1// SPDX-License-Identifier: GPL-2.0-only
2/*
3 * Copyright (C) 2011-2014 NVIDIA CORPORATION.  All rights reserved.
4 */
5
6#include <linux/bitops.h>
7#include <linux/debugfs.h>
8#include <linux/err.h>
9#include <linux/iommu.h>
10#include <linux/kernel.h>
11#include <linux/of.h>
12#include <linux/of_platform.h>
13#include <linux/pci.h>
14#include <linux/platform_device.h>
15#include <linux/slab.h>
16#include <linux/spinlock.h>
17#include <linux/dma-mapping.h>
18
19#include <soc/tegra/ahb.h>
20#include <soc/tegra/mc.h>
21
22struct tegra_smmu_group {
23	struct list_head list;
24	struct tegra_smmu *smmu;
25	const struct tegra_smmu_group_soc *soc;
26	struct iommu_group *group;
27	unsigned int swgroup;
28};
29
30struct tegra_smmu {
31	void __iomem *regs;
32	struct device *dev;
33
34	struct tegra_mc *mc;
35	const struct tegra_smmu_soc *soc;
36
37	struct list_head groups;
38
39	unsigned long pfn_mask;
40	unsigned long tlb_mask;
41
42	unsigned long *asids;
43	struct mutex lock;
44
45	struct list_head list;
46
47	struct dentry *debugfs;
48
49	struct iommu_device iommu;	/* IOMMU Core code handle */
50};
51
52struct tegra_smmu_as {
53	struct iommu_domain domain;
54	struct tegra_smmu *smmu;
55	unsigned int use_count;
56	spinlock_t lock;
57	u32 *count;
58	struct page **pts;
59	struct page *pd;
60	dma_addr_t pd_dma;
61	unsigned id;
62	u32 attr;
63};
64
65static struct tegra_smmu_as *to_smmu_as(struct iommu_domain *dom)
66{
67	return container_of(dom, struct tegra_smmu_as, domain);
68}
69
70static inline void smmu_writel(struct tegra_smmu *smmu, u32 value,
71			       unsigned long offset)
72{
73	writel(value, smmu->regs + offset);
74}
75
76static inline u32 smmu_readl(struct tegra_smmu *smmu, unsigned long offset)
77{
78	return readl(smmu->regs + offset);
79}
80
81#define SMMU_CONFIG 0x010
82#define  SMMU_CONFIG_ENABLE (1 << 0)
83
84#define SMMU_TLB_CONFIG 0x14
85#define  SMMU_TLB_CONFIG_HIT_UNDER_MISS (1 << 29)
86#define  SMMU_TLB_CONFIG_ROUND_ROBIN_ARBITRATION (1 << 28)
87#define  SMMU_TLB_CONFIG_ACTIVE_LINES(smmu) \
88	((smmu)->soc->num_tlb_lines & (smmu)->tlb_mask)
89
90#define SMMU_PTC_CONFIG 0x18
91#define  SMMU_PTC_CONFIG_ENABLE (1 << 29)
92#define  SMMU_PTC_CONFIG_REQ_LIMIT(x) (((x) & 0x0f) << 24)
93#define  SMMU_PTC_CONFIG_INDEX_MAP(x) ((x) & 0x3f)
94
95#define SMMU_PTB_ASID 0x01c
96#define  SMMU_PTB_ASID_VALUE(x) ((x) & 0x7f)
97
98#define SMMU_PTB_DATA 0x020
99#define  SMMU_PTB_DATA_VALUE(dma, attr) ((dma) >> 12 | (attr))
100
101#define SMMU_MK_PDE(dma, attr) ((dma) >> SMMU_PTE_SHIFT | (attr))
102
103#define SMMU_TLB_FLUSH 0x030
104#define  SMMU_TLB_FLUSH_VA_MATCH_ALL     (0 << 0)
105#define  SMMU_TLB_FLUSH_VA_MATCH_SECTION (2 << 0)
106#define  SMMU_TLB_FLUSH_VA_MATCH_GROUP   (3 << 0)
107#define  SMMU_TLB_FLUSH_VA_SECTION(addr) ((((addr) & 0xffc00000) >> 12) | \
108					  SMMU_TLB_FLUSH_VA_MATCH_SECTION)
109#define  SMMU_TLB_FLUSH_VA_GROUP(addr)   ((((addr) & 0xffffc000) >> 12) | \
110					  SMMU_TLB_FLUSH_VA_MATCH_GROUP)
111#define  SMMU_TLB_FLUSH_ASID_MATCH       (1 << 31)
112
113#define SMMU_PTC_FLUSH 0x034
114#define  SMMU_PTC_FLUSH_TYPE_ALL (0 << 0)
115#define  SMMU_PTC_FLUSH_TYPE_ADR (1 << 0)
116
117#define SMMU_PTC_FLUSH_HI 0x9b8
118#define  SMMU_PTC_FLUSH_HI_MASK 0x3
119
120/* per-SWGROUP SMMU_*_ASID register */
121#define SMMU_ASID_ENABLE (1 << 31)
122#define SMMU_ASID_MASK 0x7f
123#define SMMU_ASID_VALUE(x) ((x) & SMMU_ASID_MASK)
124
125/* page table definitions */
126#define SMMU_NUM_PDE 1024
127#define SMMU_NUM_PTE 1024
128
129#define SMMU_SIZE_PD (SMMU_NUM_PDE * 4)
130#define SMMU_SIZE_PT (SMMU_NUM_PTE * 4)
131
132#define SMMU_PDE_SHIFT 22
133#define SMMU_PTE_SHIFT 12
134
135#define SMMU_PAGE_MASK		(~(SMMU_SIZE_PT-1))
136#define SMMU_OFFSET_IN_PAGE(x)	((unsigned long)(x) & ~SMMU_PAGE_MASK)
137#define SMMU_PFN_PHYS(x)	((phys_addr_t)(x) << SMMU_PTE_SHIFT)
138#define SMMU_PHYS_PFN(x)	((unsigned long)((x) >> SMMU_PTE_SHIFT))
139
140#define SMMU_PD_READABLE	(1 << 31)
141#define SMMU_PD_WRITABLE	(1 << 30)
142#define SMMU_PD_NONSECURE	(1 << 29)
143
144#define SMMU_PDE_READABLE	(1 << 31)
145#define SMMU_PDE_WRITABLE	(1 << 30)
146#define SMMU_PDE_NONSECURE	(1 << 29)
147#define SMMU_PDE_NEXT		(1 << 28)
148
149#define SMMU_PTE_READABLE	(1 << 31)
150#define SMMU_PTE_WRITABLE	(1 << 30)
151#define SMMU_PTE_NONSECURE	(1 << 29)
152
153#define SMMU_PDE_ATTR		(SMMU_PDE_READABLE | SMMU_PDE_WRITABLE | \
154				 SMMU_PDE_NONSECURE)
155
156static unsigned int iova_pd_index(unsigned long iova)
157{
158	return (iova >> SMMU_PDE_SHIFT) & (SMMU_NUM_PDE - 1);
159}
160
161static unsigned int iova_pt_index(unsigned long iova)
162{
163	return (iova >> SMMU_PTE_SHIFT) & (SMMU_NUM_PTE - 1);
164}
165
166static bool smmu_dma_addr_valid(struct tegra_smmu *smmu, dma_addr_t addr)
167{
168	addr >>= 12;
169	return (addr & smmu->pfn_mask) == addr;
170}
171
172static dma_addr_t smmu_pde_to_dma(struct tegra_smmu *smmu, u32 pde)
173{
174	return (dma_addr_t)(pde & smmu->pfn_mask) << 12;
175}
176
177static void smmu_flush_ptc_all(struct tegra_smmu *smmu)
178{
179	smmu_writel(smmu, SMMU_PTC_FLUSH_TYPE_ALL, SMMU_PTC_FLUSH);
180}
181
182static inline void smmu_flush_ptc(struct tegra_smmu *smmu, dma_addr_t dma,
183				  unsigned long offset)
184{
185	u32 value;
186
187	offset &= ~(smmu->mc->soc->atom_size - 1);
188
189	if (smmu->mc->soc->num_address_bits > 32) {
190#ifdef CONFIG_ARCH_DMA_ADDR_T_64BIT
191		value = (dma >> 32) & SMMU_PTC_FLUSH_HI_MASK;
192#else
193		value = 0;
194#endif
195		smmu_writel(smmu, value, SMMU_PTC_FLUSH_HI);
196	}
197
198	value = (dma + offset) | SMMU_PTC_FLUSH_TYPE_ADR;
199	smmu_writel(smmu, value, SMMU_PTC_FLUSH);
200}
201
202static inline void smmu_flush_tlb(struct tegra_smmu *smmu)
203{
204	smmu_writel(smmu, SMMU_TLB_FLUSH_VA_MATCH_ALL, SMMU_TLB_FLUSH);
205}
206
207static inline void smmu_flush_tlb_asid(struct tegra_smmu *smmu,
208				       unsigned long asid)
209{
210	u32 value;
211
212	if (smmu->soc->num_asids == 4)
213		value = (asid & 0x3) << 29;
214	else
215		value = (asid & 0x7f) << 24;
216
217	value |= SMMU_TLB_FLUSH_ASID_MATCH | SMMU_TLB_FLUSH_VA_MATCH_ALL;
218	smmu_writel(smmu, value, SMMU_TLB_FLUSH);
219}
220
221static inline void smmu_flush_tlb_section(struct tegra_smmu *smmu,
222					  unsigned long asid,
223					  unsigned long iova)
224{
225	u32 value;
226
227	if (smmu->soc->num_asids == 4)
228		value = (asid & 0x3) << 29;
229	else
230		value = (asid & 0x7f) << 24;
231
232	value |= SMMU_TLB_FLUSH_ASID_MATCH | SMMU_TLB_FLUSH_VA_SECTION(iova);
233	smmu_writel(smmu, value, SMMU_TLB_FLUSH);
234}
235
236static inline void smmu_flush_tlb_group(struct tegra_smmu *smmu,
237					unsigned long asid,
238					unsigned long iova)
239{
240	u32 value;
241
242	if (smmu->soc->num_asids == 4)
243		value = (asid & 0x3) << 29;
244	else
245		value = (asid & 0x7f) << 24;
246
247	value |= SMMU_TLB_FLUSH_ASID_MATCH | SMMU_TLB_FLUSH_VA_GROUP(iova);
248	smmu_writel(smmu, value, SMMU_TLB_FLUSH);
249}
250
251static inline void smmu_flush(struct tegra_smmu *smmu)
252{
253	smmu_readl(smmu, SMMU_PTB_ASID);
254}
255
256static int tegra_smmu_alloc_asid(struct tegra_smmu *smmu, unsigned int *idp)
257{
258	unsigned long id;
259
260	id = find_first_zero_bit(smmu->asids, smmu->soc->num_asids);
261	if (id >= smmu->soc->num_asids)
262		return -ENOSPC;
263
264	set_bit(id, smmu->asids);
265	*idp = id;
266
267	return 0;
268}
269
270static void tegra_smmu_free_asid(struct tegra_smmu *smmu, unsigned int id)
271{
272	clear_bit(id, smmu->asids);
273}
274
275static struct iommu_domain *tegra_smmu_domain_alloc(unsigned type)
276{
277	struct tegra_smmu_as *as;
278
279	if (type != IOMMU_DOMAIN_UNMANAGED)
280		return NULL;
281
282	as = kzalloc(sizeof(*as), GFP_KERNEL);
283	if (!as)
284		return NULL;
285
286	as->attr = SMMU_PD_READABLE | SMMU_PD_WRITABLE | SMMU_PD_NONSECURE;
287
288	as->pd = alloc_page(GFP_KERNEL | __GFP_DMA | __GFP_ZERO);
289	if (!as->pd) {
290		kfree(as);
291		return NULL;
292	}
293
294	as->count = kcalloc(SMMU_NUM_PDE, sizeof(u32), GFP_KERNEL);
295	if (!as->count) {
296		__free_page(as->pd);
297		kfree(as);
298		return NULL;
299	}
300
301	as->pts = kcalloc(SMMU_NUM_PDE, sizeof(*as->pts), GFP_KERNEL);
302	if (!as->pts) {
303		kfree(as->count);
304		__free_page(as->pd);
305		kfree(as);
306		return NULL;
307	}
308
309	spin_lock_init(&as->lock);
310
311	/* setup aperture */
312	as->domain.geometry.aperture_start = 0;
313	as->domain.geometry.aperture_end = 0xffffffff;
314	as->domain.geometry.force_aperture = true;
315
316	return &as->domain;
317}
318
319static void tegra_smmu_domain_free(struct iommu_domain *domain)
320{
321	struct tegra_smmu_as *as = to_smmu_as(domain);
322
323	/* TODO: free page directory and page tables */
324
325	WARN_ON_ONCE(as->use_count);
326	kfree(as->count);
327	kfree(as->pts);
328	kfree(as);
329}
330
331static const struct tegra_smmu_swgroup *
332tegra_smmu_find_swgroup(struct tegra_smmu *smmu, unsigned int swgroup)
333{
334	const struct tegra_smmu_swgroup *group = NULL;
335	unsigned int i;
336
337	for (i = 0; i < smmu->soc->num_swgroups; i++) {
338		if (smmu->soc->swgroups[i].swgroup == swgroup) {
339			group = &smmu->soc->swgroups[i];
340			break;
341		}
342	}
343
344	return group;
345}
346
347static void tegra_smmu_enable(struct tegra_smmu *smmu, unsigned int swgroup,
348			      unsigned int asid)
349{
350	const struct tegra_smmu_swgroup *group;
351	unsigned int i;
352	u32 value;
353
354	group = tegra_smmu_find_swgroup(smmu, swgroup);
355	if (group) {
356		value = smmu_readl(smmu, group->reg);
357		value &= ~SMMU_ASID_MASK;
358		value |= SMMU_ASID_VALUE(asid);
359		value |= SMMU_ASID_ENABLE;
360		smmu_writel(smmu, value, group->reg);
361	} else {
362		pr_warn("%s group from swgroup %u not found\n", __func__,
363				swgroup);
364		/* No point moving ahead if group was not found */
365		return;
366	}
367
368	for (i = 0; i < smmu->soc->num_clients; i++) {
369		const struct tegra_mc_client *client = &smmu->soc->clients[i];
370
371		if (client->swgroup != swgroup)
372			continue;
373
374		value = smmu_readl(smmu, client->regs.smmu.reg);
375		value |= BIT(client->regs.smmu.bit);
376		smmu_writel(smmu, value, client->regs.smmu.reg);
377	}
378}
379
380static void tegra_smmu_disable(struct tegra_smmu *smmu, unsigned int swgroup,
381			       unsigned int asid)
382{
383	const struct tegra_smmu_swgroup *group;
384	unsigned int i;
385	u32 value;
386
387	group = tegra_smmu_find_swgroup(smmu, swgroup);
388	if (group) {
389		value = smmu_readl(smmu, group->reg);
390		value &= ~SMMU_ASID_MASK;
391		value |= SMMU_ASID_VALUE(asid);
392		value &= ~SMMU_ASID_ENABLE;
393		smmu_writel(smmu, value, group->reg);
394	}
395
396	for (i = 0; i < smmu->soc->num_clients; i++) {
397		const struct tegra_mc_client *client = &smmu->soc->clients[i];
398
399		if (client->swgroup != swgroup)
400			continue;
401
402		value = smmu_readl(smmu, client->regs.smmu.reg);
403		value &= ~BIT(client->regs.smmu.bit);
404		smmu_writel(smmu, value, client->regs.smmu.reg);
405	}
406}
407
408static int tegra_smmu_as_prepare(struct tegra_smmu *smmu,
409				 struct tegra_smmu_as *as)
410{
411	u32 value;
412	int err = 0;
413
414	mutex_lock(&smmu->lock);
415
416	if (as->use_count > 0) {
417		as->use_count++;
418		goto unlock;
419	}
420
421	as->pd_dma = dma_map_page(smmu->dev, as->pd, 0, SMMU_SIZE_PD,
422				  DMA_TO_DEVICE);
423	if (dma_mapping_error(smmu->dev, as->pd_dma)) {
424		err = -ENOMEM;
425		goto unlock;
426	}
427
428	/* We can't handle 64-bit DMA addresses */
429	if (!smmu_dma_addr_valid(smmu, as->pd_dma)) {
430		err = -ENOMEM;
431		goto err_unmap;
432	}
433
434	err = tegra_smmu_alloc_asid(smmu, &as->id);
435	if (err < 0)
436		goto err_unmap;
437
438	smmu_flush_ptc(smmu, as->pd_dma, 0);
439	smmu_flush_tlb_asid(smmu, as->id);
440
441	smmu_writel(smmu, as->id & 0x7f, SMMU_PTB_ASID);
442	value = SMMU_PTB_DATA_VALUE(as->pd_dma, as->attr);
443	smmu_writel(smmu, value, SMMU_PTB_DATA);
444	smmu_flush(smmu);
445
446	as->smmu = smmu;
447	as->use_count++;
448
449	mutex_unlock(&smmu->lock);
450
451	return 0;
452
453err_unmap:
454	dma_unmap_page(smmu->dev, as->pd_dma, SMMU_SIZE_PD, DMA_TO_DEVICE);
455unlock:
456	mutex_unlock(&smmu->lock);
457
458	return err;
459}
460
461static void tegra_smmu_as_unprepare(struct tegra_smmu *smmu,
462				    struct tegra_smmu_as *as)
463{
464	mutex_lock(&smmu->lock);
465
466	if (--as->use_count > 0) {
467		mutex_unlock(&smmu->lock);
468		return;
469	}
470
471	tegra_smmu_free_asid(smmu, as->id);
472
473	dma_unmap_page(smmu->dev, as->pd_dma, SMMU_SIZE_PD, DMA_TO_DEVICE);
474
475	as->smmu = NULL;
476
477	mutex_unlock(&smmu->lock);
478}
479
480static int tegra_smmu_attach_dev(struct iommu_domain *domain,
481				 struct device *dev)
482{
483	struct iommu_fwspec *fwspec = dev_iommu_fwspec_get(dev);
484	struct tegra_smmu *smmu = dev_iommu_priv_get(dev);
485	struct tegra_smmu_as *as = to_smmu_as(domain);
486	unsigned int index;
487	int err;
488
489	if (!fwspec)
490		return -ENOENT;
491
492	for (index = 0; index < fwspec->num_ids; index++) {
493		err = tegra_smmu_as_prepare(smmu, as);
494		if (err)
495			goto disable;
496
497		tegra_smmu_enable(smmu, fwspec->ids[index], as->id);
498	}
499
500	if (index == 0)
501		return -ENODEV;
502
503	return 0;
504
505disable:
506	while (index--) {
507		tegra_smmu_disable(smmu, fwspec->ids[index], as->id);
508		tegra_smmu_as_unprepare(smmu, as);
509	}
510
511	return err;
512}
513
514static void tegra_smmu_set_platform_dma(struct device *dev)
515{
516	struct iommu_domain *domain = iommu_get_domain_for_dev(dev);
517	struct iommu_fwspec *fwspec = dev_iommu_fwspec_get(dev);
518	struct tegra_smmu_as *as = to_smmu_as(domain);
519	struct tegra_smmu *smmu = as->smmu;
520	unsigned int index;
521
522	if (!fwspec)
523		return;
524
525	for (index = 0; index < fwspec->num_ids; index++) {
526		tegra_smmu_disable(smmu, fwspec->ids[index], as->id);
527		tegra_smmu_as_unprepare(smmu, as);
528	}
529}
530
531static void tegra_smmu_set_pde(struct tegra_smmu_as *as, unsigned long iova,
532			       u32 value)
533{
534	unsigned int pd_index = iova_pd_index(iova);
535	struct tegra_smmu *smmu = as->smmu;
536	u32 *pd = page_address(as->pd);
537	unsigned long offset = pd_index * sizeof(*pd);
538
539	/* Set the page directory entry first */
540	pd[pd_index] = value;
541
542	/* The flush the page directory entry from caches */
543	dma_sync_single_range_for_device(smmu->dev, as->pd_dma, offset,
544					 sizeof(*pd), DMA_TO_DEVICE);
545
546	/* And flush the iommu */
547	smmu_flush_ptc(smmu, as->pd_dma, offset);
548	smmu_flush_tlb_section(smmu, as->id, iova);
549	smmu_flush(smmu);
550}
551
552static u32 *tegra_smmu_pte_offset(struct page *pt_page, unsigned long iova)
553{
554	u32 *pt = page_address(pt_page);
555
556	return pt + iova_pt_index(iova);
557}
558
559static u32 *tegra_smmu_pte_lookup(struct tegra_smmu_as *as, unsigned long iova,
560				  dma_addr_t *dmap)
561{
562	unsigned int pd_index = iova_pd_index(iova);
563	struct tegra_smmu *smmu = as->smmu;
564	struct page *pt_page;
565	u32 *pd;
566
567	pt_page = as->pts[pd_index];
568	if (!pt_page)
569		return NULL;
570
571	pd = page_address(as->pd);
572	*dmap = smmu_pde_to_dma(smmu, pd[pd_index]);
573
574	return tegra_smmu_pte_offset(pt_page, iova);
575}
576
577static u32 *as_get_pte(struct tegra_smmu_as *as, dma_addr_t iova,
578		       dma_addr_t *dmap, struct page *page)
579{
580	unsigned int pde = iova_pd_index(iova);
581	struct tegra_smmu *smmu = as->smmu;
582
583	if (!as->pts[pde]) {
584		dma_addr_t dma;
585
586		dma = dma_map_page(smmu->dev, page, 0, SMMU_SIZE_PT,
587				   DMA_TO_DEVICE);
588		if (dma_mapping_error(smmu->dev, dma)) {
589			__free_page(page);
590			return NULL;
591		}
592
593		if (!smmu_dma_addr_valid(smmu, dma)) {
594			dma_unmap_page(smmu->dev, dma, SMMU_SIZE_PT,
595				       DMA_TO_DEVICE);
596			__free_page(page);
597			return NULL;
598		}
599
600		as->pts[pde] = page;
601
602		tegra_smmu_set_pde(as, iova, SMMU_MK_PDE(dma, SMMU_PDE_ATTR |
603							      SMMU_PDE_NEXT));
604
605		*dmap = dma;
606	} else {
607		u32 *pd = page_address(as->pd);
608
609		*dmap = smmu_pde_to_dma(smmu, pd[pde]);
610	}
611
612	return tegra_smmu_pte_offset(as->pts[pde], iova);
613}
614
615static void tegra_smmu_pte_get_use(struct tegra_smmu_as *as, unsigned long iova)
616{
617	unsigned int pd_index = iova_pd_index(iova);
618
619	as->count[pd_index]++;
620}
621
622static void tegra_smmu_pte_put_use(struct tegra_smmu_as *as, unsigned long iova)
623{
624	unsigned int pde = iova_pd_index(iova);
625	struct page *page = as->pts[pde];
626
627	/*
628	 * When no entries in this page table are used anymore, return the
629	 * memory page to the system.
630	 */
631	if (--as->count[pde] == 0) {
632		struct tegra_smmu *smmu = as->smmu;
633		u32 *pd = page_address(as->pd);
634		dma_addr_t pte_dma = smmu_pde_to_dma(smmu, pd[pde]);
635
636		tegra_smmu_set_pde(as, iova, 0);
637
638		dma_unmap_page(smmu->dev, pte_dma, SMMU_SIZE_PT, DMA_TO_DEVICE);
639		__free_page(page);
640		as->pts[pde] = NULL;
641	}
642}
643
644static void tegra_smmu_set_pte(struct tegra_smmu_as *as, unsigned long iova,
645			       u32 *pte, dma_addr_t pte_dma, u32 val)
646{
647	struct tegra_smmu *smmu = as->smmu;
648	unsigned long offset = SMMU_OFFSET_IN_PAGE(pte);
649
650	*pte = val;
651
652	dma_sync_single_range_for_device(smmu->dev, pte_dma, offset,
653					 4, DMA_TO_DEVICE);
654	smmu_flush_ptc(smmu, pte_dma, offset);
655	smmu_flush_tlb_group(smmu, as->id, iova);
656	smmu_flush(smmu);
657}
658
659static struct page *as_get_pde_page(struct tegra_smmu_as *as,
660				    unsigned long iova, gfp_t gfp,
661				    unsigned long *flags)
662{
663	unsigned int pde = iova_pd_index(iova);
664	struct page *page = as->pts[pde];
665
666	/* at first check whether allocation needs to be done at all */
667	if (page)
668		return page;
669
670	/*
671	 * In order to prevent exhaustion of the atomic memory pool, we
672	 * allocate page in a sleeping context if GFP flags permit. Hence
673	 * spinlock needs to be unlocked and re-locked after allocation.
674	 */
675	if (gfpflags_allow_blocking(gfp))
676		spin_unlock_irqrestore(&as->lock, *flags);
677
678	page = alloc_page(gfp | __GFP_DMA | __GFP_ZERO);
679
680	if (gfpflags_allow_blocking(gfp))
681		spin_lock_irqsave(&as->lock, *flags);
682
683	/*
684	 * In a case of blocking allocation, a concurrent mapping may win
685	 * the PDE allocation. In this case the allocated page isn't needed
686	 * if allocation succeeded and the allocation failure isn't fatal.
687	 */
688	if (as->pts[pde]) {
689		if (page)
690			__free_page(page);
691
692		page = as->pts[pde];
693	}
694
695	return page;
696}
697
698static int
699__tegra_smmu_map(struct iommu_domain *domain, unsigned long iova,
700		 phys_addr_t paddr, size_t size, int prot, gfp_t gfp,
701		 unsigned long *flags)
702{
703	struct tegra_smmu_as *as = to_smmu_as(domain);
704	dma_addr_t pte_dma;
705	struct page *page;
706	u32 pte_attrs;
707	u32 *pte;
708
709	page = as_get_pde_page(as, iova, gfp, flags);
710	if (!page)
711		return -ENOMEM;
712
713	pte = as_get_pte(as, iova, &pte_dma, page);
714	if (!pte)
715		return -ENOMEM;
716
717	/* If we aren't overwriting a pre-existing entry, increment use */
718	if (*pte == 0)
719		tegra_smmu_pte_get_use(as, iova);
720
721	pte_attrs = SMMU_PTE_NONSECURE;
722
723	if (prot & IOMMU_READ)
724		pte_attrs |= SMMU_PTE_READABLE;
725
726	if (prot & IOMMU_WRITE)
727		pte_attrs |= SMMU_PTE_WRITABLE;
728
729	tegra_smmu_set_pte(as, iova, pte, pte_dma,
730			   SMMU_PHYS_PFN(paddr) | pte_attrs);
731
732	return 0;
733}
734
735static size_t
736__tegra_smmu_unmap(struct iommu_domain *domain, unsigned long iova,
737		   size_t size, struct iommu_iotlb_gather *gather)
738{
739	struct tegra_smmu_as *as = to_smmu_as(domain);
740	dma_addr_t pte_dma;
741	u32 *pte;
742
743	pte = tegra_smmu_pte_lookup(as, iova, &pte_dma);
744	if (!pte || !*pte)
745		return 0;
746
747	tegra_smmu_set_pte(as, iova, pte, pte_dma, 0);
748	tegra_smmu_pte_put_use(as, iova);
749
750	return size;
751}
752
753static int tegra_smmu_map(struct iommu_domain *domain, unsigned long iova,
754			  phys_addr_t paddr, size_t size, int prot, gfp_t gfp)
755{
756	struct tegra_smmu_as *as = to_smmu_as(domain);
757	unsigned long flags;
758	int ret;
759
760	spin_lock_irqsave(&as->lock, flags);
761	ret = __tegra_smmu_map(domain, iova, paddr, size, prot, gfp, &flags);
762	spin_unlock_irqrestore(&as->lock, flags);
763
764	return ret;
765}
766
767static size_t tegra_smmu_unmap(struct iommu_domain *domain, unsigned long iova,
768			       size_t size, struct iommu_iotlb_gather *gather)
769{
770	struct tegra_smmu_as *as = to_smmu_as(domain);
771	unsigned long flags;
772
773	spin_lock_irqsave(&as->lock, flags);
774	size = __tegra_smmu_unmap(domain, iova, size, gather);
775	spin_unlock_irqrestore(&as->lock, flags);
776
777	return size;
778}
779
780static phys_addr_t tegra_smmu_iova_to_phys(struct iommu_domain *domain,
781					   dma_addr_t iova)
782{
783	struct tegra_smmu_as *as = to_smmu_as(domain);
784	unsigned long pfn;
785	dma_addr_t pte_dma;
786	u32 *pte;
787
788	pte = tegra_smmu_pte_lookup(as, iova, &pte_dma);
789	if (!pte || !*pte)
790		return 0;
791
792	pfn = *pte & as->smmu->pfn_mask;
793
794	return SMMU_PFN_PHYS(pfn) + SMMU_OFFSET_IN_PAGE(iova);
795}
796
797static struct tegra_smmu *tegra_smmu_find(struct device_node *np)
798{
799	struct platform_device *pdev;
800	struct tegra_mc *mc;
801
802	pdev = of_find_device_by_node(np);
803	if (!pdev)
804		return NULL;
805
806	mc = platform_get_drvdata(pdev);
807	if (!mc) {
808		put_device(&pdev->dev);
809		return NULL;
810	}
811
812	return mc->smmu;
813}
814
815static int tegra_smmu_configure(struct tegra_smmu *smmu, struct device *dev,
816				struct of_phandle_args *args)
817{
818	const struct iommu_ops *ops = smmu->iommu.ops;
819	int err;
820
821	err = iommu_fwspec_init(dev, &dev->of_node->fwnode, ops);
822	if (err < 0) {
823		dev_err(dev, "failed to initialize fwspec: %d\n", err);
824		return err;
825	}
826
827	err = ops->of_xlate(dev, args);
828	if (err < 0) {
829		dev_err(dev, "failed to parse SW group ID: %d\n", err);
830		iommu_fwspec_free(dev);
831		return err;
832	}
833
834	return 0;
835}
836
837static struct iommu_device *tegra_smmu_probe_device(struct device *dev)
838{
839	struct device_node *np = dev->of_node;
840	struct tegra_smmu *smmu = NULL;
841	struct of_phandle_args args;
842	unsigned int index = 0;
843	int err;
844
845	while (of_parse_phandle_with_args(np, "iommus", "#iommu-cells", index,
846					  &args) == 0) {
847		smmu = tegra_smmu_find(args.np);
848		if (smmu) {
849			err = tegra_smmu_configure(smmu, dev, &args);
850
851			if (err < 0) {
852				of_node_put(args.np);
853				return ERR_PTR(err);
854			}
855		}
856
857		of_node_put(args.np);
858		index++;
859	}
860
861	smmu = dev_iommu_priv_get(dev);
862	if (!smmu)
863		return ERR_PTR(-ENODEV);
864
865	return &smmu->iommu;
866}
867
868static const struct tegra_smmu_group_soc *
869tegra_smmu_find_group(struct tegra_smmu *smmu, unsigned int swgroup)
870{
871	unsigned int i, j;
872
873	for (i = 0; i < smmu->soc->num_groups; i++)
874		for (j = 0; j < smmu->soc->groups[i].num_swgroups; j++)
875			if (smmu->soc->groups[i].swgroups[j] == swgroup)
876				return &smmu->soc->groups[i];
877
878	return NULL;
879}
880
881static void tegra_smmu_group_release(void *iommu_data)
882{
883	struct tegra_smmu_group *group = iommu_data;
884	struct tegra_smmu *smmu = group->smmu;
885
886	mutex_lock(&smmu->lock);
887	list_del(&group->list);
888	mutex_unlock(&smmu->lock);
889}
890
891static struct iommu_group *tegra_smmu_device_group(struct device *dev)
892{
893	struct iommu_fwspec *fwspec = dev_iommu_fwspec_get(dev);
894	struct tegra_smmu *smmu = dev_iommu_priv_get(dev);
895	const struct tegra_smmu_group_soc *soc;
896	unsigned int swgroup = fwspec->ids[0];
897	struct tegra_smmu_group *group;
898	struct iommu_group *grp;
899
900	/* Find group_soc associating with swgroup */
901	soc = tegra_smmu_find_group(smmu, swgroup);
902
903	mutex_lock(&smmu->lock);
904
905	/* Find existing iommu_group associating with swgroup or group_soc */
906	list_for_each_entry(group, &smmu->groups, list)
907		if ((group->swgroup == swgroup) || (soc && group->soc == soc)) {
908			grp = iommu_group_ref_get(group->group);
909			mutex_unlock(&smmu->lock);
910			return grp;
911		}
912
913	group = devm_kzalloc(smmu->dev, sizeof(*group), GFP_KERNEL);
914	if (!group) {
915		mutex_unlock(&smmu->lock);
916		return NULL;
917	}
918
919	INIT_LIST_HEAD(&group->list);
920	group->swgroup = swgroup;
921	group->smmu = smmu;
922	group->soc = soc;
923
924	if (dev_is_pci(dev))
925		group->group = pci_device_group(dev);
926	else
927		group->group = generic_device_group(dev);
928
929	if (IS_ERR(group->group)) {
930		devm_kfree(smmu->dev, group);
931		mutex_unlock(&smmu->lock);
932		return NULL;
933	}
934
935	iommu_group_set_iommudata(group->group, group, tegra_smmu_group_release);
936	if (soc)
937		iommu_group_set_name(group->group, soc->name);
938	list_add_tail(&group->list, &smmu->groups);
939	mutex_unlock(&smmu->lock);
940
941	return group->group;
942}
943
944static int tegra_smmu_of_xlate(struct device *dev,
945			       struct of_phandle_args *args)
946{
947	struct platform_device *iommu_pdev = of_find_device_by_node(args->np);
948	struct tegra_mc *mc = platform_get_drvdata(iommu_pdev);
949	u32 id = args->args[0];
950
951	/*
952	 * Note: we are here releasing the reference of &iommu_pdev->dev, which
953	 * is mc->dev. Although some functions in tegra_smmu_ops may keep using
954	 * its private data beyond this point, it's still safe to do so because
955	 * the SMMU parent device is the same as the MC, so the reference count
956	 * isn't strictly necessary.
957	 */
958	put_device(&iommu_pdev->dev);
959
960	dev_iommu_priv_set(dev, mc->smmu);
961
962	return iommu_fwspec_add_ids(dev, &id, 1);
963}
964
965static const struct iommu_ops tegra_smmu_ops = {
966	.domain_alloc = tegra_smmu_domain_alloc,
967	.probe_device = tegra_smmu_probe_device,
968	.device_group = tegra_smmu_device_group,
969	.set_platform_dma_ops = tegra_smmu_set_platform_dma,
970	.of_xlate = tegra_smmu_of_xlate,
971	.pgsize_bitmap = SZ_4K,
972	.default_domain_ops = &(const struct iommu_domain_ops) {
973		.attach_dev	= tegra_smmu_attach_dev,
974		.map		= tegra_smmu_map,
975		.unmap		= tegra_smmu_unmap,
976		.iova_to_phys	= tegra_smmu_iova_to_phys,
977		.free		= tegra_smmu_domain_free,
978	}
979};
980
981static void tegra_smmu_ahb_enable(void)
982{
983	static const struct of_device_id ahb_match[] = {
984		{ .compatible = "nvidia,tegra30-ahb", },
985		{ }
986	};
987	struct device_node *ahb;
988
989	ahb = of_find_matching_node(NULL, ahb_match);
990	if (ahb) {
991		tegra_ahb_enable_smmu(ahb);
992		of_node_put(ahb);
993	}
994}
995
996static int tegra_smmu_swgroups_show(struct seq_file *s, void *data)
997{
998	struct tegra_smmu *smmu = s->private;
999	unsigned int i;
1000	u32 value;
1001
1002	seq_printf(s, "swgroup    enabled  ASID\n");
1003	seq_printf(s, "------------------------\n");
1004
1005	for (i = 0; i < smmu->soc->num_swgroups; i++) {
1006		const struct tegra_smmu_swgroup *group = &smmu->soc->swgroups[i];
1007		const char *status;
1008		unsigned int asid;
1009
1010		value = smmu_readl(smmu, group->reg);
1011
1012		if (value & SMMU_ASID_ENABLE)
1013			status = "yes";
1014		else
1015			status = "no";
1016
1017		asid = value & SMMU_ASID_MASK;
1018
1019		seq_printf(s, "%-9s  %-7s  %#04x\n", group->name, status,
1020			   asid);
1021	}
1022
1023	return 0;
1024}
1025
1026DEFINE_SHOW_ATTRIBUTE(tegra_smmu_swgroups);
1027
1028static int tegra_smmu_clients_show(struct seq_file *s, void *data)
1029{
1030	struct tegra_smmu *smmu = s->private;
1031	unsigned int i;
1032	u32 value;
1033
1034	seq_printf(s, "client       enabled\n");
1035	seq_printf(s, "--------------------\n");
1036
1037	for (i = 0; i < smmu->soc->num_clients; i++) {
1038		const struct tegra_mc_client *client = &smmu->soc->clients[i];
1039		const char *status;
1040
1041		value = smmu_readl(smmu, client->regs.smmu.reg);
1042
1043		if (value & BIT(client->regs.smmu.bit))
1044			status = "yes";
1045		else
1046			status = "no";
1047
1048		seq_printf(s, "%-12s %s\n", client->name, status);
1049	}
1050
1051	return 0;
1052}
1053
1054DEFINE_SHOW_ATTRIBUTE(tegra_smmu_clients);
1055
1056static void tegra_smmu_debugfs_init(struct tegra_smmu *smmu)
1057{
1058	smmu->debugfs = debugfs_create_dir("smmu", NULL);
1059	if (!smmu->debugfs)
1060		return;
1061
1062	debugfs_create_file("swgroups", S_IRUGO, smmu->debugfs, smmu,
1063			    &tegra_smmu_swgroups_fops);
1064	debugfs_create_file("clients", S_IRUGO, smmu->debugfs, smmu,
1065			    &tegra_smmu_clients_fops);
1066}
1067
1068static void tegra_smmu_debugfs_exit(struct tegra_smmu *smmu)
1069{
1070	debugfs_remove_recursive(smmu->debugfs);
1071}
1072
1073struct tegra_smmu *tegra_smmu_probe(struct device *dev,
1074				    const struct tegra_smmu_soc *soc,
1075				    struct tegra_mc *mc)
1076{
1077	struct tegra_smmu *smmu;
1078	u32 value;
1079	int err;
1080
1081	smmu = devm_kzalloc(dev, sizeof(*smmu), GFP_KERNEL);
1082	if (!smmu)
1083		return ERR_PTR(-ENOMEM);
1084
1085	/*
1086	 * This is a bit of a hack. Ideally we'd want to simply return this
1087	 * value. However iommu_device_register() will attempt to add
1088	 * all devices to the IOMMU before we get that far. In order
1089	 * not to rely on global variables to track the IOMMU instance, we
1090	 * set it here so that it can be looked up from the .probe_device()
1091	 * callback via the IOMMU device's .drvdata field.
1092	 */
1093	mc->smmu = smmu;
1094
1095	smmu->asids = devm_bitmap_zalloc(dev, soc->num_asids, GFP_KERNEL);
1096	if (!smmu->asids)
1097		return ERR_PTR(-ENOMEM);
1098
1099	INIT_LIST_HEAD(&smmu->groups);
1100	mutex_init(&smmu->lock);
1101
1102	smmu->regs = mc->regs;
1103	smmu->soc = soc;
1104	smmu->dev = dev;
1105	smmu->mc = mc;
1106
1107	smmu->pfn_mask =
1108		BIT_MASK(mc->soc->num_address_bits - SMMU_PTE_SHIFT) - 1;
1109	dev_dbg(dev, "address bits: %u, PFN mask: %#lx\n",
1110		mc->soc->num_address_bits, smmu->pfn_mask);
1111	smmu->tlb_mask = (1 << fls(smmu->soc->num_tlb_lines)) - 1;
1112	dev_dbg(dev, "TLB lines: %u, mask: %#lx\n", smmu->soc->num_tlb_lines,
1113		smmu->tlb_mask);
1114
1115	value = SMMU_PTC_CONFIG_ENABLE | SMMU_PTC_CONFIG_INDEX_MAP(0x3f);
1116
1117	if (soc->supports_request_limit)
1118		value |= SMMU_PTC_CONFIG_REQ_LIMIT(8);
1119
1120	smmu_writel(smmu, value, SMMU_PTC_CONFIG);
1121
1122	value = SMMU_TLB_CONFIG_HIT_UNDER_MISS |
1123		SMMU_TLB_CONFIG_ACTIVE_LINES(smmu);
1124
1125	if (soc->supports_round_robin_arbitration)
1126		value |= SMMU_TLB_CONFIG_ROUND_ROBIN_ARBITRATION;
1127
1128	smmu_writel(smmu, value, SMMU_TLB_CONFIG);
1129
1130	smmu_flush_ptc_all(smmu);
1131	smmu_flush_tlb(smmu);
1132	smmu_writel(smmu, SMMU_CONFIG_ENABLE, SMMU_CONFIG);
1133	smmu_flush(smmu);
1134
1135	tegra_smmu_ahb_enable();
1136
1137	err = iommu_device_sysfs_add(&smmu->iommu, dev, NULL, dev_name(dev));
1138	if (err)
1139		return ERR_PTR(err);
1140
1141	err = iommu_device_register(&smmu->iommu, &tegra_smmu_ops, dev);
1142	if (err) {
1143		iommu_device_sysfs_remove(&smmu->iommu);
1144		return ERR_PTR(err);
1145	}
1146
1147	if (IS_ENABLED(CONFIG_DEBUG_FS))
1148		tegra_smmu_debugfs_init(smmu);
1149
1150	return smmu;
1151}
1152
1153void tegra_smmu_remove(struct tegra_smmu *smmu)
1154{
1155	iommu_device_unregister(&smmu->iommu);
1156	iommu_device_sysfs_remove(&smmu->iommu);
1157
1158	if (IS_ENABLED(CONFIG_DEBUG_FS))
1159		tegra_smmu_debugfs_exit(smmu);
1160}
1161