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
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
|
// SPDX-License-Identifier: GPL-2.0+
/* Copyright (C) 2015-2018 Broadcom */
/**
* DOC: V3D GEM BO management support
*
* Compared to VC4 (V3D 2.x), V3D 3.3 introduces an MMU between the
* GPU and the bus, allowing us to use shmem objects for our storage
* instead of CMA.
*
* Physically contiguous objects may still be imported to V3D, but the
* driver doesn't allocate physically contiguous objects on its own.
* Display engines requiring physically contiguous allocations should
* look into Mesa's "renderonly" support (as used by the Mesa pl111
* driver) for an example of how to integrate with V3D.
*
* Long term, we should support evicting pages from the MMU when under
* memory pressure (thus the v3d_bo_get_pages() refcounting), but
* that's not a high priority since our systems tend to not have swap.
*/
#include <linux/dma-buf.h>
#include <linux/pfn_t.h>
#include "v3d_drv.h"
#include "uapi/drm/v3d_drm.h"
/* Pins the shmem pages, fills in the .pages and .sgt fields of the BO, and maps
* it for DMA.
*/
static int
v3d_bo_get_pages(struct v3d_bo *bo)
{
struct drm_gem_object *obj = &bo->base;
struct drm_device *dev = obj->dev;
int npages = obj->size >> PAGE_SHIFT;
int ret = 0;
mutex_lock(&bo->lock);
if (bo->pages_refcount++ != 0)
goto unlock;
if (!obj->import_attach) {
bo->pages = drm_gem_get_pages(obj);
if (IS_ERR(bo->pages)) {
ret = PTR_ERR(bo->pages);
goto unlock;
}
bo->sgt = drm_prime_pages_to_sg(bo->pages, npages);
if (IS_ERR(bo->sgt)) {
ret = PTR_ERR(bo->sgt);
goto put_pages;
}
/* Map the pages for use by the GPU. */
dma_map_sg(dev->dev, bo->sgt->sgl,
bo->sgt->nents, DMA_BIDIRECTIONAL);
} else {
bo->pages = kcalloc(npages, sizeof(*bo->pages), GFP_KERNEL);
if (!bo->pages)
goto put_pages;
drm_prime_sg_to_page_addr_arrays(bo->sgt, bo->pages,
NULL, npages);
/* Note that dma-bufs come in mapped. */
}
mutex_unlock(&bo->lock);
return 0;
put_pages:
drm_gem_put_pages(obj, bo->pages, true, true);
bo->pages = NULL;
unlock:
bo->pages_refcount--;
mutex_unlock(&bo->lock);
return ret;
}
static void
v3d_bo_put_pages(struct v3d_bo *bo)
{
struct drm_gem_object *obj = &bo->base;
mutex_lock(&bo->lock);
if (--bo->pages_refcount == 0) {
if (!obj->import_attach) {
dma_unmap_sg(obj->dev->dev, bo->sgt->sgl,
bo->sgt->nents, DMA_BIDIRECTIONAL);
sg_free_table(bo->sgt);
kfree(bo->sgt);
drm_gem_put_pages(obj, bo->pages, true, true);
} else {
kfree(bo->pages);
}
}
mutex_unlock(&bo->lock);
}
static struct v3d_bo *v3d_bo_create_struct(struct drm_device *dev,
size_t unaligned_size)
{
struct v3d_dev *v3d = to_v3d_dev(dev);
struct drm_gem_object *obj;
struct v3d_bo *bo;
size_t size = roundup(unaligned_size, PAGE_SIZE);
int ret;
if (size == 0)
return ERR_PTR(-EINVAL);
bo = kzalloc(sizeof(*bo), GFP_KERNEL);
if (!bo)
return ERR_PTR(-ENOMEM);
obj = &bo->base;
INIT_LIST_HEAD(&bo->vmas);
INIT_LIST_HEAD(&bo->unref_head);
mutex_init(&bo->lock);
ret = drm_gem_object_init(dev, obj, size);
if (ret)
goto free_bo;
spin_lock(&v3d->mm_lock);
ret = drm_mm_insert_node_generic(&v3d->mm, &bo->node,
obj->size >> PAGE_SHIFT,
GMP_GRANULARITY >> PAGE_SHIFT, 0, 0);
spin_unlock(&v3d->mm_lock);
if (ret)
goto free_obj;
return bo;
free_obj:
drm_gem_object_release(obj);
free_bo:
kfree(bo);
return ERR_PTR(ret);
}
struct v3d_bo *v3d_bo_create(struct drm_device *dev, struct drm_file *file_priv,
size_t unaligned_size)
{
struct v3d_dev *v3d = to_v3d_dev(dev);
struct drm_gem_object *obj;
struct v3d_bo *bo;
int ret;
bo = v3d_bo_create_struct(dev, unaligned_size);
if (IS_ERR(bo))
return bo;
obj = &bo->base;
ret = v3d_bo_get_pages(bo);
if (ret)
goto free_mm;
v3d_mmu_insert_ptes(bo);
mutex_lock(&v3d->bo_lock);
v3d->bo_stats.num_allocated++;
v3d->bo_stats.pages_allocated += obj->size >> PAGE_SHIFT;
mutex_unlock(&v3d->bo_lock);
return bo;
free_mm:
spin_lock(&v3d->mm_lock);
drm_mm_remove_node(&bo->node);
spin_unlock(&v3d->mm_lock);
drm_gem_object_release(obj);
kfree(bo);
return ERR_PTR(ret);
}
/* Called DRM core on the last userspace/kernel unreference of the
* BO.
*/
void v3d_free_object(struct drm_gem_object *obj)
{
struct v3d_dev *v3d = to_v3d_dev(obj->dev);
struct v3d_bo *bo = to_v3d_bo(obj);
mutex_lock(&v3d->bo_lock);
v3d->bo_stats.num_allocated--;
v3d->bo_stats.pages_allocated -= obj->size >> PAGE_SHIFT;
mutex_unlock(&v3d->bo_lock);
v3d_bo_put_pages(bo);
if (obj->import_attach)
drm_prime_gem_destroy(obj, bo->sgt);
v3d_mmu_remove_ptes(bo);
spin_lock(&v3d->mm_lock);
drm_mm_remove_node(&bo->node);
spin_unlock(&v3d->mm_lock);
mutex_destroy(&bo->lock);
drm_gem_object_release(obj);
kfree(bo);
}
static void
v3d_set_mmap_vma_flags(struct vm_area_struct *vma)
{
vma->vm_flags &= ~VM_PFNMAP;
vma->vm_flags |= VM_MIXEDMAP;
vma->vm_page_prot = pgprot_writecombine(vm_get_page_prot(vma->vm_flags));
}
vm_fault_t v3d_gem_fault(struct vm_fault *vmf)
{
struct vm_area_struct *vma = vmf->vma;
struct drm_gem_object *obj = vma->vm_private_data;
struct v3d_bo *bo = to_v3d_bo(obj);
pfn_t pfn;
pgoff_t pgoff;
/* We don't use vmf->pgoff since that has the fake offset: */
pgoff = (vmf->address - vma->vm_start) >> PAGE_SHIFT;
pfn = __pfn_to_pfn_t(page_to_pfn(bo->pages[pgoff]), PFN_DEV);
return vmf_insert_mixed(vma, vmf->address, pfn);
}
int v3d_mmap(struct file *filp, struct vm_area_struct *vma)
{
int ret;
ret = drm_gem_mmap(filp, vma);
if (ret)
return ret;
v3d_set_mmap_vma_flags(vma);
return ret;
}
int v3d_prime_mmap(struct drm_gem_object *obj, struct vm_area_struct *vma)
{
int ret;
ret = drm_gem_mmap_obj(obj, obj->size, vma);
if (ret < 0)
return ret;
v3d_set_mmap_vma_flags(vma);
return 0;
}
struct sg_table *
v3d_prime_get_sg_table(struct drm_gem_object *obj)
{
struct v3d_bo *bo = to_v3d_bo(obj);
int npages = obj->size >> PAGE_SHIFT;
return drm_prime_pages_to_sg(bo->pages, npages);
}
struct drm_gem_object *
v3d_prime_import_sg_table(struct drm_device *dev,
struct dma_buf_attachment *attach,
struct sg_table *sgt)
{
struct v3d_dev *v3d = to_v3d_dev(dev);
struct drm_gem_object *obj;
struct v3d_bo *bo;
bo = v3d_bo_create_struct(dev, attach->dmabuf->size);
if (IS_ERR(bo))
return ERR_CAST(bo);
obj = &bo->base;
obj->resv = attach->dmabuf->resv;
bo->sgt = sgt;
obj->import_attach = attach;
v3d_bo_get_pages(bo);
mutex_lock(&v3d->bo_lock);
v3d->bo_stats.num_allocated++;
v3d->bo_stats.pages_allocated += obj->size >> PAGE_SHIFT;
mutex_unlock(&v3d->bo_lock);
v3d_mmu_insert_ptes(bo);
return obj;
}
int v3d_create_bo_ioctl(struct drm_device *dev, void *data,
struct drm_file *file_priv)
{
struct drm_v3d_create_bo *args = data;
struct v3d_bo *bo = NULL;
int ret;
if (args->flags != 0) {
DRM_INFO("unknown create_bo flags: %d\n", args->flags);
return -EINVAL;
}
bo = v3d_bo_create(dev, file_priv, PAGE_ALIGN(args->size));
if (IS_ERR(bo))
return PTR_ERR(bo);
args->offset = bo->node.start << PAGE_SHIFT;
ret = drm_gem_handle_create(file_priv, &bo->base, &args->handle);
drm_gem_object_put_unlocked(&bo->base);
return ret;
}
int v3d_mmap_bo_ioctl(struct drm_device *dev, void *data,
struct drm_file *file_priv)
{
struct drm_v3d_mmap_bo *args = data;
struct drm_gem_object *gem_obj;
int ret;
if (args->flags != 0) {
DRM_INFO("unknown mmap_bo flags: %d\n", args->flags);
return -EINVAL;
}
gem_obj = drm_gem_object_lookup(file_priv, args->handle);
if (!gem_obj) {
DRM_DEBUG("Failed to look up GEM BO %d\n", args->handle);
return -ENOENT;
}
ret = drm_gem_create_mmap_offset(gem_obj);
if (ret == 0)
args->offset = drm_vma_node_offset_addr(&gem_obj->vma_node);
drm_gem_object_put_unlocked(gem_obj);
return ret;
}
int v3d_get_bo_offset_ioctl(struct drm_device *dev, void *data,
struct drm_file *file_priv)
{
struct drm_v3d_get_bo_offset *args = data;
struct drm_gem_object *gem_obj;
struct v3d_bo *bo;
gem_obj = drm_gem_object_lookup(file_priv, args->handle);
if (!gem_obj) {
DRM_DEBUG("Failed to look up GEM BO %d\n", args->handle);
return -ENOENT;
}
bo = to_v3d_bo(gem_obj);
args->offset = bo->node.start << PAGE_SHIFT;
drm_gem_object_put_unlocked(gem_obj);
return 0;
}
|