xref: /illumos-gate/usr/src/cmd/fm/eversholt/files/i386/i86pc/amd64.esc (revision 20c794b3)

/*
 * CDDL HEADER START
 *
 * The contents of this file are subject to the terms of the
 * Common Development and Distribution License (the "License").
 * You may not use this file except in compliance with the License.
 *
 * You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
 * or http://www.opensolaris.org/os/licensing.
 * See the License for the specific language governing permissions
 * and limitations under the License.
 *
 * When distributing Covered Code, include this CDDL HEADER in each
 * file and include the License file at usr/src/OPENSOLARIS.LICENSE.
 * If applicable, add the following below this CDDL HEADER, with the
 * fields enclosed by brackets "[]" replaced with your own identifying
 * information: Portions Copyright [yyyy] [name of copyright owner]
 *
 * CDDL HEADER END
 */

/*
 * Copyright 2007 Sun Microsystems, Inc.  All rights reserved.
 * Use is subject to license terms.
 */

#pragma ident	"%Z%%M%	%I%	%E% SMI"

#pragma dictionary "AMD"

/*
 * Eversholt rules for the AMD Opteron CPU/Memory
 */

fru motherboard;
fru chip;
fru dimm;

asru chip/cpu;
asru dimm;
asru dimm/rank;
asru dram-channel;
asru chip/memory-controller/chip-select;

#define	MAX(x, y) ((x) >= (y) ? (x) : (y))
#define	MIN(x, y) ((x) <= (y) ? (x) : (y))

/*
 * SET_ADDR and SET_OFFSET are used to set a payload value in the fault that
 * we diagnose for page faults, to record the physical address of the faulting
 * page.  The "asru-" prefix is hooked in the "rewrite-ASRU" confcalls made on
 * diagnosis of associated faults when the libtopo mem scheme rewrites the
 * asru in "mem" scheme.
 */
#define	SET_ADDR (setpayloadprop("asru-physaddr", payloadprop("IA32_MCi_ADDR")))

#define	SET_OFFSET (setpayloadprop("asru-offset", \
	payloadprop("resource[0].hc-specific.offset")))

/*
 * RESOURCE_EXISTS is true if a member with name "resource" exists in the
 * payload - regardless of type (e.g., nvlist or nvlist array) or value.
 */
#define	RESOURCE_EXISTS	(payloadprop_defined("resource"))

/*
 * CONTAINS_RANK is true if the "resource" nvlist array (as used in memory
 * ereports) exists and one if its members matches the path for the
 * rank node.  Our memory propogation are of the form
 *
 * "prop foo@chip/memory-controller/dimm/rank -> blah@chip/cpu"
 *
 * since cpus detect memory errors;  in eversholt such a propogation, where
 * the lhs path and rhs path do not match, expands to the cross-product of
 * all dimms, ranks and cpus on the same chip (since chip appears in the
 * path on both sides).  We use CONTAINS_RANK to constrain the propogation
 * such that it only happens if the payload resource matches the rank.
 */
#define	CONTAINS_RANK (payloadprop_contains("resource", \
	asru(chip/memory-controller/dimm/rank)) \
	|| payloadprop_contains("resource", \
	asru(chip/memory-controller/dimm)))

/*
 * The following will tell us whether a syndrome that is known to be
 * correctable (from a mem_ce ereport) is single-bit or multi-bit.  For a
 * correctable ChipKill syndrome the number of bits set in the lowest
 * nibble indicates how many bits were in error.
 */

#define	CBITMASK(synd) ((synd) & 0xf)

#define	CKSINGLE(synd)							\
	((synd) == 0 ||							\
	(CBITMASK(synd) == 0x1 || CBITMASK(synd) == 0x2 ||		\
	CBITMASK(synd) == 0x4 || CBITMASK(synd) == 0x8))

#define	SINGLE_BIT_CE							\
	(payloadprop("syndrome-type") == "E" ||				\
	(payloadprop("syndrome-type") == "C" &&				\
	CKSINGLE(payloadprop("syndrome"))))

#define	MULTI_BIT_CE							\
	(payloadprop("syndrome-type") == "C" &&				\
	!CKSINGLE(payloadprop("syndrome")))

/*
 * A single bit fault in a memory rank can cause:
 *
 *  - mem_ce : reported by nb
 *  - inf_sys_ecc1: reported by ic or dc; inf_sys_ecc1 errors detected at the
 *    ic do not record a syndrome; these errors will not be triggered in
 *    ChipKill ECC mode (the NB corrects all ECC errors in that mode)
 *  - s_ecc1: reported by bu; this error will not be triggered in ChipKill
 *    ECC mode (the NB corrects all ECC in that mode)
 *
 * Single-bit errors are fed into a per-rank SERD engine; if a SERD engine
 * trips we diagnose a fault.memory.page so that the response agent can
 * retire the page that caused the trip.  If the total number of pages
 * faulted in this way on a single rank exceeds a threshold we will
 * diagnose a fault.memory.dimm_sb against the containing.
 *
 * Multibit ChipKill-correctable errors are treated identically to
 * single-bit errors, but via separate serd engines to allow distinct
 * parameters if desired.
 *
 * Uncorrectable errors produce an immediate page fault and corresponding
 * fault.memory.dimm_ue.
 *
 * Page faults are essentially internal - action is only required when
 * they are accompanied by a dimm fault.  As such we include message=0
 * on page faults.
 */

event ereport.cpu.amd.ic.inf_sys_ecc1@chip/cpu{within(5s)};
event ereport.cpu.amd.dc.inf_sys_ecc1@chip/cpu{within(5s)};
event ereport.cpu.amd.bu.s_ecc1@chip/cpu{within(5s)};
event ereport.cpu.amd.nb.mem_ce@chip/cpu{within(5s)};

/*
 * If the address is not valid then no resource member will be included
 * in a nb.mem_ce or nb.mem_ue ereport.  These cases should be rare.
 * We will also discard all inf_sys_ecc1 events detected at the ic since they
 * have no syndrome and therefore no resource information.
 * We will discard such ereports.  An alternative may be to SERD them
 * on a per MC basis and trip if we see too many such events.
 */

event upset.memory.discard1@chip/cpu;

/*								#PAGE#
 * Single-bit correctable errors are diagnosed as upsets and feed into per-rank
 * SERD engines which diagnose fault.memory.page_sb if they trip.
 *
 * Multi-bit correctable (via ChipKill) errors are diagnosed as upsets and feed
 * into additional per-rank SERD engines which diagnose fault.memory.page_ck
 * if they trip.
 *
 * The number of fault.memory.page and fault.memory.page_ck diagnosed is
 * counted in stat engines for each type.  These are used in deciding
 * whether to declare a dimm faulty after repeated page faults.
 */

#define PAGE_FIT		1
#define PAGE_SB_COUNT		2
#define PAGE_SB_TIME		72h
#define	PAGE_CK_COUNT		2
#define	PAGE_CK_TIME		72h

/*
 * The fraction of pages on a single rank that must be diagnosed as faulty
 * with single correctable unit faults before we will fault the rank.
 * Once we have faulted the rank we will continue to diagnose any further page
 * faults on the rank up to some maximum multiple of the threshold at which
 * we faulted the dimm.  This allows us to potentially contain some fairly
 * far-reaching but still limited-extent fault (such as a partial column
 * failure) without getting carried away and allowing a single faulty rank to
 * use up the entire system-imposed page retirenment limit (which, once
 * reached, causes retirement request to have no effect other than to fill
 * the fault manager cache and logs).
 *
 * This fraction is specified in basis points, where 100 basis points are
 * equivalent to 1 percent.  It is applied on a per-rank basis.
 *
 * The system imposes an absolute maximum on the number of pages it will
 * retire;  the current value is 10 basis points, or 0.1% of 'physmem'.  Note
 * that 'physmem' is reduced from installed memory pages by an amount
 * reflecting permanent kernel memory allocations.  This system page retire
 * limit bounds the maximum real response to page faults across all ranks
 * that fault manager response agents can effect, but it should not be confused
 * with any diagnosis threshold (i.e., the number of faulty pages we are
 * prepared to tolerate from a single rank before faulting the rank is
 * distinct from the total number of pages we are prepared to retire from use
 * in response to that and other faults).  It is, however, desirable to
 * arrange that the maximum number of pages we are prepared to fault from
 * any one rank is less than the system-wide quota.
 */
#define	PAGE_RETIRE_LIMIT_BPS	5		/* or 0.05%; ~ 131 pages/GB %/

/*
 * A macro to manipulate the above fraction.  Given a size in bytes convert
 * this to pages (4K pagesize) and calculate the number of those pages
 * indicated by PAGE_RETIRE_LIMIT_BPS basis points.
 */
#define	_BPS_PGCNT(totalbytes) \
	((((totalbytes) / 4096 ) * PAGE_RETIRE_LIMIT_BPS) / 10000)

/*
 * The single-correctable-unit threshold at which number of faulted pages
 * on a rank we we fault the rank.  We insist that this be at least 128 and
 * never more than 512.
 */
#define	RANK_THRESH MIN(512, MAX(128, \
	_BPS_PGCNT(confprop(asru(chip/memory-controller/dimm/rank), "size"))))

/*
 * The maximum number of single-correctable-unit page faults we will diagnose
 * on a single rank (must be greater than RANK_THRESH).  We set
 * this at twice the rank fault threshold.
 */
#define	RANK_PGFLT_MAX (2 * RANK_THRESH)

engine stat.sbpgflt@chip/memory-controller/dimm/rank;
engine stat.ckpgflt@chip/memory-controller/dimm/rank;

event fault.memory.page_sb@chip/memory-controller/dimm/rank,
    FITrate=PAGE_FIT, ASRU=dimm/rank, message=0,
    count=stat.sbpgflt@chip/memory-controller/dimm/rank,
    action=confcall("rewrite-ASRU"); /* rewrite ASRU to identify page in rank */

#define	SB_PGFLTS (count(stat.sbpgflt@chip/memory-controller/dimm/rank))

event fault.memory.page_ck@chip/memory-controller/dimm/rank,
    FITrate=PAGE_FIT, ASRU=dimm/rank, message=0,
    count=stat.ckpgflt@chip/memory-controller/dimm/rank,
    action=confcall("rewrite-ASRU"); /* rewrite ASRU to identify page in rank */

#define	CK_PGFLTS (count(stat.ckpgflt@chip/memory-controller/dimm/rank))

#define	RANK_PGFLT_LIMIT_REACHED \
    (SB_PGFLTS + CK_PGFLTS > RANK_PGFLT_MAX)

event ereport.memory.page_sb_trip@chip/memory-controller/dimm/rank{within(5s)};
engine serd.memory.page_sb@chip/memory-controller/dimm/rank,
    N=PAGE_SB_COUNT, T=PAGE_SB_TIME, method=persistent,
    trip=ereport.memory.page_sb_trip@chip/memory-controller/dimm/rank;
event upset.memory.page_sb@chip/memory-controller/dimm/rank,
    engine=serd.memory.page_sb@chip/memory-controller/dimm/rank;

event ereport.memory.page_ck_trip@chip/memory-controller/dimm/rank{within(5s)};
engine serd.memory.page_ck@chip/memory-controller/dimm/rank,
    N=PAGE_CK_COUNT, T=PAGE_CK_TIME, method=persistent,
    trip=ereport.memory.page_ck_trip@chip/memory-controller/dimm/rank;
event upset.memory.page_ck@chip/memory-controller/dimm/rank,
    engine=serd.memory.page_ck@chip/memory-controller/dimm/rank;

event upset.memory.overpgfltlimit@chip/memory-controller/dimm/rank;

/*
 * If we have not reached the per-rank limit on faulted pages then
 * continue to explain ereport observations as upsets which can lead
 * lead to page fault diagnoses if the serd engine trips.
 */
prop upset.memory.page_sb@chip/memory-controller/dimm/rank
    { CONTAINS_RANK && SINGLE_BIT_CE && !RANK_PGFLT_LIMIT_REACHED } (0)->
    ereport.cpu.amd.dc.inf_sys_ecc1@chip/cpu,
    ereport.cpu.amd.bu.s_ecc1@chip/cpu,
    ereport.cpu.amd.nb.mem_ce@chip/cpu;

prop upset.memory.page_ck@chip/memory-controller/dimm/rank
    { CONTAINS_RANK && MULTI_BIT_CE && !RANK_PGFLT_LIMIT_REACHED } (0)->
    /* no dc.inf_sys_ecc1 or bu.s_ecc1 in ChipKill mode */
    ereport.cpu.amd.nb.mem_ce@chip/cpu;

/*
 * If we have reached the per-rank limit on faulted pages then diagnose
 * further observations on the rank to a engine-less upset (i.e., discard
 * them).
 */
prop upset.memory.overpgfltlimit@chip/memory-controller/dimm/rank
    { CONTAINS_RANK && RANK_PGFLT_LIMIT_REACHED } (1)->
    ereport.cpu.amd.dc.inf_sys_ecc1@chip/cpu,
    ereport.cpu.amd.bu.s_ecc1@chip/cpu,
    ereport.cpu.amd.nb.mem_ce@chip/cpu;

prop fault.memory.page_sb@chip/memory-controller/dimm/rank (1)->
    ereport.memory.page_sb_trip@chip/memory-controller/dimm/rank;

prop fault.memory.page_ck@chip/memory-controller/dimm/rank (1)->
    ereport.memory.page_ck_trip@chip/memory-controller/dimm/rank;

prop fault.memory.page_sb@chip/memory-controller/dimm/rank
    { CONTAINS_RANK && SET_ADDR && SET_OFFSET } (0)->
    ereport.cpu.amd.dc.inf_sys_ecc1@chip/cpu,
    ereport.cpu.amd.bu.s_ecc1@chip/cpu,
    ereport.cpu.amd.nb.mem_ce@chip/cpu;

prop fault.memory.page_ck@chip/memory-controller/dimm/rank
    { CONTAINS_RANK && SET_ADDR && SET_OFFSET } (0)->
    ereport.cpu.amd.dc.inf_sys_ecc1@chip/cpu,
    ereport.cpu.amd.bu.s_ecc1@chip/cpu,
    ereport.cpu.amd.nb.mem_ce@chip/cpu;

/*
 * Discard memory ereports that do not indicate a resource.
 */
prop upset.memory.discard1@chip/cpu
    { !RESOURCE_EXISTS } (1)->
    ereport.cpu.amd.ic.inf_sys_ecc1@chip/cpu,
    ereport.cpu.amd.dc.inf_sys_ecc1@chip/cpu,
    ereport.cpu.amd.bu.s_ecc1@chip/cpu,
    ereport.cpu.amd.nb.mem_ce@chip/cpu;

/*								#DIMM_SCU#
 * "Single-correctable-unit" DIMM faults are diagnosed when the total number of
 * page faults (diagnosed from repeated single-bit or multibit-chipkills)
 * from any one rank on that DIMM reaches a threshold.  A "correctable unit"
 * is a single bit in normal 64/8 ECC mode, or a single symbol in ChipKill
 * 128/16 mode (i.e., nibble-aligned nibble for the code used on Opteron).
 *
 * We do not stop diagnosing further single-bit page faults once we have
 * declared a single-bit DIMM fault - we continue diagnosing them and
 * response agents can continue to retire those pages up to the system-imposed
 * retirement limit.
 *
 * Two distinct fault types may be diagnosed - fault.memory.dimm_sb and
 * fault.memory.dimm_ck.  Which one is diagnosed depends on whether we
 * have reached the threshold for a majority of single-bit page faults or
 * multibit page faults.
 *
 * Implementation: we maintain parallel SERD engines to the page_sb and
 * page_ck engines, which trip in unison.  On trip it generates a distinct
 * ereport which we diagnose to a fault if the threshold has been
 * reached, or to a throwaway upset if not.
 *
 */

#define DIMM_SB_FIT		2000
#define DIMM_CK_FIT		4000

event fault.memory.dimm_sb@chip/memory-controller/dimm/rank,
    FITrate=DIMM_SB_FIT, FRU=dimm, ASRU=dimm,
    action=confcall("rewrite-ASRU"); /* rewrite non-leaf ASRU in mem scheme */

event fault.memory.dimm_ck@chip/memory-controller/dimm/rank,
    FITrate=DIMM_CK_FIT, FRU=dimm, ASRU=dimm,
    action=confcall("rewrite-ASRU"); /* rewrite non-leaf ASRU in mem scheme */

event ereport.memory.dimm_sb_trip@chip/memory-controller/dimm/rank
	{ within(5s) };
engine serd.memory.dimm_sb@chip/memory-controller/dimm/rank,
    N=PAGE_SB_COUNT, T=PAGE_SB_TIME, method=persistent,
    trip=ereport.memory.dimm_sb_trip@chip/memory-controller/dimm/rank;
event upset.memory.dimm_sb@chip/memory-controller/dimm/rank,
    engine=serd.memory.dimm_sb@chip/memory-controller/dimm/rank;

event ereport.memory.dimm_ck_trip@chip/memory-controller/dimm/rank
	{ within(5s) };
engine serd.memory.dimm_ck@chip/memory-controller/dimm/rank,
    N=PAGE_CK_COUNT, T=PAGE_CK_TIME, method=persistent,
    trip=ereport.memory.dimm_ck_trip@chip/memory-controller/dimm/rank;
event upset.memory.dimm_ck@chip/memory-controller/dimm/rank,
    engine=serd.memory.dimm_ck@chip/memory-controller/dimm/rank;

event upset.memory.discard2@chip/memory-controller/dimm/rank;

prop upset.memory.dimm_sb@chip/memory-controller/dimm/rank
    { CONTAINS_RANK && SINGLE_BIT_CE } (0)->
    ereport.cpu.amd.dc.inf_sys_ecc1@chip/cpu,
    ereport.cpu.amd.bu.s_ecc1@chip/cpu,
    ereport.cpu.amd.nb.mem_ce@chip/cpu;

prop upset.memory.dimm_ck@chip/memory-controller/dimm/rank
    { CONTAINS_RANK && MULTI_BIT_CE } (0)->
    ereport.cpu.amd.nb.mem_ce@chip/cpu;

/*
 * The following two propogations diagnose a fault.memory.dimm_sb when
 * either the dimm_sb or dimm_ck engine trips (for a new page fault)
 * and the total number of page faults (sb and ck) exceeds the threshold
 * value with the majority being from sb page faults.
 */
prop fault.memory.dimm_sb@chip/memory-controller/dimm/rank (0)->
    ereport.memory.dimm_sb_trip@chip/memory-controller/dimm/rank
    { SB_PGFLTS + CK_PGFLTS > RANK_THRESH && SB_PGFLTS > RANK_THRESH / 2 };

prop fault.memory.dimm_sb@chip/memory-controller/dimm/rank (0)->
    ereport.memory.dimm_ck_trip@chip/memory-controller/dimm/rank
    { SB_PGFLTS + CK_PGFLTS > RANK_THRESH && SB_PGFLTS > RANK_THRESH / 2 };

/*
 * The following two propogation diagnose a fault.memory.dimm_ck when
 * either the dimm_sb or dimm_ck engine trip (for a new page fault)
 * and the total number of page faults (sb and ck) exceeds the threshold
 * value with the majority  being from ck page faults.
 */
prop fault.memory.dimm_ck@chip/memory-controller/dimm/rank (0)->
    ereport.memory.dimm_sb_trip@chip/memory-controller/dimm/rank
    { SB_PGFLTS + CK_PGFLTS > RANK_THRESH && CK_PGFLTS > RANK_THRESH / 2 };

prop fault.memory.dimm_ck@chip/memory-controller/dimm/rank (0)->
    ereport.memory.dimm_ck_trip@chip/memory-controller/dimm/rank
    { SB_PGFLTS + CK_PGFLTS > RANK_THRESH && CK_PGFLTS > RANK_THRESH / 2 };

prop upset.memory.discard2@chip/memory-controller/dimm/rank (1)->
    ereport.memory.dimm_sb_trip@chip/memory-controller/dimm/rank,
    ereport.memory.dimm_ck_trip@chip/memory-controller/dimm/rank;

/* 								#DIMM_UE#
 *								#PAGE_UE#
 * An uncorrectable multi-bit fault in a memory dimm can cause:
 *
 *  - mem_ue    	   : reported by nb for an access from a remote cpu
 *  - inf_sys_eccm : reported by ic or dc; the ic does not report a syndrome
 *  - s_eccm	   : reported by bu
 *
 * Note we use a SERD engine here simply as a way of ensuring that we get
 * both dimm and page faults reported.
 *
 * Since on production systems we force HT Sync Flood on uncorrectable
 * memory errors (if not already set as such by the BIOS, as it should be)
 * we won't actually receive these ereports since the system will be reset.
 */

#define DIMM_UE_FIT		6000

event ereport.cpu.amd.ic.inf_sys_eccm@chip/cpu{within(5s)};
event ereport.cpu.amd.dc.inf_sys_eccm@chip/cpu{within(5s)};
event ereport.cpu.amd.bu.s_eccm@chip/cpu{within(5s)};
event ereport.cpu.amd.nb.mem_ue@chip/cpu{within(5s)};

event fault.memory.dimm_ue@chip/memory-controller/dimm/rank,
    FITrate=DIMM_UE_FIT, FRU=dimm, ASRU=dimm,
    action=confcall("rewrite-ASRU"); /* rewrite non-leaf ASRU in mem scheme */

event fault.memory.page_ue@chip/memory-controller/dimm/rank,
    FITrate=PAGE_FIT, ASRU=dimm/rank, message=0,
    action=confcall("rewrite-ASRU"); /* rewrite ASRU to identify page in rank */

event ereport.memory.dimm_ue_trip@chip/memory-controller/dimm/rank{within(5s)};
engine serd.memory.dimm_ue@chip/memory-controller/dimm/rank,
    N=0, T=1h, method=persistent,
    trip=ereport.memory.dimm_ue_trip@chip/memory-controller/dimm/rank;
event upset.memory.dimm_ue@chip/memory-controller/dimm/rank,
    engine=serd.memory.dimm_ue@chip/memory-controller/dimm/rank;

event ereport.memory.page_ue_trip@chip/memory-controller/dimm/rank{within(5s)};
engine serd.memory.page_ue@chip/memory-controller/dimm/rank,
    N=0, T=1h, method=persistent,
    trip=ereport.memory.page_ue_trip@chip/memory-controller/dimm/rank;
event upset.memory.page_ue@chip/memory-controller/dimm/rank,
    engine=serd.memory.page_ue@chip/memory-controller/dimm/rank;

event upset.memory.discard3@chip/cpu;

prop upset.memory.page_ue@chip/memory-controller/dimm/rank
    { CONTAINS_RANK } (0)->
    ereport.cpu.amd.ic.inf_sys_eccm@chip/cpu,
    ereport.cpu.amd.dc.inf_sys_eccm@chip/cpu,
    ereport.cpu.amd.bu.s_eccm@chip/cpu,
    ereport.cpu.amd.nb.mem_ue@chip/cpu;

prop upset.memory.dimm_ue@chip/memory-controller/dimm/rank
    { CONTAINS_RANK } (0)->
    ereport.cpu.amd.ic.inf_sys_eccm@chip/cpu,
    ereport.cpu.amd.dc.inf_sys_eccm@chip/cpu,
    ereport.cpu.amd.bu.s_eccm@chip/cpu,
    ereport.cpu.amd.nb.mem_ue@chip/cpu;

prop fault.memory.page_ue@chip/memory-controller/dimm/rank (1)->
    ereport.memory.page_ue_trip@chip/memory-controller/dimm/rank;

prop fault.memory.page_ue@chip/memory-controller/dimm/rank
    { CONTAINS_RANK && SET_ADDR && SET_OFFSET } (0)->
    ereport.cpu.amd.ic.inf_sys_eccm@chip/cpu,
    ereport.cpu.amd.dc.inf_sys_eccm@chip/cpu,
    ereport.cpu.amd.bu.s_eccm@chip/cpu,
    ereport.cpu.amd.nb.mem_ue@chip/cpu;

prop fault.memory.dimm_ue@chip/memory-controller/dimm/rank (1)->
    ereport.memory.dimm_ue_trip@chip/memory-controller/dimm/rank;

prop upset.memory.discard3@chip/cpu
    { !RESOURCE_EXISTS } (1)->
    ereport.cpu.amd.ic.inf_sys_eccm@chip/cpu,
    ereport.cpu.amd.dc.inf_sys_eccm@chip/cpu,
    ereport.cpu.amd.bu.s_eccm@chip/cpu,
    ereport.cpu.amd.nb.mem_ue@chip/cpu;

/*								#CSTESTFAIL#
 * If the BIOS fails a chip-select during POST, or perhaps after a
 * sync flood from an uncorrectable error, then on revision F and G it
 * should mark that chip-select as TestFail in the CS Base register.
 * When the memory-controller driver discovers all the MC configuration
 * it notes such failed chip-selects and creates topology nodes for the
 * chip-select and associated dimms and ranks, and produces an ereport for each
 * failed chip-select with detector set to the memory-controller node
 * and resource indicating the failed chip-select.
 */

event ereport.cpu.amd.mc.cs_testfail@chip/memory-controller{within(5s)};

event fault.memory.dimm_testfail@chip/memory-controller/dimm/rank,
    FITrate=1000, ASRU=dimm, FRU=dimm,
    action=confcall("rewrite-ASRU"); /* rewrite non-leaf ASRU in mem scheme */

event error.memory.cs_testfail@chip/memory-controller/chip-select;

#define	CONTAINS_CS (payloadprop_contains("resource", \
	asru(chip/memory-controller/chip-select)))

prop error.memory.cs_testfail@chip/memory-controller/chip-select (1)->
    ereport.cpu.amd.mc.cs_testfail@chip/memory-controller
    { CONTAINS_CS };

#define CSMATCH(s) \
	(confprop_defined(asru(chip/memory-controller/chip-select), s) && \
	confprop(asru(chip/memory-controller/chip-select), s) == \
	confprop(asru(chip/memory-controller/dimm/rank), "csname"))

prop fault.memory.dimm_testfail@chip/memory-controller/dimm/rank (1)->
    error.memory.cs_testfail@chip/memory-controller/chip-select
    { CSMATCH("dimm1-csname") || CSMATCH("dimm2-csname")};

/*								#ADDRPAR#
 * DRAM Command/Address Parity Errors.
 *
 *  - dramaddr_par : reported by the nb; the NB status register includes
 *    a bit indicating which dram controller channel (A or B) experienced
 *    the error.
 */

event ereport.cpu.amd.nb.dramaddr_par@chip/cpu{within(5s)};

event fault.cpu.amd.dramchannel@chip/memory-controller/dram-channel,
    FITrate=1000, ASRU=dram-channel;

#define GET_CHANNEL ($chan = (payloadprop("IA32_MCi_STATUS") >> 32 & 0x200) ? \
    1 : 0)

prop fault.cpu.amd.dramchannel@chip/memory-controller/dram-channel[y] (0)->
    ereport.cpu.amd.nb.dramaddr_par@chip/cpu { GET_CHANNEL && $chan == y };

/*
 * l2 cache data errors.
 */

#define L2CACHEDATA_FIT		1000
#define L2CACHEDATA_SB_COUNT	3
#define L2CACHEDATA_SB_TIME	12h

event fault.cpu.amd.l2cachedata@chip/cpu, FITrate=L2CACHEDATA_FIT,
	FRU=chip, ASRU=chip/cpu;
event error.cpu.amd.l2cachedata_sb@chip/cpu;
event error.cpu.amd.l2cachedata_mb@chip/cpu;

prop fault.cpu.amd.l2cachedata@chip/cpu (1)->
    error.cpu.amd.l2cachedata_sb@chip/cpu,
    error.cpu.amd.l2cachedata_mb@chip/cpu;

/* 								#L2D_SINGLE#
 * A single bit data array fault in an l2 cache can cause:
 *
 *  - inf_l2_ecc1 : reported by ic on this cpu
 *  - inf_l2_ecc1 : reported by dc on this cpu
 *  - l2d_ecc1 : reported by bu on copyback or on snoop from another cpu
 *
 * Single-bit errors are diagnosed to cache upsets.  SERD engines are used
 * to count upsets resulting from CEs.
 */

event ereport.cpu.amd.ic.inf_l2_ecc1@chip/cpu{within(5s)};
event ereport.cpu.amd.dc.inf_l2_ecc1@chip/cpu{within(5s)};
event ereport.cpu.amd.bu.l2d_ecc1@chip/cpu{within(5s)};
event ereport.cpu.amd.l2d_sb_trip@chip/cpu{within(5s)};

engine serd.cpu.amd.l2d_sb@chip/cpu,
    N=L2CACHEDATA_SB_COUNT, T=L2CACHEDATA_SB_TIME, method=persistent,
    trip=ereport.cpu.amd.l2d_sb_trip@chip/cpu;

event upset.cpu.amd.l2d_sb@chip/cpu,
	engine=serd.cpu.amd.l2d_sb@chip/cpu;

prop upset.cpu.amd.l2d_sb@chip/cpu (1)->
    ereport.cpu.amd.ic.inf_l2_ecc1@chip/cpu,
    ereport.cpu.amd.dc.inf_l2_ecc1@chip/cpu,
    ereport.cpu.amd.bu.l2d_ecc1@chip/cpu;

prop error.cpu.amd.l2cachedata_sb@chip/cpu (1)->
    ereport.cpu.amd.l2d_sb_trip@chip/cpu;

prop fault.cpu.amd.l2cachedata@chip/cpu (0)->
    ereport.cpu.amd.ic.inf_l2_ecc1@chip/cpu,
    ereport.cpu.amd.dc.inf_l2_ecc1@chip/cpu,
    ereport.cpu.amd.bu.l2d_ecc1@chip/cpu;

/* 								#L2D_MULTI#
 * A multi-bit data array fault in an l2 cache can cause:
 *
 *  - inf_l2_eccm : reported by ic on this cpu
 *  - inf_l2_eccm : reported by dc on this cpu
 *  - l2d_eccm : reported by bu on copyback or on snoop from another cpu
 */

event ereport.cpu.amd.ic.inf_l2_eccm@chip/cpu{within(5s)};
event ereport.cpu.amd.dc.inf_l2_eccm@chip/cpu{within(5s)};
event ereport.cpu.amd.bu.l2d_eccm@chip/cpu{within(5s)};

prop error.cpu.amd.l2cachedata_mb@chip/cpu (1)->
    ereport.cpu.amd.ic.inf_l2_eccm@chip/cpu,
    ereport.cpu.amd.dc.inf_l2_eccm@chip/cpu,
    ereport.cpu.amd.bu.l2d_eccm@chip/cpu;

prop fault.cpu.amd.l2cachedata@chip/cpu (0)->
    ereport.cpu.amd.ic.inf_l2_eccm@chip/cpu,
    ereport.cpu.amd.dc.inf_l2_eccm@chip/cpu,
    ereport.cpu.amd.bu.l2d_eccm@chip/cpu;

/*
 * l2 cache main tag errors
 */

#define L2CACHETAG_FIT		1000
#define L2CACHETAG_SB_COUNT	3
#define L2CACHETAG_SB_TIME	12h

event fault.cpu.amd.l2cachetag@chip/cpu, FITrate=L2CACHETAG_FIT,
	FRU=chip, ASRU=chip/cpu;
event error.cpu.amd.l2cachetag_sb@chip/cpu;
event error.cpu.amd.l2cachetag_mb@chip/cpu;

prop fault.cpu.amd.l2cachetag@chip/cpu (1)->
    error.cpu.amd.l2cachetag_sb@chip/cpu,
    error.cpu.amd.l2cachetag_mb@chip/cpu;

/* 								#L2T_SINGLE#
 * A single bit tag array fault in an l2 cache can cause:
 *
 *  - l2t_ecc1 : reported by bu on this cpu when detected during snoop
 *  - l2t_par : reported by bu on this cpu when detected other than during snoop
 *
 * Note that the bu.l2t_par ereport could be due to a single bit or multi bit
 * event. If the l2t_sb_trip has already triggered it will be treated as another
 * ce, otherwise it will be treated as a ue event.
 */

event ereport.cpu.amd.bu.l2t_ecc1@chip/cpu{within(5s)};
event ereport.cpu.amd.bu.l2t_par@chip/cpu{within(5s)};
event ereport.cpu.amd.l2t_sb_trip@chip/cpu{within(5s)};

engine serd.cpu.amd.l2t_sb@chip/cpu,
    N=L2CACHETAG_SB_COUNT, T=L2CACHETAG_SB_TIME, method=persistent,
    trip=ereport.cpu.amd.l2t_sb_trip@chip/cpu;

event upset.cpu.amd.l2t_sb@chip/cpu,
	engine=serd.cpu.amd.l2t_sb@chip/cpu;

prop upset.cpu.amd.l2t_sb@chip/cpu (1)->
    ereport.cpu.amd.bu.l2t_ecc1@chip/cpu,
    ereport.cpu.amd.bu.l2t_par@chip/cpu;

prop error.cpu.amd.l2cachetag_sb@chip/cpu (1)->
    ereport.cpu.amd.l2t_sb_trip@chip/cpu;

prop fault.cpu.amd.l2cachetag@chip/cpu (0)->
    ereport.cpu.amd.bu.l2t_ecc1@chip/cpu,
    ereport.cpu.amd.bu.l2t_par@chip/cpu;

/* 								#L2T_MULTI#
 * A multi-bit tag array fault in an l2 cache can cause:
 *
 *  - l2t_eccm : reported by bu on this cpu when detected during snoop
 *  - l2t_par : reported by bu on this cpu when detected other than during snoop
 */

event ereport.cpu.amd.bu.l2t_eccm@chip/cpu{within(5s)};

prop error.cpu.amd.l2cachetag_mb@chip/cpu (1)->
    ereport.cpu.amd.bu.l2t_eccm@chip/cpu,
    ereport.cpu.amd.bu.l2t_par@chip/cpu;

prop fault.cpu.amd.l2cachetag@chip/cpu (0)->
    ereport.cpu.amd.bu.l2t_eccm@chip/cpu,
    ereport.cpu.amd.bu.l2t_par@chip/cpu;

/* 								#ICD_PAR#
 * A data array parity fault in an I cache can cause:
 *
 *  - data_par : reported by ic on this cpu
 */

#define ICACHEDATA_FIT		1000
#define ICACHEDATA_SB_COUNT	2
#define ICACHEDATA_SB_TIME	168h

event ereport.cpu.amd.ic.data_par@chip/cpu{within(5s)};
event ereport.cpu.amd.ic_dp_trip@chip/cpu{within(5s)};

event fault.cpu.amd.icachedata@chip/cpu, FITrate=ICACHEDATA_FIT,
	FRU=chip, ASRU=chip/cpu;

engine serd.cpu.amd.icachedata@chip/cpu,
    N=ICACHEDATA_SB_COUNT, T=ICACHEDATA_SB_TIME, method=persistent,
    trip=ereport.cpu.amd.ic_dp_trip@chip/cpu;

event upset.cpu.amd.icachedata@chip/cpu,
	engine=serd.cpu.amd.icachedata@chip/cpu;

prop upset.cpu.amd.icachedata@chip/cpu (1)->
    ereport.cpu.amd.ic.data_par@chip/cpu;

prop fault.cpu.amd.icachedata@chip/cpu (1)->
    ereport.cpu.amd.ic_dp_trip@chip/cpu;

prop fault.cpu.amd.icachedata@chip/cpu (0)->
    ereport.cpu.amd.ic.data_par@chip/cpu;

/* 								#ICT_PAR#
 * A tag array parity fault in an I cache can cause:
 *
 *  - tag_par : reported by ic on this cpu
 */

#define ICACHETAG_FIT		1000
#define ICACHETAG_SB_COUNT	2
#define ICACHETAG_SB_TIME	168h

event ereport.cpu.amd.ic.tag_par@chip/cpu{within(5s)};
event ereport.cpu.amd.ic_tp_trip@chip/cpu{within(5s)};

event fault.cpu.amd.icachetag@chip/cpu, FITrate=ICACHETAG_FIT,
	FRU=chip, ASRU=chip/cpu;

engine serd.cpu.amd.icachetag@chip/cpu,
    N=ICACHETAG_SB_COUNT, T=ICACHETAG_SB_TIME, method=persistent,
    trip=ereport.cpu.amd.ic_tp_trip@chip/cpu;

event upset.cpu.amd.icachetag@chip/cpu,
	engine=serd.cpu.amd.icachetag@chip/cpu;

prop upset.cpu.amd.icachetag@chip/cpu (1)->
    ereport.cpu.amd.ic.tag_par@chip/cpu;

prop fault.cpu.amd.icachetag@chip/cpu (1)->
    ereport.cpu.amd.ic_tp_trip@chip/cpu;

prop fault.cpu.amd.icachetag@chip/cpu (0)->
    ereport.cpu.amd.ic.tag_par@chip/cpu;

/* 								#ICT_SNOOP#
 * A snoop tag array parity fault in an I cache can cause:
 *
 *  - stag_par : reported by ic on this cpu
 */

#define ICACHESTAG_FIT		1000

event ereport.cpu.amd.ic.stag_par@chip/cpu{within(5s)};

event fault.cpu.amd.icachestag@chip/cpu, FITrate=ICACHESTAG_FIT,
	FRU=chip, ASRU=chip/cpu;

prop fault.cpu.amd.icachestag@chip/cpu (1)->
    ereport.cpu.amd.ic.stag_par@chip/cpu;

/* 								#ICTLB_1#
 * An l1tlb parity fault in an I cache can cause:
 *
 *  - l1tlb_par : reported by ic on this cpu
 */

#define ICACHEL1TLB_FIT		1000
#define ICACHEL1TLB_SB_COUNT	2
#define ICACHEL1TLB_SB_TIME	168h

event ereport.cpu.amd.ic.l1tlb_par@chip/cpu{within(5s)};
event ereport.cpu.amd.ic_l1tlb_trip@chip/cpu{within(5s)};

event fault.cpu.amd.l1itlb@chip/cpu, FITrate=ICACHEL1TLB_FIT,
	FRU=chip, ASRU=chip/cpu;

engine serd.cpu.amd.l1itlb@chip/cpu,
    N=ICACHEL1TLB_SB_COUNT, T=ICACHEL1TLB_SB_TIME, method=persistent,
    trip=ereport.cpu.amd.ic_l1tlb_trip@chip/cpu;

event upset.cpu.amd.l1itlb@chip/cpu,
	engine=serd.cpu.amd.l1itlb@chip/cpu;

prop upset.cpu.amd.l1itlb@chip/cpu (1)->
    ereport.cpu.amd.ic.l1tlb_par@chip/cpu;

prop fault.cpu.amd.l1itlb@chip/cpu (1)->
    ereport.cpu.amd.ic_l1tlb_trip@chip/cpu;

prop fault.cpu.amd.l1itlb@chip/cpu (0)->
    ereport.cpu.amd.ic.l1tlb_par@chip/cpu;

/* 								#ICTLB_2#
 * An l2tlb parity fault in an I cache can cause:
 *
 *  - l2tlb_par : reported by ic on this cpu
 */

#define ICACHEL2TLB_FIT		1000
#define ICACHEL2TLB_SB_COUNT	2
#define ICACHEL2TLB_SB_TIME	168h

event ereport.cpu.amd.ic.l2tlb_par@chip/cpu{within(5s)};
event ereport.cpu.amd.ic_l2tlb_trip@chip/cpu{within(5s)};

event fault.cpu.amd.l2itlb@chip/cpu, FITrate=ICACHEL2TLB_FIT,
	FRU=chip, ASRU=chip/cpu;

engine serd.cpu.amd.l2itlb@chip/cpu,
    N=ICACHEL2TLB_SB_COUNT, T=ICACHEL2TLB_SB_TIME, method=persistent,
    trip=ereport.cpu.amd.ic_l2tlb_trip@chip/cpu;

event upset.cpu.amd.l2itlb@chip/cpu,
	engine=serd.cpu.amd.l2itlb@chip/cpu;

prop upset.cpu.amd.l2itlb@chip/cpu (1)->
    ereport.cpu.amd.ic.l2tlb_par@chip/cpu;

prop fault.cpu.amd.l2itlb@chip/cpu (1)->
    ereport.cpu.amd.ic_l2tlb_trip@chip/cpu;

prop fault.cpu.amd.l2itlb@chip/cpu (0)->
    ereport.cpu.amd.ic.l2tlb_par@chip/cpu;

/*
 * dcache data errors
 */

#define DCACHEDATA_FIT		1000
#define DCACHEDATA_SB_COUNT	2
#define DCACHEDATA_SB_TIME	168h

event fault.cpu.amd.dcachedata@chip/cpu, FITrate=DCACHEDATA_FIT,
	FRU=chip, ASRU=chip/cpu;
event error.cpu.amd.dcachedata_sb@chip/cpu;
event error.cpu.amd.dcachedata_mb@chip/cpu;

prop fault.cpu.amd.dcachedata@chip/cpu (1)->
    error.cpu.amd.dcachedata_sb@chip/cpu,
    error.cpu.amd.dcachedata_mb@chip/cpu;

/* 								#DCD_SINGLE#
 * A single bit data array fault in an D cache can cause:
 *
 *  - data_ecc1 : reported by dc on this cpu by scrubber
 *  - data_ecc1_uc : reported by dc on this cpu other than by scrubber
 *
 * Make data_ecc1_uc fault immediately as it may have caused a panic
 */

event ereport.cpu.amd.dc.data_ecc1@chip/cpu{within(5s)};
event ereport.cpu.amd.dc.data_ecc1_uc@chip/cpu{within(5s)};
event ereport.cpu.amd.dc_sb_trip@chip/cpu{within(5s)};

engine serd.cpu.amd.dc_sb@chip/cpu,
    N=DCACHEDATA_SB_COUNT, T=DCACHEDATA_SB_TIME, method=persistent,
    trip=ereport.cpu.amd.dc_sb_trip@chip/cpu;

engine serd.cpu.amd.dc_sb_uc@chip/cpu,
    N=0, T=1hr, method=persistent,
    trip=ereport.cpu.amd.dc_sb_trip@chip/cpu;

event upset.cpu.amd.dc_sb@chip/cpu,
	engine=serd.cpu.amd.dc_sb@chip/cpu;

event upset.cpu.amd.dc_sb_uc@chip/cpu,
	engine=serd.cpu.amd.dc_sb_uc@chip/cpu;

prop upset.cpu.amd.dc_sb@chip/cpu (1)->
    ereport.cpu.amd.dc.data_ecc1@chip/cpu;

prop upset.cpu.amd.dc_sb_uc@chip/cpu (1)->
    ereport.cpu.amd.dc.data_ecc1_uc@chip/cpu;

prop error.cpu.amd.dcachedata_sb@chip/cpu (1)->
    ereport.cpu.amd.dc_sb_trip@chip/cpu;

prop fault.cpu.amd.dcachedata@chip/cpu (0)->
    ereport.cpu.amd.dc.data_ecc1@chip/cpu,
    ereport.cpu.amd.dc.data_ecc1_uc@chip/cpu;

/* 								#DCD_MULTI#
 * A multi-bit data array fault in an D cache can cause:
 *
 *  - data_eccm : reported by dc on this cpu
 */

event ereport.cpu.amd.dc.data_eccm@chip/cpu{within(5s)};

prop error.cpu.amd.dcachedata_mb@chip/cpu (1)->
    ereport.cpu.amd.dc.data_eccm@chip/cpu;

prop fault.cpu.amd.dcachedata@chip/cpu (0)->
    ereport.cpu.amd.dc.data_eccm@chip/cpu;

/* 								#DCT_PAR#
 * A tag array parity fault in an D cache can cause:
 *
 *  - tag_par : reported by dc on this cpu
 */

#define DCACHETAG_FIT		1000

event ereport.cpu.amd.dc.tag_par@chip/cpu{within(5s)};

event fault.cpu.amd.dcachetag@chip/cpu, FITrate=DCACHETAG_FIT,
	FRU=chip, ASRU=chip/cpu;

prop fault.cpu.amd.dcachetag@chip/cpu (1)->
    ereport.cpu.amd.dc.tag_par@chip/cpu;

/* 								#DCT_SNOOP#
 * A snoop tag array parity fault in an D cache can cause:
 *
 *  - stag_par : reported by dc on this cpu
 */

#define DCACHESTAG_FIT		1000

event ereport.cpu.amd.dc.stag_par@chip/cpu{within(5s)};

event fault.cpu.amd.dcachestag@chip/cpu, FITrate=DCACHESTAG_FIT,
	FRU=chip, ASRU=chip/cpu;

prop fault.cpu.amd.dcachestag@chip/cpu (1)->
    ereport.cpu.amd.dc.stag_par@chip/cpu;

/* 								#DCTLB_1#
 * An l1tlb parity fault in an D cache can cause:
 *
 *  - l1tlb_par : reported by dc on this cpu
 */

#define L1DTLB_FIT		1000

event ereport.cpu.amd.dc.l1tlb_par@chip/cpu{within(5s)};

event fault.cpu.amd.l1dtlb@chip/cpu, FITrate=L1DTLB_FIT,
	FRU=chip, ASRU=chip/cpu;

prop fault.cpu.amd.l1dtlb@chip/cpu (1)->
    ereport.cpu.amd.dc.l1tlb_par@chip/cpu;

/* 								#DCTLB_2#
 * An l2tlb parity fault in an D cache can cause:
 *
 *  - l2tlb_par : reported by dc on this cpu
 */

#define L2DTLB_FIT		1000

event ereport.cpu.amd.dc.l2tlb_par@chip/cpu{within(5s)};

event fault.cpu.amd.l2dtlb@chip/cpu, FITrate=L2DTLB_FIT,
	FRU=chip, ASRU=chip/cpu;

prop fault.cpu.amd.l2dtlb@chip/cpu (1)->
    ereport.cpu.amd.dc.l2tlb_par@chip/cpu;

/*								#MISC#
 * Ereports that should not normally happen and which we will discard
 * without diagnosis if they do.  These fall into a few categories:
 *
 *	- the corresponding detector is not enabled, typically because
 *	  detection/handling of the event is taking place elsewhere
 *	  (nb.ma, nb.ta, ls.rde, ic.rdde, bu.s_rde, nb.gart_walk)
 *	- the event is associated with a sync flood so even if the detector is
 *	  enabled we will never handle the event and generate an ereport *and*
 *	  even if the ereport did arrive we could perform no useful diagnosis
 *	  e.g., the NB can be configured for sync flood on nb.mem_eccm
 *	  but we don't choose to discard that ereport here since we could have
 *	  made a useful diagnosis from it had it been delivered
 *	  (nb.ht_sync, nb.ht_crc)
 *	- events that will be accompanied by an immediate panic and
 *	  delivery of the ereport during subsequent reboot but from
 *	  which no useful diagnosis can be made. (nb.rmw, nb.wdog)
 *
 * Ereports for all of these can be generated by error simulation and
 * injection.  We will perform a null diagnosos of all these ereports in order
 * to avoid "no subscription" complaints during test harness runs.
 */

event ereport.cpu.amd.nb.ma@cpu{within(5s)};
event ereport.cpu.amd.nb.ta@cpu{within(5s)};
event ereport.cpu.amd.ls.s_rde@cpu{within(5s)};
event ereport.cpu.amd.ic.rdde@cpu{within(5s)};
event ereport.cpu.amd.bu.s_rde@cpu{within(5s)};
event ereport.cpu.amd.nb.gart_walk@cpu{within(5s)};
event ereport.cpu.amd.nb.ht_sync@cpu{within(5s)};
event ereport.cpu.amd.nb.ht_crc@cpu{within(5s)};
event ereport.cpu.amd.nb.rmw@cpu{within(5s)};
event ereport.cpu.amd.nb.wdog@cpu{within(5s)};
event ereport.cpu.amd.unknown@cpu{within(5s)};

event upset.null_diag@cpu;

prop upset.null_diag@cpu (1)->
    ereport.cpu.amd.nb.ma@cpu,
    ereport.cpu.amd.nb.ta@cpu,
    ereport.cpu.amd.ls.s_rde@cpu,
    ereport.cpu.amd.ic.rdde@cpu,
    ereport.cpu.amd.bu.s_rde@cpu,
    ereport.cpu.amd.nb.gart_walk@cpu,
    ereport.cpu.amd.nb.ht_sync@cpu,
    ereport.cpu.amd.nb.ht_crc@cpu,
    ereport.cpu.amd.nb.rmw@cpu,
    ereport.cpu.amd.nb.wdog@cpu,
    ereport.cpu.amd.unknown@cpu;