usr/src/cmd/fm/modules/common/cpumem-retire/cma_cache.c

   1 /*
   2  * CDDL HEADER START
   3  *
   4  * The contents of this file are subject to the terms of the
   5  * Common Development and Distribution License (the "License").
   6  * You may not use this file except in compliance with the License.
   7  *
   8  * You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE
   9  * or http://www.opensolaris.org/os/licensing.
  10  * See the License for the specific language governing permissions
  11  * and limitations under the License.
  12  *
  13  * When distributing Covered Code, include this CDDL HEADER in each
  14  * file and include the License file at usr/src/OPENSOLARIS.LICENSE.
  15  * If applicable, add the following below this CDDL HEADER, with the
  16  * fields enclosed by brackets "[]" replaced with your own identifying
  17  * information: Portions Copyright [yyyy] [name of copyright owner]
  18  *
  19  * CDDL HEADER END
  20  */
  21 /*
  22  * Copyright 2009 Sun Microsystems, Inc.  All rights reserved.
  23  * Use is subject to license terms.
  24  */
  25
  26 #include <cma.h>
  27 #include <fcntl.h>
  28 #include <strings.h>
  29 #include <unistd.h>
  30 #include <errno.h>
  31 #include <fm/fmd_api.h>
  32 #include <sys/fm/protocol.h>
  33 #include <sys/mem_cache.h>
  34
  35 int
  36 /* ARGSUSED 4 */
  37 cma_cache_way_retire(fmd_hdl_t *hdl, nvlist_t *nvl, nvlist_t *asru,
  38     const char *uuid, boolean_t repair)
  39 {
  40         uint_t          cpuid;
  41         uint32_t        way;
  42         uint32_t        index;
  43         uint16_t        bit = 0;
  44         uint8_t         type;
  45         cache_info_t    cache_info;
  46         int ret, fd;
  47         int     err;
  48
  49         fmd_hdl_debug(hdl, "cpu cache *line* fault processing\n");
  50         fmd_hdl_debug(hdl, "asru %lx\n", asru);
  51
  52         /*
  53          * This added expansion is needed to cover the situation where a
  54          * cpu fault from the resource cache is replayed at fmd restart,
  55          * and the cpu resource has been remapped or replaced.  The stored
  56          * FMRI is expanded, but may have stale data.
  57          */
  58
  59         if (fmd_nvl_fmri_expand(hdl, asru) < 0) {
  60                 fmd_hdl_debug(hdl, "failed to expand cpu asru\n");
  61                 cma_stats.bad_flts.fmds_value.ui64++;
  62                 return (CMA_RA_FAILURE);
  63         }
  64
  65         if (nvlist_lookup_uint32(asru, FM_FMRI_CPU_ID, &cpuid) != 0) {
  66                 fmd_hdl_debug(hdl, "cpu cache fault missing '%s'\n",
  67                     FM_FMRI_CPU_ID);
  68                 cma_stats.bad_flts.fmds_value.ui64++;
  69                 return (CMA_RA_FAILURE);
  70         }
  71
  72         if (nvlist_lookup_uint32(asru, FM_FMRI_CPU_CACHE_INDEX, &index) != 0) {
  73                 fmd_hdl_debug(hdl, "cpu cache fault missing '%s'\n",
  74                     FM_FMRI_CPU_CACHE_INDEX);
  75                 cma_stats.bad_flts.fmds_value.ui64++;
  76                 return (CMA_RA_FAILURE);
  77         }
  78
  79         if (nvlist_lookup_uint32(asru, FM_FMRI_CPU_CACHE_WAY, &way) != 0) {
  80                 fmd_hdl_debug(hdl, "cpu cache fault missing '%s'\n",
  81                     FM_FMRI_CPU_CACHE_WAY);
  82                 cma_stats.bad_flts.fmds_value.ui64++;
  83                 return (CMA_RA_FAILURE);
  84         }
  85
  86         if (nvlist_lookup_uint8(asru, FM_FMRI_CPU_CACHE_TYPE, &type) != 0) {
  87                 fmd_hdl_debug(hdl, "cpu cache fault missing '%s'\n",
  88                     FM_FMRI_CPU_CACHE_TYPE);
  89                 cma_stats.bad_flts.fmds_value.ui64++;
  90                 return (CMA_RA_FAILURE);
  91         }
  92
  93         /*
  94          * Tag faults will use it to set the bit to a stable state
  95          */
  96
  97         if (nvlist_lookup_uint16(asru, FM_FMRI_CPU_CACHE_BIT, &bit) != 0) {
  98                 fmd_hdl_debug(hdl, "cpu cache fault missing '%s'\n",
  99                     FM_FMRI_CPU_CACHE_BIT);
 100                 cma_stats.bad_flts.fmds_value.ui64++;
 101                 return (CMA_RA_FAILURE);
 102         }
 103         if (repair) {
 104                 fmd_hdl_debug(hdl,
 105                     "cpu %d: Unretire for index %d, way %d\n bit %d"
 106                     " type 0x%02x is being called now. We will not unretire"
 107                     " cacheline. This message is for information.\n",
 108                     cpuid, index, way, bit, type);
 109                 /*
 110                  * The DE will do the actual unretire.
 111                  * The agent is called because the DE informs the fmd
 112                  * that the resource is repaired.
 113                  */
 114                 return (CMA_RA_SUCCESS);
 115         }
 116         fmd_hdl_debug(hdl,
 117             "cpu %d: Retiring index %d, way %d\n bit %d"
 118             " type 0x%02x", cpuid, index, way, bit, type);
 119         fd = open(mem_cache_device, O_RDWR);
 120         if (fd == -1) {
 121                 fmd_hdl_debug(hdl, "Driver open failed\n");
 122                 return (CMA_RA_FAILURE);
 123         }
 124         cache_info.cpu_id = cpuid;
 125         cache_info.way = way;
 126         cache_info.bit = bit;
 127         cache_info.index = index;
 128         cache_info.cache = type == FM_FMRI_CPU_CACHE_TYPE_L3 ?
 129             L3_CACHE_DATA : L2_CACHE_DATA;
 130
 131         ret = ioctl(fd, MEM_CACHE_RETIRE, &cache_info);
 132         /*
 133          * save errno before we call close.
 134          * Need errno to display the error if ioctl fails.
 135          */
 136         err = errno;
 137         (void) close(fd);
 138         if (ret == -1) {
 139                 fmd_hdl_debug(hdl, "Driver call failed errno = %d\n", err);
 140                 return (CMA_RA_FAILURE);
 141         }
 142         return (CMA_RA_SUCCESS);
 143 }