ci_lpmaint.c

<B>Digital的Unix操作系统VAX 4.2源码</B>

#ifndef	lint
static char *sccsid = "@(#)ci_lpmaint.c	4.1	(ULTRIX)	7/2/90";
#endif	lint

/************************************************************************
 *                                                                      *
 *                      Copyright (c) 1988 - 1989 by                    *
 *              Digital Equipment Corporation, Maynard, MA              *
 *                      All rights reserved.                            *
 *                                                                      *
 *   This software is furnished under a license and may be used and     *
 *   copied  only  in accordance with the terms of such license and     *
 *   with the  inclusion  of  the  above  copyright  notice.   This     *
 *   software  or  any  other copies thereof may not be provided or     *
 *   otherwise made available to any other person.  No title to and     *
 *   ownership of the software is hereby transferred.                   *
 *                                                                      *
 *   The information in this software is subject to change  without     *
 *   notice  and should not be construed as a commitment by Digital     *
 *   Equipment Corporation.                                             *
 *                                                                      *
 *   Digital assumes no responsibility for the use  or  reliability     *
 *   of its software on equipment which is not supplied by Digital.     *
 *                                                                      *
 ************************************************************************
 *
 *
 *   Facility:	Systems Communication Architecture
 *		Computer Interconnect Port Driver
 *
 *   Abstract:	This module contains Computer Interconnect Port Driver( CI )
 *		local port maintenance service functions.
 *
 *   Creator:   Todd M. Katz    Creation Date:  May 6, 1985
 *
 *   Function/Routines:
 *
 *   ci_crash_lport		Crash Local CI Port
 *   ci_remote_reset		Reset Remote Port and System
 *   ci_remote_start		Start Remote Port and System
 *   ci_shutdown		Inform Known Systems of Local Shutdown
 *
 *   Modification History:
 *
 *   06-Jun-1990	Pete Keilty
 *	Added six new error cases to ci_crash_lport that the CIXCD 
 *	can generate. They are SE_INVPA, SE_INVSN, SE_INVDDL, SE_IRESVCD,
 *	SE_IRESEQ, SE_DISCVCPKT.
 *
 *   19-Sep-1989	Pete Keilty
 *	1. Add pccb to macro Format_gvph & Pd_to_ppd.
 *	2. Add XCD removed XCB.
 *
 *   06-Apr-1989	Pete Keilty
 *	Added include file smp_lock.h
 *
 *   17-Jan-1989	Todd M. Katz		TMK0005
 *	1. The macro Scaaddr_lol() has been renamed to Scaaddr_low().  It now
 *	   accesses only the low order word( instead of low order longword ) of
 *	   a SCA system address.
 *	2. Include header file ../vaxmsi/msisysap.h.
 *
 *   18-Aug-1988	Todd M. Katz		TMK0004
 *	1. SCA event codes have been completed revised.  All former local port
 *	   crash codes are now defined as either severe or fatal error events.
 *	2. LPC_BROKEN has been renamed -> SE_PPDSANITY.  This event is now
 *	   logged by ci_log_packet() instead of by ci_log_dev_attn().
 *	3. Modify ci_crash_lport() to apply the local port crash severity
 *	   modifier( ESM_LPC ) to the local port crash reason code whenever
 *	   the local port is not in the process of being cleaned up from a
 *	   previous crash.
 *	4. Modify ci_shutdown() to explicitly take offline the local CI port it
 *	   is shutting down.  This action effectively disables all CI PPD port
 *	   polling on the local CI port.
 *	5. Refer to error logging as event logging.
 *
 *   03-Jun-1988	Todd M. Katz		TMK0003
 *	1. Create a single unified hierarchical set of naming conventions for
 *	   use within the CI port driver and describe them within ciport.h.
 *	   Apply these conventions to all names( routine, macro, constant, and
 *	   data structure ) used within the driver.  Restructure the driver to
 *	   segregate most CI family and port type specific code into separate
 *	   routines.  Such restructuring requires ci_crash_lport() and
 *	   ci_shutdown() to indirectly invoke appropriate CI family specific
 *	   local CI port disablement routines instead of directly invoking
 *	   ci_disable_port() as was previously done.
 *	2. Modify ci_crash_lport() to correctly ascertain local port state when
 *	   processing exhausted message free queue local port crash reasons(
 *	   LPC_MFQE ).  The incorrect assumption was being made that ports are
 *	   not in the Enabled port state when they report this error.
 *	3. Modify the circumstances under which local ports are unmapped.
 *	   Unmapping is now done only when the local port adapter itself loses
 *	   power( CI750/CI780/CIBCI only ) or the port is marked broken and is
 *	   permanently shutdown.  Formerly, unmapping was done whenever a local
 *	   port was crashed, but before its re-initialization commenced; and
 *	   just immediately prior to its initial initialization.  This changes
 *	   requires appropriate modifications to the comments of the routine
 *	   ci_crash_port().
 *	4. The process under which the CI7b family of CI ports undergoes power
 *	   failure recovery has drastically changed.  Modify the comments to
 *	   ci_crash_port() to reflect this change and add comments describing
 *	   how broken ports are crashed.
 *	5. CI device attention event packets used to reserve space for fields
 *	   initialized only in the case of certain events.  This has changed.
 *	   All CI event packets are now tailered exactly to the event being
 *	   logged.  Therefore, there is no longer any need to initialize the
 *	   PCCB optional device attention information structure ( using the
 *	   macro Init_dattnopt()) since this structure is no longer blindly
 *	   copied into each and every device attention event log packet.
 *	   ci_crash_lport() has been changed appropriately.
 *	6. Modify ci_shutdown() to mark each local CI port broken prior to
 *	   disablement.  This forces complete disablement of the local port
 *	   including unmapping of the local port adapter I/O space.
 *
 *   02-Jun-1988	Ricky S. Palmer
 *      Removed inclusion of header file ../vaxmsi/msisysap.h
 *
 *   09-Apr-1988	Todd M. Katz		TMK0002
 *	1. Add support for the CIBCA-BA hardware port type.  Differentiate
 *	   CIBCA-BA from CIBCA-AA hardware ports when necessary; otherwise,
 *	   refer to both types as just CIBCA ports.
 *	2. Add use of Pccb_fork() macro in place of straight-line code.
 *	3. Modify ci_crash_lport() to optionally panic on requests to crash
 *	   local ports.  Error recovery is the default action taken; however,
 *	   panicing occurs whenever the CI configuration variable ci_lpc_panic
 *	   has been set.
 *
 *   08-Jan-1988	Todd M. Katz		TMK0001
 *	Formated module, revised comments, increased generality and
 *	robustness, made CI PPD and GVP completely independent from underlying
 *	port drivers, restructured code paths, and added SMP support.
 */

/* Libraries and Include Files.
 */
#include		"../h/types.h"
#include		"../h/param.h"
#include		"../h/ksched.h"
#include		"../h/time.h"
#include		"../h/errlog.h"
#include		"../h/smp_lock.h"
#include		"../io/scs/sca.h"
#include		"../io/scs/scaparam.h"
#include		"../io/ci/cippdsysap.h"
#include		"../io/ci/cisysap.h"
#include		"../io/msi/msisysap.h"
#include		"../io/bi/bvpsysap.h"
#include		"../io/gvp/gvpsysap.h"
#include		"../io/uba/uqsysap.h"
#include		"../io/sysap/sysap.h"
#include		"../io/ci/cippdscs.h"
#include		"../io/ci/ciscs.h"
#include		"../io/msi/msiscs.h"
#include		"../io/bi/bvpscs.h"
#include		"../io/gvp/gvpscs.h"
#include		"../io/uba/uqscs.h"
#include		"../io/scs/scs.h"
#include		"../io/gvp/gvp.h"
#include		"../io/ci/cippd.h"
#include		"../io/ci/ciport.h"

/* External Variables and Routines.
 */
extern	u_long		gvp_queue_retry;
extern	u_short		ci_lpc_panic, ci_max_reinits;
extern	GVPH		*ci_alloc_pkt();
extern	PB		*cippd_get_pb();
extern	void		ci_cleanup_port(), ci_dealloc_pkt(), ci_log_dev_attn(),
			ci_log_packet();

/*   Name:	ci_crash_lport	- Crash CI Local Port
 *
 *   Abstract:	This routine crashes a specific local CI port and initiates the
 *		process by which it is disabled, cleaned up, and
 *		re-initialized.  It may be sequentially invoked any number of
 *		times to crash a specific port; however, only its first
 *		invocation actually results in crashing of the port.  All
 *		subsequent invocations only log crash requests and dispose of
 *		optional buffers.  This situation persists until after the next
 *		incarnation of the port begins following completion of port
 *		clean up and re-initialization.
 *
 *		Power failure recovery of the CI7B family of CI ports( CI750,
 *		CI780, and CIBCI ) deserves special mentioning because of the
 *		potential of the ports themselves to power fail independently
 *		of the rest of the system.  An independent loss of power is
 *		detected by the port, triggers an interrupt, and results in
 *		crashing of the port( reason code == SE_POWER ) by the
 *		appropriate interrupt service routine.  The local port is
 *		disabled and cleaned up as it would be for any other event.
 *		Unfortunately, the local port can not be re-initialized while
 *		there is no power.
 *
 *		Restoration of power to the port may occur at either of the
 *		following two times:
 *
 *		1. Before port disablement and clean up completes.
 *		2. Sometime after port disablement and clean up completes.
 *
 *		At both times it is a special routine( ci_unmapped_isr()) which
 *		services the interrupt triggered by detection of power
 *		restoration by the local port.  This special routine is in
 *		position to handle the interrupt instead of the routine
 *		normally employed by the local port because at the time the
 *		local port was disabled it was also unmapped as protection
 *		against extraneous machine checks.  This unmapping included
 *		switching the interrupt service handler for the local port to
 *		this special routine.
 *
 *		When the special routine is invoked it re-maps the local port.
 *		This action switches the interrupt service hander for the local
 *		port back to the routine appropriate for the local port's
 *		hardware port type.  The special routine also marks the local
 *		port as having power.  What it does next is dependent upon when
 *		power was restored to the local port.  It does nothing more if
 *		port disablement and clean up are still in progress.  Port
 *		re-initialization occurs immediately following port clean up
 *		as it would for any other event.  However, if port clean up has
 *		completed then the special routine explicitly schedules port
 *		re-initialization before dismissing the interrupt.  The end
 *		result is the same regardless of when local port power is
 *		restored, successful recovery from independent failure through
 *		re-initialization of the local port.
 *
 *		The CIBX family of CI ports( CIBCA, CIXCD ) is not subject to
 *		such Byzantine recovery scenarios.  This is because such ports
 *		are unable to detect independent losses of power; and thus, may
 *		not power fail independently of the rest of the system.
 *
 *		An additional explanation is also warranted for those instances
 *		when this routine is invoked to crash a local port and take it
 *		permanently offline.  Such instances are indicated by the local
 *		port marked broken.  The specific fatal event maybe any one of
 *		a number of codes.  The actions taken are the same as for any
 *		other event, the local port is disabled, cleaned up, and the
 *		initialization routine is scheduled.  However, the local port
 *		initialization routine detects that the local port is marked
 *		broken, aborts port re-initialization, and takes the local port
 *		permanently offline.  Furthermore, when the local port was
 *		disabled it was also unmapped because it was marked broken at
 *		the time.  This act of unmapping included switching the
 *		interrupt service handler for the broken port to a special
 *		routine, ci_unmapped_isr().  Note that this is the same routine
 *		used as the interrupt service handler during power failure
 *		recovery of the CI7B family of CI ports.
 *
 *		NOTE: All error recovery is bypassed in favor of immediately
 *		      panicing the system whenever the CI configuration
 *		      variable flag ci_lpc_panic is set.  Normally, this flag
 *		      is 0 and error recovery is the default action taken.  It
 *		      is set only when special debugging activity is required.
 *