1 /* $NetBSD: memerr.c,v 1.14 2002/10/02 16:02:26 thorpej Exp $ */ 2 3 /* 4 * Copyright (c) 1992, 1993 5 * The Regents of the University of California. All rights reserved. 6 * 7 * This software was developed by the Computer Systems Engineering group 8 * at Lawrence Berkeley Laboratory under DARPA contract BG 91-66 and 9 * contributed to Berkeley. 10 * 11 * All advertising materials mentioning features or use of this software 12 * must display the following acknowledgement: 13 * This product includes software developed by the University of 14 * California, Lawrence Berkeley Laboratory. 15 * 16 * Redistribution and use in source and binary forms, with or without 17 * modification, are permitted provided that the following conditions 18 * are met: 19 * 1. Redistributions of source code must retain the above copyright 20 * notice, this list of conditions and the following disclaimer. 21 * 2. Redistributions in binary form must reproduce the above copyright 22 * notice, this list of conditions and the following disclaimer in the 23 * documentation and/or other materials provided with the distribution. 24 * 3. All advertising materials mentioning features or use of this software 25 * must display the following acknowledgement: 26 * This product includes software developed by the University of 27 * California, Berkeley and its contributors. 28 * 4. Neither the name of the University nor the names of its contributors 29 * may be used to endorse or promote products derived from this software 30 * without specific prior written permission. 31 * 32 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND 33 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 34 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 35 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE 36 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 37 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 38 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 39 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 40 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 41 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 42 * SUCH DAMAGE. 43 * 44 * @(#)memreg.c 8.1 (Berkeley) 6/11/93 45 */ 46 47 #include <sys/param.h> 48 #include <sys/systm.h> 49 #include <sys/device.h> 50 51 #include <machine/autoconf.h> 52 #include <machine/cpu.h> 53 #include <machine/idprom.h> 54 #include <machine/pte.h> 55 56 #include <sun3/sun3/machdep.h> 57 #include <sun3/dev/memerr.h> 58 /* #include <sun3/dev/eccreg.h> - not yet */ 59 60 #define ME_PRI 7 /* Interrupt level (NMI) */ 61 62 extern unsigned char cpu_machine_id; 63 64 enum memerr_type { ME_PAR = 0, ME_ECC = 1 }; 65 66 struct memerr_softc { 67 struct device sc_dev; 68 struct memerr *sc_reg; 69 enum memerr_type sc_type; 70 char *sc_typename; /* "Parity" or "ECC" */ 71 char *sc_csrbits; /* how to print csr bits */ 72 /* XXX: counters? */ 73 }; 74 75 static int memerr_match __P((struct device *, struct cfdata *, void *)); 76 static void memerr_attach __P((struct device *, struct device *, void *)); 77 static int memerr_interrupt __P((void *)); 78 static void memerr_correctable __P((struct memerr_softc *)); 79 80 CFATTACH_DECL(memerr, sizeof(struct memerr_softc), 81 memerr_match, memerr_attach, NULL, NULL); 82 83 int memerr_attached; 84 85 static int 86 memerr_match(parent, cf, args) 87 struct device *parent; 88 struct cfdata *cf; 89 void *args; 90 { 91 struct confargs *ca = args; 92 93 /* This driver only supports one instance. */ 94 if (memerr_attached) 95 return (0); 96 97 /* Make sure there is something there... */ 98 if (bus_peek(ca->ca_bustype, ca->ca_paddr, 1) == -1) 99 return (0); 100 101 /* Default interrupt priority. */ 102 if (ca->ca_intpri == -1) 103 ca->ca_intpri = ME_PRI; 104 105 return (1); 106 } 107 108 static void 109 memerr_attach(parent, self, args) 110 struct device *parent; 111 struct device *self; 112 void *args; 113 { 114 struct memerr_softc *sc = (void *)self; 115 struct confargs *ca = args; 116 struct memerr *mer; 117 118 /* 119 * Which type of memory subsystem do we have? 120 */ 121 switch (cpu_machine_id) { 122 case SUN3_MACH_160: /* XXX: correct? */ 123 case SUN3_MACH_260: 124 case SUN3X_MACH_470: 125 sc->sc_type = ME_ECC; 126 sc->sc_typename = "ECC"; 127 sc->sc_csrbits = ME_ECC_STR; 128 break; 129 130 default: 131 sc->sc_type = ME_PAR; 132 sc->sc_typename = "Parity"; 133 sc->sc_csrbits = ME_PAR_STR; 134 break; 135 } 136 printf(": (%s memory)\n", sc->sc_typename); 137 138 mer = bus_mapin(ca->ca_bustype, ca->ca_paddr, sizeof(*mer)); 139 if (mer == NULL) 140 panic("memerr: can not map register"); 141 sc->sc_reg = mer; 142 143 /* Install interrupt handler. */ 144 isr_add_autovect(memerr_interrupt, sc, ca->ca_intpri); 145 146 /* Enable error interrupt (and checking). */ 147 if (sc->sc_type == ME_PAR) 148 mer->me_csr = ME_CSR_IENA | ME_PAR_CHECK; 149 else { 150 /* 151 * XXX: Some day, figure out how to decode 152 * correctable errors and set ME_ECC_CE_ENA 153 * here so we can log them... 154 */ 155 mer->me_csr = ME_CSR_IENA; /* | ME_ECC_CE_ENA */ 156 } 157 memerr_attached = 1; 158 } 159 160 /***************************************************************** 161 * Functions for ECC memory 162 *****************************************************************/ 163 164 static int 165 memerr_interrupt(arg) 166 void *arg; 167 { 168 struct memerr_softc *sc = arg; 169 volatile struct memerr *me = sc->sc_reg; 170 u_char csr, ctx; 171 u_int pa, va; 172 int pte; 173 char bits[64]; 174 175 csr = me->me_csr; 176 if ((csr & ME_CSR_IPEND) == 0) 177 return (0); 178 179 va = me->me_vaddr; 180 ctx = (va >> 28) & 0xF; 181 va &= 0x0FFFffff; 182 pte = get_pte(va); 183 pa = PG_PA(pte); 184 185 printf("\nMemory error on %s cycle!\n", 186 (ctx & 8) ? "DVMA" : "CPU"); 187 printf(" ctx=%d, vaddr=0x%x, paddr=0x%x\n", 188 (ctx & 7), va, pa); 189 printf(" csr=%s\n", bitmask_snprintf(csr, sc->sc_csrbits, 190 bits, sizeof(bits))); 191 192 /* 193 * If we have parity-checked memory, there is 194 * not much to be done. Any error is fatal. 195 */ 196 if (sc->sc_type == ME_PAR) { 197 if (csr & ME_PAR_EMASK) { 198 /* Parity errors are fatal. */ 199 goto die; 200 } 201 /* The IPEND bit was set, but no error bits. */ 202 goto noerror; 203 } 204 205 /* 206 * We have ECC memory. More complicated... 207 */ 208 if (csr & (ME_ECC_WBTMO | ME_ECC_WBERR)) { 209 printf(" write-back failed, pte=0x%x\n", pte); 210 goto die; 211 } 212 if (csr & ME_ECC_UE) { 213 printf(" uncorrectable ECC error\n"); 214 goto die; 215 } 216 if (csr & ME_ECC_CE) { 217 /* Just log this and continue. */ 218 memerr_correctable(sc); 219 goto recover; 220 } 221 /* The IPEND bit was set, but no error bits. */ 222 goto noerror; 223 224 die: 225 panic("all bets are off..."); 226 227 noerror: 228 printf("memerr: no error bits set?\n"); 229 230 recover: 231 /* Clear the error by writing the address register. */ 232 me->me_vaddr = 0; 233 return (1); 234 } 235 236 /* 237 * Announce (and log) a correctable ECC error. 238 * Need to look at the ECC syndrome register on 239 * the memory board that caused the error... 240 */ 241 void 242 memerr_correctable(sc) 243 struct memerr_softc *sc; 244 { 245 /* XXX: Not yet... */ 246 } 247