Staging: netwave: delete the driver
[linux/fpc-iii.git] / drivers / pci / dmar.c
blob33ead97f0c4b24e2e030ae913bef26cdf435d533
1 /*
2 * Copyright (c) 2006, Intel Corporation.
4 * This program is free software; you can redistribute it and/or modify it
5 * under the terms and conditions of the GNU General Public License,
6 * version 2, as published by the Free Software Foundation.
8 * This program is distributed in the hope it will be useful, but WITHOUT
9 * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
10 * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for
11 * more details.
13 * You should have received a copy of the GNU General Public License along with
14 * this program; if not, write to the Free Software Foundation, Inc., 59 Temple
15 * Place - Suite 330, Boston, MA 02111-1307 USA.
17 * Copyright (C) 2006-2008 Intel Corporation
18 * Author: Ashok Raj <ashok.raj@intel.com>
19 * Author: Shaohua Li <shaohua.li@intel.com>
20 * Author: Anil S Keshavamurthy <anil.s.keshavamurthy@intel.com>
22 * This file implements early detection/parsing of Remapping Devices
23 * reported to OS through BIOS via DMA remapping reporting (DMAR) ACPI
24 * tables.
26 * These routines are used by both DMA-remapping and Interrupt-remapping
29 #include <linux/pci.h>
30 #include <linux/dmar.h>
31 #include <linux/iova.h>
32 #include <linux/intel-iommu.h>
33 #include <linux/timer.h>
34 #include <linux/irq.h>
35 #include <linux/interrupt.h>
36 #include <linux/tboot.h>
37 #include <linux/dmi.h>
38 #include <linux/slab.h>
40 #define PREFIX "DMAR: "
42 /* No locks are needed as DMA remapping hardware unit
43 * list is constructed at boot time and hotplug of
44 * these units are not supported by the architecture.
46 LIST_HEAD(dmar_drhd_units);
48 static struct acpi_table_header * __initdata dmar_tbl;
49 static acpi_size dmar_tbl_size;
51 static void __init dmar_register_drhd_unit(struct dmar_drhd_unit *drhd)
54 * add INCLUDE_ALL at the tail, so scan the list will find it at
55 * the very end.
57 if (drhd->include_all)
58 list_add_tail(&drhd->list, &dmar_drhd_units);
59 else
60 list_add(&drhd->list, &dmar_drhd_units);
63 static int __init dmar_parse_one_dev_scope(struct acpi_dmar_device_scope *scope,
64 struct pci_dev **dev, u16 segment)
66 struct pci_bus *bus;
67 struct pci_dev *pdev = NULL;
68 struct acpi_dmar_pci_path *path;
69 int count;
71 bus = pci_find_bus(segment, scope->bus);
72 path = (struct acpi_dmar_pci_path *)(scope + 1);
73 count = (scope->length - sizeof(struct acpi_dmar_device_scope))
74 / sizeof(struct acpi_dmar_pci_path);
76 while (count) {
77 if (pdev)
78 pci_dev_put(pdev);
80 * Some BIOSes list non-exist devices in DMAR table, just
81 * ignore it
83 if (!bus) {
84 printk(KERN_WARNING
85 PREFIX "Device scope bus [%d] not found\n",
86 scope->bus);
87 break;
89 pdev = pci_get_slot(bus, PCI_DEVFN(path->dev, path->fn));
90 if (!pdev) {
91 printk(KERN_WARNING PREFIX
92 "Device scope device [%04x:%02x:%02x.%02x] not found\n",
93 segment, bus->number, path->dev, path->fn);
94 break;
96 path ++;
97 count --;
98 bus = pdev->subordinate;
100 if (!pdev) {
101 printk(KERN_WARNING PREFIX
102 "Device scope device [%04x:%02x:%02x.%02x] not found\n",
103 segment, scope->bus, path->dev, path->fn);
104 *dev = NULL;
105 return 0;
107 if ((scope->entry_type == ACPI_DMAR_SCOPE_TYPE_ENDPOINT && \
108 pdev->subordinate) || (scope->entry_type == \
109 ACPI_DMAR_SCOPE_TYPE_BRIDGE && !pdev->subordinate)) {
110 pci_dev_put(pdev);
111 printk(KERN_WARNING PREFIX
112 "Device scope type does not match for %s\n",
113 pci_name(pdev));
114 return -EINVAL;
116 *dev = pdev;
117 return 0;
120 static int __init dmar_parse_dev_scope(void *start, void *end, int *cnt,
121 struct pci_dev ***devices, u16 segment)
123 struct acpi_dmar_device_scope *scope;
124 void * tmp = start;
125 int index;
126 int ret;
128 *cnt = 0;
129 while (start < end) {
130 scope = start;
131 if (scope->entry_type == ACPI_DMAR_SCOPE_TYPE_ENDPOINT ||
132 scope->entry_type == ACPI_DMAR_SCOPE_TYPE_BRIDGE)
133 (*cnt)++;
134 else
135 printk(KERN_WARNING PREFIX
136 "Unsupported device scope\n");
137 start += scope->length;
139 if (*cnt == 0)
140 return 0;
142 *devices = kcalloc(*cnt, sizeof(struct pci_dev *), GFP_KERNEL);
143 if (!*devices)
144 return -ENOMEM;
146 start = tmp;
147 index = 0;
148 while (start < end) {
149 scope = start;
150 if (scope->entry_type == ACPI_DMAR_SCOPE_TYPE_ENDPOINT ||
151 scope->entry_type == ACPI_DMAR_SCOPE_TYPE_BRIDGE) {
152 ret = dmar_parse_one_dev_scope(scope,
153 &(*devices)[index], segment);
154 if (ret) {
155 kfree(*devices);
156 return ret;
158 index ++;
160 start += scope->length;
163 return 0;
167 * dmar_parse_one_drhd - parses exactly one DMA remapping hardware definition
168 * structure which uniquely represent one DMA remapping hardware unit
169 * present in the platform
171 static int __init
172 dmar_parse_one_drhd(struct acpi_dmar_header *header)
174 struct acpi_dmar_hardware_unit *drhd;
175 struct dmar_drhd_unit *dmaru;
176 int ret = 0;
178 drhd = (struct acpi_dmar_hardware_unit *)header;
179 dmaru = kzalloc(sizeof(*dmaru), GFP_KERNEL);
180 if (!dmaru)
181 return -ENOMEM;
183 dmaru->hdr = header;
184 dmaru->reg_base_addr = drhd->address;
185 dmaru->segment = drhd->segment;
186 dmaru->include_all = drhd->flags & 0x1; /* BIT0: INCLUDE_ALL */
188 ret = alloc_iommu(dmaru);
189 if (ret) {
190 kfree(dmaru);
191 return ret;
193 dmar_register_drhd_unit(dmaru);
194 return 0;
197 static int __init dmar_parse_dev(struct dmar_drhd_unit *dmaru)
199 struct acpi_dmar_hardware_unit *drhd;
200 int ret = 0;
202 drhd = (struct acpi_dmar_hardware_unit *) dmaru->hdr;
204 if (dmaru->include_all)
205 return 0;
207 ret = dmar_parse_dev_scope((void *)(drhd + 1),
208 ((void *)drhd) + drhd->header.length,
209 &dmaru->devices_cnt, &dmaru->devices,
210 drhd->segment);
211 if (ret) {
212 list_del(&dmaru->list);
213 kfree(dmaru);
215 return ret;
218 #ifdef CONFIG_DMAR
219 LIST_HEAD(dmar_rmrr_units);
221 static void __init dmar_register_rmrr_unit(struct dmar_rmrr_unit *rmrr)
223 list_add(&rmrr->list, &dmar_rmrr_units);
227 static int __init
228 dmar_parse_one_rmrr(struct acpi_dmar_header *header)
230 struct acpi_dmar_reserved_memory *rmrr;
231 struct dmar_rmrr_unit *rmrru;
233 rmrru = kzalloc(sizeof(*rmrru), GFP_KERNEL);
234 if (!rmrru)
235 return -ENOMEM;
237 rmrru->hdr = header;
238 rmrr = (struct acpi_dmar_reserved_memory *)header;
239 rmrru->base_address = rmrr->base_address;
240 rmrru->end_address = rmrr->end_address;
242 dmar_register_rmrr_unit(rmrru);
243 return 0;
246 static int __init
247 rmrr_parse_dev(struct dmar_rmrr_unit *rmrru)
249 struct acpi_dmar_reserved_memory *rmrr;
250 int ret;
252 rmrr = (struct acpi_dmar_reserved_memory *) rmrru->hdr;
253 ret = dmar_parse_dev_scope((void *)(rmrr + 1),
254 ((void *)rmrr) + rmrr->header.length,
255 &rmrru->devices_cnt, &rmrru->devices, rmrr->segment);
257 if (ret || (rmrru->devices_cnt == 0)) {
258 list_del(&rmrru->list);
259 kfree(rmrru);
261 return ret;
264 static LIST_HEAD(dmar_atsr_units);
266 static int __init dmar_parse_one_atsr(struct acpi_dmar_header *hdr)
268 struct acpi_dmar_atsr *atsr;
269 struct dmar_atsr_unit *atsru;
271 atsr = container_of(hdr, struct acpi_dmar_atsr, header);
272 atsru = kzalloc(sizeof(*atsru), GFP_KERNEL);
273 if (!atsru)
274 return -ENOMEM;
276 atsru->hdr = hdr;
277 atsru->include_all = atsr->flags & 0x1;
279 list_add(&atsru->list, &dmar_atsr_units);
281 return 0;
284 static int __init atsr_parse_dev(struct dmar_atsr_unit *atsru)
286 int rc;
287 struct acpi_dmar_atsr *atsr;
289 if (atsru->include_all)
290 return 0;
292 atsr = container_of(atsru->hdr, struct acpi_dmar_atsr, header);
293 rc = dmar_parse_dev_scope((void *)(atsr + 1),
294 (void *)atsr + atsr->header.length,
295 &atsru->devices_cnt, &atsru->devices,
296 atsr->segment);
297 if (rc || !atsru->devices_cnt) {
298 list_del(&atsru->list);
299 kfree(atsru);
302 return rc;
305 int dmar_find_matched_atsr_unit(struct pci_dev *dev)
307 int i;
308 struct pci_bus *bus;
309 struct acpi_dmar_atsr *atsr;
310 struct dmar_atsr_unit *atsru;
312 list_for_each_entry(atsru, &dmar_atsr_units, list) {
313 atsr = container_of(atsru->hdr, struct acpi_dmar_atsr, header);
314 if (atsr->segment == pci_domain_nr(dev->bus))
315 goto found;
318 return 0;
320 found:
321 for (bus = dev->bus; bus; bus = bus->parent) {
322 struct pci_dev *bridge = bus->self;
324 if (!bridge || !pci_is_pcie(bridge) ||
325 bridge->pcie_type == PCI_EXP_TYPE_PCI_BRIDGE)
326 return 0;
328 if (bridge->pcie_type == PCI_EXP_TYPE_ROOT_PORT) {
329 for (i = 0; i < atsru->devices_cnt; i++)
330 if (atsru->devices[i] == bridge)
331 return 1;
332 break;
336 if (atsru->include_all)
337 return 1;
339 return 0;
341 #endif
343 #ifdef CONFIG_ACPI_NUMA
344 static int __init
345 dmar_parse_one_rhsa(struct acpi_dmar_header *header)
347 struct acpi_dmar_rhsa *rhsa;
348 struct dmar_drhd_unit *drhd;
350 rhsa = (struct acpi_dmar_rhsa *)header;
351 for_each_drhd_unit(drhd) {
352 if (drhd->reg_base_addr == rhsa->base_address) {
353 int node = acpi_map_pxm_to_node(rhsa->proximity_domain);
355 if (!node_online(node))
356 node = -1;
357 drhd->iommu->node = node;
358 return 0;
361 WARN(1, "Your BIOS is broken; RHSA refers to non-existent DMAR unit at %llx\n"
362 "BIOS vendor: %s; Ver: %s; Product Version: %s\n",
363 drhd->reg_base_addr,
364 dmi_get_system_info(DMI_BIOS_VENDOR),
365 dmi_get_system_info(DMI_BIOS_VERSION),
366 dmi_get_system_info(DMI_PRODUCT_VERSION));
368 return 0;
370 #endif
372 static void __init
373 dmar_table_print_dmar_entry(struct acpi_dmar_header *header)
375 struct acpi_dmar_hardware_unit *drhd;
376 struct acpi_dmar_reserved_memory *rmrr;
377 struct acpi_dmar_atsr *atsr;
378 struct acpi_dmar_rhsa *rhsa;
380 switch (header->type) {
381 case ACPI_DMAR_TYPE_HARDWARE_UNIT:
382 drhd = container_of(header, struct acpi_dmar_hardware_unit,
383 header);
384 printk (KERN_INFO PREFIX
385 "DRHD base: %#016Lx flags: %#x\n",
386 (unsigned long long)drhd->address, drhd->flags);
387 break;
388 case ACPI_DMAR_TYPE_RESERVED_MEMORY:
389 rmrr = container_of(header, struct acpi_dmar_reserved_memory,
390 header);
391 printk (KERN_INFO PREFIX
392 "RMRR base: %#016Lx end: %#016Lx\n",
393 (unsigned long long)rmrr->base_address,
394 (unsigned long long)rmrr->end_address);
395 break;
396 case ACPI_DMAR_TYPE_ATSR:
397 atsr = container_of(header, struct acpi_dmar_atsr, header);
398 printk(KERN_INFO PREFIX "ATSR flags: %#x\n", atsr->flags);
399 break;
400 case ACPI_DMAR_HARDWARE_AFFINITY:
401 rhsa = container_of(header, struct acpi_dmar_rhsa, header);
402 printk(KERN_INFO PREFIX "RHSA base: %#016Lx proximity domain: %#x\n",
403 (unsigned long long)rhsa->base_address,
404 rhsa->proximity_domain);
405 break;
410 * dmar_table_detect - checks to see if the platform supports DMAR devices
412 static int __init dmar_table_detect(void)
414 acpi_status status = AE_OK;
416 /* if we could find DMAR table, then there are DMAR devices */
417 status = acpi_get_table_with_size(ACPI_SIG_DMAR, 0,
418 (struct acpi_table_header **)&dmar_tbl,
419 &dmar_tbl_size);
421 if (ACPI_SUCCESS(status) && !dmar_tbl) {
422 printk (KERN_WARNING PREFIX "Unable to map DMAR\n");
423 status = AE_NOT_FOUND;
426 return (ACPI_SUCCESS(status) ? 1 : 0);
430 * parse_dmar_table - parses the DMA reporting table
432 static int __init
433 parse_dmar_table(void)
435 struct acpi_table_dmar *dmar;
436 struct acpi_dmar_header *entry_header;
437 int ret = 0;
440 * Do it again, earlier dmar_tbl mapping could be mapped with
441 * fixed map.
443 dmar_table_detect();
446 * ACPI tables may not be DMA protected by tboot, so use DMAR copy
447 * SINIT saved in SinitMleData in TXT heap (which is DMA protected)
449 dmar_tbl = tboot_get_dmar_table(dmar_tbl);
451 dmar = (struct acpi_table_dmar *)dmar_tbl;
452 if (!dmar)
453 return -ENODEV;
455 if (dmar->width < PAGE_SHIFT - 1) {
456 printk(KERN_WARNING PREFIX "Invalid DMAR haw\n");
457 return -EINVAL;
460 printk (KERN_INFO PREFIX "Host address width %d\n",
461 dmar->width + 1);
463 entry_header = (struct acpi_dmar_header *)(dmar + 1);
464 while (((unsigned long)entry_header) <
465 (((unsigned long)dmar) + dmar_tbl->length)) {
466 /* Avoid looping forever on bad ACPI tables */
467 if (entry_header->length == 0) {
468 printk(KERN_WARNING PREFIX
469 "Invalid 0-length structure\n");
470 ret = -EINVAL;
471 break;
474 dmar_table_print_dmar_entry(entry_header);
476 switch (entry_header->type) {
477 case ACPI_DMAR_TYPE_HARDWARE_UNIT:
478 ret = dmar_parse_one_drhd(entry_header);
479 break;
480 case ACPI_DMAR_TYPE_RESERVED_MEMORY:
481 #ifdef CONFIG_DMAR
482 ret = dmar_parse_one_rmrr(entry_header);
483 #endif
484 break;
485 case ACPI_DMAR_TYPE_ATSR:
486 #ifdef CONFIG_DMAR
487 ret = dmar_parse_one_atsr(entry_header);
488 #endif
489 break;
490 case ACPI_DMAR_HARDWARE_AFFINITY:
491 #ifdef CONFIG_ACPI_NUMA
492 ret = dmar_parse_one_rhsa(entry_header);
493 #endif
494 break;
495 default:
496 printk(KERN_WARNING PREFIX
497 "Unknown DMAR structure type %d\n",
498 entry_header->type);
499 ret = 0; /* for forward compatibility */
500 break;
502 if (ret)
503 break;
505 entry_header = ((void *)entry_header + entry_header->length);
507 return ret;
510 int dmar_pci_device_match(struct pci_dev *devices[], int cnt,
511 struct pci_dev *dev)
513 int index;
515 while (dev) {
516 for (index = 0; index < cnt; index++)
517 if (dev == devices[index])
518 return 1;
520 /* Check our parent */
521 dev = dev->bus->self;
524 return 0;
527 struct dmar_drhd_unit *
528 dmar_find_matched_drhd_unit(struct pci_dev *dev)
530 struct dmar_drhd_unit *dmaru = NULL;
531 struct acpi_dmar_hardware_unit *drhd;
533 list_for_each_entry(dmaru, &dmar_drhd_units, list) {
534 drhd = container_of(dmaru->hdr,
535 struct acpi_dmar_hardware_unit,
536 header);
538 if (dmaru->include_all &&
539 drhd->segment == pci_domain_nr(dev->bus))
540 return dmaru;
542 if (dmar_pci_device_match(dmaru->devices,
543 dmaru->devices_cnt, dev))
544 return dmaru;
547 return NULL;
550 int __init dmar_dev_scope_init(void)
552 struct dmar_drhd_unit *drhd, *drhd_n;
553 int ret = -ENODEV;
555 list_for_each_entry_safe(drhd, drhd_n, &dmar_drhd_units, list) {
556 ret = dmar_parse_dev(drhd);
557 if (ret)
558 return ret;
561 #ifdef CONFIG_DMAR
563 struct dmar_rmrr_unit *rmrr, *rmrr_n;
564 struct dmar_atsr_unit *atsr, *atsr_n;
566 list_for_each_entry_safe(rmrr, rmrr_n, &dmar_rmrr_units, list) {
567 ret = rmrr_parse_dev(rmrr);
568 if (ret)
569 return ret;
572 list_for_each_entry_safe(atsr, atsr_n, &dmar_atsr_units, list) {
573 ret = atsr_parse_dev(atsr);
574 if (ret)
575 return ret;
578 #endif
580 return ret;
584 int __init dmar_table_init(void)
586 static int dmar_table_initialized;
587 int ret;
589 if (dmar_table_initialized)
590 return 0;
592 dmar_table_initialized = 1;
594 ret = parse_dmar_table();
595 if (ret) {
596 if (ret != -ENODEV)
597 printk(KERN_INFO PREFIX "parse DMAR table failure.\n");
598 return ret;
601 if (list_empty(&dmar_drhd_units)) {
602 printk(KERN_INFO PREFIX "No DMAR devices found\n");
603 return -ENODEV;
606 #ifdef CONFIG_DMAR
607 if (list_empty(&dmar_rmrr_units))
608 printk(KERN_INFO PREFIX "No RMRR found\n");
610 if (list_empty(&dmar_atsr_units))
611 printk(KERN_INFO PREFIX "No ATSR found\n");
612 #endif
614 return 0;
617 static int bios_warned;
619 int __init check_zero_address(void)
621 struct acpi_table_dmar *dmar;
622 struct acpi_dmar_header *entry_header;
623 struct acpi_dmar_hardware_unit *drhd;
625 dmar = (struct acpi_table_dmar *)dmar_tbl;
626 entry_header = (struct acpi_dmar_header *)(dmar + 1);
628 while (((unsigned long)entry_header) <
629 (((unsigned long)dmar) + dmar_tbl->length)) {
630 /* Avoid looping forever on bad ACPI tables */
631 if (entry_header->length == 0) {
632 printk(KERN_WARNING PREFIX
633 "Invalid 0-length structure\n");
634 return 0;
637 if (entry_header->type == ACPI_DMAR_TYPE_HARDWARE_UNIT) {
638 void __iomem *addr;
639 u64 cap, ecap;
641 drhd = (void *)entry_header;
642 if (!drhd->address) {
643 /* Promote an attitude of violence to a BIOS engineer today */
644 WARN(1, "Your BIOS is broken; DMAR reported at address zero!\n"
645 "BIOS vendor: %s; Ver: %s; Product Version: %s\n",
646 dmi_get_system_info(DMI_BIOS_VENDOR),
647 dmi_get_system_info(DMI_BIOS_VERSION),
648 dmi_get_system_info(DMI_PRODUCT_VERSION));
649 bios_warned = 1;
650 goto failed;
653 addr = early_ioremap(drhd->address, VTD_PAGE_SIZE);
654 if (!addr ) {
655 printk("IOMMU: can't validate: %llx\n", drhd->address);
656 goto failed;
658 cap = dmar_readq(addr + DMAR_CAP_REG);
659 ecap = dmar_readq(addr + DMAR_ECAP_REG);
660 early_iounmap(addr, VTD_PAGE_SIZE);
661 if (cap == (uint64_t)-1 && ecap == (uint64_t)-1) {
662 /* Promote an attitude of violence to a BIOS engineer today */
663 WARN(1, "Your BIOS is broken; DMAR reported at address %llx returns all ones!\n"
664 "BIOS vendor: %s; Ver: %s; Product Version: %s\n",
665 drhd->address,
666 dmi_get_system_info(DMI_BIOS_VENDOR),
667 dmi_get_system_info(DMI_BIOS_VERSION),
668 dmi_get_system_info(DMI_PRODUCT_VERSION));
669 bios_warned = 1;
670 goto failed;
674 entry_header = ((void *)entry_header + entry_header->length);
676 return 1;
678 failed:
679 #ifdef CONFIG_DMAR
680 dmar_disabled = 1;
681 #endif
682 return 0;
685 void __init detect_intel_iommu(void)
687 int ret;
689 ret = dmar_table_detect();
690 if (ret)
691 ret = check_zero_address();
693 #ifdef CONFIG_INTR_REMAP
694 struct acpi_table_dmar *dmar;
696 * for now we will disable dma-remapping when interrupt
697 * remapping is enabled.
698 * When support for queued invalidation for IOTLB invalidation
699 * is added, we will not need this any more.
701 dmar = (struct acpi_table_dmar *) dmar_tbl;
702 if (ret && cpu_has_x2apic && dmar->flags & 0x1)
703 printk(KERN_INFO
704 "Queued invalidation will be enabled to support "
705 "x2apic and Intr-remapping.\n");
706 #endif
707 #ifdef CONFIG_DMAR
708 if (ret && !no_iommu && !iommu_detected && !dmar_disabled) {
709 iommu_detected = 1;
710 /* Make sure ACS will be enabled */
711 pci_request_acs();
713 #endif
714 #ifdef CONFIG_X86
715 if (ret)
716 x86_init.iommu.iommu_init = intel_iommu_init;
717 #endif
719 early_acpi_os_unmap_memory(dmar_tbl, dmar_tbl_size);
720 dmar_tbl = NULL;
724 int alloc_iommu(struct dmar_drhd_unit *drhd)
726 struct intel_iommu *iommu;
727 int map_size;
728 u32 ver;
729 static int iommu_allocated = 0;
730 int agaw = 0;
731 int msagaw = 0;
733 if (!drhd->reg_base_addr) {
734 if (!bios_warned) {
735 WARN(1, "Your BIOS is broken; DMAR reported at address zero!\n"
736 "BIOS vendor: %s; Ver: %s; Product Version: %s\n",
737 dmi_get_system_info(DMI_BIOS_VENDOR),
738 dmi_get_system_info(DMI_BIOS_VERSION),
739 dmi_get_system_info(DMI_PRODUCT_VERSION));
740 bios_warned = 1;
742 return -EINVAL;
745 iommu = kzalloc(sizeof(*iommu), GFP_KERNEL);
746 if (!iommu)
747 return -ENOMEM;
749 iommu->seq_id = iommu_allocated++;
750 sprintf (iommu->name, "dmar%d", iommu->seq_id);
752 iommu->reg = ioremap(drhd->reg_base_addr, VTD_PAGE_SIZE);
753 if (!iommu->reg) {
754 printk(KERN_ERR "IOMMU: can't map the region\n");
755 goto error;
757 iommu->cap = dmar_readq(iommu->reg + DMAR_CAP_REG);
758 iommu->ecap = dmar_readq(iommu->reg + DMAR_ECAP_REG);
760 if (iommu->cap == (uint64_t)-1 && iommu->ecap == (uint64_t)-1) {
761 if (!bios_warned) {
762 /* Promote an attitude of violence to a BIOS engineer today */
763 WARN(1, "Your BIOS is broken; DMAR reported at address %llx returns all ones!\n"
764 "BIOS vendor: %s; Ver: %s; Product Version: %s\n",
765 drhd->reg_base_addr,
766 dmi_get_system_info(DMI_BIOS_VENDOR),
767 dmi_get_system_info(DMI_BIOS_VERSION),
768 dmi_get_system_info(DMI_PRODUCT_VERSION));
769 bios_warned = 1;
771 goto err_unmap;
774 #ifdef CONFIG_DMAR
775 agaw = iommu_calculate_agaw(iommu);
776 if (agaw < 0) {
777 printk(KERN_ERR
778 "Cannot get a valid agaw for iommu (seq_id = %d)\n",
779 iommu->seq_id);
780 goto err_unmap;
782 msagaw = iommu_calculate_max_sagaw(iommu);
783 if (msagaw < 0) {
784 printk(KERN_ERR
785 "Cannot get a valid max agaw for iommu (seq_id = %d)\n",
786 iommu->seq_id);
787 goto err_unmap;
789 #endif
790 iommu->agaw = agaw;
791 iommu->msagaw = msagaw;
793 iommu->node = -1;
795 /* the registers might be more than one page */
796 map_size = max_t(int, ecap_max_iotlb_offset(iommu->ecap),
797 cap_max_fault_reg_offset(iommu->cap));
798 map_size = VTD_PAGE_ALIGN(map_size);
799 if (map_size > VTD_PAGE_SIZE) {
800 iounmap(iommu->reg);
801 iommu->reg = ioremap(drhd->reg_base_addr, map_size);
802 if (!iommu->reg) {
803 printk(KERN_ERR "IOMMU: can't map the region\n");
804 goto error;
808 ver = readl(iommu->reg + DMAR_VER_REG);
809 pr_info("IOMMU %llx: ver %d:%d cap %llx ecap %llx\n",
810 (unsigned long long)drhd->reg_base_addr,
811 DMAR_VER_MAJOR(ver), DMAR_VER_MINOR(ver),
812 (unsigned long long)iommu->cap,
813 (unsigned long long)iommu->ecap);
815 spin_lock_init(&iommu->register_lock);
817 drhd->iommu = iommu;
818 return 0;
820 err_unmap:
821 iounmap(iommu->reg);
822 error:
823 kfree(iommu);
824 return -1;
827 void free_iommu(struct intel_iommu *iommu)
829 if (!iommu)
830 return;
832 #ifdef CONFIG_DMAR
833 free_dmar_iommu(iommu);
834 #endif
836 if (iommu->reg)
837 iounmap(iommu->reg);
838 kfree(iommu);
842 * Reclaim all the submitted descriptors which have completed its work.
844 static inline void reclaim_free_desc(struct q_inval *qi)
846 while (qi->desc_status[qi->free_tail] == QI_DONE ||
847 qi->desc_status[qi->free_tail] == QI_ABORT) {
848 qi->desc_status[qi->free_tail] = QI_FREE;
849 qi->free_tail = (qi->free_tail + 1) % QI_LENGTH;
850 qi->free_cnt++;
854 static int qi_check_fault(struct intel_iommu *iommu, int index)
856 u32 fault;
857 int head, tail;
858 struct q_inval *qi = iommu->qi;
859 int wait_index = (index + 1) % QI_LENGTH;
861 if (qi->desc_status[wait_index] == QI_ABORT)
862 return -EAGAIN;
864 fault = readl(iommu->reg + DMAR_FSTS_REG);
867 * If IQE happens, the head points to the descriptor associated
868 * with the error. No new descriptors are fetched until the IQE
869 * is cleared.
871 if (fault & DMA_FSTS_IQE) {
872 head = readl(iommu->reg + DMAR_IQH_REG);
873 if ((head >> DMAR_IQ_SHIFT) == index) {
874 printk(KERN_ERR "VT-d detected invalid descriptor: "
875 "low=%llx, high=%llx\n",
876 (unsigned long long)qi->desc[index].low,
877 (unsigned long long)qi->desc[index].high);
878 memcpy(&qi->desc[index], &qi->desc[wait_index],
879 sizeof(struct qi_desc));
880 __iommu_flush_cache(iommu, &qi->desc[index],
881 sizeof(struct qi_desc));
882 writel(DMA_FSTS_IQE, iommu->reg + DMAR_FSTS_REG);
883 return -EINVAL;
888 * If ITE happens, all pending wait_desc commands are aborted.
889 * No new descriptors are fetched until the ITE is cleared.
891 if (fault & DMA_FSTS_ITE) {
892 head = readl(iommu->reg + DMAR_IQH_REG);
893 head = ((head >> DMAR_IQ_SHIFT) - 1 + QI_LENGTH) % QI_LENGTH;
894 head |= 1;
895 tail = readl(iommu->reg + DMAR_IQT_REG);
896 tail = ((tail >> DMAR_IQ_SHIFT) - 1 + QI_LENGTH) % QI_LENGTH;
898 writel(DMA_FSTS_ITE, iommu->reg + DMAR_FSTS_REG);
900 do {
901 if (qi->desc_status[head] == QI_IN_USE)
902 qi->desc_status[head] = QI_ABORT;
903 head = (head - 2 + QI_LENGTH) % QI_LENGTH;
904 } while (head != tail);
906 if (qi->desc_status[wait_index] == QI_ABORT)
907 return -EAGAIN;
910 if (fault & DMA_FSTS_ICE)
911 writel(DMA_FSTS_ICE, iommu->reg + DMAR_FSTS_REG);
913 return 0;
917 * Submit the queued invalidation descriptor to the remapping
918 * hardware unit and wait for its completion.
920 int qi_submit_sync(struct qi_desc *desc, struct intel_iommu *iommu)
922 int rc;
923 struct q_inval *qi = iommu->qi;
924 struct qi_desc *hw, wait_desc;
925 int wait_index, index;
926 unsigned long flags;
928 if (!qi)
929 return 0;
931 hw = qi->desc;
933 restart:
934 rc = 0;
936 spin_lock_irqsave(&qi->q_lock, flags);
937 while (qi->free_cnt < 3) {
938 spin_unlock_irqrestore(&qi->q_lock, flags);
939 cpu_relax();
940 spin_lock_irqsave(&qi->q_lock, flags);
943 index = qi->free_head;
944 wait_index = (index + 1) % QI_LENGTH;
946 qi->desc_status[index] = qi->desc_status[wait_index] = QI_IN_USE;
948 hw[index] = *desc;
950 wait_desc.low = QI_IWD_STATUS_DATA(QI_DONE) |
951 QI_IWD_STATUS_WRITE | QI_IWD_TYPE;
952 wait_desc.high = virt_to_phys(&qi->desc_status[wait_index]);
954 hw[wait_index] = wait_desc;
956 __iommu_flush_cache(iommu, &hw[index], sizeof(struct qi_desc));
957 __iommu_flush_cache(iommu, &hw[wait_index], sizeof(struct qi_desc));
959 qi->free_head = (qi->free_head + 2) % QI_LENGTH;
960 qi->free_cnt -= 2;
963 * update the HW tail register indicating the presence of
964 * new descriptors.
966 writel(qi->free_head << DMAR_IQ_SHIFT, iommu->reg + DMAR_IQT_REG);
968 while (qi->desc_status[wait_index] != QI_DONE) {
970 * We will leave the interrupts disabled, to prevent interrupt
971 * context to queue another cmd while a cmd is already submitted
972 * and waiting for completion on this cpu. This is to avoid
973 * a deadlock where the interrupt context can wait indefinitely
974 * for free slots in the queue.
976 rc = qi_check_fault(iommu, index);
977 if (rc)
978 break;
980 spin_unlock(&qi->q_lock);
981 cpu_relax();
982 spin_lock(&qi->q_lock);
985 qi->desc_status[index] = QI_DONE;
987 reclaim_free_desc(qi);
988 spin_unlock_irqrestore(&qi->q_lock, flags);
990 if (rc == -EAGAIN)
991 goto restart;
993 return rc;
997 * Flush the global interrupt entry cache.
999 void qi_global_iec(struct intel_iommu *iommu)
1001 struct qi_desc desc;
1003 desc.low = QI_IEC_TYPE;
1004 desc.high = 0;
1006 /* should never fail */
1007 qi_submit_sync(&desc, iommu);
1010 void qi_flush_context(struct intel_iommu *iommu, u16 did, u16 sid, u8 fm,
1011 u64 type)
1013 struct qi_desc desc;
1015 desc.low = QI_CC_FM(fm) | QI_CC_SID(sid) | QI_CC_DID(did)
1016 | QI_CC_GRAN(type) | QI_CC_TYPE;
1017 desc.high = 0;
1019 qi_submit_sync(&desc, iommu);
1022 void qi_flush_iotlb(struct intel_iommu *iommu, u16 did, u64 addr,
1023 unsigned int size_order, u64 type)
1025 u8 dw = 0, dr = 0;
1027 struct qi_desc desc;
1028 int ih = 0;
1030 if (cap_write_drain(iommu->cap))
1031 dw = 1;
1033 if (cap_read_drain(iommu->cap))
1034 dr = 1;
1036 desc.low = QI_IOTLB_DID(did) | QI_IOTLB_DR(dr) | QI_IOTLB_DW(dw)
1037 | QI_IOTLB_GRAN(type) | QI_IOTLB_TYPE;
1038 desc.high = QI_IOTLB_ADDR(addr) | QI_IOTLB_IH(ih)
1039 | QI_IOTLB_AM(size_order);
1041 qi_submit_sync(&desc, iommu);
1044 void qi_flush_dev_iotlb(struct intel_iommu *iommu, u16 sid, u16 qdep,
1045 u64 addr, unsigned mask)
1047 struct qi_desc desc;
1049 if (mask) {
1050 BUG_ON(addr & ((1 << (VTD_PAGE_SHIFT + mask)) - 1));
1051 addr |= (1 << (VTD_PAGE_SHIFT + mask - 1)) - 1;
1052 desc.high = QI_DEV_IOTLB_ADDR(addr) | QI_DEV_IOTLB_SIZE;
1053 } else
1054 desc.high = QI_DEV_IOTLB_ADDR(addr);
1056 if (qdep >= QI_DEV_IOTLB_MAX_INVS)
1057 qdep = 0;
1059 desc.low = QI_DEV_IOTLB_SID(sid) | QI_DEV_IOTLB_QDEP(qdep) |
1060 QI_DIOTLB_TYPE;
1062 qi_submit_sync(&desc, iommu);
1066 * Disable Queued Invalidation interface.
1068 void dmar_disable_qi(struct intel_iommu *iommu)
1070 unsigned long flags;
1071 u32 sts;
1072 cycles_t start_time = get_cycles();
1074 if (!ecap_qis(iommu->ecap))
1075 return;
1077 spin_lock_irqsave(&iommu->register_lock, flags);
1079 sts = dmar_readq(iommu->reg + DMAR_GSTS_REG);
1080 if (!(sts & DMA_GSTS_QIES))
1081 goto end;
1084 * Give a chance to HW to complete the pending invalidation requests.
1086 while ((readl(iommu->reg + DMAR_IQT_REG) !=
1087 readl(iommu->reg + DMAR_IQH_REG)) &&
1088 (DMAR_OPERATION_TIMEOUT > (get_cycles() - start_time)))
1089 cpu_relax();
1091 iommu->gcmd &= ~DMA_GCMD_QIE;
1092 writel(iommu->gcmd, iommu->reg + DMAR_GCMD_REG);
1094 IOMMU_WAIT_OP(iommu, DMAR_GSTS_REG, readl,
1095 !(sts & DMA_GSTS_QIES), sts);
1096 end:
1097 spin_unlock_irqrestore(&iommu->register_lock, flags);
1101 * Enable queued invalidation.
1103 static void __dmar_enable_qi(struct intel_iommu *iommu)
1105 u32 sts;
1106 unsigned long flags;
1107 struct q_inval *qi = iommu->qi;
1109 qi->free_head = qi->free_tail = 0;
1110 qi->free_cnt = QI_LENGTH;
1112 spin_lock_irqsave(&iommu->register_lock, flags);
1114 /* write zero to the tail reg */
1115 writel(0, iommu->reg + DMAR_IQT_REG);
1117 dmar_writeq(iommu->reg + DMAR_IQA_REG, virt_to_phys(qi->desc));
1119 iommu->gcmd |= DMA_GCMD_QIE;
1120 writel(iommu->gcmd, iommu->reg + DMAR_GCMD_REG);
1122 /* Make sure hardware complete it */
1123 IOMMU_WAIT_OP(iommu, DMAR_GSTS_REG, readl, (sts & DMA_GSTS_QIES), sts);
1125 spin_unlock_irqrestore(&iommu->register_lock, flags);
1129 * Enable Queued Invalidation interface. This is a must to support
1130 * interrupt-remapping. Also used by DMA-remapping, which replaces
1131 * register based IOTLB invalidation.
1133 int dmar_enable_qi(struct intel_iommu *iommu)
1135 struct q_inval *qi;
1136 struct page *desc_page;
1138 if (!ecap_qis(iommu->ecap))
1139 return -ENOENT;
1142 * queued invalidation is already setup and enabled.
1144 if (iommu->qi)
1145 return 0;
1147 iommu->qi = kmalloc(sizeof(*qi), GFP_ATOMIC);
1148 if (!iommu->qi)
1149 return -ENOMEM;
1151 qi = iommu->qi;
1154 desc_page = alloc_pages_node(iommu->node, GFP_ATOMIC | __GFP_ZERO, 0);
1155 if (!desc_page) {
1156 kfree(qi);
1157 iommu->qi = 0;
1158 return -ENOMEM;
1161 qi->desc = page_address(desc_page);
1163 qi->desc_status = kmalloc(QI_LENGTH * sizeof(int), GFP_ATOMIC);
1164 if (!qi->desc_status) {
1165 free_page((unsigned long) qi->desc);
1166 kfree(qi);
1167 iommu->qi = 0;
1168 return -ENOMEM;
1171 qi->free_head = qi->free_tail = 0;
1172 qi->free_cnt = QI_LENGTH;
1174 spin_lock_init(&qi->q_lock);
1176 __dmar_enable_qi(iommu);
1178 return 0;
1181 /* iommu interrupt handling. Most stuff are MSI-like. */
1183 enum faulttype {
1184 DMA_REMAP,
1185 INTR_REMAP,
1186 UNKNOWN,
1189 static const char *dma_remap_fault_reasons[] =
1191 "Software",
1192 "Present bit in root entry is clear",
1193 "Present bit in context entry is clear",
1194 "Invalid context entry",
1195 "Access beyond MGAW",
1196 "PTE Write access is not set",
1197 "PTE Read access is not set",
1198 "Next page table ptr is invalid",
1199 "Root table address invalid",
1200 "Context table ptr is invalid",
1201 "non-zero reserved fields in RTP",
1202 "non-zero reserved fields in CTP",
1203 "non-zero reserved fields in PTE",
1206 static const char *intr_remap_fault_reasons[] =
1208 "Detected reserved fields in the decoded interrupt-remapped request",
1209 "Interrupt index exceeded the interrupt-remapping table size",
1210 "Present field in the IRTE entry is clear",
1211 "Error accessing interrupt-remapping table pointed by IRTA_REG",
1212 "Detected reserved fields in the IRTE entry",
1213 "Blocked a compatibility format interrupt request",
1214 "Blocked an interrupt request due to source-id verification failure",
1217 #define MAX_FAULT_REASON_IDX (ARRAY_SIZE(fault_reason_strings) - 1)
1219 const char *dmar_get_fault_reason(u8 fault_reason, int *fault_type)
1221 if (fault_reason >= 0x20 && (fault_reason <= 0x20 +
1222 ARRAY_SIZE(intr_remap_fault_reasons))) {
1223 *fault_type = INTR_REMAP;
1224 return intr_remap_fault_reasons[fault_reason - 0x20];
1225 } else if (fault_reason < ARRAY_SIZE(dma_remap_fault_reasons)) {
1226 *fault_type = DMA_REMAP;
1227 return dma_remap_fault_reasons[fault_reason];
1228 } else {
1229 *fault_type = UNKNOWN;
1230 return "Unknown";
1234 void dmar_msi_unmask(unsigned int irq)
1236 struct intel_iommu *iommu = get_irq_data(irq);
1237 unsigned long flag;
1239 /* unmask it */
1240 spin_lock_irqsave(&iommu->register_lock, flag);
1241 writel(0, iommu->reg + DMAR_FECTL_REG);
1242 /* Read a reg to force flush the post write */
1243 readl(iommu->reg + DMAR_FECTL_REG);
1244 spin_unlock_irqrestore(&iommu->register_lock, flag);
1247 void dmar_msi_mask(unsigned int irq)
1249 unsigned long flag;
1250 struct intel_iommu *iommu = get_irq_data(irq);
1252 /* mask it */
1253 spin_lock_irqsave(&iommu->register_lock, flag);
1254 writel(DMA_FECTL_IM, iommu->reg + DMAR_FECTL_REG);
1255 /* Read a reg to force flush the post write */
1256 readl(iommu->reg + DMAR_FECTL_REG);
1257 spin_unlock_irqrestore(&iommu->register_lock, flag);
1260 void dmar_msi_write(int irq, struct msi_msg *msg)
1262 struct intel_iommu *iommu = get_irq_data(irq);
1263 unsigned long flag;
1265 spin_lock_irqsave(&iommu->register_lock, flag);
1266 writel(msg->data, iommu->reg + DMAR_FEDATA_REG);
1267 writel(msg->address_lo, iommu->reg + DMAR_FEADDR_REG);
1268 writel(msg->address_hi, iommu->reg + DMAR_FEUADDR_REG);
1269 spin_unlock_irqrestore(&iommu->register_lock, flag);
1272 void dmar_msi_read(int irq, struct msi_msg *msg)
1274 struct intel_iommu *iommu = get_irq_data(irq);
1275 unsigned long flag;
1277 spin_lock_irqsave(&iommu->register_lock, flag);
1278 msg->data = readl(iommu->reg + DMAR_FEDATA_REG);
1279 msg->address_lo = readl(iommu->reg + DMAR_FEADDR_REG);
1280 msg->address_hi = readl(iommu->reg + DMAR_FEUADDR_REG);
1281 spin_unlock_irqrestore(&iommu->register_lock, flag);
1284 static int dmar_fault_do_one(struct intel_iommu *iommu, int type,
1285 u8 fault_reason, u16 source_id, unsigned long long addr)
1287 const char *reason;
1288 int fault_type;
1290 reason = dmar_get_fault_reason(fault_reason, &fault_type);
1292 if (fault_type == INTR_REMAP)
1293 printk(KERN_ERR "INTR-REMAP: Request device [[%02x:%02x.%d] "
1294 "fault index %llx\n"
1295 "INTR-REMAP:[fault reason %02d] %s\n",
1296 (source_id >> 8), PCI_SLOT(source_id & 0xFF),
1297 PCI_FUNC(source_id & 0xFF), addr >> 48,
1298 fault_reason, reason);
1299 else
1300 printk(KERN_ERR
1301 "DMAR:[%s] Request device [%02x:%02x.%d] "
1302 "fault addr %llx \n"
1303 "DMAR:[fault reason %02d] %s\n",
1304 (type ? "DMA Read" : "DMA Write"),
1305 (source_id >> 8), PCI_SLOT(source_id & 0xFF),
1306 PCI_FUNC(source_id & 0xFF), addr, fault_reason, reason);
1307 return 0;
1310 #define PRIMARY_FAULT_REG_LEN (16)
1311 irqreturn_t dmar_fault(int irq, void *dev_id)
1313 struct intel_iommu *iommu = dev_id;
1314 int reg, fault_index;
1315 u32 fault_status;
1316 unsigned long flag;
1318 spin_lock_irqsave(&iommu->register_lock, flag);
1319 fault_status = readl(iommu->reg + DMAR_FSTS_REG);
1320 if (fault_status)
1321 printk(KERN_ERR "DRHD: handling fault status reg %x\n",
1322 fault_status);
1324 /* TBD: ignore advanced fault log currently */
1325 if (!(fault_status & DMA_FSTS_PPF))
1326 goto clear_rest;
1328 fault_index = dma_fsts_fault_record_index(fault_status);
1329 reg = cap_fault_reg_offset(iommu->cap);
1330 while (1) {
1331 u8 fault_reason;
1332 u16 source_id;
1333 u64 guest_addr;
1334 int type;
1335 u32 data;
1337 /* highest 32 bits */
1338 data = readl(iommu->reg + reg +
1339 fault_index * PRIMARY_FAULT_REG_LEN + 12);
1340 if (!(data & DMA_FRCD_F))
1341 break;
1343 fault_reason = dma_frcd_fault_reason(data);
1344 type = dma_frcd_type(data);
1346 data = readl(iommu->reg + reg +
1347 fault_index * PRIMARY_FAULT_REG_LEN + 8);
1348 source_id = dma_frcd_source_id(data);
1350 guest_addr = dmar_readq(iommu->reg + reg +
1351 fault_index * PRIMARY_FAULT_REG_LEN);
1352 guest_addr = dma_frcd_page_addr(guest_addr);
1353 /* clear the fault */
1354 writel(DMA_FRCD_F, iommu->reg + reg +
1355 fault_index * PRIMARY_FAULT_REG_LEN + 12);
1357 spin_unlock_irqrestore(&iommu->register_lock, flag);
1359 dmar_fault_do_one(iommu, type, fault_reason,
1360 source_id, guest_addr);
1362 fault_index++;
1363 if (fault_index >= cap_num_fault_regs(iommu->cap))
1364 fault_index = 0;
1365 spin_lock_irqsave(&iommu->register_lock, flag);
1367 clear_rest:
1368 /* clear all the other faults */
1369 fault_status = readl(iommu->reg + DMAR_FSTS_REG);
1370 writel(fault_status, iommu->reg + DMAR_FSTS_REG);
1372 spin_unlock_irqrestore(&iommu->register_lock, flag);
1373 return IRQ_HANDLED;
1376 int dmar_set_interrupt(struct intel_iommu *iommu)
1378 int irq, ret;
1381 * Check if the fault interrupt is already initialized.
1383 if (iommu->irq)
1384 return 0;
1386 irq = create_irq();
1387 if (!irq) {
1388 printk(KERN_ERR "IOMMU: no free vectors\n");
1389 return -EINVAL;
1392 set_irq_data(irq, iommu);
1393 iommu->irq = irq;
1395 ret = arch_setup_dmar_msi(irq);
1396 if (ret) {
1397 set_irq_data(irq, NULL);
1398 iommu->irq = 0;
1399 destroy_irq(irq);
1400 return ret;
1403 ret = request_irq(irq, dmar_fault, 0, iommu->name, iommu);
1404 if (ret)
1405 printk(KERN_ERR "IOMMU: can't request irq\n");
1406 return ret;
1409 int __init enable_drhd_fault_handling(void)
1411 struct dmar_drhd_unit *drhd;
1414 * Enable fault control interrupt.
1416 for_each_drhd_unit(drhd) {
1417 int ret;
1418 struct intel_iommu *iommu = drhd->iommu;
1419 ret = dmar_set_interrupt(iommu);
1421 if (ret) {
1422 printk(KERN_ERR "DRHD %Lx: failed to enable fault, "
1423 " interrupt, ret %d\n",
1424 (unsigned long long)drhd->reg_base_addr, ret);
1425 return -1;
1429 return 0;
1433 * Re-enable Queued Invalidation interface.
1435 int dmar_reenable_qi(struct intel_iommu *iommu)
1437 if (!ecap_qis(iommu->ecap))
1438 return -ENOENT;
1440 if (!iommu->qi)
1441 return -ENOENT;
1444 * First disable queued invalidation.
1446 dmar_disable_qi(iommu);
1448 * Then enable queued invalidation again. Since there is no pending
1449 * invalidation requests now, it's safe to re-enable queued
1450 * invalidation.
1452 __dmar_enable_qi(iommu);
1454 return 0;
1458 * Check interrupt remapping support in DMAR table description.
1460 int dmar_ir_support(void)
1462 struct acpi_table_dmar *dmar;
1463 dmar = (struct acpi_table_dmar *)dmar_tbl;
1464 return dmar->flags & 0x1;