xref: /openbmc/linux/drivers/iommu/intel/dmar.c (revision 0661cb2a)
1 // SPDX-License-Identifier: GPL-2.0-only
2 /*
3  * Copyright (c) 2006, Intel Corporation.
4  *
5  * Copyright (C) 2006-2008 Intel Corporation
6  * Author: Ashok Raj <ashok.raj@intel.com>
7  * Author: Shaohua Li <shaohua.li@intel.com>
8  * Author: Anil S Keshavamurthy <anil.s.keshavamurthy@intel.com>
9  *
10  * This file implements early detection/parsing of Remapping Devices
11  * reported to OS through BIOS via DMA remapping reporting (DMAR) ACPI
12  * tables.
13  *
14  * These routines are used by both DMA-remapping and Interrupt-remapping
15  */
16 
17 #define pr_fmt(fmt)     "DMAR: " fmt
18 
19 #include <linux/pci.h>
20 #include <linux/dmar.h>
21 #include <linux/iova.h>
22 #include <linux/intel-iommu.h>
23 #include <linux/timer.h>
24 #include <linux/irq.h>
25 #include <linux/interrupt.h>
26 #include <linux/tboot.h>
27 #include <linux/dmi.h>
28 #include <linux/slab.h>
29 #include <linux/iommu.h>
30 #include <linux/numa.h>
31 #include <linux/limits.h>
32 #include <asm/irq_remapping.h>
33 #include <asm/iommu_table.h>
34 #include <trace/events/intel_iommu.h>
35 
36 #include "../irq_remapping.h"
37 #include "perf.h"
38 
39 typedef int (*dmar_res_handler_t)(struct acpi_dmar_header *, void *);
40 struct dmar_res_callback {
41 	dmar_res_handler_t	cb[ACPI_DMAR_TYPE_RESERVED];
42 	void			*arg[ACPI_DMAR_TYPE_RESERVED];
43 	bool			ignore_unhandled;
44 	bool			print_entry;
45 };
46 
47 /*
48  * Assumptions:
49  * 1) The hotplug framework guarentees that DMAR unit will be hot-added
50  *    before IO devices managed by that unit.
51  * 2) The hotplug framework guarantees that DMAR unit will be hot-removed
52  *    after IO devices managed by that unit.
53  * 3) Hotplug events are rare.
54  *
55  * Locking rules for DMA and interrupt remapping related global data structures:
56  * 1) Use dmar_global_lock in process context
57  * 2) Use RCU in interrupt context
58  */
59 DECLARE_RWSEM(dmar_global_lock);
60 LIST_HEAD(dmar_drhd_units);
61 
62 struct acpi_table_header * __initdata dmar_tbl;
63 static int dmar_dev_scope_status = 1;
64 static unsigned long dmar_seq_ids[BITS_TO_LONGS(DMAR_UNITS_SUPPORTED)];
65 
66 static int alloc_iommu(struct dmar_drhd_unit *drhd);
67 static void free_iommu(struct intel_iommu *iommu);
68 
69 extern const struct iommu_ops intel_iommu_ops;
70 
71 static void dmar_register_drhd_unit(struct dmar_drhd_unit *drhd)
72 {
73 	/*
74 	 * add INCLUDE_ALL at the tail, so scan the list will find it at
75 	 * the very end.
76 	 */
77 	if (drhd->include_all)
78 		list_add_tail_rcu(&drhd->list, &dmar_drhd_units);
79 	else
80 		list_add_rcu(&drhd->list, &dmar_drhd_units);
81 }
82 
83 void *dmar_alloc_dev_scope(void *start, void *end, int *cnt)
84 {
85 	struct acpi_dmar_device_scope *scope;
86 
87 	*cnt = 0;
88 	while (start < end) {
89 		scope = start;
90 		if (scope->entry_type == ACPI_DMAR_SCOPE_TYPE_NAMESPACE ||
91 		    scope->entry_type == ACPI_DMAR_SCOPE_TYPE_ENDPOINT ||
92 		    scope->entry_type == ACPI_DMAR_SCOPE_TYPE_BRIDGE)
93 			(*cnt)++;
94 		else if (scope->entry_type != ACPI_DMAR_SCOPE_TYPE_IOAPIC &&
95 			scope->entry_type != ACPI_DMAR_SCOPE_TYPE_HPET) {
96 			pr_warn("Unsupported device scope\n");
97 		}
98 		start += scope->length;
99 	}
100 	if (*cnt == 0)
101 		return NULL;
102 
103 	return kcalloc(*cnt, sizeof(struct dmar_dev_scope), GFP_KERNEL);
104 }
105 
106 void dmar_free_dev_scope(struct dmar_dev_scope **devices, int *cnt)
107 {
108 	int i;
109 	struct device *tmp_dev;
110 
111 	if (*devices && *cnt) {
112 		for_each_active_dev_scope(*devices, *cnt, i, tmp_dev)
113 			put_device(tmp_dev);
114 		kfree(*devices);
115 	}
116 
117 	*devices = NULL;
118 	*cnt = 0;
119 }
120 
121 /* Optimize out kzalloc()/kfree() for normal cases */
122 static char dmar_pci_notify_info_buf[64];
123 
124 static struct dmar_pci_notify_info *
125 dmar_alloc_pci_notify_info(struct pci_dev *dev, unsigned long event)
126 {
127 	int level = 0;
128 	size_t size;
129 	struct pci_dev *tmp;
130 	struct dmar_pci_notify_info *info;
131 
132 	BUG_ON(dev->is_virtfn);
133 
134 	/*
135 	 * Ignore devices that have a domain number higher than what can
136 	 * be looked up in DMAR, e.g. VMD subdevices with domain 0x10000
137 	 */
138 	if (pci_domain_nr(dev->bus) > U16_MAX)
139 		return NULL;
140 
141 	/* Only generate path[] for device addition event */
142 	if (event == BUS_NOTIFY_ADD_DEVICE)
143 		for (tmp = dev; tmp; tmp = tmp->bus->self)
144 			level++;
145 
146 	size = struct_size(info, path, level);
147 	if (size <= sizeof(dmar_pci_notify_info_buf)) {
148 		info = (struct dmar_pci_notify_info *)dmar_pci_notify_info_buf;
149 	} else {
150 		info = kzalloc(size, GFP_KERNEL);
151 		if (!info) {
152 			pr_warn("Out of memory when allocating notify_info "
153 				"for %s.\n", pci_name(dev));
154 			if (dmar_dev_scope_status == 0)
155 				dmar_dev_scope_status = -ENOMEM;
156 			return NULL;
157 		}
158 	}
159 
160 	info->event = event;
161 	info->dev = dev;
162 	info->seg = pci_domain_nr(dev->bus);
163 	info->level = level;
164 	if (event == BUS_NOTIFY_ADD_DEVICE) {
165 		for (tmp = dev; tmp; tmp = tmp->bus->self) {
166 			level--;
167 			info->path[level].bus = tmp->bus->number;
168 			info->path[level].device = PCI_SLOT(tmp->devfn);
169 			info->path[level].function = PCI_FUNC(tmp->devfn);
170 			if (pci_is_root_bus(tmp->bus))
171 				info->bus = tmp->bus->number;
172 		}
173 	}
174 
175 	return info;
176 }
177 
178 static inline void dmar_free_pci_notify_info(struct dmar_pci_notify_info *info)
179 {
180 	if ((void *)info != dmar_pci_notify_info_buf)
181 		kfree(info);
182 }
183 
184 static bool dmar_match_pci_path(struct dmar_pci_notify_info *info, int bus,
185 				struct acpi_dmar_pci_path *path, int count)
186 {
187 	int i;
188 
189 	if (info->bus != bus)
190 		goto fallback;
191 	if (info->level != count)
192 		goto fallback;
193 
194 	for (i = 0; i < count; i++) {
195 		if (path[i].device != info->path[i].device ||
196 		    path[i].function != info->path[i].function)
197 			goto fallback;
198 	}
199 
200 	return true;
201 
202 fallback:
203 
204 	if (count != 1)
205 		return false;
206 
207 	i = info->level - 1;
208 	if (bus              == info->path[i].bus &&
209 	    path[0].device   == info->path[i].device &&
210 	    path[0].function == info->path[i].function) {
211 		pr_info(FW_BUG "RMRR entry for device %02x:%02x.%x is broken - applying workaround\n",
212 			bus, path[0].device, path[0].function);
213 		return true;
214 	}
215 
216 	return false;
217 }
218 
219 /* Return: > 0 if match found, 0 if no match found, < 0 if error happens */
220 int dmar_insert_dev_scope(struct dmar_pci_notify_info *info,
221 			  void *start, void*end, u16 segment,
222 			  struct dmar_dev_scope *devices,
223 			  int devices_cnt)
224 {
225 	int i, level;
226 	struct device *tmp, *dev = &info->dev->dev;
227 	struct acpi_dmar_device_scope *scope;
228 	struct acpi_dmar_pci_path *path;
229 
230 	if (segment != info->seg)
231 		return 0;
232 
233 	for (; start < end; start += scope->length) {
234 		scope = start;
235 		if (scope->entry_type != ACPI_DMAR_SCOPE_TYPE_ENDPOINT &&
236 		    scope->entry_type != ACPI_DMAR_SCOPE_TYPE_BRIDGE)
237 			continue;
238 
239 		path = (struct acpi_dmar_pci_path *)(scope + 1);
240 		level = (scope->length - sizeof(*scope)) / sizeof(*path);
241 		if (!dmar_match_pci_path(info, scope->bus, path, level))
242 			continue;
243 
244 		/*
245 		 * We expect devices with endpoint scope to have normal PCI
246 		 * headers, and devices with bridge scope to have bridge PCI
247 		 * headers.  However PCI NTB devices may be listed in the
248 		 * DMAR table with bridge scope, even though they have a
249 		 * normal PCI header.  NTB devices are identified by class
250 		 * "BRIDGE_OTHER" (0680h) - we don't declare a socpe mismatch
251 		 * for this special case.
252 		 */
253 		if ((scope->entry_type == ACPI_DMAR_SCOPE_TYPE_ENDPOINT &&
254 		     info->dev->hdr_type != PCI_HEADER_TYPE_NORMAL) ||
255 		    (scope->entry_type == ACPI_DMAR_SCOPE_TYPE_BRIDGE &&
256 		     (info->dev->hdr_type == PCI_HEADER_TYPE_NORMAL &&
257 		      info->dev->class >> 16 != PCI_BASE_CLASS_BRIDGE))) {
258 			pr_warn("Device scope type does not match for %s\n",
259 				pci_name(info->dev));
260 			return -EINVAL;
261 		}
262 
263 		for_each_dev_scope(devices, devices_cnt, i, tmp)
264 			if (tmp == NULL) {
265 				devices[i].bus = info->dev->bus->number;
266 				devices[i].devfn = info->dev->devfn;
267 				rcu_assign_pointer(devices[i].dev,
268 						   get_device(dev));
269 				return 1;
270 			}
271 		BUG_ON(i >= devices_cnt);
272 	}
273 
274 	return 0;
275 }
276 
277 int dmar_remove_dev_scope(struct dmar_pci_notify_info *info, u16 segment,
278 			  struct dmar_dev_scope *devices, int count)
279 {
280 	int index;
281 	struct device *tmp;
282 
283 	if (info->seg != segment)
284 		return 0;
285 
286 	for_each_active_dev_scope(devices, count, index, tmp)
287 		if (tmp == &info->dev->dev) {
288 			RCU_INIT_POINTER(devices[index].dev, NULL);
289 			synchronize_rcu();
290 			put_device(tmp);
291 			return 1;
292 		}
293 
294 	return 0;
295 }
296 
297 static int dmar_pci_bus_add_dev(struct dmar_pci_notify_info *info)
298 {
299 	int ret = 0;
300 	struct dmar_drhd_unit *dmaru;
301 	struct acpi_dmar_hardware_unit *drhd;
302 
303 	for_each_drhd_unit(dmaru) {
304 		if (dmaru->include_all)
305 			continue;
306 
307 		drhd = container_of(dmaru->hdr,
308 				    struct acpi_dmar_hardware_unit, header);
309 		ret = dmar_insert_dev_scope(info, (void *)(drhd + 1),
310 				((void *)drhd) + drhd->header.length,
311 				dmaru->segment,
312 				dmaru->devices, dmaru->devices_cnt);
313 		if (ret)
314 			break;
315 	}
316 	if (ret >= 0)
317 		ret = dmar_iommu_notify_scope_dev(info);
318 	if (ret < 0 && dmar_dev_scope_status == 0)
319 		dmar_dev_scope_status = ret;
320 
321 	if (ret >= 0)
322 		intel_irq_remap_add_device(info);
323 
324 	return ret;
325 }
326 
327 static void  dmar_pci_bus_del_dev(struct dmar_pci_notify_info *info)
328 {
329 	struct dmar_drhd_unit *dmaru;
330 
331 	for_each_drhd_unit(dmaru)
332 		if (dmar_remove_dev_scope(info, dmaru->segment,
333 			dmaru->devices, dmaru->devices_cnt))
334 			break;
335 	dmar_iommu_notify_scope_dev(info);
336 }
337 
338 static inline void vf_inherit_msi_domain(struct pci_dev *pdev)
339 {
340 	struct pci_dev *physfn = pci_physfn(pdev);
341 
342 	dev_set_msi_domain(&pdev->dev, dev_get_msi_domain(&physfn->dev));
343 }
344 
345 static int dmar_pci_bus_notifier(struct notifier_block *nb,
346 				 unsigned long action, void *data)
347 {
348 	struct pci_dev *pdev = to_pci_dev(data);
349 	struct dmar_pci_notify_info *info;
350 
351 	/* Only care about add/remove events for physical functions.
352 	 * For VFs we actually do the lookup based on the corresponding
353 	 * PF in device_to_iommu() anyway. */
354 	if (pdev->is_virtfn) {
355 		/*
356 		 * Ensure that the VF device inherits the irq domain of the
357 		 * PF device. Ideally the device would inherit the domain
358 		 * from the bus, but DMAR can have multiple units per bus
359 		 * which makes this impossible. The VF 'bus' could inherit
360 		 * from the PF device, but that's yet another x86'sism to
361 		 * inflict on everybody else.
362 		 */
363 		if (action == BUS_NOTIFY_ADD_DEVICE)
364 			vf_inherit_msi_domain(pdev);
365 		return NOTIFY_DONE;
366 	}
367 
368 	if (action != BUS_NOTIFY_ADD_DEVICE &&
369 	    action != BUS_NOTIFY_REMOVED_DEVICE)
370 		return NOTIFY_DONE;
371 
372 	info = dmar_alloc_pci_notify_info(pdev, action);
373 	if (!info)
374 		return NOTIFY_DONE;
375 
376 	down_write(&dmar_global_lock);
377 	if (action == BUS_NOTIFY_ADD_DEVICE)
378 		dmar_pci_bus_add_dev(info);
379 	else if (action == BUS_NOTIFY_REMOVED_DEVICE)
380 		dmar_pci_bus_del_dev(info);
381 	up_write(&dmar_global_lock);
382 
383 	dmar_free_pci_notify_info(info);
384 
385 	return NOTIFY_OK;
386 }
387 
388 static struct notifier_block dmar_pci_bus_nb = {
389 	.notifier_call = dmar_pci_bus_notifier,
390 	.priority = INT_MIN,
391 };
392 
393 static struct dmar_drhd_unit *
394 dmar_find_dmaru(struct acpi_dmar_hardware_unit *drhd)
395 {
396 	struct dmar_drhd_unit *dmaru;
397 
398 	list_for_each_entry_rcu(dmaru, &dmar_drhd_units, list,
399 				dmar_rcu_check())
400 		if (dmaru->segment == drhd->segment &&
401 		    dmaru->reg_base_addr == drhd->address)
402 			return dmaru;
403 
404 	return NULL;
405 }
406 
407 /*
408  * dmar_parse_one_drhd - parses exactly one DMA remapping hardware definition
409  * structure which uniquely represent one DMA remapping hardware unit
410  * present in the platform
411  */
412 static int dmar_parse_one_drhd(struct acpi_dmar_header *header, void *arg)
413 {
414 	struct acpi_dmar_hardware_unit *drhd;
415 	struct dmar_drhd_unit *dmaru;
416 	int ret;
417 
418 	drhd = (struct acpi_dmar_hardware_unit *)header;
419 	dmaru = dmar_find_dmaru(drhd);
420 	if (dmaru)
421 		goto out;
422 
423 	dmaru = kzalloc(sizeof(*dmaru) + header->length, GFP_KERNEL);
424 	if (!dmaru)
425 		return -ENOMEM;
426 
427 	/*
428 	 * If header is allocated from slab by ACPI _DSM method, we need to
429 	 * copy the content because the memory buffer will be freed on return.
430 	 */
431 	dmaru->hdr = (void *)(dmaru + 1);
432 	memcpy(dmaru->hdr, header, header->length);
433 	dmaru->reg_base_addr = drhd->address;
434 	dmaru->segment = drhd->segment;
435 	dmaru->include_all = drhd->flags & 0x1; /* BIT0: INCLUDE_ALL */
436 	dmaru->devices = dmar_alloc_dev_scope((void *)(drhd + 1),
437 					      ((void *)drhd) + drhd->header.length,
438 					      &dmaru->devices_cnt);
439 	if (dmaru->devices_cnt && dmaru->devices == NULL) {
440 		kfree(dmaru);
441 		return -ENOMEM;
442 	}
443 
444 	ret = alloc_iommu(dmaru);
445 	if (ret) {
446 		dmar_free_dev_scope(&dmaru->devices,
447 				    &dmaru->devices_cnt);
448 		kfree(dmaru);
449 		return ret;
450 	}
451 	dmar_register_drhd_unit(dmaru);
452 
453 out:
454 	if (arg)
455 		(*(int *)arg)++;
456 
457 	return 0;
458 }
459 
460 static void dmar_free_drhd(struct dmar_drhd_unit *dmaru)
461 {
462 	if (dmaru->devices && dmaru->devices_cnt)
463 		dmar_free_dev_scope(&dmaru->devices, &dmaru->devices_cnt);
464 	if (dmaru->iommu)
465 		free_iommu(dmaru->iommu);
466 	kfree(dmaru);
467 }
468 
469 static int __init dmar_parse_one_andd(struct acpi_dmar_header *header,
470 				      void *arg)
471 {
472 	struct acpi_dmar_andd *andd = (void *)header;
473 
474 	/* Check for NUL termination within the designated length */
475 	if (strnlen(andd->device_name, header->length - 8) == header->length - 8) {
476 		pr_warn(FW_BUG
477 			   "Your BIOS is broken; ANDD object name is not NUL-terminated\n"
478 			   "BIOS vendor: %s; Ver: %s; Product Version: %s\n",
479 			   dmi_get_system_info(DMI_BIOS_VENDOR),
480 			   dmi_get_system_info(DMI_BIOS_VERSION),
481 			   dmi_get_system_info(DMI_PRODUCT_VERSION));
482 		add_taint(TAINT_FIRMWARE_WORKAROUND, LOCKDEP_STILL_OK);
483 		return -EINVAL;
484 	}
485 	pr_info("ANDD device: %x name: %s\n", andd->device_number,
486 		andd->device_name);
487 
488 	return 0;
489 }
490 
491 #ifdef CONFIG_ACPI_NUMA
492 static int dmar_parse_one_rhsa(struct acpi_dmar_header *header, void *arg)
493 {
494 	struct acpi_dmar_rhsa *rhsa;
495 	struct dmar_drhd_unit *drhd;
496 
497 	rhsa = (struct acpi_dmar_rhsa *)header;
498 	for_each_drhd_unit(drhd) {
499 		if (drhd->reg_base_addr == rhsa->base_address) {
500 			int node = pxm_to_node(rhsa->proximity_domain);
501 
502 			if (!node_online(node))
503 				node = NUMA_NO_NODE;
504 			drhd->iommu->node = node;
505 			return 0;
506 		}
507 	}
508 	pr_warn(FW_BUG
509 		"Your BIOS is broken; RHSA refers to non-existent DMAR unit at %llx\n"
510 		"BIOS vendor: %s; Ver: %s; Product Version: %s\n",
511 		rhsa->base_address,
512 		dmi_get_system_info(DMI_BIOS_VENDOR),
513 		dmi_get_system_info(DMI_BIOS_VERSION),
514 		dmi_get_system_info(DMI_PRODUCT_VERSION));
515 	add_taint(TAINT_FIRMWARE_WORKAROUND, LOCKDEP_STILL_OK);
516 
517 	return 0;
518 }
519 #else
520 #define	dmar_parse_one_rhsa		dmar_res_noop
521 #endif
522 
523 static void
524 dmar_table_print_dmar_entry(struct acpi_dmar_header *header)
525 {
526 	struct acpi_dmar_hardware_unit *drhd;
527 	struct acpi_dmar_reserved_memory *rmrr;
528 	struct acpi_dmar_atsr *atsr;
529 	struct acpi_dmar_rhsa *rhsa;
530 	struct acpi_dmar_satc *satc;
531 
532 	switch (header->type) {
533 	case ACPI_DMAR_TYPE_HARDWARE_UNIT:
534 		drhd = container_of(header, struct acpi_dmar_hardware_unit,
535 				    header);
536 		pr_info("DRHD base: %#016Lx flags: %#x\n",
537 			(unsigned long long)drhd->address, drhd->flags);
538 		break;
539 	case ACPI_DMAR_TYPE_RESERVED_MEMORY:
540 		rmrr = container_of(header, struct acpi_dmar_reserved_memory,
541 				    header);
542 		pr_info("RMRR base: %#016Lx end: %#016Lx\n",
543 			(unsigned long long)rmrr->base_address,
544 			(unsigned long long)rmrr->end_address);
545 		break;
546 	case ACPI_DMAR_TYPE_ROOT_ATS:
547 		atsr = container_of(header, struct acpi_dmar_atsr, header);
548 		pr_info("ATSR flags: %#x\n", atsr->flags);
549 		break;
550 	case ACPI_DMAR_TYPE_HARDWARE_AFFINITY:
551 		rhsa = container_of(header, struct acpi_dmar_rhsa, header);
552 		pr_info("RHSA base: %#016Lx proximity domain: %#x\n",
553 		       (unsigned long long)rhsa->base_address,
554 		       rhsa->proximity_domain);
555 		break;
556 	case ACPI_DMAR_TYPE_NAMESPACE:
557 		/* We don't print this here because we need to sanity-check
558 		   it first. So print it in dmar_parse_one_andd() instead. */
559 		break;
560 	case ACPI_DMAR_TYPE_SATC:
561 		satc = container_of(header, struct acpi_dmar_satc, header);
562 		pr_info("SATC flags: 0x%x\n", satc->flags);
563 		break;
564 	}
565 }
566 
567 /**
568  * dmar_table_detect - checks to see if the platform supports DMAR devices
569  */
570 static int __init dmar_table_detect(void)
571 {
572 	acpi_status status = AE_OK;
573 
574 	/* if we could find DMAR table, then there are DMAR devices */
575 	status = acpi_get_table(ACPI_SIG_DMAR, 0, &dmar_tbl);
576 
577 	if (ACPI_SUCCESS(status) && !dmar_tbl) {
578 		pr_warn("Unable to map DMAR\n");
579 		status = AE_NOT_FOUND;
580 	}
581 
582 	return ACPI_SUCCESS(status) ? 0 : -ENOENT;
583 }
584 
585 static int dmar_walk_remapping_entries(struct acpi_dmar_header *start,
586 				       size_t len, struct dmar_res_callback *cb)
587 {
588 	struct acpi_dmar_header *iter, *next;
589 	struct acpi_dmar_header *end = ((void *)start) + len;
590 
591 	for (iter = start; iter < end; iter = next) {
592 		next = (void *)iter + iter->length;
593 		if (iter->length == 0) {
594 			/* Avoid looping forever on bad ACPI tables */
595 			pr_debug(FW_BUG "Invalid 0-length structure\n");
596 			break;
597 		} else if (next > end) {
598 			/* Avoid passing table end */
599 			pr_warn(FW_BUG "Record passes table end\n");
600 			return -EINVAL;
601 		}
602 
603 		if (cb->print_entry)
604 			dmar_table_print_dmar_entry(iter);
605 
606 		if (iter->type >= ACPI_DMAR_TYPE_RESERVED) {
607 			/* continue for forward compatibility */
608 			pr_debug("Unknown DMAR structure type %d\n",
609 				 iter->type);
610 		} else if (cb->cb[iter->type]) {
611 			int ret;
612 
613 			ret = cb->cb[iter->type](iter, cb->arg[iter->type]);
614 			if (ret)
615 				return ret;
616 		} else if (!cb->ignore_unhandled) {
617 			pr_warn("No handler for DMAR structure type %d\n",
618 				iter->type);
619 			return -EINVAL;
620 		}
621 	}
622 
623 	return 0;
624 }
625 
626 static inline int dmar_walk_dmar_table(struct acpi_table_dmar *dmar,
627 				       struct dmar_res_callback *cb)
628 {
629 	return dmar_walk_remapping_entries((void *)(dmar + 1),
630 			dmar->header.length - sizeof(*dmar), cb);
631 }
632 
633 /**
634  * parse_dmar_table - parses the DMA reporting table
635  */
636 static int __init
637 parse_dmar_table(void)
638 {
639 	struct acpi_table_dmar *dmar;
640 	int drhd_count = 0;
641 	int ret;
642 	struct dmar_res_callback cb = {
643 		.print_entry = true,
644 		.ignore_unhandled = true,
645 		.arg[ACPI_DMAR_TYPE_HARDWARE_UNIT] = &drhd_count,
646 		.cb[ACPI_DMAR_TYPE_HARDWARE_UNIT] = &dmar_parse_one_drhd,
647 		.cb[ACPI_DMAR_TYPE_RESERVED_MEMORY] = &dmar_parse_one_rmrr,
648 		.cb[ACPI_DMAR_TYPE_ROOT_ATS] = &dmar_parse_one_atsr,
649 		.cb[ACPI_DMAR_TYPE_HARDWARE_AFFINITY] = &dmar_parse_one_rhsa,
650 		.cb[ACPI_DMAR_TYPE_NAMESPACE] = &dmar_parse_one_andd,
651 		.cb[ACPI_DMAR_TYPE_SATC] = &dmar_parse_one_satc,
652 	};
653 
654 	/*
655 	 * Do it again, earlier dmar_tbl mapping could be mapped with
656 	 * fixed map.
657 	 */
658 	dmar_table_detect();
659 
660 	/*
661 	 * ACPI tables may not be DMA protected by tboot, so use DMAR copy
662 	 * SINIT saved in SinitMleData in TXT heap (which is DMA protected)
663 	 */
664 	dmar_tbl = tboot_get_dmar_table(dmar_tbl);
665 
666 	dmar = (struct acpi_table_dmar *)dmar_tbl;
667 	if (!dmar)
668 		return -ENODEV;
669 
670 	if (dmar->width < PAGE_SHIFT - 1) {
671 		pr_warn("Invalid DMAR haw\n");
672 		return -EINVAL;
673 	}
674 
675 	pr_info("Host address width %d\n", dmar->width + 1);
676 	ret = dmar_walk_dmar_table(dmar, &cb);
677 	if (ret == 0 && drhd_count == 0)
678 		pr_warn(FW_BUG "No DRHD structure found in DMAR table\n");
679 
680 	return ret;
681 }
682 
683 static int dmar_pci_device_match(struct dmar_dev_scope devices[],
684 				 int cnt, struct pci_dev *dev)
685 {
686 	int index;
687 	struct device *tmp;
688 
689 	while (dev) {
690 		for_each_active_dev_scope(devices, cnt, index, tmp)
691 			if (dev_is_pci(tmp) && dev == to_pci_dev(tmp))
692 				return 1;
693 
694 		/* Check our parent */
695 		dev = dev->bus->self;
696 	}
697 
698 	return 0;
699 }
700 
701 struct dmar_drhd_unit *
702 dmar_find_matched_drhd_unit(struct pci_dev *dev)
703 {
704 	struct dmar_drhd_unit *dmaru;
705 	struct acpi_dmar_hardware_unit *drhd;
706 
707 	dev = pci_physfn(dev);
708 
709 	rcu_read_lock();
710 	for_each_drhd_unit(dmaru) {
711 		drhd = container_of(dmaru->hdr,
712 				    struct acpi_dmar_hardware_unit,
713 				    header);
714 
715 		if (dmaru->include_all &&
716 		    drhd->segment == pci_domain_nr(dev->bus))
717 			goto out;
718 
719 		if (dmar_pci_device_match(dmaru->devices,
720 					  dmaru->devices_cnt, dev))
721 			goto out;
722 	}
723 	dmaru = NULL;
724 out:
725 	rcu_read_unlock();
726 
727 	return dmaru;
728 }
729 
730 static void __init dmar_acpi_insert_dev_scope(u8 device_number,
731 					      struct acpi_device *adev)
732 {
733 	struct dmar_drhd_unit *dmaru;
734 	struct acpi_dmar_hardware_unit *drhd;
735 	struct acpi_dmar_device_scope *scope;
736 	struct device *tmp;
737 	int i;
738 	struct acpi_dmar_pci_path *path;
739 
740 	for_each_drhd_unit(dmaru) {
741 		drhd = container_of(dmaru->hdr,
742 				    struct acpi_dmar_hardware_unit,
743 				    header);
744 
745 		for (scope = (void *)(drhd + 1);
746 		     (unsigned long)scope < ((unsigned long)drhd) + drhd->header.length;
747 		     scope = ((void *)scope) + scope->length) {
748 			if (scope->entry_type != ACPI_DMAR_SCOPE_TYPE_NAMESPACE)
749 				continue;
750 			if (scope->enumeration_id != device_number)
751 				continue;
752 
753 			path = (void *)(scope + 1);
754 			pr_info("ACPI device \"%s\" under DMAR at %llx as %02x:%02x.%d\n",
755 				dev_name(&adev->dev), dmaru->reg_base_addr,
756 				scope->bus, path->device, path->function);
757 			for_each_dev_scope(dmaru->devices, dmaru->devices_cnt, i, tmp)
758 				if (tmp == NULL) {
759 					dmaru->devices[i].bus = scope->bus;
760 					dmaru->devices[i].devfn = PCI_DEVFN(path->device,
761 									    path->function);
762 					rcu_assign_pointer(dmaru->devices[i].dev,
763 							   get_device(&adev->dev));
764 					return;
765 				}
766 			BUG_ON(i >= dmaru->devices_cnt);
767 		}
768 	}
769 	pr_warn("No IOMMU scope found for ANDD enumeration ID %d (%s)\n",
770 		device_number, dev_name(&adev->dev));
771 }
772 
773 static int __init dmar_acpi_dev_scope_init(void)
774 {
775 	struct acpi_dmar_andd *andd;
776 
777 	if (dmar_tbl == NULL)
778 		return -ENODEV;
779 
780 	for (andd = (void *)dmar_tbl + sizeof(struct acpi_table_dmar);
781 	     ((unsigned long)andd) < ((unsigned long)dmar_tbl) + dmar_tbl->length;
782 	     andd = ((void *)andd) + andd->header.length) {
783 		if (andd->header.type == ACPI_DMAR_TYPE_NAMESPACE) {
784 			acpi_handle h;
785 			struct acpi_device *adev;
786 
787 			if (!ACPI_SUCCESS(acpi_get_handle(ACPI_ROOT_OBJECT,
788 							  andd->device_name,
789 							  &h))) {
790 				pr_err("Failed to find handle for ACPI object %s\n",
791 				       andd->device_name);
792 				continue;
793 			}
794 			if (acpi_bus_get_device(h, &adev)) {
795 				pr_err("Failed to get device for ACPI object %s\n",
796 				       andd->device_name);
797 				continue;
798 			}
799 			dmar_acpi_insert_dev_scope(andd->device_number, adev);
800 		}
801 	}
802 	return 0;
803 }
804 
805 int __init dmar_dev_scope_init(void)
806 {
807 	struct pci_dev *dev = NULL;
808 	struct dmar_pci_notify_info *info;
809 
810 	if (dmar_dev_scope_status != 1)
811 		return dmar_dev_scope_status;
812 
813 	if (list_empty(&dmar_drhd_units)) {
814 		dmar_dev_scope_status = -ENODEV;
815 	} else {
816 		dmar_dev_scope_status = 0;
817 
818 		dmar_acpi_dev_scope_init();
819 
820 		for_each_pci_dev(dev) {
821 			if (dev->is_virtfn)
822 				continue;
823 
824 			info = dmar_alloc_pci_notify_info(dev,
825 					BUS_NOTIFY_ADD_DEVICE);
826 			if (!info) {
827 				return dmar_dev_scope_status;
828 			} else {
829 				dmar_pci_bus_add_dev(info);
830 				dmar_free_pci_notify_info(info);
831 			}
832 		}
833 	}
834 
835 	return dmar_dev_scope_status;
836 }
837 
838 void __init dmar_register_bus_notifier(void)
839 {
840 	bus_register_notifier(&pci_bus_type, &dmar_pci_bus_nb);
841 }
842 
843 
844 int __init dmar_table_init(void)
845 {
846 	static int dmar_table_initialized;
847 	int ret;
848 
849 	if (dmar_table_initialized == 0) {
850 		ret = parse_dmar_table();
851 		if (ret < 0) {
852 			if (ret != -ENODEV)
853 				pr_info("Parse DMAR table failure.\n");
854 		} else  if (list_empty(&dmar_drhd_units)) {
855 			pr_info("No DMAR devices found\n");
856 			ret = -ENODEV;
857 		}
858 
859 		if (ret < 0)
860 			dmar_table_initialized = ret;
861 		else
862 			dmar_table_initialized = 1;
863 	}
864 
865 	return dmar_table_initialized < 0 ? dmar_table_initialized : 0;
866 }
867 
868 static void warn_invalid_dmar(u64 addr, const char *message)
869 {
870 	pr_warn_once(FW_BUG
871 		"Your BIOS is broken; DMAR reported at address %llx%s!\n"
872 		"BIOS vendor: %s; Ver: %s; Product Version: %s\n",
873 		addr, message,
874 		dmi_get_system_info(DMI_BIOS_VENDOR),
875 		dmi_get_system_info(DMI_BIOS_VERSION),
876 		dmi_get_system_info(DMI_PRODUCT_VERSION));
877 	add_taint(TAINT_FIRMWARE_WORKAROUND, LOCKDEP_STILL_OK);
878 }
879 
880 static int __ref
881 dmar_validate_one_drhd(struct acpi_dmar_header *entry, void *arg)
882 {
883 	struct acpi_dmar_hardware_unit *drhd;
884 	void __iomem *addr;
885 	u64 cap, ecap;
886 
887 	drhd = (void *)entry;
888 	if (!drhd->address) {
889 		warn_invalid_dmar(0, "");
890 		return -EINVAL;
891 	}
892 
893 	if (arg)
894 		addr = ioremap(drhd->address, VTD_PAGE_SIZE);
895 	else
896 		addr = early_ioremap(drhd->address, VTD_PAGE_SIZE);
897 	if (!addr) {
898 		pr_warn("Can't validate DRHD address: %llx\n", drhd->address);
899 		return -EINVAL;
900 	}
901 
902 	cap = dmar_readq(addr + DMAR_CAP_REG);
903 	ecap = dmar_readq(addr + DMAR_ECAP_REG);
904 
905 	if (arg)
906 		iounmap(addr);
907 	else
908 		early_iounmap(addr, VTD_PAGE_SIZE);
909 
910 	if (cap == (uint64_t)-1 && ecap == (uint64_t)-1) {
911 		warn_invalid_dmar(drhd->address, " returns all ones");
912 		return -EINVAL;
913 	}
914 
915 	return 0;
916 }
917 
918 int __init detect_intel_iommu(void)
919 {
920 	int ret;
921 	struct dmar_res_callback validate_drhd_cb = {
922 		.cb[ACPI_DMAR_TYPE_HARDWARE_UNIT] = &dmar_validate_one_drhd,
923 		.ignore_unhandled = true,
924 	};
925 
926 	down_write(&dmar_global_lock);
927 	ret = dmar_table_detect();
928 	if (!ret)
929 		ret = dmar_walk_dmar_table((struct acpi_table_dmar *)dmar_tbl,
930 					   &validate_drhd_cb);
931 	if (!ret && !no_iommu && !iommu_detected &&
932 	    (!dmar_disabled || dmar_platform_optin())) {
933 		iommu_detected = 1;
934 		/* Make sure ACS will be enabled */
935 		pci_request_acs();
936 	}
937 
938 #ifdef CONFIG_X86
939 	if (!ret) {
940 		x86_init.iommu.iommu_init = intel_iommu_init;
941 		x86_platform.iommu_shutdown = intel_iommu_shutdown;
942 	}
943 
944 #endif
945 
946 	if (dmar_tbl) {
947 		acpi_put_table(dmar_tbl);
948 		dmar_tbl = NULL;
949 	}
950 	up_write(&dmar_global_lock);
951 
952 	return ret ? ret : 1;
953 }
954 
955 static void unmap_iommu(struct intel_iommu *iommu)
956 {
957 	iounmap(iommu->reg);
958 	release_mem_region(iommu->reg_phys, iommu->reg_size);
959 }
960 
961 /**
962  * map_iommu: map the iommu's registers
963  * @iommu: the iommu to map
964  * @phys_addr: the physical address of the base resgister
965  *
966  * Memory map the iommu's registers.  Start w/ a single page, and
967  * possibly expand if that turns out to be insufficent.
968  */
969 static int map_iommu(struct intel_iommu *iommu, u64 phys_addr)
970 {
971 	int map_size, err=0;
972 
973 	iommu->reg_phys = phys_addr;
974 	iommu->reg_size = VTD_PAGE_SIZE;
975 
976 	if (!request_mem_region(iommu->reg_phys, iommu->reg_size, iommu->name)) {
977 		pr_err("Can't reserve memory\n");
978 		err = -EBUSY;
979 		goto out;
980 	}
981 
982 	iommu->reg = ioremap(iommu->reg_phys, iommu->reg_size);
983 	if (!iommu->reg) {
984 		pr_err("Can't map the region\n");
985 		err = -ENOMEM;
986 		goto release;
987 	}
988 
989 	iommu->cap = dmar_readq(iommu->reg + DMAR_CAP_REG);
990 	iommu->ecap = dmar_readq(iommu->reg + DMAR_ECAP_REG);
991 
992 	if (iommu->cap == (uint64_t)-1 && iommu->ecap == (uint64_t)-1) {
993 		err = -EINVAL;
994 		warn_invalid_dmar(phys_addr, " returns all ones");
995 		goto unmap;
996 	}
997 	if (ecap_vcs(iommu->ecap))
998 		iommu->vccap = dmar_readq(iommu->reg + DMAR_VCCAP_REG);
999 
1000 	/* the registers might be more than one page */
1001 	map_size = max_t(int, ecap_max_iotlb_offset(iommu->ecap),
1002 			 cap_max_fault_reg_offset(iommu->cap));
1003 	map_size = VTD_PAGE_ALIGN(map_size);
1004 	if (map_size > iommu->reg_size) {
1005 		iounmap(iommu->reg);
1006 		release_mem_region(iommu->reg_phys, iommu->reg_size);
1007 		iommu->reg_size = map_size;
1008 		if (!request_mem_region(iommu->reg_phys, iommu->reg_size,
1009 					iommu->name)) {
1010 			pr_err("Can't reserve memory\n");
1011 			err = -EBUSY;
1012 			goto out;
1013 		}
1014 		iommu->reg = ioremap(iommu->reg_phys, iommu->reg_size);
1015 		if (!iommu->reg) {
1016 			pr_err("Can't map the region\n");
1017 			err = -ENOMEM;
1018 			goto release;
1019 		}
1020 	}
1021 	err = 0;
1022 	goto out;
1023 
1024 unmap:
1025 	iounmap(iommu->reg);
1026 release:
1027 	release_mem_region(iommu->reg_phys, iommu->reg_size);
1028 out:
1029 	return err;
1030 }
1031 
1032 static int dmar_alloc_seq_id(struct intel_iommu *iommu)
1033 {
1034 	iommu->seq_id = find_first_zero_bit(dmar_seq_ids,
1035 					    DMAR_UNITS_SUPPORTED);
1036 	if (iommu->seq_id >= DMAR_UNITS_SUPPORTED) {
1037 		iommu->seq_id = -1;
1038 	} else {
1039 		set_bit(iommu->seq_id, dmar_seq_ids);
1040 		sprintf(iommu->name, "dmar%d", iommu->seq_id);
1041 	}
1042 
1043 	return iommu->seq_id;
1044 }
1045 
1046 static void dmar_free_seq_id(struct intel_iommu *iommu)
1047 {
1048 	if (iommu->seq_id >= 0) {
1049 		clear_bit(iommu->seq_id, dmar_seq_ids);
1050 		iommu->seq_id = -1;
1051 	}
1052 }
1053 
1054 static int alloc_iommu(struct dmar_drhd_unit *drhd)
1055 {
1056 	struct intel_iommu *iommu;
1057 	u32 ver, sts;
1058 	int agaw = -1;
1059 	int msagaw = -1;
1060 	int err;
1061 
1062 	if (!drhd->reg_base_addr) {
1063 		warn_invalid_dmar(0, "");
1064 		return -EINVAL;
1065 	}
1066 
1067 	iommu = kzalloc(sizeof(*iommu), GFP_KERNEL);
1068 	if (!iommu)
1069 		return -ENOMEM;
1070 
1071 	if (dmar_alloc_seq_id(iommu) < 0) {
1072 		pr_err("Failed to allocate seq_id\n");
1073 		err = -ENOSPC;
1074 		goto error;
1075 	}
1076 
1077 	err = map_iommu(iommu, drhd->reg_base_addr);
1078 	if (err) {
1079 		pr_err("Failed to map %s\n", iommu->name);
1080 		goto error_free_seq_id;
1081 	}
1082 
1083 	err = -EINVAL;
1084 	if (cap_sagaw(iommu->cap) == 0) {
1085 		pr_info("%s: No supported address widths. Not attempting DMA translation.\n",
1086 			iommu->name);
1087 		drhd->ignored = 1;
1088 	}
1089 
1090 	if (!drhd->ignored) {
1091 		agaw = iommu_calculate_agaw(iommu);
1092 		if (agaw < 0) {
1093 			pr_err("Cannot get a valid agaw for iommu (seq_id = %d)\n",
1094 			       iommu->seq_id);
1095 			drhd->ignored = 1;
1096 		}
1097 	}
1098 	if (!drhd->ignored) {
1099 		msagaw = iommu_calculate_max_sagaw(iommu);
1100 		if (msagaw < 0) {
1101 			pr_err("Cannot get a valid max agaw for iommu (seq_id = %d)\n",
1102 			       iommu->seq_id);
1103 			drhd->ignored = 1;
1104 			agaw = -1;
1105 		}
1106 	}
1107 	iommu->agaw = agaw;
1108 	iommu->msagaw = msagaw;
1109 	iommu->segment = drhd->segment;
1110 
1111 	iommu->node = NUMA_NO_NODE;
1112 
1113 	ver = readl(iommu->reg + DMAR_VER_REG);
1114 	pr_info("%s: reg_base_addr %llx ver %d:%d cap %llx ecap %llx\n",
1115 		iommu->name,
1116 		(unsigned long long)drhd->reg_base_addr,
1117 		DMAR_VER_MAJOR(ver), DMAR_VER_MINOR(ver),
1118 		(unsigned long long)iommu->cap,
1119 		(unsigned long long)iommu->ecap);
1120 
1121 	/* Reflect status in gcmd */
1122 	sts = readl(iommu->reg + DMAR_GSTS_REG);
1123 	if (sts & DMA_GSTS_IRES)
1124 		iommu->gcmd |= DMA_GCMD_IRE;
1125 	if (sts & DMA_GSTS_TES)
1126 		iommu->gcmd |= DMA_GCMD_TE;
1127 	if (sts & DMA_GSTS_QIES)
1128 		iommu->gcmd |= DMA_GCMD_QIE;
1129 
1130 	raw_spin_lock_init(&iommu->register_lock);
1131 
1132 	/*
1133 	 * This is only for hotplug; at boot time intel_iommu_enabled won't
1134 	 * be set yet. When intel_iommu_init() runs, it registers the units
1135 	 * present at boot time, then sets intel_iommu_enabled.
1136 	 */
1137 	if (intel_iommu_enabled && !drhd->ignored) {
1138 		err = iommu_device_sysfs_add(&iommu->iommu, NULL,
1139 					     intel_iommu_groups,
1140 					     "%s", iommu->name);
1141 		if (err)
1142 			goto err_unmap;
1143 
1144 		err = iommu_device_register(&iommu->iommu, &intel_iommu_ops, NULL);
1145 		if (err)
1146 			goto err_sysfs;
1147 	}
1148 
1149 	drhd->iommu = iommu;
1150 	iommu->drhd = drhd;
1151 
1152 	return 0;
1153 
1154 err_sysfs:
1155 	iommu_device_sysfs_remove(&iommu->iommu);
1156 err_unmap:
1157 	unmap_iommu(iommu);
1158 error_free_seq_id:
1159 	dmar_free_seq_id(iommu);
1160 error:
1161 	kfree(iommu);
1162 	return err;
1163 }
1164 
1165 static void free_iommu(struct intel_iommu *iommu)
1166 {
1167 	if (intel_iommu_enabled && !iommu->drhd->ignored) {
1168 		iommu_device_unregister(&iommu->iommu);
1169 		iommu_device_sysfs_remove(&iommu->iommu);
1170 	}
1171 
1172 	if (iommu->irq) {
1173 		if (iommu->pr_irq) {
1174 			free_irq(iommu->pr_irq, iommu);
1175 			dmar_free_hwirq(iommu->pr_irq);
1176 			iommu->pr_irq = 0;
1177 		}
1178 		free_irq(iommu->irq, iommu);
1179 		dmar_free_hwirq(iommu->irq);
1180 		iommu->irq = 0;
1181 	}
1182 
1183 	if (iommu->qi) {
1184 		free_page((unsigned long)iommu->qi->desc);
1185 		kfree(iommu->qi->desc_status);
1186 		kfree(iommu->qi);
1187 	}
1188 
1189 	if (iommu->reg)
1190 		unmap_iommu(iommu);
1191 
1192 	dmar_free_seq_id(iommu);
1193 	kfree(iommu);
1194 }
1195 
1196 /*
1197  * Reclaim all the submitted descriptors which have completed its work.
1198  */
1199 static inline void reclaim_free_desc(struct q_inval *qi)
1200 {
1201 	while (qi->desc_status[qi->free_tail] == QI_DONE ||
1202 	       qi->desc_status[qi->free_tail] == QI_ABORT) {
1203 		qi->desc_status[qi->free_tail] = QI_FREE;
1204 		qi->free_tail = (qi->free_tail + 1) % QI_LENGTH;
1205 		qi->free_cnt++;
1206 	}
1207 }
1208 
1209 static const char *qi_type_string(u8 type)
1210 {
1211 	switch (type) {
1212 	case QI_CC_TYPE:
1213 		return "Context-cache Invalidation";
1214 	case QI_IOTLB_TYPE:
1215 		return "IOTLB Invalidation";
1216 	case QI_DIOTLB_TYPE:
1217 		return "Device-TLB Invalidation";
1218 	case QI_IEC_TYPE:
1219 		return "Interrupt Entry Cache Invalidation";
1220 	case QI_IWD_TYPE:
1221 		return "Invalidation Wait";
1222 	case QI_EIOTLB_TYPE:
1223 		return "PASID-based IOTLB Invalidation";
1224 	case QI_PC_TYPE:
1225 		return "PASID-cache Invalidation";
1226 	case QI_DEIOTLB_TYPE:
1227 		return "PASID-based Device-TLB Invalidation";
1228 	case QI_PGRP_RESP_TYPE:
1229 		return "Page Group Response";
1230 	default:
1231 		return "UNKNOWN";
1232 	}
1233 }
1234 
1235 static void qi_dump_fault(struct intel_iommu *iommu, u32 fault)
1236 {
1237 	unsigned int head = dmar_readl(iommu->reg + DMAR_IQH_REG);
1238 	u64 iqe_err = dmar_readq(iommu->reg + DMAR_IQER_REG);
1239 	struct qi_desc *desc = iommu->qi->desc + head;
1240 
1241 	if (fault & DMA_FSTS_IQE)
1242 		pr_err("VT-d detected Invalidation Queue Error: Reason %llx",
1243 		       DMAR_IQER_REG_IQEI(iqe_err));
1244 	if (fault & DMA_FSTS_ITE)
1245 		pr_err("VT-d detected Invalidation Time-out Error: SID %llx",
1246 		       DMAR_IQER_REG_ITESID(iqe_err));
1247 	if (fault & DMA_FSTS_ICE)
1248 		pr_err("VT-d detected Invalidation Completion Error: SID %llx",
1249 		       DMAR_IQER_REG_ICESID(iqe_err));
1250 
1251 	pr_err("QI HEAD: %s qw0 = 0x%llx, qw1 = 0x%llx\n",
1252 	       qi_type_string(desc->qw0 & 0xf),
1253 	       (unsigned long long)desc->qw0,
1254 	       (unsigned long long)desc->qw1);
1255 
1256 	head = ((head >> qi_shift(iommu)) + QI_LENGTH - 1) % QI_LENGTH;
1257 	head <<= qi_shift(iommu);
1258 	desc = iommu->qi->desc + head;
1259 
1260 	pr_err("QI PRIOR: %s qw0 = 0x%llx, qw1 = 0x%llx\n",
1261 	       qi_type_string(desc->qw0 & 0xf),
1262 	       (unsigned long long)desc->qw0,
1263 	       (unsigned long long)desc->qw1);
1264 }
1265 
1266 static int qi_check_fault(struct intel_iommu *iommu, int index, int wait_index)
1267 {
1268 	u32 fault;
1269 	int head, tail;
1270 	struct q_inval *qi = iommu->qi;
1271 	int shift = qi_shift(iommu);
1272 
1273 	if (qi->desc_status[wait_index] == QI_ABORT)
1274 		return -EAGAIN;
1275 
1276 	fault = readl(iommu->reg + DMAR_FSTS_REG);
1277 	if (fault & (DMA_FSTS_IQE | DMA_FSTS_ITE | DMA_FSTS_ICE))
1278 		qi_dump_fault(iommu, fault);
1279 
1280 	/*
1281 	 * If IQE happens, the head points to the descriptor associated
1282 	 * with the error. No new descriptors are fetched until the IQE
1283 	 * is cleared.
1284 	 */
1285 	if (fault & DMA_FSTS_IQE) {
1286 		head = readl(iommu->reg + DMAR_IQH_REG);
1287 		if ((head >> shift) == index) {
1288 			struct qi_desc *desc = qi->desc + head;
1289 
1290 			/*
1291 			 * desc->qw2 and desc->qw3 are either reserved or
1292 			 * used by software as private data. We won't print
1293 			 * out these two qw's for security consideration.
1294 			 */
1295 			memcpy(desc, qi->desc + (wait_index << shift),
1296 			       1 << shift);
1297 			writel(DMA_FSTS_IQE, iommu->reg + DMAR_FSTS_REG);
1298 			pr_info("Invalidation Queue Error (IQE) cleared\n");
1299 			return -EINVAL;
1300 		}
1301 	}
1302 
1303 	/*
1304 	 * If ITE happens, all pending wait_desc commands are aborted.
1305 	 * No new descriptors are fetched until the ITE is cleared.
1306 	 */
1307 	if (fault & DMA_FSTS_ITE) {
1308 		head = readl(iommu->reg + DMAR_IQH_REG);
1309 		head = ((head >> shift) - 1 + QI_LENGTH) % QI_LENGTH;
1310 		head |= 1;
1311 		tail = readl(iommu->reg + DMAR_IQT_REG);
1312 		tail = ((tail >> shift) - 1 + QI_LENGTH) % QI_LENGTH;
1313 
1314 		writel(DMA_FSTS_ITE, iommu->reg + DMAR_FSTS_REG);
1315 		pr_info("Invalidation Time-out Error (ITE) cleared\n");
1316 
1317 		do {
1318 			if (qi->desc_status[head] == QI_IN_USE)
1319 				qi->desc_status[head] = QI_ABORT;
1320 			head = (head - 2 + QI_LENGTH) % QI_LENGTH;
1321 		} while (head != tail);
1322 
1323 		if (qi->desc_status[wait_index] == QI_ABORT)
1324 			return -EAGAIN;
1325 	}
1326 
1327 	if (fault & DMA_FSTS_ICE) {
1328 		writel(DMA_FSTS_ICE, iommu->reg + DMAR_FSTS_REG);
1329 		pr_info("Invalidation Completion Error (ICE) cleared\n");
1330 	}
1331 
1332 	return 0;
1333 }
1334 
1335 /*
1336  * Function to submit invalidation descriptors of all types to the queued
1337  * invalidation interface(QI). Multiple descriptors can be submitted at a
1338  * time, a wait descriptor will be appended to each submission to ensure
1339  * hardware has completed the invalidation before return. Wait descriptors
1340  * can be part of the submission but it will not be polled for completion.
1341  */
1342 int qi_submit_sync(struct intel_iommu *iommu, struct qi_desc *desc,
1343 		   unsigned int count, unsigned long options)
1344 {
1345 	struct q_inval *qi = iommu->qi;
1346 	s64 devtlb_start_ktime = 0;
1347 	s64 iotlb_start_ktime = 0;
1348 	s64 iec_start_ktime = 0;
1349 	struct qi_desc wait_desc;
1350 	int wait_index, index;
1351 	unsigned long flags;
1352 	int offset, shift;
1353 	int rc, i;
1354 	u64 type;
1355 
1356 	if (!qi)
1357 		return 0;
1358 
1359 	type = desc->qw0 & GENMASK_ULL(3, 0);
1360 
1361 	if ((type == QI_IOTLB_TYPE || type == QI_EIOTLB_TYPE) &&
1362 	    dmar_latency_enabled(iommu, DMAR_LATENCY_INV_IOTLB))
1363 		iotlb_start_ktime = ktime_to_ns(ktime_get());
1364 
1365 	if ((type == QI_DIOTLB_TYPE || type == QI_DEIOTLB_TYPE) &&
1366 	    dmar_latency_enabled(iommu, DMAR_LATENCY_INV_DEVTLB))
1367 		devtlb_start_ktime = ktime_to_ns(ktime_get());
1368 
1369 	if (type == QI_IEC_TYPE &&
1370 	    dmar_latency_enabled(iommu, DMAR_LATENCY_INV_IEC))
1371 		iec_start_ktime = ktime_to_ns(ktime_get());
1372 
1373 restart:
1374 	rc = 0;
1375 
1376 	raw_spin_lock_irqsave(&qi->q_lock, flags);
1377 	/*
1378 	 * Check if we have enough empty slots in the queue to submit,
1379 	 * the calculation is based on:
1380 	 * # of desc + 1 wait desc + 1 space between head and tail
1381 	 */
1382 	while (qi->free_cnt < count + 2) {
1383 		raw_spin_unlock_irqrestore(&qi->q_lock, flags);
1384 		cpu_relax();
1385 		raw_spin_lock_irqsave(&qi->q_lock, flags);
1386 	}
1387 
1388 	index = qi->free_head;
1389 	wait_index = (index + count) % QI_LENGTH;
1390 	shift = qi_shift(iommu);
1391 
1392 	for (i = 0; i < count; i++) {
1393 		offset = ((index + i) % QI_LENGTH) << shift;
1394 		memcpy(qi->desc + offset, &desc[i], 1 << shift);
1395 		qi->desc_status[(index + i) % QI_LENGTH] = QI_IN_USE;
1396 		trace_qi_submit(iommu, desc[i].qw0, desc[i].qw1,
1397 				desc[i].qw2, desc[i].qw3);
1398 	}
1399 	qi->desc_status[wait_index] = QI_IN_USE;
1400 
1401 	wait_desc.qw0 = QI_IWD_STATUS_DATA(QI_DONE) |
1402 			QI_IWD_STATUS_WRITE | QI_IWD_TYPE;
1403 	if (options & QI_OPT_WAIT_DRAIN)
1404 		wait_desc.qw0 |= QI_IWD_PRQ_DRAIN;
1405 	wait_desc.qw1 = virt_to_phys(&qi->desc_status[wait_index]);
1406 	wait_desc.qw2 = 0;
1407 	wait_desc.qw3 = 0;
1408 
1409 	offset = wait_index << shift;
1410 	memcpy(qi->desc + offset, &wait_desc, 1 << shift);
1411 
1412 	qi->free_head = (qi->free_head + count + 1) % QI_LENGTH;
1413 	qi->free_cnt -= count + 1;
1414 
1415 	/*
1416 	 * update the HW tail register indicating the presence of
1417 	 * new descriptors.
1418 	 */
1419 	writel(qi->free_head << shift, iommu->reg + DMAR_IQT_REG);
1420 
1421 	while (qi->desc_status[wait_index] != QI_DONE) {
1422 		/*
1423 		 * We will leave the interrupts disabled, to prevent interrupt
1424 		 * context to queue another cmd while a cmd is already submitted
1425 		 * and waiting for completion on this cpu. This is to avoid
1426 		 * a deadlock where the interrupt context can wait indefinitely
1427 		 * for free slots in the queue.
1428 		 */
1429 		rc = qi_check_fault(iommu, index, wait_index);
1430 		if (rc)
1431 			break;
1432 
1433 		raw_spin_unlock(&qi->q_lock);
1434 		cpu_relax();
1435 		raw_spin_lock(&qi->q_lock);
1436 	}
1437 
1438 	for (i = 0; i < count; i++)
1439 		qi->desc_status[(index + i) % QI_LENGTH] = QI_DONE;
1440 
1441 	reclaim_free_desc(qi);
1442 	raw_spin_unlock_irqrestore(&qi->q_lock, flags);
1443 
1444 	if (rc == -EAGAIN)
1445 		goto restart;
1446 
1447 	if (iotlb_start_ktime)
1448 		dmar_latency_update(iommu, DMAR_LATENCY_INV_IOTLB,
1449 				ktime_to_ns(ktime_get()) - iotlb_start_ktime);
1450 
1451 	if (devtlb_start_ktime)
1452 		dmar_latency_update(iommu, DMAR_LATENCY_INV_DEVTLB,
1453 				ktime_to_ns(ktime_get()) - devtlb_start_ktime);
1454 
1455 	if (iec_start_ktime)
1456 		dmar_latency_update(iommu, DMAR_LATENCY_INV_IEC,
1457 				ktime_to_ns(ktime_get()) - iec_start_ktime);
1458 
1459 	return rc;
1460 }
1461 
1462 /*
1463  * Flush the global interrupt entry cache.
1464  */
1465 void qi_global_iec(struct intel_iommu *iommu)
1466 {
1467 	struct qi_desc desc;
1468 
1469 	desc.qw0 = QI_IEC_TYPE;
1470 	desc.qw1 = 0;
1471 	desc.qw2 = 0;
1472 	desc.qw3 = 0;
1473 
1474 	/* should never fail */
1475 	qi_submit_sync(iommu, &desc, 1, 0);
1476 }
1477 
1478 void qi_flush_context(struct intel_iommu *iommu, u16 did, u16 sid, u8 fm,
1479 		      u64 type)
1480 {
1481 	struct qi_desc desc;
1482 
1483 	desc.qw0 = QI_CC_FM(fm) | QI_CC_SID(sid) | QI_CC_DID(did)
1484 			| QI_CC_GRAN(type) | QI_CC_TYPE;
1485 	desc.qw1 = 0;
1486 	desc.qw2 = 0;
1487 	desc.qw3 = 0;
1488 
1489 	qi_submit_sync(iommu, &desc, 1, 0);
1490 }
1491 
1492 void qi_flush_iotlb(struct intel_iommu *iommu, u16 did, u64 addr,
1493 		    unsigned int size_order, u64 type)
1494 {
1495 	u8 dw = 0, dr = 0;
1496 
1497 	struct qi_desc desc;
1498 	int ih = 0;
1499 
1500 	if (cap_write_drain(iommu->cap))
1501 		dw = 1;
1502 
1503 	if (cap_read_drain(iommu->cap))
1504 		dr = 1;
1505 
1506 	desc.qw0 = QI_IOTLB_DID(did) | QI_IOTLB_DR(dr) | QI_IOTLB_DW(dw)
1507 		| QI_IOTLB_GRAN(type) | QI_IOTLB_TYPE;
1508 	desc.qw1 = QI_IOTLB_ADDR(addr) | QI_IOTLB_IH(ih)
1509 		| QI_IOTLB_AM(size_order);
1510 	desc.qw2 = 0;
1511 	desc.qw3 = 0;
1512 
1513 	qi_submit_sync(iommu, &desc, 1, 0);
1514 }
1515 
1516 void qi_flush_dev_iotlb(struct intel_iommu *iommu, u16 sid, u16 pfsid,
1517 			u16 qdep, u64 addr, unsigned mask)
1518 {
1519 	struct qi_desc desc;
1520 
1521 	if (mask) {
1522 		addr |= (1ULL << (VTD_PAGE_SHIFT + mask - 1)) - 1;
1523 		desc.qw1 = QI_DEV_IOTLB_ADDR(addr) | QI_DEV_IOTLB_SIZE;
1524 	} else
1525 		desc.qw1 = QI_DEV_IOTLB_ADDR(addr);
1526 
1527 	if (qdep >= QI_DEV_IOTLB_MAX_INVS)
1528 		qdep = 0;
1529 
1530 	desc.qw0 = QI_DEV_IOTLB_SID(sid) | QI_DEV_IOTLB_QDEP(qdep) |
1531 		   QI_DIOTLB_TYPE | QI_DEV_IOTLB_PFSID(pfsid);
1532 	desc.qw2 = 0;
1533 	desc.qw3 = 0;
1534 
1535 	qi_submit_sync(iommu, &desc, 1, 0);
1536 }
1537 
1538 /* PASID-based IOTLB invalidation */
1539 void qi_flush_piotlb(struct intel_iommu *iommu, u16 did, u32 pasid, u64 addr,
1540 		     unsigned long npages, bool ih)
1541 {
1542 	struct qi_desc desc = {.qw2 = 0, .qw3 = 0};
1543 
1544 	/*
1545 	 * npages == -1 means a PASID-selective invalidation, otherwise,
1546 	 * a positive value for Page-selective-within-PASID invalidation.
1547 	 * 0 is not a valid input.
1548 	 */
1549 	if (WARN_ON(!npages)) {
1550 		pr_err("Invalid input npages = %ld\n", npages);
1551 		return;
1552 	}
1553 
1554 	if (npages == -1) {
1555 		desc.qw0 = QI_EIOTLB_PASID(pasid) |
1556 				QI_EIOTLB_DID(did) |
1557 				QI_EIOTLB_GRAN(QI_GRAN_NONG_PASID) |
1558 				QI_EIOTLB_TYPE;
1559 		desc.qw1 = 0;
1560 	} else {
1561 		int mask = ilog2(__roundup_pow_of_two(npages));
1562 		unsigned long align = (1ULL << (VTD_PAGE_SHIFT + mask));
1563 
1564 		if (WARN_ON_ONCE(!IS_ALIGNED(addr, align)))
1565 			addr = ALIGN_DOWN(addr, align);
1566 
1567 		desc.qw0 = QI_EIOTLB_PASID(pasid) |
1568 				QI_EIOTLB_DID(did) |
1569 				QI_EIOTLB_GRAN(QI_GRAN_PSI_PASID) |
1570 				QI_EIOTLB_TYPE;
1571 		desc.qw1 = QI_EIOTLB_ADDR(addr) |
1572 				QI_EIOTLB_IH(ih) |
1573 				QI_EIOTLB_AM(mask);
1574 	}
1575 
1576 	qi_submit_sync(iommu, &desc, 1, 0);
1577 }
1578 
1579 /* PASID-based device IOTLB Invalidate */
1580 void qi_flush_dev_iotlb_pasid(struct intel_iommu *iommu, u16 sid, u16 pfsid,
1581 			      u32 pasid,  u16 qdep, u64 addr, unsigned int size_order)
1582 {
1583 	unsigned long mask = 1UL << (VTD_PAGE_SHIFT + size_order - 1);
1584 	struct qi_desc desc = {.qw1 = 0, .qw2 = 0, .qw3 = 0};
1585 
1586 	desc.qw0 = QI_DEV_EIOTLB_PASID(pasid) | QI_DEV_EIOTLB_SID(sid) |
1587 		QI_DEV_EIOTLB_QDEP(qdep) | QI_DEIOTLB_TYPE |
1588 		QI_DEV_IOTLB_PFSID(pfsid);
1589 
1590 	/*
1591 	 * If S bit is 0, we only flush a single page. If S bit is set,
1592 	 * The least significant zero bit indicates the invalidation address
1593 	 * range. VT-d spec 6.5.2.6.
1594 	 * e.g. address bit 12[0] indicates 8KB, 13[0] indicates 16KB.
1595 	 * size order = 0 is PAGE_SIZE 4KB
1596 	 * Max Invs Pending (MIP) is set to 0 for now until we have DIT in
1597 	 * ECAP.
1598 	 */
1599 	if (!IS_ALIGNED(addr, VTD_PAGE_SIZE << size_order))
1600 		pr_warn_ratelimited("Invalidate non-aligned address %llx, order %d\n",
1601 				    addr, size_order);
1602 
1603 	/* Take page address */
1604 	desc.qw1 = QI_DEV_EIOTLB_ADDR(addr);
1605 
1606 	if (size_order) {
1607 		/*
1608 		 * Existing 0s in address below size_order may be the least
1609 		 * significant bit, we must set them to 1s to avoid having
1610 		 * smaller size than desired.
1611 		 */
1612 		desc.qw1 |= GENMASK_ULL(size_order + VTD_PAGE_SHIFT - 1,
1613 					VTD_PAGE_SHIFT);
1614 		/* Clear size_order bit to indicate size */
1615 		desc.qw1 &= ~mask;
1616 		/* Set the S bit to indicate flushing more than 1 page */
1617 		desc.qw1 |= QI_DEV_EIOTLB_SIZE;
1618 	}
1619 
1620 	qi_submit_sync(iommu, &desc, 1, 0);
1621 }
1622 
1623 void qi_flush_pasid_cache(struct intel_iommu *iommu, u16 did,
1624 			  u64 granu, u32 pasid)
1625 {
1626 	struct qi_desc desc = {.qw1 = 0, .qw2 = 0, .qw3 = 0};
1627 
1628 	desc.qw0 = QI_PC_PASID(pasid) | QI_PC_DID(did) |
1629 			QI_PC_GRAN(granu) | QI_PC_TYPE;
1630 	qi_submit_sync(iommu, &desc, 1, 0);
1631 }
1632 
1633 /*
1634  * Disable Queued Invalidation interface.
1635  */
1636 void dmar_disable_qi(struct intel_iommu *iommu)
1637 {
1638 	unsigned long flags;
1639 	u32 sts;
1640 	cycles_t start_time = get_cycles();
1641 
1642 	if (!ecap_qis(iommu->ecap))
1643 		return;
1644 
1645 	raw_spin_lock_irqsave(&iommu->register_lock, flags);
1646 
1647 	sts =  readl(iommu->reg + DMAR_GSTS_REG);
1648 	if (!(sts & DMA_GSTS_QIES))
1649 		goto end;
1650 
1651 	/*
1652 	 * Give a chance to HW to complete the pending invalidation requests.
1653 	 */
1654 	while ((readl(iommu->reg + DMAR_IQT_REG) !=
1655 		readl(iommu->reg + DMAR_IQH_REG)) &&
1656 		(DMAR_OPERATION_TIMEOUT > (get_cycles() - start_time)))
1657 		cpu_relax();
1658 
1659 	iommu->gcmd &= ~DMA_GCMD_QIE;
1660 	writel(iommu->gcmd, iommu->reg + DMAR_GCMD_REG);
1661 
1662 	IOMMU_WAIT_OP(iommu, DMAR_GSTS_REG, readl,
1663 		      !(sts & DMA_GSTS_QIES), sts);
1664 end:
1665 	raw_spin_unlock_irqrestore(&iommu->register_lock, flags);
1666 }
1667 
1668 /*
1669  * Enable queued invalidation.
1670  */
1671 static void __dmar_enable_qi(struct intel_iommu *iommu)
1672 {
1673 	u32 sts;
1674 	unsigned long flags;
1675 	struct q_inval *qi = iommu->qi;
1676 	u64 val = virt_to_phys(qi->desc);
1677 
1678 	qi->free_head = qi->free_tail = 0;
1679 	qi->free_cnt = QI_LENGTH;
1680 
1681 	/*
1682 	 * Set DW=1 and QS=1 in IQA_REG when Scalable Mode capability
1683 	 * is present.
1684 	 */
1685 	if (ecap_smts(iommu->ecap))
1686 		val |= (1 << 11) | 1;
1687 
1688 	raw_spin_lock_irqsave(&iommu->register_lock, flags);
1689 
1690 	/* write zero to the tail reg */
1691 	writel(0, iommu->reg + DMAR_IQT_REG);
1692 
1693 	dmar_writeq(iommu->reg + DMAR_IQA_REG, val);
1694 
1695 	iommu->gcmd |= DMA_GCMD_QIE;
1696 	writel(iommu->gcmd, iommu->reg + DMAR_GCMD_REG);
1697 
1698 	/* Make sure hardware complete it */
1699 	IOMMU_WAIT_OP(iommu, DMAR_GSTS_REG, readl, (sts & DMA_GSTS_QIES), sts);
1700 
1701 	raw_spin_unlock_irqrestore(&iommu->register_lock, flags);
1702 }
1703 
1704 /*
1705  * Enable Queued Invalidation interface. This is a must to support
1706  * interrupt-remapping. Also used by DMA-remapping, which replaces
1707  * register based IOTLB invalidation.
1708  */
1709 int dmar_enable_qi(struct intel_iommu *iommu)
1710 {
1711 	struct q_inval *qi;
1712 	struct page *desc_page;
1713 
1714 	if (!ecap_qis(iommu->ecap))
1715 		return -ENOENT;
1716 
1717 	/*
1718 	 * queued invalidation is already setup and enabled.
1719 	 */
1720 	if (iommu->qi)
1721 		return 0;
1722 
1723 	iommu->qi = kmalloc(sizeof(*qi), GFP_ATOMIC);
1724 	if (!iommu->qi)
1725 		return -ENOMEM;
1726 
1727 	qi = iommu->qi;
1728 
1729 	/*
1730 	 * Need two pages to accommodate 256 descriptors of 256 bits each
1731 	 * if the remapping hardware supports scalable mode translation.
1732 	 */
1733 	desc_page = alloc_pages_node(iommu->node, GFP_ATOMIC | __GFP_ZERO,
1734 				     !!ecap_smts(iommu->ecap));
1735 	if (!desc_page) {
1736 		kfree(qi);
1737 		iommu->qi = NULL;
1738 		return -ENOMEM;
1739 	}
1740 
1741 	qi->desc = page_address(desc_page);
1742 
1743 	qi->desc_status = kcalloc(QI_LENGTH, sizeof(int), GFP_ATOMIC);
1744 	if (!qi->desc_status) {
1745 		free_page((unsigned long) qi->desc);
1746 		kfree(qi);
1747 		iommu->qi = NULL;
1748 		return -ENOMEM;
1749 	}
1750 
1751 	raw_spin_lock_init(&qi->q_lock);
1752 
1753 	__dmar_enable_qi(iommu);
1754 
1755 	return 0;
1756 }
1757 
1758 /* iommu interrupt handling. Most stuff are MSI-like. */
1759 
1760 enum faulttype {
1761 	DMA_REMAP,
1762 	INTR_REMAP,
1763 	UNKNOWN,
1764 };
1765 
1766 static const char *dma_remap_fault_reasons[] =
1767 {
1768 	"Software",
1769 	"Present bit in root entry is clear",
1770 	"Present bit in context entry is clear",
1771 	"Invalid context entry",
1772 	"Access beyond MGAW",
1773 	"PTE Write access is not set",
1774 	"PTE Read access is not set",
1775 	"Next page table ptr is invalid",
1776 	"Root table address invalid",
1777 	"Context table ptr is invalid",
1778 	"non-zero reserved fields in RTP",
1779 	"non-zero reserved fields in CTP",
1780 	"non-zero reserved fields in PTE",
1781 	"PCE for translation request specifies blocking",
1782 };
1783 
1784 static const char * const dma_remap_sm_fault_reasons[] = {
1785 	"SM: Invalid Root Table Address",
1786 	"SM: TTM 0 for request with PASID",
1787 	"SM: TTM 0 for page group request",
1788 	"Unknown", "Unknown", "Unknown", "Unknown", "Unknown", /* 0x33-0x37 */
1789 	"SM: Error attempting to access Root Entry",
1790 	"SM: Present bit in Root Entry is clear",
1791 	"SM: Non-zero reserved field set in Root Entry",
1792 	"Unknown", "Unknown", "Unknown", "Unknown", "Unknown", /* 0x3B-0x3F */
1793 	"SM: Error attempting to access Context Entry",
1794 	"SM: Present bit in Context Entry is clear",
1795 	"SM: Non-zero reserved field set in the Context Entry",
1796 	"SM: Invalid Context Entry",
1797 	"SM: DTE field in Context Entry is clear",
1798 	"SM: PASID Enable field in Context Entry is clear",
1799 	"SM: PASID is larger than the max in Context Entry",
1800 	"SM: PRE field in Context-Entry is clear",
1801 	"SM: RID_PASID field error in Context-Entry",
1802 	"Unknown", "Unknown", "Unknown", "Unknown", "Unknown", "Unknown", "Unknown", /* 0x49-0x4F */
1803 	"SM: Error attempting to access the PASID Directory Entry",
1804 	"SM: Present bit in Directory Entry is clear",
1805 	"SM: Non-zero reserved field set in PASID Directory Entry",
1806 	"Unknown", "Unknown", "Unknown", "Unknown", "Unknown", /* 0x53-0x57 */
1807 	"SM: Error attempting to access PASID Table Entry",
1808 	"SM: Present bit in PASID Table Entry is clear",
1809 	"SM: Non-zero reserved field set in PASID Table Entry",
1810 	"SM: Invalid Scalable-Mode PASID Table Entry",
1811 	"SM: ERE field is clear in PASID Table Entry",
1812 	"SM: SRE field is clear in PASID Table Entry",
1813 	"Unknown", "Unknown",/* 0x5E-0x5F */
1814 	"Unknown", "Unknown", "Unknown", "Unknown", "Unknown", "Unknown", "Unknown", "Unknown", /* 0x60-0x67 */
1815 	"Unknown", "Unknown", "Unknown", "Unknown", "Unknown", "Unknown", "Unknown", "Unknown", /* 0x68-0x6F */
1816 	"SM: Error attempting to access first-level paging entry",
1817 	"SM: Present bit in first-level paging entry is clear",
1818 	"SM: Non-zero reserved field set in first-level paging entry",
1819 	"SM: Error attempting to access FL-PML4 entry",
1820 	"SM: First-level entry address beyond MGAW in Nested translation",
1821 	"SM: Read permission error in FL-PML4 entry in Nested translation",
1822 	"SM: Read permission error in first-level paging entry in Nested translation",
1823 	"SM: Write permission error in first-level paging entry in Nested translation",
1824 	"SM: Error attempting to access second-level paging entry",
1825 	"SM: Read/Write permission error in second-level paging entry",
1826 	"SM: Non-zero reserved field set in second-level paging entry",
1827 	"SM: Invalid second-level page table pointer",
1828 	"SM: A/D bit update needed in second-level entry when set up in no snoop",
1829 	"Unknown", "Unknown", "Unknown", /* 0x7D-0x7F */
1830 	"SM: Address in first-level translation is not canonical",
1831 	"SM: U/S set 0 for first-level translation with user privilege",
1832 	"SM: No execute permission for request with PASID and ER=1",
1833 	"SM: Address beyond the DMA hardware max",
1834 	"SM: Second-level entry address beyond the max",
1835 	"SM: No write permission for Write/AtomicOp request",
1836 	"SM: No read permission for Read/AtomicOp request",
1837 	"SM: Invalid address-interrupt address",
1838 	"Unknown", "Unknown", "Unknown", "Unknown", "Unknown", "Unknown", "Unknown", "Unknown", /* 0x88-0x8F */
1839 	"SM: A/D bit update needed in first-level entry when set up in no snoop",
1840 };
1841 
1842 static const char *irq_remap_fault_reasons[] =
1843 {
1844 	"Detected reserved fields in the decoded interrupt-remapped request",
1845 	"Interrupt index exceeded the interrupt-remapping table size",
1846 	"Present field in the IRTE entry is clear",
1847 	"Error accessing interrupt-remapping table pointed by IRTA_REG",
1848 	"Detected reserved fields in the IRTE entry",
1849 	"Blocked a compatibility format interrupt request",
1850 	"Blocked an interrupt request due to source-id verification failure",
1851 };
1852 
1853 static const char *dmar_get_fault_reason(u8 fault_reason, int *fault_type)
1854 {
1855 	if (fault_reason >= 0x20 && (fault_reason - 0x20 <
1856 					ARRAY_SIZE(irq_remap_fault_reasons))) {
1857 		*fault_type = INTR_REMAP;
1858 		return irq_remap_fault_reasons[fault_reason - 0x20];
1859 	} else if (fault_reason >= 0x30 && (fault_reason - 0x30 <
1860 			ARRAY_SIZE(dma_remap_sm_fault_reasons))) {
1861 		*fault_type = DMA_REMAP;
1862 		return dma_remap_sm_fault_reasons[fault_reason - 0x30];
1863 	} else if (fault_reason < ARRAY_SIZE(dma_remap_fault_reasons)) {
1864 		*fault_type = DMA_REMAP;
1865 		return dma_remap_fault_reasons[fault_reason];
1866 	} else {
1867 		*fault_type = UNKNOWN;
1868 		return "Unknown";
1869 	}
1870 }
1871 
1872 
1873 static inline int dmar_msi_reg(struct intel_iommu *iommu, int irq)
1874 {
1875 	if (iommu->irq == irq)
1876 		return DMAR_FECTL_REG;
1877 	else if (iommu->pr_irq == irq)
1878 		return DMAR_PECTL_REG;
1879 	else
1880 		BUG();
1881 }
1882 
1883 void dmar_msi_unmask(struct irq_data *data)
1884 {
1885 	struct intel_iommu *iommu = irq_data_get_irq_handler_data(data);
1886 	int reg = dmar_msi_reg(iommu, data->irq);
1887 	unsigned long flag;
1888 
1889 	/* unmask it */
1890 	raw_spin_lock_irqsave(&iommu->register_lock, flag);
1891 	writel(0, iommu->reg + reg);
1892 	/* Read a reg to force flush the post write */
1893 	readl(iommu->reg + reg);
1894 	raw_spin_unlock_irqrestore(&iommu->register_lock, flag);
1895 }
1896 
1897 void dmar_msi_mask(struct irq_data *data)
1898 {
1899 	struct intel_iommu *iommu = irq_data_get_irq_handler_data(data);
1900 	int reg = dmar_msi_reg(iommu, data->irq);
1901 	unsigned long flag;
1902 
1903 	/* mask it */
1904 	raw_spin_lock_irqsave(&iommu->register_lock, flag);
1905 	writel(DMA_FECTL_IM, iommu->reg + reg);
1906 	/* Read a reg to force flush the post write */
1907 	readl(iommu->reg + reg);
1908 	raw_spin_unlock_irqrestore(&iommu->register_lock, flag);
1909 }
1910 
1911 void dmar_msi_write(int irq, struct msi_msg *msg)
1912 {
1913 	struct intel_iommu *iommu = irq_get_handler_data(irq);
1914 	int reg = dmar_msi_reg(iommu, irq);
1915 	unsigned long flag;
1916 
1917 	raw_spin_lock_irqsave(&iommu->register_lock, flag);
1918 	writel(msg->data, iommu->reg + reg + 4);
1919 	writel(msg->address_lo, iommu->reg + reg + 8);
1920 	writel(msg->address_hi, iommu->reg + reg + 12);
1921 	raw_spin_unlock_irqrestore(&iommu->register_lock, flag);
1922 }
1923 
1924 void dmar_msi_read(int irq, struct msi_msg *msg)
1925 {
1926 	struct intel_iommu *iommu = irq_get_handler_data(irq);
1927 	int reg = dmar_msi_reg(iommu, irq);
1928 	unsigned long flag;
1929 
1930 	raw_spin_lock_irqsave(&iommu->register_lock, flag);
1931 	msg->data = readl(iommu->reg + reg + 4);
1932 	msg->address_lo = readl(iommu->reg + reg + 8);
1933 	msg->address_hi = readl(iommu->reg + reg + 12);
1934 	raw_spin_unlock_irqrestore(&iommu->register_lock, flag);
1935 }
1936 
1937 static int dmar_fault_do_one(struct intel_iommu *iommu, int type,
1938 		u8 fault_reason, u32 pasid, u16 source_id,
1939 		unsigned long long addr)
1940 {
1941 	const char *reason;
1942 	int fault_type;
1943 
1944 	reason = dmar_get_fault_reason(fault_reason, &fault_type);
1945 
1946 	if (fault_type == INTR_REMAP)
1947 		pr_err("[INTR-REMAP] Request device [0x%02x:0x%02x.%d] fault index 0x%llx [fault reason 0x%02x] %s\n",
1948 		       source_id >> 8, PCI_SLOT(source_id & 0xFF),
1949 		       PCI_FUNC(source_id & 0xFF), addr >> 48,
1950 		       fault_reason, reason);
1951 	else if (pasid == INVALID_IOASID)
1952 		pr_err("[%s NO_PASID] Request device [0x%02x:0x%02x.%d] fault addr 0x%llx [fault reason 0x%02x] %s\n",
1953 		       type ? "DMA Read" : "DMA Write",
1954 		       source_id >> 8, PCI_SLOT(source_id & 0xFF),
1955 		       PCI_FUNC(source_id & 0xFF), addr,
1956 		       fault_reason, reason);
1957 	else
1958 		pr_err("[%s PASID 0x%x] Request device [0x%02x:0x%02x.%d] fault addr 0x%llx [fault reason 0x%02x] %s\n",
1959 		       type ? "DMA Read" : "DMA Write", pasid,
1960 		       source_id >> 8, PCI_SLOT(source_id & 0xFF),
1961 		       PCI_FUNC(source_id & 0xFF), addr,
1962 		       fault_reason, reason);
1963 
1964 	return 0;
1965 }
1966 
1967 #define PRIMARY_FAULT_REG_LEN (16)
1968 irqreturn_t dmar_fault(int irq, void *dev_id)
1969 {
1970 	struct intel_iommu *iommu = dev_id;
1971 	int reg, fault_index;
1972 	u32 fault_status;
1973 	unsigned long flag;
1974 	static DEFINE_RATELIMIT_STATE(rs,
1975 				      DEFAULT_RATELIMIT_INTERVAL,
1976 				      DEFAULT_RATELIMIT_BURST);
1977 
1978 	raw_spin_lock_irqsave(&iommu->register_lock, flag);
1979 	fault_status = readl(iommu->reg + DMAR_FSTS_REG);
1980 	if (fault_status && __ratelimit(&rs))
1981 		pr_err("DRHD: handling fault status reg %x\n", fault_status);
1982 
1983 	/* TBD: ignore advanced fault log currently */
1984 	if (!(fault_status & DMA_FSTS_PPF))
1985 		goto unlock_exit;
1986 
1987 	fault_index = dma_fsts_fault_record_index(fault_status);
1988 	reg = cap_fault_reg_offset(iommu->cap);
1989 	while (1) {
1990 		/* Disable printing, simply clear the fault when ratelimited */
1991 		bool ratelimited = !__ratelimit(&rs);
1992 		u8 fault_reason;
1993 		u16 source_id;
1994 		u64 guest_addr;
1995 		u32 pasid;
1996 		int type;
1997 		u32 data;
1998 		bool pasid_present;
1999 
2000 		/* highest 32 bits */
2001 		data = readl(iommu->reg + reg +
2002 				fault_index * PRIMARY_FAULT_REG_LEN + 12);
2003 		if (!(data & DMA_FRCD_F))
2004 			break;
2005 
2006 		if (!ratelimited) {
2007 			fault_reason = dma_frcd_fault_reason(data);
2008 			type = dma_frcd_type(data);
2009 
2010 			pasid = dma_frcd_pasid_value(data);
2011 			data = readl(iommu->reg + reg +
2012 				     fault_index * PRIMARY_FAULT_REG_LEN + 8);
2013 			source_id = dma_frcd_source_id(data);
2014 
2015 			pasid_present = dma_frcd_pasid_present(data);
2016 			guest_addr = dmar_readq(iommu->reg + reg +
2017 					fault_index * PRIMARY_FAULT_REG_LEN);
2018 			guest_addr = dma_frcd_page_addr(guest_addr);
2019 		}
2020 
2021 		/* clear the fault */
2022 		writel(DMA_FRCD_F, iommu->reg + reg +
2023 			fault_index * PRIMARY_FAULT_REG_LEN + 12);
2024 
2025 		raw_spin_unlock_irqrestore(&iommu->register_lock, flag);
2026 
2027 		if (!ratelimited)
2028 			/* Using pasid -1 if pasid is not present */
2029 			dmar_fault_do_one(iommu, type, fault_reason,
2030 					  pasid_present ? pasid : INVALID_IOASID,
2031 					  source_id, guest_addr);
2032 
2033 		fault_index++;
2034 		if (fault_index >= cap_num_fault_regs(iommu->cap))
2035 			fault_index = 0;
2036 		raw_spin_lock_irqsave(&iommu->register_lock, flag);
2037 	}
2038 
2039 	writel(DMA_FSTS_PFO | DMA_FSTS_PPF | DMA_FSTS_PRO,
2040 	       iommu->reg + DMAR_FSTS_REG);
2041 
2042 unlock_exit:
2043 	raw_spin_unlock_irqrestore(&iommu->register_lock, flag);
2044 	return IRQ_HANDLED;
2045 }
2046 
2047 int dmar_set_interrupt(struct intel_iommu *iommu)
2048 {
2049 	int irq, ret;
2050 
2051 	/*
2052 	 * Check if the fault interrupt is already initialized.
2053 	 */
2054 	if (iommu->irq)
2055 		return 0;
2056 
2057 	irq = dmar_alloc_hwirq(iommu->seq_id, iommu->node, iommu);
2058 	if (irq > 0) {
2059 		iommu->irq = irq;
2060 	} else {
2061 		pr_err("No free IRQ vectors\n");
2062 		return -EINVAL;
2063 	}
2064 
2065 	ret = request_irq(irq, dmar_fault, IRQF_NO_THREAD, iommu->name, iommu);
2066 	if (ret)
2067 		pr_err("Can't request irq\n");
2068 	return ret;
2069 }
2070 
2071 int __init enable_drhd_fault_handling(void)
2072 {
2073 	struct dmar_drhd_unit *drhd;
2074 	struct intel_iommu *iommu;
2075 
2076 	/*
2077 	 * Enable fault control interrupt.
2078 	 */
2079 	for_each_iommu(iommu, drhd) {
2080 		u32 fault_status;
2081 		int ret = dmar_set_interrupt(iommu);
2082 
2083 		if (ret) {
2084 			pr_err("DRHD %Lx: failed to enable fault, interrupt, ret %d\n",
2085 			       (unsigned long long)drhd->reg_base_addr, ret);
2086 			return -1;
2087 		}
2088 
2089 		/*
2090 		 * Clear any previous faults.
2091 		 */
2092 		dmar_fault(iommu->irq, iommu);
2093 		fault_status = readl(iommu->reg + DMAR_FSTS_REG);
2094 		writel(fault_status, iommu->reg + DMAR_FSTS_REG);
2095 	}
2096 
2097 	return 0;
2098 }
2099 
2100 /*
2101  * Re-enable Queued Invalidation interface.
2102  */
2103 int dmar_reenable_qi(struct intel_iommu *iommu)
2104 {
2105 	if (!ecap_qis(iommu->ecap))
2106 		return -ENOENT;
2107 
2108 	if (!iommu->qi)
2109 		return -ENOENT;
2110 
2111 	/*
2112 	 * First disable queued invalidation.
2113 	 */
2114 	dmar_disable_qi(iommu);
2115 	/*
2116 	 * Then enable queued invalidation again. Since there is no pending
2117 	 * invalidation requests now, it's safe to re-enable queued
2118 	 * invalidation.
2119 	 */
2120 	__dmar_enable_qi(iommu);
2121 
2122 	return 0;
2123 }
2124 
2125 /*
2126  * Check interrupt remapping support in DMAR table description.
2127  */
2128 int __init dmar_ir_support(void)
2129 {
2130 	struct acpi_table_dmar *dmar;
2131 	dmar = (struct acpi_table_dmar *)dmar_tbl;
2132 	if (!dmar)
2133 		return 0;
2134 	return dmar->flags & 0x1;
2135 }
2136 
2137 /* Check whether DMAR units are in use */
2138 static inline bool dmar_in_use(void)
2139 {
2140 	return irq_remapping_enabled || intel_iommu_enabled;
2141 }
2142 
2143 static int __init dmar_free_unused_resources(void)
2144 {
2145 	struct dmar_drhd_unit *dmaru, *dmaru_n;
2146 
2147 	if (dmar_in_use())
2148 		return 0;
2149 
2150 	if (dmar_dev_scope_status != 1 && !list_empty(&dmar_drhd_units))
2151 		bus_unregister_notifier(&pci_bus_type, &dmar_pci_bus_nb);
2152 
2153 	down_write(&dmar_global_lock);
2154 	list_for_each_entry_safe(dmaru, dmaru_n, &dmar_drhd_units, list) {
2155 		list_del(&dmaru->list);
2156 		dmar_free_drhd(dmaru);
2157 	}
2158 	up_write(&dmar_global_lock);
2159 
2160 	return 0;
2161 }
2162 
2163 late_initcall(dmar_free_unused_resources);
2164 IOMMU_INIT_POST(detect_intel_iommu);
2165 
2166 /*
2167  * DMAR Hotplug Support
2168  * For more details, please refer to Intel(R) Virtualization Technology
2169  * for Directed-IO Architecture Specifiction, Rev 2.2, Section 8.8
2170  * "Remapping Hardware Unit Hot Plug".
2171  */
2172 static guid_t dmar_hp_guid =
2173 	GUID_INIT(0xD8C1A3A6, 0xBE9B, 0x4C9B,
2174 		  0x91, 0xBF, 0xC3, 0xCB, 0x81, 0xFC, 0x5D, 0xAF);
2175 
2176 /*
2177  * Currently there's only one revision and BIOS will not check the revision id,
2178  * so use 0 for safety.
2179  */
2180 #define	DMAR_DSM_REV_ID			0
2181 #define	DMAR_DSM_FUNC_DRHD		1
2182 #define	DMAR_DSM_FUNC_ATSR		2
2183 #define	DMAR_DSM_FUNC_RHSA		3
2184 #define	DMAR_DSM_FUNC_SATC		4
2185 
2186 static inline bool dmar_detect_dsm(acpi_handle handle, int func)
2187 {
2188 	return acpi_check_dsm(handle, &dmar_hp_guid, DMAR_DSM_REV_ID, 1 << func);
2189 }
2190 
2191 static int dmar_walk_dsm_resource(acpi_handle handle, int func,
2192 				  dmar_res_handler_t handler, void *arg)
2193 {
2194 	int ret = -ENODEV;
2195 	union acpi_object *obj;
2196 	struct acpi_dmar_header *start;
2197 	struct dmar_res_callback callback;
2198 	static int res_type[] = {
2199 		[DMAR_DSM_FUNC_DRHD] = ACPI_DMAR_TYPE_HARDWARE_UNIT,
2200 		[DMAR_DSM_FUNC_ATSR] = ACPI_DMAR_TYPE_ROOT_ATS,
2201 		[DMAR_DSM_FUNC_RHSA] = ACPI_DMAR_TYPE_HARDWARE_AFFINITY,
2202 		[DMAR_DSM_FUNC_SATC] = ACPI_DMAR_TYPE_SATC,
2203 	};
2204 
2205 	if (!dmar_detect_dsm(handle, func))
2206 		return 0;
2207 
2208 	obj = acpi_evaluate_dsm_typed(handle, &dmar_hp_guid, DMAR_DSM_REV_ID,
2209 				      func, NULL, ACPI_TYPE_BUFFER);
2210 	if (!obj)
2211 		return -ENODEV;
2212 
2213 	memset(&callback, 0, sizeof(callback));
2214 	callback.cb[res_type[func]] = handler;
2215 	callback.arg[res_type[func]] = arg;
2216 	start = (struct acpi_dmar_header *)obj->buffer.pointer;
2217 	ret = dmar_walk_remapping_entries(start, obj->buffer.length, &callback);
2218 
2219 	ACPI_FREE(obj);
2220 
2221 	return ret;
2222 }
2223 
2224 static int dmar_hp_add_drhd(struct acpi_dmar_header *header, void *arg)
2225 {
2226 	int ret;
2227 	struct dmar_drhd_unit *dmaru;
2228 
2229 	dmaru = dmar_find_dmaru((struct acpi_dmar_hardware_unit *)header);
2230 	if (!dmaru)
2231 		return -ENODEV;
2232 
2233 	ret = dmar_ir_hotplug(dmaru, true);
2234 	if (ret == 0)
2235 		ret = dmar_iommu_hotplug(dmaru, true);
2236 
2237 	return ret;
2238 }
2239 
2240 static int dmar_hp_remove_drhd(struct acpi_dmar_header *header, void *arg)
2241 {
2242 	int i, ret;
2243 	struct device *dev;
2244 	struct dmar_drhd_unit *dmaru;
2245 
2246 	dmaru = dmar_find_dmaru((struct acpi_dmar_hardware_unit *)header);
2247 	if (!dmaru)
2248 		return 0;
2249 
2250 	/*
2251 	 * All PCI devices managed by this unit should have been destroyed.
2252 	 */
2253 	if (!dmaru->include_all && dmaru->devices && dmaru->devices_cnt) {
2254 		for_each_active_dev_scope(dmaru->devices,
2255 					  dmaru->devices_cnt, i, dev)
2256 			return -EBUSY;
2257 	}
2258 
2259 	ret = dmar_ir_hotplug(dmaru, false);
2260 	if (ret == 0)
2261 		ret = dmar_iommu_hotplug(dmaru, false);
2262 
2263 	return ret;
2264 }
2265 
2266 static int dmar_hp_release_drhd(struct acpi_dmar_header *header, void *arg)
2267 {
2268 	struct dmar_drhd_unit *dmaru;
2269 
2270 	dmaru = dmar_find_dmaru((struct acpi_dmar_hardware_unit *)header);
2271 	if (dmaru) {
2272 		list_del_rcu(&dmaru->list);
2273 		synchronize_rcu();
2274 		dmar_free_drhd(dmaru);
2275 	}
2276 
2277 	return 0;
2278 }
2279 
2280 static int dmar_hotplug_insert(acpi_handle handle)
2281 {
2282 	int ret;
2283 	int drhd_count = 0;
2284 
2285 	ret = dmar_walk_dsm_resource(handle, DMAR_DSM_FUNC_DRHD,
2286 				     &dmar_validate_one_drhd, (void *)1);
2287 	if (ret)
2288 		goto out;
2289 
2290 	ret = dmar_walk_dsm_resource(handle, DMAR_DSM_FUNC_DRHD,
2291 				     &dmar_parse_one_drhd, (void *)&drhd_count);
2292 	if (ret == 0 && drhd_count == 0) {
2293 		pr_warn(FW_BUG "No DRHD structures in buffer returned by _DSM method\n");
2294 		goto out;
2295 	} else if (ret) {
2296 		goto release_drhd;
2297 	}
2298 
2299 	ret = dmar_walk_dsm_resource(handle, DMAR_DSM_FUNC_RHSA,
2300 				     &dmar_parse_one_rhsa, NULL);
2301 	if (ret)
2302 		goto release_drhd;
2303 
2304 	ret = dmar_walk_dsm_resource(handle, DMAR_DSM_FUNC_ATSR,
2305 				     &dmar_parse_one_atsr, NULL);
2306 	if (ret)
2307 		goto release_atsr;
2308 
2309 	ret = dmar_walk_dsm_resource(handle, DMAR_DSM_FUNC_DRHD,
2310 				     &dmar_hp_add_drhd, NULL);
2311 	if (!ret)
2312 		return 0;
2313 
2314 	dmar_walk_dsm_resource(handle, DMAR_DSM_FUNC_DRHD,
2315 			       &dmar_hp_remove_drhd, NULL);
2316 release_atsr:
2317 	dmar_walk_dsm_resource(handle, DMAR_DSM_FUNC_ATSR,
2318 			       &dmar_release_one_atsr, NULL);
2319 release_drhd:
2320 	dmar_walk_dsm_resource(handle, DMAR_DSM_FUNC_DRHD,
2321 			       &dmar_hp_release_drhd, NULL);
2322 out:
2323 	return ret;
2324 }
2325 
2326 static int dmar_hotplug_remove(acpi_handle handle)
2327 {
2328 	int ret;
2329 
2330 	ret = dmar_walk_dsm_resource(handle, DMAR_DSM_FUNC_ATSR,
2331 				     &dmar_check_one_atsr, NULL);
2332 	if (ret)
2333 		return ret;
2334 
2335 	ret = dmar_walk_dsm_resource(handle, DMAR_DSM_FUNC_DRHD,
2336 				     &dmar_hp_remove_drhd, NULL);
2337 	if (ret == 0) {
2338 		WARN_ON(dmar_walk_dsm_resource(handle, DMAR_DSM_FUNC_ATSR,
2339 					       &dmar_release_one_atsr, NULL));
2340 		WARN_ON(dmar_walk_dsm_resource(handle, DMAR_DSM_FUNC_DRHD,
2341 					       &dmar_hp_release_drhd, NULL));
2342 	} else {
2343 		dmar_walk_dsm_resource(handle, DMAR_DSM_FUNC_DRHD,
2344 				       &dmar_hp_add_drhd, NULL);
2345 	}
2346 
2347 	return ret;
2348 }
2349 
2350 static acpi_status dmar_get_dsm_handle(acpi_handle handle, u32 lvl,
2351 				       void *context, void **retval)
2352 {
2353 	acpi_handle *phdl = retval;
2354 
2355 	if (dmar_detect_dsm(handle, DMAR_DSM_FUNC_DRHD)) {
2356 		*phdl = handle;
2357 		return AE_CTRL_TERMINATE;
2358 	}
2359 
2360 	return AE_OK;
2361 }
2362 
2363 static int dmar_device_hotplug(acpi_handle handle, bool insert)
2364 {
2365 	int ret;
2366 	acpi_handle tmp = NULL;
2367 	acpi_status status;
2368 
2369 	if (!dmar_in_use())
2370 		return 0;
2371 
2372 	if (dmar_detect_dsm(handle, DMAR_DSM_FUNC_DRHD)) {
2373 		tmp = handle;
2374 	} else {
2375 		status = acpi_walk_namespace(ACPI_TYPE_DEVICE, handle,
2376 					     ACPI_UINT32_MAX,
2377 					     dmar_get_dsm_handle,
2378 					     NULL, NULL, &tmp);
2379 		if (ACPI_FAILURE(status)) {
2380 			pr_warn("Failed to locate _DSM method.\n");
2381 			return -ENXIO;
2382 		}
2383 	}
2384 	if (tmp == NULL)
2385 		return 0;
2386 
2387 	down_write(&dmar_global_lock);
2388 	if (insert)
2389 		ret = dmar_hotplug_insert(tmp);
2390 	else
2391 		ret = dmar_hotplug_remove(tmp);
2392 	up_write(&dmar_global_lock);
2393 
2394 	return ret;
2395 }
2396 
2397 int dmar_device_add(acpi_handle handle)
2398 {
2399 	return dmar_device_hotplug(handle, true);
2400 }
2401 
2402 int dmar_device_remove(acpi_handle handle)
2403 {
2404 	return dmar_device_hotplug(handle, false);
2405 }
2406 
2407 /*
2408  * dmar_platform_optin - Is %DMA_CTRL_PLATFORM_OPT_IN_FLAG set in DMAR table
2409  *
2410  * Returns true if the platform has %DMA_CTRL_PLATFORM_OPT_IN_FLAG set in
2411  * the ACPI DMAR table. This means that the platform boot firmware has made
2412  * sure no device can issue DMA outside of RMRR regions.
2413  */
2414 bool dmar_platform_optin(void)
2415 {
2416 	struct acpi_table_dmar *dmar;
2417 	acpi_status status;
2418 	bool ret;
2419 
2420 	status = acpi_get_table(ACPI_SIG_DMAR, 0,
2421 				(struct acpi_table_header **)&dmar);
2422 	if (ACPI_FAILURE(status))
2423 		return false;
2424 
2425 	ret = !!(dmar->flags & DMAR_PLATFORM_OPT_IN);
2426 	acpi_put_table((struct acpi_table_header *)dmar);
2427 
2428 	return ret;
2429 }
2430 EXPORT_SYMBOL_GPL(dmar_platform_optin);
2431