cma_main.c revision 7aec1d6e253b21f9e9b7ef68b4d81ab9859b51fe
/*
* CDDL HEADER START
*
* The contents of this file are subject to the terms of the
* Common Development and Distribution License, Version 1.0 only
* (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 2006 Sun Microsystems, Inc. All rights reserved.
* Use is subject to license terms.
*/
#pragma ident "%Z%%M% %I% %E% SMI"
#include <cma.h>
#include <strings.h>
#include <errno.h>
#include <time.h>
#include <fm/fmd_api.h>
#include <sys/fm/protocol.h>
cma_t cma;
cma_stats_t cma_stats = {
{ "cpu_flts", FMD_TYPE_UINT64, "cpu faults resolved" },
{ "cpu_fails", FMD_TYPE_UINT64, "cpu faults unresolveable" },
{ "cpu_blfails", FMD_TYPE_UINT64, "failed cpu blacklists" },
{ "cpu_supp", FMD_TYPE_UINT64, "cpu offlines suppressed" },
{ "cpu_blsupp", FMD_TYPE_UINT64, "cpu blacklists suppressed" },
{ "page_flts", FMD_TYPE_UINT64, "page faults resolved" },
{ "page_fails", FMD_TYPE_UINT64, "page faults unresolveable" },
{ "page_supp", FMD_TYPE_UINT64, "page retires suppressed" },
{ "page_nonent", FMD_TYPE_UINT64, "retires for non-existent fmris" },
{ "page_retmax", FMD_TYPE_UINT64, "hit max retries for page retire" },
{ "bad_flts", FMD_TYPE_UINT64, "invalid fault events received" },
{ "nop_flts", FMD_TYPE_UINT64, "inapplicable fault events received" },
{ "auto_flts", FMD_TYPE_UINT64, "auto-close faults received" }
};
typedef struct cma_subscriber {
const char *subr_class;
const char *subr_sname;
uint_t subr_svers;
void (*subr_func)(fmd_hdl_t *, nvlist_t *, nvlist_t *, const char *);
} cma_subscriber_t;
static const cma_subscriber_t cma_subrs[] = {
{ "fault.cpu.*", FM_FMRI_SCHEME_HC, FM_HC_SCHEME_VERSION,
cma_cpu_retire },
{ "fault.memory.page", FM_FMRI_SCHEME_MEM, FM_MEM_SCHEME_VERSION,
cma_page_retire },
{ "fault.memory.dimm", FM_FMRI_SCHEME_MEM, FM_MEM_SCHEME_VERSION,
NULL },
{ "fault.memory.dimm_sb", FM_FMRI_SCHEME_HC, FM_HC_SCHEME_VERSION,
NULL },
{ "fault.memory.dimm_ck", FM_FMRI_SCHEME_HC, FM_HC_SCHEME_VERSION,
NULL },
{ "fault.memory.dimm_ue", FM_FMRI_SCHEME_HC, FM_HC_SCHEME_VERSION,
NULL },
{ "fault.memory.bank", FM_FMRI_SCHEME_MEM, FM_MEM_SCHEME_VERSION,
NULL },
{ "fault.memory.datapath", FM_FMRI_SCHEME_MEM, FM_MEM_SCHEME_VERSION,
NULL },
/*
* The following ultraSPARC-T1 faults do NOT retire a cpu thread,
* and therefore must be intercepted before
* the default "fault.cpu.*" dispatch to cma_cpu_retire.
*/
{ "fault.cpu.ultraSPARC-T1.freg", FM_FMRI_SCHEME_CPU,
FM_CPU_SCHEME_VERSION, NULL },
{ "fault.cpu.ultraSPARC-T1.l2cachedata", FM_FMRI_SCHEME_CPU,
FM_CPU_SCHEME_VERSION, NULL },
{ "fault.cpu.ultraSPARC-T1.l2cachetag", FM_FMRI_SCHEME_CPU,
FM_CPU_SCHEME_VERSION, NULL },
{ "fault.cpu.ultraSPARC-T1.l2cachectl", FM_FMRI_SCHEME_CPU,
FM_CPU_SCHEME_VERSION, NULL },
{ "fault.cpu.ultraSPARC-T1.mau", FM_FMRI_SCHEME_CPU,
FM_CPU_SCHEME_VERSION, NULL },
{ "fault.cpu.*", FM_FMRI_SCHEME_CPU, FM_CPU_SCHEME_VERSION,
cma_cpu_retire },
{ NULL, NULL, 0, NULL }
};
static const cma_subscriber_t *
nvl2subr(fmd_hdl_t *hdl, nvlist_t *nvl, nvlist_t **asrup)
{
const cma_subscriber_t *sp;
nvlist_t *asru;
char *scheme;
uint8_t version;
if (nvlist_lookup_nvlist(nvl, FM_FAULT_ASRU, &asru) != 0 ||
nvlist_lookup_string(asru, FM_FMRI_SCHEME, &scheme) != 0 ||
nvlist_lookup_uint8(asru, FM_VERSION, &version) != 0) {
cma_stats.bad_flts.fmds_value.ui64++;
return (NULL);
}
for (sp = cma_subrs; sp->subr_class != NULL; sp++) {
if (fmd_nvl_class_match(hdl, nvl, sp->subr_class) &&
strcmp(scheme, sp->subr_sname) == 0 &&
version <= sp->subr_svers) {
*asrup = asru;
return (sp);
}
}
cma_stats.nop_flts.fmds_value.ui64++;
return (NULL);
}
static void
cma_recv_list(fmd_hdl_t *hdl, nvlist_t *nvl)
{
char *uuid = NULL;
nvlist_t **nva;
uint_t nvc = 0;
int err = 0;
err |= nvlist_lookup_string(nvl, FM_SUSPECT_UUID, &uuid);
err |= nvlist_lookup_nvlist_array(nvl, FM_SUSPECT_FAULT_LIST,
&nva, &nvc);
if (err != 0) {
cma_stats.bad_flts.fmds_value.ui64++;
return;
}
while (nvc-- != 0 && !fmd_case_uuclosed(hdl, uuid)) {
nvlist_t *nvl = *nva++;
const cma_subscriber_t *subr;
nvlist_t *asru;
if ((subr = nvl2subr(hdl, nvl, &asru)) == NULL)
continue;
if (subr->subr_func != NULL)
subr->subr_func(hdl, nvl, asru, uuid);
}
}
static void
cma_recv_one(fmd_hdl_t *hdl, nvlist_t *nvl)
{
const cma_subscriber_t *subr;
nvlist_t *asru;
if ((subr = nvl2subr(hdl, nvl, &asru)) == NULL)
return;
if (subr->subr_func != NULL)
subr->subr_func(hdl, nvl, asru, NULL);
}
/*ARGSUSED*/
static void
cma_recv(fmd_hdl_t *hdl, fmd_event_t *ep, nvlist_t *nvl, const char *class)
{
fmd_hdl_debug(hdl, "received %s\n", class);
if (strcmp(class, FM_LIST_SUSPECT_CLASS) == 0)
cma_recv_list(hdl, nvl);
else
cma_recv_one(hdl, nvl);
}
/*ARGSUSED*/
static void
cma_timeout(fmd_hdl_t *hdl, id_t id, void *arg)
{
if (id == cma.cma_page_timerid)
cma_page_retry(hdl);
}
static const fmd_hdl_ops_t fmd_ops = {
cma_recv, /* fmdo_recv */
cma_timeout, /* fmdo_timeout */
NULL, /* fmdo_close */
NULL, /* fmdo_stats */
NULL, /* fmdo_gc */
};
static const fmd_prop_t fmd_props[] = {
{ "cpu_tries", FMD_TYPE_UINT32, "10" },
{ "cpu_delay", FMD_TYPE_TIME, "1sec" },
{ "cpu_offline_enable", FMD_TYPE_BOOL, "true" },
{ "cpu_forced_offline", FMD_TYPE_BOOL, "true" },
{ "cpu_blacklist_enable", FMD_TYPE_BOOL, "true" },
{ "page_ret_mindelay", FMD_TYPE_TIME, "1sec" },
{ "page_ret_maxdelay", FMD_TYPE_TIME, "5min" },
{ "page_retire_enable", FMD_TYPE_BOOL, "true" },
#ifdef i386
/*
* On i386, leaving cases open while we retry the
* retire can cause the eft module to use large amounts
* of memory. Until eft is fixed, we set a maximum number
* of retries on page retires, after which the case will
* be closed.
*/
{ "page_retire_maxretries", FMD_TYPE_UINT32, "8" },
#else
{ "page_retire_maxretries", FMD_TYPE_UINT32, "0" },
#endif /* i386 */
{ NULL, 0, NULL }
};
static const fmd_hdl_info_t fmd_info = {
"CPU/Memory Retire Agent", CMA_VERSION, &fmd_ops, fmd_props
};
void
_fmd_init(fmd_hdl_t *hdl)
{
hrtime_t nsec;
if (fmd_hdl_register(hdl, FMD_API_VERSION, &fmd_info) != 0)
return; /* invalid data in configuration file */
fmd_hdl_subscribe(hdl, "fault.cpu.*");
fmd_hdl_subscribe(hdl, "fault.memory.*");
(void) fmd_stat_create(hdl, FMD_STAT_NOALLOC, sizeof (cma_stats) /
sizeof (fmd_stat_t), (fmd_stat_t *)&cma_stats);
cma.cma_cpu_tries = fmd_prop_get_int32(hdl, "cpu_tries");
nsec = fmd_prop_get_int64(hdl, "cpu_delay");
cma.cma_cpu_delay.tv_sec = nsec / NANOSEC;
cma.cma_cpu_delay.tv_nsec = nsec % NANOSEC;
cma.cma_page_mindelay = fmd_prop_get_int64(hdl, "page_ret_mindelay");
cma.cma_page_maxdelay = fmd_prop_get_int64(hdl, "page_ret_maxdelay");
cma.cma_cpu_dooffline = fmd_prop_get_int32(hdl, "cpu_offline_enable");
cma.cma_cpu_forcedoffline = fmd_prop_get_int32(hdl,
"cpu_forced_offline");
cma.cma_cpu_doblacklist = fmd_prop_get_int32(hdl,
"cpu_blacklist_enable");
cma.cma_page_doretire = fmd_prop_get_int32(hdl, "page_retire_enable");
cma.cma_page_maxretries =
fmd_prop_get_int32(hdl, "page_retire_maxretries");
if (cma.cma_page_maxdelay < cma.cma_page_mindelay)
fmd_hdl_abort(hdl, "page retirement delays conflict\n");
}
void
_fmd_fini(fmd_hdl_t *hdl)
{
cma_page_fini(hdl);
}