| 1 | /* $NetBSD: nouveau_core_gpuobj.c,v 1.1.1.1 2014/08/06 12:36:23 riastradh Exp $ */ |
| 2 | |
| 3 | /* |
| 4 | * Copyright 2012 Red Hat Inc. |
| 5 | * |
| 6 | * Permission is hereby granted, free of charge, to any person obtaining a |
| 7 | * copy of this software and associated documentation files (the "Software"), |
| 8 | * to deal in the Software without restriction, including without limitation |
| 9 | * the rights to use, copy, modify, merge, publish, distribute, sublicense, |
| 10 | * and/or sell copies of the Software, and to permit persons to whom the |
| 11 | * Software is furnished to do so, subject to the following conditions: |
| 12 | * |
| 13 | * The above copyright notice and this permission notice shall be included in |
| 14 | * all copies or substantial portions of the Software. |
| 15 | * |
| 16 | * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR |
| 17 | * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY, |
| 18 | * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL |
| 19 | * THE COPYRIGHT HOLDER(S) OR AUTHOR(S) BE LIABLE FOR ANY CLAIM, DAMAGES OR |
| 20 | * OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, |
| 21 | * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR |
| 22 | * OTHER DEALINGS IN THE SOFTWARE. |
| 23 | * |
| 24 | * Authors: Ben Skeggs |
| 25 | */ |
| 26 | |
| 27 | #include <sys/cdefs.h> |
| 28 | __KERNEL_RCSID(0, "$NetBSD: nouveau_core_gpuobj.c,v 1.1.1.1 2014/08/06 12:36:23 riastradh Exp $" ); |
| 29 | |
| 30 | #include <core/object.h> |
| 31 | #include <core/gpuobj.h> |
| 32 | |
| 33 | #include <subdev/instmem.h> |
| 34 | #include <subdev/bar.h> |
| 35 | #include <subdev/vm.h> |
| 36 | |
| 37 | void |
| 38 | nouveau_gpuobj_destroy(struct nouveau_gpuobj *gpuobj) |
| 39 | { |
| 40 | int i; |
| 41 | |
| 42 | if (gpuobj->flags & NVOBJ_FLAG_ZERO_FREE) { |
| 43 | for (i = 0; i < gpuobj->size; i += 4) |
| 44 | nv_wo32(gpuobj, i, 0x00000000); |
| 45 | } |
| 46 | |
| 47 | if (gpuobj->node) { |
| 48 | nouveau_mm_free(&nv_gpuobj(gpuobj->parent)->heap, |
| 49 | &gpuobj->node); |
| 50 | } |
| 51 | |
| 52 | if (gpuobj->heap.block_size) |
| 53 | nouveau_mm_fini(&gpuobj->heap); |
| 54 | |
| 55 | nouveau_object_destroy(&gpuobj->base); |
| 56 | } |
| 57 | |
| 58 | int |
| 59 | nouveau_gpuobj_create_(struct nouveau_object *parent, |
| 60 | struct nouveau_object *engine, |
| 61 | struct nouveau_oclass *oclass, u32 pclass, |
| 62 | struct nouveau_object *pargpu, |
| 63 | u32 size, u32 align, u32 flags, |
| 64 | int length, void **pobject) |
| 65 | { |
| 66 | struct nouveau_instmem *imem = nouveau_instmem(parent); |
| 67 | struct nouveau_bar *bar = nouveau_bar(parent); |
| 68 | struct nouveau_gpuobj *gpuobj; |
| 69 | struct nouveau_mm *heap = NULL; |
| 70 | int ret, i; |
| 71 | u64 addr; |
| 72 | |
| 73 | *pobject = NULL; |
| 74 | |
| 75 | if (pargpu) { |
| 76 | while ((pargpu = nv_pclass(pargpu, NV_GPUOBJ_CLASS))) { |
| 77 | if (nv_gpuobj(pargpu)->heap.block_size) |
| 78 | break; |
| 79 | pargpu = pargpu->parent; |
| 80 | } |
| 81 | |
| 82 | if (unlikely(pargpu == NULL)) { |
| 83 | nv_error(parent, "no gpuobj heap\n" ); |
| 84 | return -EINVAL; |
| 85 | } |
| 86 | |
| 87 | addr = nv_gpuobj(pargpu)->addr; |
| 88 | heap = &nv_gpuobj(pargpu)->heap; |
| 89 | atomic_inc(&parent->refcount); |
| 90 | } else { |
| 91 | ret = imem->alloc(imem, parent, size, align, &parent); |
| 92 | pargpu = parent; |
| 93 | if (ret) |
| 94 | return ret; |
| 95 | |
| 96 | addr = nv_memobj(pargpu)->addr; |
| 97 | size = nv_memobj(pargpu)->size; |
| 98 | |
| 99 | if (bar && bar->alloc) { |
| 100 | struct nouveau_instobj *iobj = (void *)parent; |
| 101 | struct nouveau_mem **mem = (void *)(iobj + 1); |
| 102 | struct nouveau_mem *node = *mem; |
| 103 | if (!bar->alloc(bar, parent, node, &pargpu)) { |
| 104 | nouveau_object_ref(NULL, &parent); |
| 105 | parent = pargpu; |
| 106 | } |
| 107 | } |
| 108 | } |
| 109 | |
| 110 | ret = nouveau_object_create_(parent, engine, oclass, pclass | |
| 111 | NV_GPUOBJ_CLASS, length, pobject); |
| 112 | nouveau_object_ref(NULL, &parent); |
| 113 | gpuobj = *pobject; |
| 114 | if (ret) |
| 115 | return ret; |
| 116 | |
| 117 | gpuobj->parent = pargpu; |
| 118 | gpuobj->flags = flags; |
| 119 | gpuobj->addr = addr; |
| 120 | gpuobj->size = size; |
| 121 | |
| 122 | if (heap) { |
| 123 | ret = nouveau_mm_head(heap, 1, size, size, |
| 124 | max(align, (u32)1), &gpuobj->node); |
| 125 | if (ret) |
| 126 | return ret; |
| 127 | |
| 128 | gpuobj->addr += gpuobj->node->offset; |
| 129 | } |
| 130 | |
| 131 | if (gpuobj->flags & NVOBJ_FLAG_HEAP) { |
| 132 | ret = nouveau_mm_init(&gpuobj->heap, 0, gpuobj->size, 1); |
| 133 | if (ret) |
| 134 | return ret; |
| 135 | } |
| 136 | |
| 137 | if (flags & NVOBJ_FLAG_ZERO_ALLOC) { |
| 138 | for (i = 0; i < gpuobj->size; i += 4) |
| 139 | nv_wo32(gpuobj, i, 0x00000000); |
| 140 | } |
| 141 | |
| 142 | return ret; |
| 143 | } |
| 144 | |
| 145 | struct nouveau_gpuobj_class { |
| 146 | struct nouveau_object *pargpu; |
| 147 | u64 size; |
| 148 | u32 align; |
| 149 | u32 flags; |
| 150 | }; |
| 151 | |
| 152 | static int |
| 153 | _nouveau_gpuobj_ctor(struct nouveau_object *parent, |
| 154 | struct nouveau_object *engine, |
| 155 | struct nouveau_oclass *oclass, void *data, u32 size, |
| 156 | struct nouveau_object **pobject) |
| 157 | { |
| 158 | struct nouveau_gpuobj_class *args = data; |
| 159 | struct nouveau_gpuobj *object; |
| 160 | int ret; |
| 161 | |
| 162 | ret = nouveau_gpuobj_create(parent, engine, oclass, 0, args->pargpu, |
| 163 | args->size, args->align, args->flags, |
| 164 | &object); |
| 165 | *pobject = nv_object(object); |
| 166 | if (ret) |
| 167 | return ret; |
| 168 | |
| 169 | return 0; |
| 170 | } |
| 171 | |
| 172 | void |
| 173 | _nouveau_gpuobj_dtor(struct nouveau_object *object) |
| 174 | { |
| 175 | nouveau_gpuobj_destroy(nv_gpuobj(object)); |
| 176 | } |
| 177 | |
| 178 | int |
| 179 | _nouveau_gpuobj_init(struct nouveau_object *object) |
| 180 | { |
| 181 | return nouveau_gpuobj_init(nv_gpuobj(object)); |
| 182 | } |
| 183 | |
| 184 | int |
| 185 | _nouveau_gpuobj_fini(struct nouveau_object *object, bool suspend) |
| 186 | { |
| 187 | return nouveau_gpuobj_fini(nv_gpuobj(object), suspend); |
| 188 | } |
| 189 | |
| 190 | u32 |
| 191 | _nouveau_gpuobj_rd32(struct nouveau_object *object, u64 addr) |
| 192 | { |
| 193 | struct nouveau_gpuobj *gpuobj = nv_gpuobj(object); |
| 194 | struct nouveau_ofuncs *pfuncs = nv_ofuncs(gpuobj->parent); |
| 195 | if (gpuobj->node) |
| 196 | addr += gpuobj->node->offset; |
| 197 | return pfuncs->rd32(gpuobj->parent, addr); |
| 198 | } |
| 199 | |
| 200 | void |
| 201 | _nouveau_gpuobj_wr32(struct nouveau_object *object, u64 addr, u32 data) |
| 202 | { |
| 203 | struct nouveau_gpuobj *gpuobj = nv_gpuobj(object); |
| 204 | struct nouveau_ofuncs *pfuncs = nv_ofuncs(gpuobj->parent); |
| 205 | if (gpuobj->node) |
| 206 | addr += gpuobj->node->offset; |
| 207 | pfuncs->wr32(gpuobj->parent, addr, data); |
| 208 | } |
| 209 | |
| 210 | static struct nouveau_oclass |
| 211 | _nouveau_gpuobj_oclass = { |
| 212 | .handle = 0x00000000, |
| 213 | .ofuncs = &(struct nouveau_ofuncs) { |
| 214 | .ctor = _nouveau_gpuobj_ctor, |
| 215 | .dtor = _nouveau_gpuobj_dtor, |
| 216 | .init = _nouveau_gpuobj_init, |
| 217 | .fini = _nouveau_gpuobj_fini, |
| 218 | .rd32 = _nouveau_gpuobj_rd32, |
| 219 | .wr32 = _nouveau_gpuobj_wr32, |
| 220 | }, |
| 221 | }; |
| 222 | |
| 223 | int |
| 224 | nouveau_gpuobj_new(struct nouveau_object *parent, struct nouveau_object *pargpu, |
| 225 | u32 size, u32 align, u32 flags, |
| 226 | struct nouveau_gpuobj **pgpuobj) |
| 227 | { |
| 228 | struct nouveau_object *engine = parent; |
| 229 | struct nouveau_gpuobj_class args = { |
| 230 | .pargpu = pargpu, |
| 231 | .size = size, |
| 232 | .align = align, |
| 233 | .flags = flags, |
| 234 | }; |
| 235 | |
| 236 | if (!nv_iclass(engine, NV_SUBDEV_CLASS)) |
| 237 | engine = engine->engine; |
| 238 | BUG_ON(engine == NULL); |
| 239 | |
| 240 | return nouveau_object_ctor(parent, engine, &_nouveau_gpuobj_oclass, |
| 241 | &args, sizeof(args), |
| 242 | (struct nouveau_object **)pgpuobj); |
| 243 | } |
| 244 | |
| 245 | int |
| 246 | nouveau_gpuobj_map(struct nouveau_gpuobj *gpuobj, u32 access, |
| 247 | struct nouveau_vma *vma) |
| 248 | { |
| 249 | struct nouveau_bar *bar = nouveau_bar(gpuobj); |
| 250 | int ret = -EINVAL; |
| 251 | |
| 252 | if (bar && bar->umap) { |
| 253 | struct nouveau_instobj *iobj = (void *) |
| 254 | nv_pclass(nv_object(gpuobj), NV_MEMOBJ_CLASS); |
| 255 | struct nouveau_mem **mem = (void *)(iobj + 1); |
| 256 | ret = bar->umap(bar, *mem, access, vma); |
| 257 | } |
| 258 | |
| 259 | return ret; |
| 260 | } |
| 261 | |
| 262 | int |
| 263 | nouveau_gpuobj_map_vm(struct nouveau_gpuobj *gpuobj, struct nouveau_vm *vm, |
| 264 | u32 access, struct nouveau_vma *vma) |
| 265 | { |
| 266 | struct nouveau_instobj *iobj = (void *) |
| 267 | nv_pclass(nv_object(gpuobj), NV_MEMOBJ_CLASS); |
| 268 | struct nouveau_mem **mem = (void *)(iobj + 1); |
| 269 | int ret; |
| 270 | |
| 271 | ret = nouveau_vm_get(vm, gpuobj->size, 12, access, vma); |
| 272 | if (ret) |
| 273 | return ret; |
| 274 | |
| 275 | nouveau_vm_map(vma, *mem); |
| 276 | return 0; |
| 277 | } |
| 278 | |
| 279 | void |
| 280 | nouveau_gpuobj_unmap(struct nouveau_vma *vma) |
| 281 | { |
| 282 | if (vma->node) { |
| 283 | nouveau_vm_unmap(vma); |
| 284 | nouveau_vm_put(vma); |
| 285 | } |
| 286 | } |
| 287 | |
| 288 | /* the below is basically only here to support sharing the paged dma object |
| 289 | * for PCI(E)GART on <=nv4x chipsets, and should *not* be expected to work |
| 290 | * anywhere else. |
| 291 | */ |
| 292 | |
| 293 | static void |
| 294 | nouveau_gpudup_dtor(struct nouveau_object *object) |
| 295 | { |
| 296 | struct nouveau_gpuobj *gpuobj = (void *)object; |
| 297 | nouveau_object_ref(NULL, &gpuobj->parent); |
| 298 | nouveau_object_destroy(&gpuobj->base); |
| 299 | } |
| 300 | |
| 301 | static struct nouveau_oclass |
| 302 | nouveau_gpudup_oclass = { |
| 303 | .handle = NV_GPUOBJ_CLASS, |
| 304 | .ofuncs = &(struct nouveau_ofuncs) { |
| 305 | .dtor = nouveau_gpudup_dtor, |
| 306 | .init = nouveau_object_init, |
| 307 | .fini = nouveau_object_fini, |
| 308 | }, |
| 309 | }; |
| 310 | |
| 311 | int |
| 312 | nouveau_gpuobj_dup(struct nouveau_object *parent, struct nouveau_gpuobj *base, |
| 313 | struct nouveau_gpuobj **pgpuobj) |
| 314 | { |
| 315 | struct nouveau_gpuobj *gpuobj; |
| 316 | int ret; |
| 317 | |
| 318 | ret = nouveau_object_create(parent, parent->engine, |
| 319 | &nouveau_gpudup_oclass, 0, &gpuobj); |
| 320 | *pgpuobj = gpuobj; |
| 321 | if (ret) |
| 322 | return ret; |
| 323 | |
| 324 | nouveau_object_ref(nv_object(base), &gpuobj->parent); |
| 325 | gpuobj->addr = base->addr; |
| 326 | gpuobj->size = base->size; |
| 327 | return 0; |
| 328 | } |
| 329 | |