xref: /freebsd/sys/x86/iommu/intel_fault.c (revision 1e66f787c838b5af7de716e266caf4e5d190d54b)
1 /*-
2  * SPDX-License-Identifier: BSD-2-Clause-FreeBSD
3  *
4  * Copyright (c) 2013 The FreeBSD Foundation
5  * All rights reserved.
6  *
7  * This software was developed by Konstantin Belousov <kib@FreeBSD.org>
8  * under sponsorship from the FreeBSD Foundation.
9  *
10  * Redistribution and use in source and binary forms, with or without
11  * modification, are permitted provided that the following conditions
12  * are met:
13  * 1. Redistributions of source code must retain the above copyright
14  *    notice, this list of conditions and the following disclaimer.
15  * 2. Redistributions in binary form must reproduce the above copyright
16  *    notice, this list of conditions and the following disclaimer in the
17  *    documentation and/or other materials provided with the distribution.
18  *
19  * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
20  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
21  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
22  * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
23  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
24  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
25  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
26  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
27  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
28  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
29  * SUCH DAMAGE.
30  */
31 
32 #include <sys/cdefs.h>
33 __FBSDID("$FreeBSD$");
34 
35 #include "opt_acpi.h"
36 
37 #include <sys/param.h>
38 #include <sys/bus.h>
39 #include <sys/kernel.h>
40 #include <sys/malloc.h>
41 #include <sys/memdesc.h>
42 #include <sys/module.h>
43 #include <sys/rman.h>
44 #include <sys/taskqueue.h>
45 #include <sys/tree.h>
46 #include <sys/vmem.h>
47 #include <machine/bus.h>
48 #include <contrib/dev/acpica/include/acpi.h>
49 #include <contrib/dev/acpica/include/accommon.h>
50 #include <dev/acpica/acpivar.h>
51 #include <dev/pci/pcireg.h>
52 #include <dev/pci/pcivar.h>
53 #include <vm/vm.h>
54 #include <vm/vm_extern.h>
55 #include <vm/vm_kern.h>
56 #include <vm/vm_page.h>
57 #include <vm/vm_map.h>
58 #include <x86/include/busdma_impl.h>
59 #include <x86/iommu/intel_reg.h>
60 #include <x86/iommu/busdma_dmar.h>
61 #include <x86/iommu/intel_dmar.h>
62 
63 /*
64  * Fault interrupt handling for DMARs.  If advanced fault logging is
65  * not implemented by hardware, the code emulates it.  Fast interrupt
66  * handler flushes the fault registers into circular buffer at
67  * unit->fault_log, and schedules a task.
68  *
69  * The fast handler is used since faults usually come in bursts, and
70  * number of fault log registers is limited, e.g. down to one for 5400
71  * MCH.  We are trying to reduce the latency for clearing the fault
72  * register file.  The task is usually long-running, since printf() is
73  * slow, but this is not problematic because bursts are rare.
74  *
75  * For the same reason, each translation unit task is executed in its
76  * own thread.
77  *
78  * XXXKIB It seems there is no hardware available which implements
79  * advanced fault logging, so the code to handle AFL is not written.
80  */
81 
82 static int
83 dmar_fault_next(struct dmar_unit *unit, int faultp)
84 {
85 
86 	faultp += 2;
87 	if (faultp == unit->fault_log_size)
88 		faultp = 0;
89 	return (faultp);
90 }
91 
92 static void
93 dmar_fault_intr_clear(struct dmar_unit *unit, uint32_t fsts)
94 {
95 	uint32_t clear;
96 
97 	clear = 0;
98 	if ((fsts & DMAR_FSTS_ITE) != 0) {
99 		printf("DMAR%d: Invalidation timed out\n", unit->unit);
100 		clear |= DMAR_FSTS_ITE;
101 	}
102 	if ((fsts & DMAR_FSTS_ICE) != 0) {
103 		printf("DMAR%d: Invalidation completion error\n",
104 		    unit->unit);
105 		clear |= DMAR_FSTS_ICE;
106 	}
107 	if ((fsts & DMAR_FSTS_IQE) != 0) {
108 		printf("DMAR%d: Invalidation queue error\n",
109 		    unit->unit);
110 		clear |= DMAR_FSTS_IQE;
111 	}
112 	if ((fsts & DMAR_FSTS_APF) != 0) {
113 		printf("DMAR%d: Advanced pending fault\n", unit->unit);
114 		clear |= DMAR_FSTS_APF;
115 	}
116 	if ((fsts & DMAR_FSTS_AFO) != 0) {
117 		printf("DMAR%d: Advanced fault overflow\n", unit->unit);
118 		clear |= DMAR_FSTS_AFO;
119 	}
120 	if (clear != 0)
121 		dmar_write4(unit, DMAR_FSTS_REG, clear);
122 }
123 
124 int
125 dmar_fault_intr(void *arg)
126 {
127 	struct dmar_unit *unit;
128 	uint64_t fault_rec[2];
129 	uint32_t fsts;
130 	int fri, frir, faultp;
131 	bool enqueue;
132 
133 	unit = arg;
134 	enqueue = false;
135 	fsts = dmar_read4(unit, DMAR_FSTS_REG);
136 	dmar_fault_intr_clear(unit, fsts);
137 
138 	if ((fsts & DMAR_FSTS_PPF) == 0)
139 		goto done;
140 
141 	fri = DMAR_FSTS_FRI(fsts);
142 	for (;;) {
143 		frir = (DMAR_CAP_FRO(unit->hw_cap) + fri) * 16;
144 		fault_rec[1] = dmar_read8(unit, frir + 8);
145 		if ((fault_rec[1] & DMAR_FRCD2_F) == 0)
146 			break;
147 		fault_rec[0] = dmar_read8(unit, frir);
148 		dmar_write4(unit, frir + 12, DMAR_FRCD2_F32);
149 		DMAR_FAULT_LOCK(unit);
150 		faultp = unit->fault_log_head;
151 		if (dmar_fault_next(unit, faultp) == unit->fault_log_tail) {
152 			/* XXXKIB log overflow */
153 		} else {
154 			unit->fault_log[faultp] = fault_rec[0];
155 			unit->fault_log[faultp + 1] = fault_rec[1];
156 			unit->fault_log_head = dmar_fault_next(unit, faultp);
157 			enqueue = true;
158 		}
159 		DMAR_FAULT_UNLOCK(unit);
160 		fri += 1;
161 		if (fri >= DMAR_CAP_NFR(unit->hw_cap))
162 			fri = 0;
163 	}
164 
165 done:
166 	/*
167 	 * On SandyBridge, due to errata BJ124, IvyBridge errata
168 	 * BV100, and Haswell errata HSD40, "Spurious Intel VT-d
169 	 * Interrupts May Occur When the PFO Bit is Set".  Handle the
170 	 * cases by clearing overflow bit even if no fault is
171 	 * reported.
172 	 *
173 	 * On IvyBridge, errata BV30 states that clearing clear
174 	 * DMAR_FRCD2_F bit in the fault register causes spurious
175 	 * interrupt.  Do nothing.
176 	 *
177 	 */
178 	if ((fsts & DMAR_FSTS_PFO) != 0) {
179 		printf("DMAR%d: Fault Overflow\n", unit->unit);
180 		dmar_write4(unit, DMAR_FSTS_REG, DMAR_FSTS_PFO);
181 	}
182 
183 	if (enqueue) {
184 		taskqueue_enqueue(unit->fault_taskqueue,
185 		    &unit->fault_task);
186 	}
187 	return (FILTER_HANDLED);
188 }
189 
190 static void
191 dmar_fault_task(void *arg, int pending __unused)
192 {
193 	struct dmar_unit *unit;
194 	struct dmar_ctx *ctx;
195 	uint64_t fault_rec[2];
196 	int sid, bus, slot, func, faultp;
197 
198 	unit = arg;
199 	DMAR_FAULT_LOCK(unit);
200 	for (;;) {
201 		faultp = unit->fault_log_tail;
202 		if (faultp == unit->fault_log_head)
203 			break;
204 
205 		fault_rec[0] = unit->fault_log[faultp];
206 		fault_rec[1] = unit->fault_log[faultp + 1];
207 		unit->fault_log_tail = dmar_fault_next(unit, faultp);
208 		DMAR_FAULT_UNLOCK(unit);
209 
210 		sid = DMAR_FRCD2_SID(fault_rec[1]);
211 		printf("DMAR%d: ", unit->unit);
212 		DMAR_LOCK(unit);
213 		ctx = dmar_find_ctx_locked(unit, sid);
214 		if (ctx == NULL) {
215 			printf("<unknown dev>:");
216 
217 			/*
218 			 * Note that the slot and function will not be correct
219 			 * if ARI is in use, but without a ctx entry we have
220 			 * no way of knowing whether ARI is in use or not.
221 			 */
222 			bus = PCI_RID2BUS(sid);
223 			slot = PCI_RID2SLOT(sid);
224 			func = PCI_RID2FUNC(sid);
225 		} else {
226 			ctx->flags |= DMAR_CTX_FAULTED;
227 			ctx->last_fault_rec[0] = fault_rec[0];
228 			ctx->last_fault_rec[1] = fault_rec[1];
229 			device_print_prettyname(ctx->ctx_tag.owner);
230 			bus = pci_get_bus(ctx->ctx_tag.owner);
231 			slot = pci_get_slot(ctx->ctx_tag.owner);
232 			func = pci_get_function(ctx->ctx_tag.owner);
233 		}
234 		DMAR_UNLOCK(unit);
235 		printf(
236 		    "pci%d:%d:%d sid %x fault acc %x adt 0x%x reason 0x%x "
237 		    "addr %jx\n",
238 		    bus, slot, func, sid, DMAR_FRCD2_T(fault_rec[1]),
239 		    DMAR_FRCD2_AT(fault_rec[1]), DMAR_FRCD2_FR(fault_rec[1]),
240 		    (uintmax_t)fault_rec[0]);
241 		DMAR_FAULT_LOCK(unit);
242 	}
243 	DMAR_FAULT_UNLOCK(unit);
244 }
245 
246 static void
247 dmar_clear_faults(struct dmar_unit *unit)
248 {
249 	uint32_t frec, frir, fsts;
250 	int i;
251 
252 	for (i = 0; i < DMAR_CAP_NFR(unit->hw_cap); i++) {
253 		frir = (DMAR_CAP_FRO(unit->hw_cap) + i) * 16;
254 		frec = dmar_read4(unit, frir + 12);
255 		if ((frec & DMAR_FRCD2_F32) == 0)
256 			continue;
257 		dmar_write4(unit, frir + 12, DMAR_FRCD2_F32);
258 	}
259 	fsts = dmar_read4(unit, DMAR_FSTS_REG);
260 	dmar_write4(unit, DMAR_FSTS_REG, fsts);
261 }
262 
263 int
264 dmar_init_fault_log(struct dmar_unit *unit)
265 {
266 
267 	mtx_init(&unit->fault_lock, "dmarflt", NULL, MTX_SPIN);
268 	unit->fault_log_size = 256; /* 128 fault log entries */
269 	TUNABLE_INT_FETCH("hw.dmar.fault_log_size", &unit->fault_log_size);
270 	if (unit->fault_log_size % 2 != 0)
271 		panic("hw.dmar_fault_log_size must be even");
272 	unit->fault_log = malloc(sizeof(uint64_t) * unit->fault_log_size,
273 	    M_DEVBUF, M_WAITOK | M_ZERO);
274 
275 	TASK_INIT(&unit->fault_task, 0, dmar_fault_task, unit);
276 	unit->fault_taskqueue = taskqueue_create_fast("dmarff", M_WAITOK,
277 	    taskqueue_thread_enqueue, &unit->fault_taskqueue);
278 	taskqueue_start_threads(&unit->fault_taskqueue, 1, PI_AV,
279 	    "dmar%d fault taskq", unit->unit);
280 
281 	DMAR_LOCK(unit);
282 	dmar_disable_fault_intr(unit);
283 	dmar_clear_faults(unit);
284 	dmar_enable_fault_intr(unit);
285 	DMAR_UNLOCK(unit);
286 
287 	return (0);
288 }
289 
290 void
291 dmar_fini_fault_log(struct dmar_unit *unit)
292 {
293 
294 	DMAR_LOCK(unit);
295 	dmar_disable_fault_intr(unit);
296 	DMAR_UNLOCK(unit);
297 
298 	if (unit->fault_taskqueue == NULL)
299 		return;
300 
301 	taskqueue_drain(unit->fault_taskqueue, &unit->fault_task);
302 	taskqueue_free(unit->fault_taskqueue);
303 	unit->fault_taskqueue = NULL;
304 	mtx_destroy(&unit->fault_lock);
305 
306 	free(unit->fault_log, M_DEVBUF);
307 	unit->fault_log = NULL;
308 	unit->fault_log_head = unit->fault_log_tail = 0;
309 }
310 
311 void
312 dmar_enable_fault_intr(struct dmar_unit *unit)
313 {
314 	uint32_t fectl;
315 
316 	DMAR_ASSERT_LOCKED(unit);
317 	fectl = dmar_read4(unit, DMAR_FECTL_REG);
318 	fectl &= ~DMAR_FECTL_IM;
319 	dmar_write4(unit, DMAR_FECTL_REG, fectl);
320 }
321 
322 void
323 dmar_disable_fault_intr(struct dmar_unit *unit)
324 {
325 	uint32_t fectl;
326 
327 	DMAR_ASSERT_LOCKED(unit);
328 	fectl = dmar_read4(unit, DMAR_FECTL_REG);
329 	dmar_write4(unit, DMAR_FECTL_REG, fectl | DMAR_FECTL_IM);
330 }
331