summaryrefslogtreecommitdiffstats
path: root/arch/mips/mm/dma-noncoherent.c
blob: b01b9a3e424fa3545a87a846b85e809b6f46ffdd (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
// SPDX-License-Identifier: GPL-2.0
/*
 * Copyright (C) 2000  Ani Joshi <ajoshi@unixbox.com>
 * Copyright (C) 2000, 2001, 06	 Ralf Baechle <ralf@linux-mips.org>
 * swiped from i386, and cloned for MIPS by Geert, polished by Ralf.
 */
#include <linux/dma-direct.h>
#include <linux/dma-noncoherent.h>
#include <linux/dma-contiguous.h>
#include <linux/highmem.h>

#include <asm/cache.h>
#include <asm/cpu-type.h>
#include <asm/dma-coherence.h>
#include <asm/io.h>

/*
 * The affected CPUs below in 'cpu_needs_post_dma_flush()' can speculatively
 * fill random cachelines with stale data at any time, requiring an extra
 * flush post-DMA.
 *
 * Warning on the terminology - Linux calls an uncached area coherent;  MIPS
 * terminology calls memory areas with hardware maintained coherency coherent.
 *
 * Note that the R14000 and R16000 should also be checked for in this condition.
 * However this function is only called on non-I/O-coherent systems and only the
 * R10000 and R12000 are used in such systems, the SGI IP28 Indigo² rsp.
 * SGI IP32 aka O2.
 */
static inline bool cpu_needs_post_dma_flush(struct device *dev)
{
	switch (boot_cpu_type()) {
	case CPU_R10000:
	case CPU_R12000:
	case CPU_BMIPS5000:
		return true;
	default:
		/*
		 * Presence of MAARs suggests that the CPU supports
		 * speculatively prefetching data, and therefore requires
		 * the post-DMA flush/invalidate.
		 */
		return cpu_has_maar;
	}
}

void *arch_dma_alloc(struct device *dev, size_t size,
		dma_addr_t *dma_handle, gfp_t gfp, unsigned long attrs)
{
	void *ret;

	ret = dma_direct_alloc_pages(dev, size, dma_handle, gfp, attrs);
	if (!ret && !(attrs & DMA_ATTR_NON_CONSISTENT)) {
		dma_cache_wback_inv((unsigned long) ret, size);
		ret = (void *)UNCAC_ADDR(ret);
	}

	return ret;
}

void arch_dma_free(struct device *dev, size_t size, void *cpu_addr,
		dma_addr_t dma_addr, unsigned long attrs)
{
	if (!(attrs & DMA_ATTR_NON_CONSISTENT))
		cpu_addr = (void *)CAC_ADDR((unsigned long)cpu_addr);
	dma_direct_free_pages(dev, size, cpu_addr, dma_addr, attrs);
}

int arch_dma_mmap(struct device *dev, struct vm_area_struct *vma,
		void *cpu_addr, dma_addr_t dma_addr, size_t size,
		unsigned long attrs)
{
	unsigned long user_count = vma_pages(vma);
	unsigned long count = PAGE_ALIGN(size) >> PAGE_SHIFT;
	unsigned long addr = CAC_ADDR((unsigned long)cpu_addr);
	unsigned long off = vma->vm_pgoff;
	unsigned long pfn = page_to_pfn(virt_to_page((void *)addr));
	int ret = -ENXIO;

	if (attrs & DMA_ATTR_WRITE_COMBINE)
		vma->vm_page_prot = pgprot_writecombine(vma->vm_page_prot);
	else
		vma->vm_page_prot = pgprot_noncached(vma->vm_page_prot);

	if (dma_mmap_from_dev_coherent(dev, vma, cpu_addr, size, &ret))
		return ret;

	if (off < count && user_count <= (count - off)) {
		ret = remap_pfn_range(vma, vma->vm_start,
				      pfn + off,
				      user_count << PAGE_SHIFT,
				      vma->vm_page_prot);
	}

	return ret;
}

static inline void dma_sync_virt(void *addr, size_t size,
		enum dma_data_direction dir)
{
	switch (dir) {
	case DMA_TO_DEVICE:
		dma_cache_wback((unsigned long)addr, size);
		break;

	case DMA_FROM_DEVICE:
		dma_cache_inv((unsigned long)addr, size);
		break;

	case DMA_BIDIRECTIONAL:
		dma_cache_wback_inv((unsigned long)addr, size);
		break;

	default:
		BUG();
	}
}

/*
 * A single sg entry may refer to multiple physically contiguous pages.  But
 * we still need to process highmem pages individually.  If highmem is not
 * configured then the bulk of this loop gets optimized out.
 */
static inline void dma_sync_phys(phys_addr_t paddr, size_t size,
		enum dma_data_direction dir)
{
	struct page *page = pfn_to_page(paddr >> PAGE_SHIFT);
	unsigned long offset = paddr & ~PAGE_MASK;
	size_t left = size;

	do {
		size_t len = left;

		if (PageHighMem(page)) {
			void *addr;

			if (offset + len > PAGE_SIZE) {
				if (offset >= PAGE_SIZE) {
					page += offset >> PAGE_SHIFT;
					offset &= ~PAGE_MASK;
				}
				len = PAGE_SIZE - offset;
			}

			addr = kmap_atomic(page);
			dma_sync_virt(addr + offset, len, dir);
			kunmap_atomic(addr);
		} else
			dma_sync_virt(page_address(page) + offset, size, dir);
		offset = 0;
		page++;
		left -= len;
	} while (left);
}

void arch_sync_dma_for_device(struct device *dev, phys_addr_t paddr,
		size_t size, enum dma_data_direction dir)
{
	dma_sync_phys(paddr, size, dir);
}

void arch_sync_dma_for_cpu(struct device *dev, phys_addr_t paddr,
		size_t size, enum dma_data_direction dir)
{
	if (cpu_needs_post_dma_flush(dev))
		dma_sync_phys(paddr, size, dir);
}

void arch_dma_cache_sync(struct device *dev, void *vaddr, size_t size,
		enum dma_data_direction direction)
{
	BUG_ON(direction == DMA_NONE);

	dma_sync_virt(vaddr, size, direction);
}