more style cleanups. Switch complex macros to being inline functions. Put @includes in a better spot. Fix many cases of 2 space indents and spaces between a function name and the parens. Use KASSERT instead of a home-rolled ASSERT. Remove some undeeded caddr casts.
2004-05-01 02:27:06 +00:00
bcopy(Reply->SenseData,
&(ccb->csio.sense_data), size);
unexpand -a everything
2002-05-14 01:09:03 +00:00
}
}
/*
* Return Reply so that it can be used for the next command
* since we have no more need for it now
*/
Convert the asr driver to use the bus_space API. This does not represent a significant functional change, but it further cleans up the code and brings it closer to being portable. Thanks to Don Bowman for helping to test this.
2004-05-05 03:42:33 +00:00
asr_set_FromFIFO(sc, ReplyOffset);
unexpand -a everything
2002-05-14 01:09:03 +00:00
if (ccb->ccb_h.path) {
xpt_done ((union ccb *)ccb);
} else {
More low-hanging fruit: kill caddr_t in calls to wakeup(9) / [mt]sleep(9).
2003-03-02 16:54:40 +00:00
wakeup (ccb);
unexpand -a everything
2002-05-14 01:09:03 +00:00
}
}
return (processed);
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
} /* asr_intr */
unexpand -a everything
2002-05-14 01:09:03 +00:00
#undef QueueSize /* Grrrr */
#undef SG_Size /* Grrrr */
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
/*
unexpand -a everything
2002-05-14 01:09:03 +00:00
* Meant to be included at the bottom of asr.c !!!
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
*/
/*
unexpand -a everything
2002-05-14 01:09:03 +00:00
* Included here as hard coded. Done because other necessary include
* files utilize C++ comment structures which make them a nuisance to
* included here just to pick up these three typedefs.
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
*/
typedef U32 DPT_TAG_T;
typedef U32 DPT_MSG_T;
typedef U32 DPT_RTN_T;
unexpand -a everything
2002-05-14 01:09:03 +00:00
#undef SCSI_RESET /* Conflicts with "scsi/scsiconf.h" defintion */
#include "dev/asr/osd_unix.h"
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
Replace all calls to minor() with dev2unit(). After I removed all the unit2minor()/minor2unit() calls from the kernel yesterday, I realised calling minor() everywhere is quite confusing. Character devices now only have the ability to store a unit number, not a minor number. Remove the confusion by using dev2unit() everywhere. This commit could also be considered as a bug fix. A lot of drivers call minor(), while they should actually be calling dev2unit(). In -CURRENT this isn't a problem, but it turns out we never had any problem reports related to that issue in the past. I suspect not many people connect more than 256 pieces of the same hardware. Reviewed by: kib
2008-09-27 08:51:18 +00:00
#define asr_unit(dev) dev2unit(dev)
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
Get rid of a bunch of useless macros. STATIC becomes static, INLINE becomes __inline where appropriate and gets nuked elsewhere, IN/OUT/INOUT go away. Reformat code affected by this.
2004-05-01 01:25:05 +00:00
static u_int8_t ASR_ctlr_held;
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
Get rid of a bunch of useless macros. STATIC becomes static, INLINE becomes __inline where appropriate and gets nuked elsewhere, IN/OUT/INOUT go away. Reformat code affected by this.
2004-05-01 01:25:05 +00:00
static int
Do the dreaded s/dev_t/struct cdev */ Bump __FreeBSD_version accordingly.
2004-06-16 09:47:26 +00:00
asr_open(struct cdev *dev, int32_t flags, int32_t ifmt, struct thread *td)
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
{
unexpand -a everything
2002-05-14 01:09:03 +00:00
int s;
Get rid of a bunch of useless macros. STATIC becomes static, INLINE becomes __inline where appropriate and gets nuked elsewhere, IN/OUT/INOUT go away. Reformat code affected by this.
2004-05-01 01:25:05 +00:00
int error;
unexpand -a everything
2002-05-14 01:09:03 +00:00
Remove ASR_get_sc() and reference the softc in the dev_t. For some nefarious reason, the I2O protocol requires knowledge of all I2O devices in the system, so we can't get rid of the evil linked-list of softc's yet.
2004-05-01 05:56:57 +00:00
if (dev->si_drv1 == NULL) {
unexpand -a everything
2002-05-14 01:09:03 +00:00
return (ENODEV);
}
s = splcam ();
if (ASR_ctlr_held) {
error = EBUSY;
Sweep kernel replacing suser(9) calls with priv(9) calls, assigning specific privilege names to a broad range of privileges. These may require some future tweaking. Sponsored by: nCircle Network Security, Inc. Obtained from: TrustedBSD Project Discussed on: arch@ Reviewed (at least in part) by: mlaier, jmg, pjd, bde, ceri, Alex Lyashkov <umka at sevcity dot net>, Skip Ford <skip dot ford at verizon dot net>, Antoine Brodin <antoine dot brodin at laposte dot net>
2006-11-06 13:42:10 +00:00
} else if ((error = priv_check(td, PRIV_DRIVER)) == 0) {
unexpand -a everything
2002-05-14 01:09:03 +00:00
++ASR_ctlr_held;
}
splx(s);
return (error);
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
} /* asr_open */
Get rid of a bunch of useless macros. STATIC becomes static, INLINE becomes __inline where appropriate and gets nuked elsewhere, IN/OUT/INOUT go away. Reformat code affected by this.
2004-05-01 01:25:05 +00:00
static int
Do the dreaded s/dev_t/struct cdev */ Bump __FreeBSD_version accordingly.
2004-06-16 09:47:26 +00:00
asr_close(struct cdev *dev, int flags, int ifmt, struct thread *td)
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
{
unexpand -a everything
2002-05-14 01:09:03 +00:00
ASR_ctlr_held = 0;
return (0);
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
} /* asr_close */
/*-------------------------------------------------------------------------*/
unexpand -a everything
2002-05-14 01:09:03 +00:00
/* Function ASR_queue_i */
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
/*-------------------------------------------------------------------------*/
unexpand -a everything
2002-05-14 01:09:03 +00:00
/* The Parameters Passed To This Function Are : */
/* Asr_softc_t * : HBA miniport driver's adapter data storage. */
/* PI2O_MESSAGE_FRAME : Msg Structure Pointer For This Command */
/* I2O_SCSI_ERROR_REPLY_MESSAGE_FRAME following the Msg Structure */
/* */
/* This Function Will Take The User Request Packet And Convert It To An */
/* I2O MSG And Send It Off To The Adapter. */
/* */
/* Return : 0 For OK, Error Code Otherwise */
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
/*-------------------------------------------------------------------------*/
Get rid of a bunch of useless macros. STATIC becomes static, INLINE becomes __inline where appropriate and gets nuked elsewhere, IN/OUT/INOUT go away. Reformat code affected by this.
2004-05-01 01:25:05 +00:00
static int
ASR_queue_i(Asr_softc_t *sc, PI2O_MESSAGE_FRAME Packet)
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
{
unexpand -a everything
2002-05-14 01:09:03 +00:00
union asr_ccb * ccb;
PI2O_SCSI_ERROR_REPLY_MESSAGE_FRAME Reply;
PI2O_MESSAGE_FRAME Message_Ptr;
PI2O_SCSI_ERROR_REPLY_MESSAGE_FRAME Reply_Ptr;
int MessageSizeInBytes;
int ReplySizeInBytes;
int error;
int s;
/* Scatter Gather buffer list */
struct ioctlSgList_S {
SLIST_ENTRY(ioctlSgList_S) link;
caddr_t UserSpace;
I2O_FLAGS_COUNT FlagsCount;
char KernelSpace[sizeof(long)];
} * elm;
/* Generates a `first' entry */
SLIST_HEAD(ioctlSgListHead_S, ioctlSgList_S) sgList;
if (ASR_getBlinkLedCode(sc)) {
debug_usr_cmd_printf ("Adapter currently in BlinkLed %x\n",
ASR_getBlinkLedCode(sc));
return (EIO);
}
/* Copy in the message into a local allocation */
if ((Message_Ptr = (PI2O_MESSAGE_FRAME)malloc (
Begin the slow, painful process of cleaning up asr. This eliminates many of the bogus casts of NULL. It also distracts me while I procrastinate on useful work.
2004-04-21 03:51:39 +00:00
sizeof(I2O_MESSAGE_FRAME), M_TEMP, M_WAITOK)) == NULL) {
unexpand -a everything
2002-05-14 01:09:03 +00:00
debug_usr_cmd_printf (
"Failed to acquire I2O_MESSAGE_FRAME memory\n");
return (ENOMEM);
}
if ((error = copyin ((caddr_t)Packet, (caddr_t)Message_Ptr,
sizeof(I2O_MESSAGE_FRAME))) != 0) {
Remove the defAlignLong and getAlignLong macros. I guess that the original intent was to make sure that message structs allocated off of the stack were 4-byte aligned. However, the macros as defined did absolutely nothing. And since I2O forces you to manually copy messages down to the hardware, there really is no point of enforced alignment anyways.
2004-05-02 03:33:18 +00:00
free(Message_Ptr, M_TEMP);
unexpand -a everything
2002-05-14 01:09:03 +00:00
debug_usr_cmd_printf ("Can't copy in packet errno=%d\n", error);
return (error);
}
/* Acquire information to determine type of packet */
MessageSizeInBytes = (I2O_MESSAGE_FRAME_getMessageSize(Message_Ptr)<<2);
/* The offset of the reply information within the user packet */
Reply = (PI2O_SCSI_ERROR_REPLY_MESSAGE_FRAME)((char *)Packet
+ MessageSizeInBytes);
/* Check if the message is a synchronous initialization command */
s = I2O_MESSAGE_FRAME_getFunction(Message_Ptr);
Remove the defAlignLong and getAlignLong macros. I guess that the original intent was to make sure that message structs allocated off of the stack were 4-byte aligned. However, the macros as defined did absolutely nothing. And since I2O forces you to manually copy messages down to the hardware, there really is no point of enforced alignment anyways.
2004-05-02 03:33:18 +00:00
free(Message_Ptr, M_TEMP);
unexpand -a everything
2002-05-14 01:09:03 +00:00
switch (s) {
case I2O_EXEC_IOP_RESET:
{ U32 status;
Convert the asr driver to use the bus_space API. This does not represent a significant functional change, but it further cleans up the code and brings it closer to being portable. Thanks to Don Bowman for helping to test this.
2004-05-05 03:42:33 +00:00
status = ASR_resetIOP(sc);
unexpand -a everything
2002-05-14 01:09:03 +00:00
ReplySizeInBytes = sizeof(status);
debug_usr_cmd_printf ("resetIOP done\n");
return (copyout ((caddr_t)&status, (caddr_t)Reply,
ReplySizeInBytes));
}
case I2O_EXEC_STATUS_GET:
Squash another use of vtophys. Instead of creating separate busdma objects for doing static memory transfers, start collecting them into a single object.
2006-02-04 17:56:17 +00:00
{ PI2O_EXEC_STATUS_GET_REPLY status;
unexpand -a everything
2002-05-14 01:09:03 +00:00
Squash another use of vtophys. Instead of creating separate busdma objects for doing static memory transfers, start collecting them into a single object.
2006-02-04 17:56:17 +00:00
status = &sc->ha_statusmem->status;
if (ASR_getStatus(sc) == NULL) {
unexpand -a everything
2002-05-14 01:09:03 +00:00
debug_usr_cmd_printf ("getStatus failed\n");
return (ENXIO);
}
ReplySizeInBytes = sizeof(status);
debug_usr_cmd_printf ("getStatus done\n");
Squash another use of vtophys. Instead of creating separate busdma objects for doing static memory transfers, start collecting them into a single object.
2006-02-04 17:56:17 +00:00
return (copyout ((caddr_t)status, (caddr_t)Reply,
unexpand -a everything
2002-05-14 01:09:03 +00:00
ReplySizeInBytes));
}
case I2O_EXEC_OUTBOUND_INIT:
{ U32 status;
status = ASR_initOutBound(sc);
ReplySizeInBytes = sizeof(status);
debug_usr_cmd_printf ("intOutBound done\n");
return (copyout ((caddr_t)&status, (caddr_t)Reply,
ReplySizeInBytes));
}
}
/* Determine if the message size is valid */
if ((MessageSizeInBytes < sizeof(I2O_MESSAGE_FRAME))
|| (MAX_INBOUND_SIZE < MessageSizeInBytes)) {
debug_usr_cmd_printf ("Packet size %d incorrect\n",
MessageSizeInBytes);
return (EINVAL);
}
if ((Message_Ptr = (PI2O_MESSAGE_FRAME)malloc (MessageSizeInBytes,
Begin the slow, painful process of cleaning up asr. This eliminates many of the bogus casts of NULL. It also distracts me while I procrastinate on useful work.
2004-04-21 03:51:39 +00:00
M_TEMP, M_WAITOK)) == NULL) {
unexpand -a everything
2002-05-14 01:09:03 +00:00
debug_usr_cmd_printf ("Failed to acquire frame[%d] memory\n",
MessageSizeInBytes);
return (ENOMEM);
}
if ((error = copyin ((caddr_t)Packet, (caddr_t)Message_Ptr,
MessageSizeInBytes)) != 0) {
Remove the defAlignLong and getAlignLong macros. I guess that the original intent was to make sure that message structs allocated off of the stack were 4-byte aligned. However, the macros as defined did absolutely nothing. And since I2O forces you to manually copy messages down to the hardware, there really is no point of enforced alignment anyways.
2004-05-02 03:33:18 +00:00
free(Message_Ptr, M_TEMP);
unexpand -a everything
2002-05-14 01:09:03 +00:00
debug_usr_cmd_printf ("Can't copy in packet[%d] errno=%d\n",
MessageSizeInBytes, error);
return (error);
}
/* Check the size of the reply frame, and start constructing */
if ((Reply_Ptr = (PI2O_SCSI_ERROR_REPLY_MESSAGE_FRAME)malloc (
Begin the slow, painful process of cleaning up asr. This eliminates many of the bogus casts of NULL. It also distracts me while I procrastinate on useful work.
2004-04-21 03:51:39 +00:00
sizeof(I2O_MESSAGE_FRAME), M_TEMP, M_WAITOK)) == NULL) {
Remove the defAlignLong and getAlignLong macros. I guess that the original intent was to make sure that message structs allocated off of the stack were 4-byte aligned. However, the macros as defined did absolutely nothing. And since I2O forces you to manually copy messages down to the hardware, there really is no point of enforced alignment anyways.
2004-05-02 03:33:18 +00:00
free(Message_Ptr, M_TEMP);
unexpand -a everything
2002-05-14 01:09:03 +00:00
debug_usr_cmd_printf (
"Failed to acquire I2O_MESSAGE_FRAME memory\n");
return (ENOMEM);
}
if ((error = copyin ((caddr_t)Reply, (caddr_t)Reply_Ptr,
sizeof(I2O_MESSAGE_FRAME))) != 0) {
Remove the defAlignLong and getAlignLong macros. I guess that the original intent was to make sure that message structs allocated off of the stack were 4-byte aligned. However, the macros as defined did absolutely nothing. And since I2O forces you to manually copy messages down to the hardware, there really is no point of enforced alignment anyways.
2004-05-02 03:33:18 +00:00
free(Reply_Ptr, M_TEMP);
free(Message_Ptr, M_TEMP);
unexpand -a everything
2002-05-14 01:09:03 +00:00
debug_usr_cmd_printf (
"Failed to copy in reply frame, errno=%d\n",
error);
return (error);
}
ReplySizeInBytes = (I2O_MESSAGE_FRAME_getMessageSize(
&(Reply_Ptr->StdReplyFrame.StdMessageFrame)) << 2);
Remove the defAlignLong and getAlignLong macros. I guess that the original intent was to make sure that message structs allocated off of the stack were 4-byte aligned. However, the macros as defined did absolutely nothing. And since I2O forces you to manually copy messages down to the hardware, there really is no point of enforced alignment anyways.
2004-05-02 03:33:18 +00:00
free(Reply_Ptr, M_TEMP);
unexpand -a everything
2002-05-14 01:09:03 +00:00
if (ReplySizeInBytes < sizeof(I2O_SINGLE_REPLY_MESSAGE_FRAME)) {
Remove the defAlignLong and getAlignLong macros. I guess that the original intent was to make sure that message structs allocated off of the stack were 4-byte aligned. However, the macros as defined did absolutely nothing. And since I2O forces you to manually copy messages down to the hardware, there really is no point of enforced alignment anyways.
2004-05-02 03:33:18 +00:00
free(Message_Ptr, M_TEMP);
unexpand -a everything
2002-05-14 01:09:03 +00:00
debug_usr_cmd_printf (
"Failed to copy in reply frame[%d], errno=%d\n",
ReplySizeInBytes, error);
return (EINVAL);
}
if ((Reply_Ptr = (PI2O_SCSI_ERROR_REPLY_MESSAGE_FRAME)malloc (
((ReplySizeInBytes > sizeof(I2O_SCSI_ERROR_REPLY_MESSAGE_FRAME))
Begin the slow, painful process of cleaning up asr. This eliminates many of the bogus casts of NULL. It also distracts me while I procrastinate on useful work.
2004-04-21 03:51:39 +00:00
? ReplySizeInBytes : sizeof(I2O_SCSI_ERROR_REPLY_MESSAGE_FRAME)),
M_TEMP, M_WAITOK)) == NULL) {
Remove the defAlignLong and getAlignLong macros. I guess that the original intent was to make sure that message structs allocated off of the stack were 4-byte aligned. However, the macros as defined did absolutely nothing. And since I2O forces you to manually copy messages down to the hardware, there really is no point of enforced alignment anyways.
2004-05-02 03:33:18 +00:00
free(Message_Ptr, M_TEMP);
unexpand -a everything
2002-05-14 01:09:03 +00:00
debug_usr_cmd_printf ("Failed to acquire frame[%d] memory\n",
ReplySizeInBytes);
return (ENOMEM);
}
Remove the defAlignLong and getAlignLong macros. I guess that the original intent was to make sure that message structs allocated off of the stack were 4-byte aligned. However, the macros as defined did absolutely nothing. And since I2O forces you to manually copy messages down to the hardware, there really is no point of enforced alignment anyways.
2004-05-02 03:33:18 +00:00
(void)ASR_fillMessage((void *)Reply_Ptr, ReplySizeInBytes);
unexpand -a everything
2002-05-14 01:09:03 +00:00
Reply_Ptr->StdReplyFrame.StdMessageFrame.InitiatorContext
= Message_Ptr->InitiatorContext;
Reply_Ptr->StdReplyFrame.TransactionContext
= ((PI2O_PRIVATE_MESSAGE_FRAME)Message_Ptr)->TransactionContext;
I2O_MESSAGE_FRAME_setMsgFlags(
&(Reply_Ptr->StdReplyFrame.StdMessageFrame),
I2O_MESSAGE_FRAME_getMsgFlags(
&(Reply_Ptr->StdReplyFrame.StdMessageFrame))
| I2O_MESSAGE_FLAGS_REPLY);
/* Check if the message is a special case command */
switch (I2O_MESSAGE_FRAME_getFunction(Message_Ptr)) {
case I2O_EXEC_SYS_TAB_SET: /* Special Case of empty Scatter Gather */
if (MessageSizeInBytes == ((I2O_MESSAGE_FRAME_getVersionOffset(
Message_Ptr) & 0xF0) >> 2)) {
Remove the defAlignLong and getAlignLong macros. I guess that the original intent was to make sure that message structs allocated off of the stack were 4-byte aligned. However, the macros as defined did absolutely nothing. And since I2O forces you to manually copy messages down to the hardware, there really is no point of enforced alignment anyways.
2004-05-02 03:33:18 +00:00
free(Message_Ptr, M_TEMP);
unexpand -a everything
2002-05-14 01:09:03 +00:00
I2O_SINGLE_REPLY_MESSAGE_FRAME_setDetailedStatusCode(
&(Reply_Ptr->StdReplyFrame),
(ASR_setSysTab(sc) != CAM_REQ_CMP));
I2O_MESSAGE_FRAME_setMessageSize(
&(Reply_Ptr->StdReplyFrame.StdMessageFrame),
sizeof(I2O_SINGLE_REPLY_MESSAGE_FRAME));
error = copyout ((caddr_t)Reply_Ptr, (caddr_t)Reply,
ReplySizeInBytes);
Remove the defAlignLong and getAlignLong macros. I guess that the original intent was to make sure that message structs allocated off of the stack were 4-byte aligned. However, the macros as defined did absolutely nothing. And since I2O forces you to manually copy messages down to the hardware, there really is no point of enforced alignment anyways.
2004-05-02 03:33:18 +00:00
free(Reply_Ptr, M_TEMP);
unexpand -a everything
2002-05-14 01:09:03 +00:00
return (error);
}
}
/* Deal in the general case */
/* First allocate and optionally copy in each scatter gather element */
SLIST_INIT(&sgList);
if ((I2O_MESSAGE_FRAME_getVersionOffset(Message_Ptr) & 0xF0) != 0) {
PI2O_SGE_SIMPLE_ELEMENT sg;
/*
* since this code is reused in several systems, code
* efficiency is greater by using a shift operation rather
* than a divide by sizeof(u_int32_t).
*/
sg = (PI2O_SGE_SIMPLE_ELEMENT)((char *)Message_Ptr
+ ((I2O_MESSAGE_FRAME_getVersionOffset(Message_Ptr) & 0xF0)
>> 2));
while (sg < (PI2O_SGE_SIMPLE_ELEMENT)(((caddr_t)Message_Ptr)
+ MessageSizeInBytes)) {
caddr_t v;
int len;
if ((I2O_FLAGS_COUNT_getFlags(&(sg->FlagsCount))
& I2O_SGL_FLAGS_SIMPLE_ADDRESS_ELEMENT) == 0) {
error = EINVAL;
break;
}
len = I2O_FLAGS_COUNT_getCount(&(sg->FlagsCount));
debug_usr_cmd_printf ("SG[%d] = %x[%d]\n",
sg - (PI2O_SGE_SIMPLE_ELEMENT)((char *)Message_Ptr
+ ((I2O_MESSAGE_FRAME_getVersionOffset(
Message_Ptr) & 0xF0) >> 2)),
I2O_SGE_SIMPLE_ELEMENT_getPhysicalAddress(sg), len);
if ((elm = (struct ioctlSgList_S *)malloc (
sizeof(*elm) - sizeof(elm->KernelSpace) + len,
Begin the slow, painful process of cleaning up asr. This eliminates many of the bogus casts of NULL. It also distracts me while I procrastinate on useful work.
2004-04-21 03:51:39 +00:00
M_TEMP, M_WAITOK)) == NULL) {
unexpand -a everything
2002-05-14 01:09:03 +00:00
debug_usr_cmd_printf (
"Failed to allocate SG[%d]\n", len);
error = ENOMEM;
break;
}
SLIST_INSERT_HEAD(&sgList, elm, link);
elm->FlagsCount = sg->FlagsCount;
elm->UserSpace = (caddr_t)
(I2O_SGE_SIMPLE_ELEMENT_getPhysicalAddress(sg));
v = elm->KernelSpace;
/* Copy in outgoing data (DIR bit could be invalid) */
if ((error = copyin (elm->UserSpace, (caddr_t)v, len))
!= 0) {
break;
}
/*
* If the buffer is not contiguous, lets
* break up the scatter/gather entries.
*/
while ((len > 0)
&& (sg < (PI2O_SGE_SIMPLE_ELEMENT)
(((caddr_t)Message_Ptr) + MAX_INBOUND_SIZE))) {
int next, base, span;
span = 0;
next = base = KVTOPHYS(v);
I2O_SGE_SIMPLE_ELEMENT_setPhysicalAddress(sg,
base);
/* How far can we go physically contiguously */
while ((len > 0) && (base == next)) {
int size;
next = trunc_page(base) + PAGE_SIZE;
size = next - base;
if (size > len) {
size = len;
}
span += size;
v += size;
len -= size;
base = KVTOPHYS(v);
}
/* Construct the Flags */
I2O_FLAGS_COUNT_setCount(&(sg->FlagsCount),
span);
{
int flags = I2O_FLAGS_COUNT_getFlags(
&(elm->FlagsCount));
/* Any remaining length? */
if (len > 0) {
flags &=
~(I2O_SGL_FLAGS_END_OF_BUFFER
| I2O_SGL_FLAGS_LAST_ELEMENT);
}
I2O_FLAGS_COUNT_setFlags(
&(sg->FlagsCount), flags);
}
debug_usr_cmd_printf ("sg[%d] = %x[%d]\n",
sg - (PI2O_SGE_SIMPLE_ELEMENT)
((char *)Message_Ptr
+ ((I2O_MESSAGE_FRAME_getVersionOffset(
Message_Ptr) & 0xF0) >> 2)),
I2O_SGE_SIMPLE_ELEMENT_getPhysicalAddress(sg),
span);
if (len <= 0) {
break;
}
/*
* Incrementing requires resizing of the
* packet, and moving up the existing SG
* elements.
*/
++sg;
MessageSizeInBytes += sizeof(*sg);
I2O_MESSAGE_FRAME_setMessageSize(Message_Ptr,
I2O_MESSAGE_FRAME_getMessageSize(Message_Ptr)
+ (sizeof(*sg) / sizeof(U32)));
{
PI2O_MESSAGE_FRAME NewMessage_Ptr;
if ((NewMessage_Ptr
= (PI2O_MESSAGE_FRAME)
malloc (MessageSizeInBytes,
Begin the slow, painful process of cleaning up asr. This eliminates many of the bogus casts of NULL. It also distracts me while I procrastinate on useful work.
2004-04-21 03:51:39 +00:00
M_TEMP, M_WAITOK)) == NULL) {
unexpand -a everything
2002-05-14 01:09:03 +00:00
debug_usr_cmd_printf (
"Failed to acquire frame[%d] memory\n",
MessageSizeInBytes);
error = ENOMEM;
break;
}
span = ((caddr_t)sg)
- (caddr_t)Message_Ptr;
Many more style cleanups. Switch complex macros to being inline functions. Put @includes in a better spot. Fix many cases of 2 space indents and spaces between a function name and the parens. Use KASSERT instead of a home-rolled ASSERT. Remove some undeeded caddr casts.
2004-05-01 02:27:06 +00:00
bcopy(Message_Ptr,NewMessage_Ptr, span);
bcopy((caddr_t)(sg-1),
unexpand -a everything
2002-05-14 01:09:03 +00:00
((caddr_t)NewMessage_Ptr) + span,
MessageSizeInBytes - span);
Remove the defAlignLong and getAlignLong macros. I guess that the original intent was to make sure that message structs allocated off of the stack were 4-byte aligned. However, the macros as defined did absolutely nothing. And since I2O forces you to manually copy messages down to the hardware, there really is no point of enforced alignment anyways.
2004-05-02 03:33:18 +00:00
free(Message_Ptr, M_TEMP);
unexpand -a everything
2002-05-14 01:09:03 +00:00
sg = (PI2O_SGE_SIMPLE_ELEMENT)
(((caddr_t)NewMessage_Ptr) + span);
Message_Ptr = NewMessage_Ptr;
}
}
if ((error)
|| ((I2O_FLAGS_COUNT_getFlags(&(sg->FlagsCount))
& I2O_SGL_FLAGS_LAST_ELEMENT) != 0)) {
break;
}
++sg;
}
if (error) {
Begin the slow, painful process of cleaning up asr. This eliminates many of the bogus casts of NULL. It also distracts me while I procrastinate on useful work.
2004-04-21 03:51:39 +00:00
while ((elm = SLIST_FIRST(&sgList)) != NULL) {
unexpand -a everything
2002-05-14 01:09:03 +00:00
SLIST_REMOVE_HEAD(&sgList, link);
Remove the defAlignLong and getAlignLong macros. I guess that the original intent was to make sure that message structs allocated off of the stack were 4-byte aligned. However, the macros as defined did absolutely nothing. And since I2O forces you to manually copy messages down to the hardware, there really is no point of enforced alignment anyways.
2004-05-02 03:33:18 +00:00
free(elm, M_TEMP);
unexpand -a everything
2002-05-14 01:09:03 +00:00
}
Remove the defAlignLong and getAlignLong macros. I guess that the original intent was to make sure that message structs allocated off of the stack were 4-byte aligned. However, the macros as defined did absolutely nothing. And since I2O forces you to manually copy messages down to the hardware, there really is no point of enforced alignment anyways.
2004-05-02 03:33:18 +00:00
free(Reply_Ptr, M_TEMP);
free(Message_Ptr, M_TEMP);
unexpand -a everything
2002-05-14 01:09:03 +00:00
return (error);
}
}
debug_usr_cmd_printf ("Inbound: ");
debug_usr_cmd_dump_message(Message_Ptr);
/* Send the command */
Begin the slow, painful process of cleaning up asr. This eliminates many of the bogus casts of NULL. It also distracts me while I procrastinate on useful work.
2004-04-21 03:51:39 +00:00
if ((ccb = asr_alloc_ccb (sc)) == NULL) {
unexpand -a everything
2002-05-14 01:09:03 +00:00
/* Free up in-kernel buffers */
Begin the slow, painful process of cleaning up asr. This eliminates many of the bogus casts of NULL. It also distracts me while I procrastinate on useful work.
2004-04-21 03:51:39 +00:00
while ((elm = SLIST_FIRST(&sgList)) != NULL) {
unexpand -a everything
2002-05-14 01:09:03 +00:00
SLIST_REMOVE_HEAD(&sgList, link);
Remove the defAlignLong and getAlignLong macros. I guess that the original intent was to make sure that message structs allocated off of the stack were 4-byte aligned. However, the macros as defined did absolutely nothing. And since I2O forces you to manually copy messages down to the hardware, there really is no point of enforced alignment anyways.
2004-05-02 03:33:18 +00:00
free(elm, M_TEMP);
unexpand -a everything
2002-05-14 01:09:03 +00:00
}
Remove the defAlignLong and getAlignLong macros. I guess that the original intent was to make sure that message structs allocated off of the stack were 4-byte aligned. However, the macros as defined did absolutely nothing. And since I2O forces you to manually copy messages down to the hardware, there really is no point of enforced alignment anyways.
2004-05-02 03:33:18 +00:00
free(Reply_Ptr, M_TEMP);
free(Message_Ptr, M_TEMP);
unexpand -a everything
2002-05-14 01:09:03 +00:00
return (ENOMEM);
}
/*
* We do not need any (optional byteswapping) method access to
* the Initiator context field.
*/
I2O_MESSAGE_FRAME_setInitiatorContext64(
(PI2O_MESSAGE_FRAME)Message_Ptr, (long)ccb);
(void)ASR_queue (sc, (PI2O_MESSAGE_FRAME)Message_Ptr);
Remove the defAlignLong and getAlignLong macros. I guess that the original intent was to make sure that message structs allocated off of the stack were 4-byte aligned. However, the macros as defined did absolutely nothing. And since I2O forces you to manually copy messages down to the hardware, there really is no point of enforced alignment anyways.
2004-05-02 03:33:18 +00:00
free(Message_Ptr, M_TEMP);
unexpand -a everything
2002-05-14 01:09:03 +00:00
/*
* Wait for the board to report a finished instruction.
*/
s = splcam();
while ((ccb->ccb_h.status & CAM_STATUS_MASK) == CAM_REQ_INPROG) {
if (ASR_getBlinkLedCode(sc)) {
/* Reset Adapter */
printf ("asr%d: Blink LED 0x%x resetting adapter\n",
cam_sim_unit(xpt_path_sim(ccb->ccb_h.path)),
ASR_getBlinkLedCode(sc));
if (ASR_reset (sc) == ENXIO) {
/* Command Cleanup */
ASR_ccbRemove(sc, ccb);
}
splx(s);
/* Free up in-kernel buffers */
Begin the slow, painful process of cleaning up asr. This eliminates many of the bogus casts of NULL. It also distracts me while I procrastinate on useful work.
2004-04-21 03:51:39 +00:00
while ((elm = SLIST_FIRST(&sgList)) != NULL) {
unexpand -a everything
2002-05-14 01:09:03 +00:00
SLIST_REMOVE_HEAD(&sgList, link);
Remove the defAlignLong and getAlignLong macros. I guess that the original intent was to make sure that message structs allocated off of the stack were 4-byte aligned. However, the macros as defined did absolutely nothing. And since I2O forces you to manually copy messages down to the hardware, there really is no point of enforced alignment anyways.
2004-05-02 03:33:18 +00:00
free(elm, M_TEMP);
unexpand -a everything
2002-05-14 01:09:03 +00:00
}
Remove the defAlignLong and getAlignLong macros. I guess that the original intent was to make sure that message structs allocated off of the stack were 4-byte aligned. However, the macros as defined did absolutely nothing. And since I2O forces you to manually copy messages down to the hardware, there really is no point of enforced alignment anyways.
2004-05-02 03:33:18 +00:00
free(Reply_Ptr, M_TEMP);
unexpand -a everything
2002-05-14 01:09:03 +00:00
asr_free_ccb(ccb);
return (EIO);
}
/* Check every second for BlinkLed */
/* There is no PRICAM, but outwardly PRIBIO is functional */
More low-hanging fruit: kill caddr_t in calls to wakeup(9) / [mt]sleep(9).
2003-03-02 16:54:40 +00:00
tsleep(ccb, PRIBIO, "asr", hz);
unexpand -a everything
2002-05-14 01:09:03 +00:00
}
splx(s);
debug_usr_cmd_printf ("Outbound: ");
debug_usr_cmd_dump_message(Reply_Ptr);
I2O_SINGLE_REPLY_MESSAGE_FRAME_setDetailedStatusCode(
&(Reply_Ptr->StdReplyFrame),
(ccb->ccb_h.status != CAM_REQ_CMP));
if (ReplySizeInBytes >= (sizeof(I2O_SCSI_ERROR_REPLY_MESSAGE_FRAME)
- I2O_SCSI_SENSE_DATA_SZ - sizeof(U32))) {
I2O_SCSI_ERROR_REPLY_MESSAGE_FRAME_setTransferCount(Reply_Ptr,
ccb->csio.dxfer_len - ccb->csio.resid);
}
if ((ccb->ccb_h.status & CAM_AUTOSNS_VALID) && (ReplySizeInBytes
> (sizeof(I2O_SCSI_ERROR_REPLY_MESSAGE_FRAME)
- I2O_SCSI_SENSE_DATA_SZ))) {
int size = ReplySizeInBytes
- sizeof(I2O_SCSI_ERROR_REPLY_MESSAGE_FRAME)
- I2O_SCSI_SENSE_DATA_SZ;
if (size > sizeof(ccb->csio.sense_data)) {
size = sizeof(ccb->csio.sense_data);
}
Many more style cleanups. Switch complex macros to being inline functions. Put @includes in a better spot. Fix many cases of 2 space indents and spaces between a function name and the parens. Use KASSERT instead of a home-rolled ASSERT. Remove some undeeded caddr casts.
2004-05-01 02:27:06 +00:00
bcopy(&(ccb->csio.sense_data), Reply_Ptr->SenseData, size);
unexpand -a everything
2002-05-14 01:09:03 +00:00
I2O_SCSI_ERROR_REPLY_MESSAGE_FRAME_setAutoSenseTransferCount(
Many more style cleanups. Switch complex macros to being inline functions. Put @includes in a better spot. Fix many cases of 2 space indents and spaces between a function name and the parens. Use KASSERT instead of a home-rolled ASSERT. Remove some undeeded caddr casts.
2004-05-01 02:27:06 +00:00
Reply_Ptr, size);
unexpand -a everything
2002-05-14 01:09:03 +00:00
}
/* Free up in-kernel buffers */
Begin the slow, painful process of cleaning up asr. This eliminates many of the bogus casts of NULL. It also distracts me while I procrastinate on useful work.
2004-04-21 03:51:39 +00:00
while ((elm = SLIST_FIRST(&sgList)) != NULL) {
unexpand -a everything
2002-05-14 01:09:03 +00:00
/* Copy out as necessary */
if ((error == 0)
/* DIR bit considered `valid', error due to ignorance works */
&& ((I2O_FLAGS_COUNT_getFlags(&(elm->FlagsCount))
& I2O_SGL_FLAGS_DIR) == 0)) {
Many more style cleanups. Switch complex macros to being inline functions. Put @includes in a better spot. Fix many cases of 2 space indents and spaces between a function name and the parens. Use KASSERT instead of a home-rolled ASSERT. Remove some undeeded caddr casts.
2004-05-01 02:27:06 +00:00
error = copyout((caddr_t)(elm->KernelSpace),
unexpand -a everything
2002-05-14 01:09:03 +00:00
elm->UserSpace,
I2O_FLAGS_COUNT_getCount(&(elm->FlagsCount)));
}
SLIST_REMOVE_HEAD(&sgList, link);
Many more style cleanups. Switch complex macros to being inline functions. Put @includes in a better spot. Fix many cases of 2 space indents and spaces between a function name and the parens. Use KASSERT instead of a home-rolled ASSERT. Remove some undeeded caddr casts.
2004-05-01 02:27:06 +00:00
free(elm, M_TEMP);
unexpand -a everything
2002-05-14 01:09:03 +00:00
}
if (error == 0) {
/* Copy reply frame to user space */
Many more style cleanups. Switch complex macros to being inline functions. Put @includes in a better spot. Fix many cases of 2 space indents and spaces between a function name and the parens. Use KASSERT instead of a home-rolled ASSERT. Remove some undeeded caddr casts.
2004-05-01 02:27:06 +00:00
error = copyout((caddr_t)Reply_Ptr, (caddr_t)Reply,
ReplySizeInBytes);
unexpand -a everything
2002-05-14 01:09:03 +00:00
}
Many more style cleanups. Switch complex macros to being inline functions. Put @includes in a better spot. Fix many cases of 2 space indents and spaces between a function name and the parens. Use KASSERT instead of a home-rolled ASSERT. Remove some undeeded caddr casts.
2004-05-01 02:27:06 +00:00
free(Reply_Ptr, M_TEMP);
unexpand -a everything
2002-05-14 01:09:03 +00:00
asr_free_ccb(ccb);
return (error);
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
} /* ASR_queue_i */
/*----------------------------------------------------------------------*/
unexpand -a everything
2002-05-14 01:09:03 +00:00
/* Function asr_ioctl */
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
/*----------------------------------------------------------------------*/
unexpand -a everything
2002-05-14 01:09:03 +00:00
/* The parameters passed to this function are : */
/* dev : Device number. */
/* cmd : Ioctl Command */
/* data : User Argument Passed In. */
/* flag : Mode Parameter */
/* proc : Process Parameter */
/* */
/* This function is the user interface into this adapter driver */
/* */
/* Return : zero if OK, error code if not */
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
/*----------------------------------------------------------------------*/
Get rid of a bunch of useless macros. STATIC becomes static, INLINE becomes __inline where appropriate and gets nuked elsewhere, IN/OUT/INOUT go away. Reformat code affected by this.
2004-05-01 01:25:05 +00:00
static int
Do the dreaded s/dev_t/struct cdev */ Bump __FreeBSD_version accordingly.
2004-06-16 09:47:26 +00:00
asr_ioctl(struct cdev *dev, u_long cmd, caddr_t data, int flag, struct thread *td)
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
{
Remove ASR_get_sc() and reference the softc in the dev_t. For some nefarious reason, the I2O protocol requires knowledge of all I2O devices in the system, so we can't get rid of the evil linked-list of softc's yet.
2004-05-01 05:56:57 +00:00
Asr_softc_t *sc = dev->si_drv1;
Roto-till the ioctl code. Remove all the groping around in the x86 BIOS segment, remove the groping around in the Option ROM segments, remove the bogus tests for bcopy vs. copyout. There really is no reason for a management app to know these things other than to create l33t info tables for the user.
2004-05-05 05:29:19 +00:00
int i, error = 0;
Bring back some of the ioctl junk that was removed in rev 1.59 as a i386-only kernel option, ASR_COMPAT, and under BURN_BRIDGES. It is really ugly, but raidutils depends on it. Discussed with: scottl
2005-03-17 01:20:49 +00:00
#ifdef ASR_IOCTL_COMPAT
int j;
#endif /* ASR_IOCTL_COMPAT */
unexpand -a everything
2002-05-14 01:09:03 +00:00
Begin the slow, painful process of cleaning up asr. This eliminates many of the bogus casts of NULL. It also distracts me while I procrastinate on useful work.
2004-04-21 03:51:39 +00:00
if (sc != NULL)
unexpand -a everything
2002-05-14 01:09:03 +00:00
switch(cmd) {
case DPT_SIGNATURE:
Bring back some of the ioctl junk that was removed in rev 1.59 as a i386-only kernel option, ASR_COMPAT, and under BURN_BRIDGES. It is really ugly, but raidutils depends on it. Discussed with: scottl
2005-03-17 01:20:49 +00:00
#ifdef ASR_IOCTL_COMPAT
#if (dsDescription_size != 50)
case DPT_SIGNATURE + ((50 - dsDescription_size) << 16):
#endif
if (cmd & 0xFFFF0000) {
bcopy(&ASR_sig, data, sizeof(dpt_sig_S));
return (0);
}
/* Traditional version of the ioctl interface */
case DPT_SIGNATURE & 0x0000FFFF:
#endif
Roto-till the ioctl code. Remove all the groping around in the x86 BIOS segment, remove the groping around in the Option ROM segments, remove the bogus tests for bcopy vs. copyout. There really is no reason for a management app to know these things other than to create l33t info tables for the user.
2004-05-05 05:29:19 +00:00
return (copyout((caddr_t)(&ASR_sig), *((caddr_t *)data),
sizeof(dpt_sig_S)));
unexpand -a everything
2002-05-14 01:09:03 +00:00
/* Traditional version of the ioctl interface */
case DPT_CTRLINFO & 0x0000FFFF:
case DPT_CTRLINFO: {
struct {
u_int16_t length;
u_int16_t drvrHBAnum;
u_int32_t baseAddr;
u_int16_t blinkState;
u_int8_t pciBusNum;
u_int8_t pciDeviceNum;
u_int16_t hbaFlags;
u_int16_t Interrupt;
u_int32_t reserved1;
u_int32_t reserved2;
u_int32_t reserved3;
} CtlrInfo;
More whitespace style cleanups, also remove unneeded (void *) casts for bzero().
2004-05-01 03:06:54 +00:00
bzero(&CtlrInfo, sizeof(CtlrInfo));
unexpand -a everything
2002-05-14 01:09:03 +00:00
CtlrInfo.length = sizeof(CtlrInfo) - sizeof(u_int16_t);
CtlrInfo.drvrHBAnum = asr_unit(dev);
Convert the asr driver to use the bus_space API. This does not represent a significant functional change, but it further cleans up the code and brings it closer to being portable. Thanks to Don Bowman for helping to test this.
2004-05-05 03:42:33 +00:00
CtlrInfo.baseAddr = sc->ha_Base;
unexpand -a everything
2002-05-14 01:09:03 +00:00
i = ASR_getBlinkLedCode (sc);
Roto-till the ioctl code. Remove all the groping around in the x86 BIOS segment, remove the groping around in the Option ROM segments, remove the bogus tests for bcopy vs. copyout. There really is no reason for a management app to know these things other than to create l33t info tables for the user.
2004-05-05 05:29:19 +00:00
if (i == -1)
unexpand -a everything
2002-05-14 01:09:03 +00:00
i = 0;
Roto-till the ioctl code. Remove all the groping around in the x86 BIOS segment, remove the groping around in the Option ROM segments, remove the bogus tests for bcopy vs. copyout. There really is no reason for a management app to know these things other than to create l33t info tables for the user.
2004-05-05 05:29:19 +00:00
unexpand -a everything
2002-05-14 01:09:03 +00:00
CtlrInfo.blinkState = i;
CtlrInfo.pciBusNum = sc->ha_pciBusNum;
CtlrInfo.pciDeviceNum = sc->ha_pciDeviceNum;
`unexpand -a' should be run _before_ sed 's/^#define /#define^I/g'.
2002-05-14 21:59:10 +00:00
#define FLG_OSD_PCI_VALID 0x0001
#define FLG_OSD_DMA 0x0002
#define FLG_OSD_I2O 0x0004
Many more style cleanups. Switch complex macros to being inline functions. Put @includes in a better spot. Fix many cases of 2 space indents and spaces between a function name and the parens. Use KASSERT instead of a home-rolled ASSERT. Remove some undeeded caddr casts.
2004-05-01 02:27:06 +00:00
CtlrInfo.hbaFlags = FLG_OSD_PCI_VALID|FLG_OSD_DMA|FLG_OSD_I2O;
unexpand -a everything
2002-05-14 01:09:03 +00:00
CtlrInfo.Interrupt = sc->ha_irq;
Bring back some of the ioctl junk that was removed in rev 1.59 as a i386-only kernel option, ASR_COMPAT, and under BURN_BRIDGES. It is really ugly, but raidutils depends on it. Discussed with: scottl
2005-03-17 01:20:49 +00:00
#ifdef ASR_IOCTL_COMPAT
if (cmd & 0xffff0000)
bcopy(&CtlrInfo, data, sizeof(CtlrInfo));
else
#endif /* ASR_IOCTL_COMPAT */
Roto-till the ioctl code. Remove all the groping around in the x86 BIOS segment, remove the groping around in the Option ROM segments, remove the bogus tests for bcopy vs. copyout. There really is no reason for a management app to know these things other than to create l33t info tables for the user.
2004-05-05 05:29:19 +00:00
error = copyout(&CtlrInfo, *(caddr_t *)data, sizeof(CtlrInfo));
unexpand -a everything
2002-05-14 01:09:03 +00:00
} return (error);
/* Traditional version of the ioctl interface */
case DPT_SYSINFO & 0x0000FFFF:
case DPT_SYSINFO: {
sysInfo_S Info;
Bring back some of the ioctl junk that was removed in rev 1.59 as a i386-only kernel option, ASR_COMPAT, and under BURN_BRIDGES. It is really ugly, but raidutils depends on it. Discussed with: scottl
2005-03-17 01:20:49 +00:00
#ifdef ASR_IOCTL_COMPAT
char * cp;
/* Kernel Specific ptok `hack' */
#define ptok(a) ((char *)(uintptr_t)(a) + KERNBASE)
bzero(&Info, sizeof(Info));
/* Appears I am the only person in the Kernel doing this */
outb (0x70, 0x12);
i = inb(0x71);
j = i >> 4;
if (i == 0x0f) {
outb (0x70, 0x19);
j = inb (0x71);
}
Info.drive0CMOS = j;
j = i & 0x0f;
if (i == 0x0f) {
outb (0x70, 0x1a);
j = inb (0x71);
}
Info.drive1CMOS = j;
Info.numDrives = *((char *)ptok(0x475));
#endif /* ASR_IOCTL_COMPAT */
unexpand -a everything
2002-05-14 01:09:03 +00:00
Many more style cleanups. Switch complex macros to being inline functions. Put @includes in a better spot. Fix many cases of 2 space indents and spaces between a function name and the parens. Use KASSERT instead of a home-rolled ASSERT. Remove some undeeded caddr casts.
2004-05-01 02:27:06 +00:00
bzero(&Info, sizeof(Info));
unexpand -a everything
2002-05-14 01:09:03 +00:00
Info.processorFamily = ASR_sig.dsProcessorFamily;
Tighten up the defined()'s.
2002-05-14 21:54:56 +00:00
#if defined(__i386__)
unexpand -a everything
2002-05-14 01:09:03 +00:00
switch (cpu) {
case CPU_386SX: case CPU_386:
Info.processorType = PROC_386; break;
case CPU_486SX: case CPU_486:
Info.processorType = PROC_486; break;
case CPU_586:
Info.processorType = PROC_PENTIUM; break;
case CPU_686:
Info.processorType = PROC_SEXIUM; break;
}
Make an attempt to get the asr driver to compile on Alpha by fixing some i386 specific bogons. Compile with -O0, as anything higher gives the compiler a fit. No idea if this driver will actually work on Alpha, though.
2001-04-01 08:33:01 +00:00
#endif
unexpand -a everything
2002-05-14 01:09:03 +00:00
Info.osType = OS_BSDI_UNIX;
Info.osMajorVersion = osrelease[0] - '0';
Info.osMinorVersion = osrelease[2] - '0';
/* Info.osRevision = 0; */
/* Info.osSubRevision = 0; */
Info.busType = SI_PCI_BUS;
Roto-till the ioctl code. Remove all the groping around in the x86 BIOS segment, remove the groping around in the Option ROM segments, remove the bogus tests for bcopy vs. copyout. There really is no reason for a management app to know these things other than to create l33t info tables for the user.
2004-05-05 05:29:19 +00:00
Info.flags = SI_OSversionValid|SI_BusTypeValid|SI_NO_SmartROM;
unexpand -a everything
2002-05-14 01:09:03 +00:00
Bring back some of the ioctl junk that was removed in rev 1.59 as a i386-only kernel option, ASR_COMPAT, and under BURN_BRIDGES. It is really ugly, but raidutils depends on it. Discussed with: scottl
2005-03-17 01:20:49 +00:00
#ifdef ASR_IOCTL_COMPAT
Info.flags |= SI_CMOS_Valid | SI_NumDrivesValid;
/* Go Out And Look For I2O SmartROM */
for(j = 0xC8000; j < 0xE0000; j += 2048) {
int k;
cp = ptok(j);
if (*((unsigned short *)cp) != 0xAA55) {
continue;
}
j += (cp[2] * 512) - 2048;
if ((*((u_long *)(cp + 6))
!= ('S' + (' ' * 256) + (' ' * 65536L)))
|| (*((u_long *)(cp + 10))
!= ('I' + ('2' * 256) + ('0' * 65536L)))) {
continue;
}
cp += 0x24;
for (k = 0; k < 64; ++k) {
if (*((unsigned short *)cp)
== (' ' + ('v' * 256))) {
break;
}
}
if (k < 64) {
Info.smartROMMajorVersion
= *((unsigned char *)(cp += 4)) - '0';
Info.smartROMMinorVersion
= *((unsigned char *)(cp += 2));
Info.smartROMRevision
= *((unsigned char *)(++cp));
Info.flags |= SI_SmartROMverValid;
Info.flags &= ~SI_NO_SmartROM;
break;
}
}
/* Get The Conventional Memory Size From CMOS */
outb (0x70, 0x16);
j = inb (0x71);
j <<= 8;
outb (0x70, 0x15);
j |= inb(0x71);
Info.conventionalMemSize = j;
/* Get The Extended Memory Found At Power On From CMOS */
outb (0x70, 0x31);
j = inb (0x71);
j <<= 8;
outb (0x70, 0x30);
j |= inb(0x71);
Info.extendedMemSize = j;
Info.flags |= SI_MemorySizeValid;
/* Copy Out The Info Structure To The User */
if (cmd & 0xFFFF0000)
bcopy(&Info, data, sizeof(Info));
else
#endif /* ASR_IOCTL_COMPAT */
Roto-till the ioctl code. Remove all the groping around in the x86 BIOS segment, remove the groping around in the Option ROM segments, remove the bogus tests for bcopy vs. copyout. There really is no reason for a management app to know these things other than to create l33t info tables for the user.
2004-05-05 05:29:19 +00:00
error = copyout(&Info, *(caddr_t *)data, sizeof(Info));
unexpand -a everything
2002-05-14 01:09:03 +00:00
return (error); }
/* Get The BlinkLED State */
case DPT_BLINKLED:
i = ASR_getBlinkLedCode (sc);
Bring back some of the ioctl junk that was removed in rev 1.59 as a i386-only kernel option, ASR_COMPAT, and under BURN_BRIDGES. It is really ugly, but raidutils depends on it. Discussed with: scottl
2005-03-17 01:20:49 +00:00
if (i == -1)
unexpand -a everything
2002-05-14 01:09:03 +00:00
i = 0;
Bring back some of the ioctl junk that was removed in rev 1.59 as a i386-only kernel option, ASR_COMPAT, and under BURN_BRIDGES. It is really ugly, but raidutils depends on it. Discussed with: scottl
2005-03-17 01:20:49 +00:00
#ifdef ASR_IOCTL_COMPAT
if (cmd & 0xffff0000)
bcopy(&i, data, sizeof(i));
else
#endif /* ASR_IOCTL_COMPAT */
Roto-till the ioctl code. Remove all the groping around in the x86 BIOS segment, remove the groping around in the Option ROM segments, remove the bogus tests for bcopy vs. copyout. There really is no reason for a management app to know these things other than to create l33t info tables for the user.
2004-05-05 05:29:19 +00:00
error = copyout(&i, *(caddr_t *)data, sizeof(i));
unexpand -a everything
2002-05-14 01:09:03 +00:00
break;
/* Send an I2O command */
case I2OUSRCMD:
Many more style cleanups. Switch complex macros to being inline functions. Put @includes in a better spot. Fix many cases of 2 space indents and spaces between a function name and the parens. Use KASSERT instead of a home-rolled ASSERT. Remove some undeeded caddr casts.
2004-05-01 02:27:06 +00:00
return (ASR_queue_i(sc, *((PI2O_MESSAGE_FRAME *)data)));
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
unexpand -a everything
2002-05-14 01:09:03 +00:00
/* Reset and re-initialize the adapter */
case I2ORESETCMD:
Many more style cleanups. Switch complex macros to being inline functions. Put @includes in a better spot. Fix many cases of 2 space indents and spaces between a function name and the parens. Use KASSERT instead of a home-rolled ASSERT. Remove some undeeded caddr casts.
2004-05-01 02:27:06 +00:00
return (ASR_reset(sc));
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
unexpand -a everything
2002-05-14 01:09:03 +00:00
/* Rescan the LCT table and resynchronize the information */
case I2ORESCANCMD:
Many more style cleanups. Switch complex macros to being inline functions. Put @includes in a better spot. Fix many cases of 2 space indents and spaces between a function name and the parens. Use KASSERT instead of a home-rolled ASSERT. Remove some undeeded caddr casts.
2004-05-01 02:27:06 +00:00
return (ASR_rescan(sc));
unexpand -a everything
2002-05-14 01:09:03 +00:00
}
return (EINVAL);
Add the 'asr' driver, supplied by Mark Salyzyn of Adaptec (nee DPT). This provides support for the Adaptec SCSI RAID controller family, as well as the DPT SmartRAID V and VI families. The driver will be maintained by Mark and Adaptec, and any changes should be referred to the MAINTAINER.
2000-09-01 07:51:25 +00:00
} /* asr_ioctl */
Reference in New Issue Copy Permalink