1 ===================== 1 =====================
2 DRM Memory Management 2 DRM Memory Management
3 ===================== 3 =====================
4 4
5 Modern Linux systems require large amount of g 5 Modern Linux systems require large amount of graphics memory to store
6 frame buffers, textures, vertices and other gr 6 frame buffers, textures, vertices and other graphics-related data. Given
7 the very dynamic nature of many of that data, 7 the very dynamic nature of many of that data, managing graphics memory
8 efficiently is thus crucial for the graphics s 8 efficiently is thus crucial for the graphics stack and plays a central
9 role in the DRM infrastructure. 9 role in the DRM infrastructure.
10 10
11 The DRM core includes two memory managers, nam !! 11 The DRM core includes two memory managers, namely Translation Table Maps
12 (TTM) and Graphics Execution Manager (GEM). TT 12 (TTM) and Graphics Execution Manager (GEM). TTM was the first DRM memory
13 manager to be developed and tried to be a one- 13 manager to be developed and tried to be a one-size-fits-them all
14 solution. It provides a single userspace API t 14 solution. It provides a single userspace API to accommodate the need of
15 all hardware, supporting both Unified Memory A 15 all hardware, supporting both Unified Memory Architecture (UMA) devices
16 and devices with dedicated video RAM (i.e. mos 16 and devices with dedicated video RAM (i.e. most discrete video cards).
17 This resulted in a large, complex piece of cod 17 This resulted in a large, complex piece of code that turned out to be
18 hard to use for driver development. 18 hard to use for driver development.
19 19
20 GEM started as an Intel-sponsored project in r 20 GEM started as an Intel-sponsored project in reaction to TTM's
21 complexity. Its design philosophy is completel 21 complexity. Its design philosophy is completely different: instead of
22 providing a solution to every graphics memory- 22 providing a solution to every graphics memory-related problems, GEM
23 identified common code between drivers and cre 23 identified common code between drivers and created a support library to
24 share it. GEM has simpler initialization and e 24 share it. GEM has simpler initialization and execution requirements than
25 TTM, but has no video RAM management capabilit 25 TTM, but has no video RAM management capabilities and is thus limited to
26 UMA devices. 26 UMA devices.
27 27
28 The Translation Table Manager (TTM) 28 The Translation Table Manager (TTM)
29 =================================== 29 ===================================
30 30
31 .. kernel-doc:: drivers/gpu/drm/ttm/ttm_module !! 31 TTM design background and information belongs here.
32 :doc: TTM <<
33 32
34 .. kernel-doc:: include/drm/ttm/ttm_caching.h !! 33 TTM initialization
35 :internal: !! 34 ------------------
36 35
37 TTM device object reference !! 36 **Warning**
38 --------------------------- !! 37 This section is outdated.
39 38
40 .. kernel-doc:: include/drm/ttm/ttm_device.h !! 39 Drivers wishing to support TTM must pass a filled :c:type:`ttm_bo_driver
41 :internal: !! 40 <ttm_bo_driver>` structure to ttm_bo_device_init, together with an
>> 41 initialized global reference to the memory manager. The ttm_bo_driver
>> 42 structure contains several fields with function pointers for
>> 43 initializing the TTM, allocating and freeing memory, waiting for command
>> 44 completion and fence synchronization, and memory migration.
42 45
43 .. kernel-doc:: drivers/gpu/drm/ttm/ttm_device !! 46 The :c:type:`struct drm_global_reference <drm_global_reference>` is made
44 :export: !! 47 up of several fields:
45 48
46 TTM resource placement reference !! 49 .. code-block:: c
47 -------------------------------- <<
48 50
49 .. kernel-doc:: include/drm/ttm/ttm_placement. !! 51 struct drm_global_reference {
50 :internal: !! 52 enum ttm_global_types global_type;
>> 53 size_t size;
>> 54 void *object;
>> 55 int (*init) (struct drm_global_reference *);
>> 56 void (*release) (struct drm_global_reference *);
>> 57 };
>> 58
>> 59
>> 60 There should be one global reference structure for your memory manager
>> 61 as a whole, and there will be others for each object created by the
>> 62 memory manager at runtime. Your global TTM should have a type of
>> 63 TTM_GLOBAL_TTM_MEM. The size field for the global object should be
>> 64 sizeof(struct ttm_mem_global), and the init and release hooks should
>> 65 point at your driver-specific init and release routines, which probably
>> 66 eventually call ttm_mem_global_init and ttm_mem_global_release,
>> 67 respectively.
51 68
52 TTM resource object reference !! 69 Once your global TTM accounting structure is set up and initialized by
53 ----------------------------- !! 70 calling ttm_global_item_ref() on it, you need to create a buffer
>> 71 object TTM to provide a pool for buffer object allocation by clients and
>> 72 the kernel itself. The type of this object should be
>> 73 TTM_GLOBAL_TTM_BO, and its size should be sizeof(struct
>> 74 ttm_bo_global). Again, driver-specific init and release functions may
>> 75 be provided, likely eventually calling ttm_bo_global_init() and
>> 76 ttm_bo_global_release(), respectively. Also, like the previous
>> 77 object, ttm_global_item_ref() is used to create an initial reference
>> 78 count for the TTM, which will call your initialization function.
54 79
55 .. kernel-doc:: include/drm/ttm/ttm_resource.h !! 80 See the radeon_ttm.c file for an example of usage.
56 :internal: <<
57 81
58 .. kernel-doc:: drivers/gpu/drm/ttm/ttm_resour !! 82 .. kernel-doc:: drivers/gpu/drm/drm_global.c
59 :export: 83 :export:
60 84
61 TTM TT object reference <<
62 ----------------------- <<
63 <<
64 .. kernel-doc:: include/drm/ttm/ttm_tt.h <<
65 :internal: <<
66 <<
67 .. kernel-doc:: drivers/gpu/drm/ttm/ttm_tt.c <<
68 :export: <<
69 <<
70 TTM page pool reference <<
71 ----------------------- <<
72 <<
73 .. kernel-doc:: include/drm/ttm/ttm_pool.h <<
74 :internal: <<
75 <<
76 .. kernel-doc:: drivers/gpu/drm/ttm/ttm_pool.c <<
77 :export: <<
78 85
79 The Graphics Execution Manager (GEM) 86 The Graphics Execution Manager (GEM)
80 ==================================== 87 ====================================
81 88
82 The GEM design approach has resulted in a memo 89 The GEM design approach has resulted in a memory manager that doesn't
83 provide full coverage of all (or even all comm 90 provide full coverage of all (or even all common) use cases in its
84 userspace or kernel API. GEM exposes a set of 91 userspace or kernel API. GEM exposes a set of standard memory-related
85 operations to userspace and a set of helper fu 92 operations to userspace and a set of helper functions to drivers, and
86 let drivers implement hardware-specific operat 93 let drivers implement hardware-specific operations with their own
87 private API. 94 private API.
88 95
89 The GEM userspace API is described in the `GEM 96 The GEM userspace API is described in the `GEM - the Graphics Execution
90 Manager <http://lwn.net/Articles/283798/>`__ a 97 Manager <http://lwn.net/Articles/283798/>`__ article on LWN. While
91 slightly outdated, the document provides a goo 98 slightly outdated, the document provides a good overview of the GEM API
92 principles. Buffer allocation and read and wri 99 principles. Buffer allocation and read and write operations, described
93 as part of the common GEM API, are currently i 100 as part of the common GEM API, are currently implemented using
94 driver-specific ioctls. 101 driver-specific ioctls.
95 102
96 GEM is data-agnostic. It manages abstract buff 103 GEM is data-agnostic. It manages abstract buffer objects without knowing
97 what individual buffers contain. APIs that req 104 what individual buffers contain. APIs that require knowledge of buffer
98 contents or purpose, such as buffer allocation 105 contents or purpose, such as buffer allocation or synchronization
99 primitives, are thus outside of the scope of G 106 primitives, are thus outside of the scope of GEM and must be implemented
100 using driver-specific ioctls. 107 using driver-specific ioctls.
101 108
102 On a fundamental level, GEM involves several o 109 On a fundamental level, GEM involves several operations:
103 110
104 - Memory allocation and freeing 111 - Memory allocation and freeing
105 - Command execution 112 - Command execution
106 - Aperture management at command execution ti 113 - Aperture management at command execution time
107 114
108 Buffer object allocation is relatively straigh 115 Buffer object allocation is relatively straightforward and largely
109 provided by Linux's shmem layer, which provide 116 provided by Linux's shmem layer, which provides memory to back each
110 object. 117 object.
111 118
112 Device-specific operations, such as command ex 119 Device-specific operations, such as command execution, pinning, buffer
113 read & write, mapping, and domain ownership tr 120 read & write, mapping, and domain ownership transfers are left to
114 driver-specific ioctls. 121 driver-specific ioctls.
115 122
116 GEM Initialization 123 GEM Initialization
117 ------------------ 124 ------------------
118 125
119 Drivers that use GEM must set the DRIVER_GEM b 126 Drivers that use GEM must set the DRIVER_GEM bit in the struct
120 :c:type:`struct drm_driver <drm_driver>` drive 127 :c:type:`struct drm_driver <drm_driver>` driver_features
121 field. The DRM core will then automatically in 128 field. The DRM core will then automatically initialize the GEM core
122 before calling the load operation. Behind the 129 before calling the load operation. Behind the scene, this will create a
123 DRM Memory Manager object which provides an ad 130 DRM Memory Manager object which provides an address space pool for
124 object allocation. 131 object allocation.
125 132
126 In a KMS configuration, drivers need to alloca 133 In a KMS configuration, drivers need to allocate and initialize a
127 command ring buffer following core GEM initial 134 command ring buffer following core GEM initialization if required by the
128 hardware. UMA devices usually have what is cal 135 hardware. UMA devices usually have what is called a "stolen" memory
129 region, which provides space for the initial f 136 region, which provides space for the initial framebuffer and large,
130 contiguous memory regions required by the devi 137 contiguous memory regions required by the device. This space is
131 typically not managed by GEM, and must be init 138 typically not managed by GEM, and must be initialized separately into
132 its own DRM MM object. 139 its own DRM MM object.
133 140
134 GEM Objects Creation 141 GEM Objects Creation
135 -------------------- 142 --------------------
136 143
137 GEM splits creation of GEM objects and allocat 144 GEM splits creation of GEM objects and allocation of the memory that
138 backs them in two distinct operations. 145 backs them in two distinct operations.
139 146
140 GEM objects are represented by an instance of 147 GEM objects are represented by an instance of struct :c:type:`struct
141 drm_gem_object <drm_gem_object>`. Drivers usua 148 drm_gem_object <drm_gem_object>`. Drivers usually need to
142 extend GEM objects with private information an 149 extend GEM objects with private information and thus create a
143 driver-specific GEM object structure type that 150 driver-specific GEM object structure type that embeds an instance of
144 struct :c:type:`struct drm_gem_object <drm_gem 151 struct :c:type:`struct drm_gem_object <drm_gem_object>`.
145 152
146 To create a GEM object, a driver allocates mem 153 To create a GEM object, a driver allocates memory for an instance of its
147 specific GEM object type and initializes the e 154 specific GEM object type and initializes the embedded struct
148 :c:type:`struct drm_gem_object <drm_gem_object 155 :c:type:`struct drm_gem_object <drm_gem_object>` with a call
149 to drm_gem_object_init(). The function takes a !! 156 to :c:func:`drm_gem_object_init()`. The function takes a pointer
150 to the DRM device, a pointer to the GEM object 157 to the DRM device, a pointer to the GEM object and the buffer object
151 size in bytes. 158 size in bytes.
152 159
153 GEM uses shmem to allocate anonymous pageable 160 GEM uses shmem to allocate anonymous pageable memory.
154 drm_gem_object_init() will create an shmfs fil !! 161 :c:func:`drm_gem_object_init()` will create an shmfs file of the
155 requested size and store it into the struct :c 162 requested size and store it into the struct :c:type:`struct
156 drm_gem_object <drm_gem_object>` filp field. T 163 drm_gem_object <drm_gem_object>` filp field. The memory is
157 used as either main storage for the object whe 164 used as either main storage for the object when the graphics hardware
158 uses system memory directly or as a backing st 165 uses system memory directly or as a backing store otherwise.
159 166
160 Drivers are responsible for the actual physica 167 Drivers are responsible for the actual physical pages allocation by
161 calling shmem_read_mapping_page_gfp() for each !! 168 calling :c:func:`shmem_read_mapping_page_gfp()` for each page.
162 Note that they can decide to allocate pages wh 169 Note that they can decide to allocate pages when initializing the GEM
163 object, or to delay allocation until the memor 170 object, or to delay allocation until the memory is needed (for instance
164 when a page fault occurs as a result of a user 171 when a page fault occurs as a result of a userspace memory access or
165 when the driver needs to start a DMA transfer 172 when the driver needs to start a DMA transfer involving the memory).
166 173
167 Anonymous pageable memory allocation is not al 174 Anonymous pageable memory allocation is not always desired, for instance
168 when the hardware requires physically contiguo 175 when the hardware requires physically contiguous system memory as is
169 often the case in embedded devices. Drivers ca 176 often the case in embedded devices. Drivers can create GEM objects with
170 no shmfs backing (called private GEM objects) !! 177 no shmfs backing (called private GEM objects) by initializing them with
171 to drm_gem_private_object_init() instead of dr !! 178 a call to :c:func:`drm_gem_private_object_init()` instead of
172 private GEM objects must be managed by drivers !! 179 :c:func:`drm_gem_object_init()`. Storage for private GEM objects
>> 180 must be managed by drivers.
173 181
174 GEM Objects Lifetime 182 GEM Objects Lifetime
175 -------------------- 183 --------------------
176 184
177 All GEM objects are reference-counted by the G 185 All GEM objects are reference-counted by the GEM core. References can be
178 acquired and release by calling drm_gem_object !! 186 acquired and release by :c:func:`calling drm_gem_object_get()` and
179 respectively. !! 187 :c:func:`drm_gem_object_put()` respectively. The caller must hold the
>> 188 :c:type:`struct drm_device <drm_device>` struct_mutex lock when calling
>> 189 :c:func:`drm_gem_object_get()`. As a convenience, GEM provides
>> 190 :c:func:`drm_gem_object_put_unlocked()` functions that can be called without
>> 191 holding the lock.
180 192
181 When the last reference to a GEM object is rel 193 When the last reference to a GEM object is released the GEM core calls
182 the :c:type:`struct drm_gem_object_funcs <gem_ !! 194 the :c:type:`struct drm_driver <drm_driver>` gem_free_object
183 operation. That operation is mandatory for GEM 195 operation. That operation is mandatory for GEM-enabled drivers and must
184 free the GEM object and all associated resourc 196 free the GEM object and all associated resources.
185 197
186 void (\*free) (struct drm_gem_object \*obj); D !! 198 void (\*gem_free_object) (struct drm_gem_object \*obj); Drivers are
187 responsible for freeing all GEM object resourc 199 responsible for freeing all GEM object resources. This includes the
188 resources created by the GEM core, which need 200 resources created by the GEM core, which need to be released with
189 drm_gem_object_release(). !! 201 :c:func:`drm_gem_object_release()`.
190 202
191 GEM Objects Naming 203 GEM Objects Naming
192 ------------------ 204 ------------------
193 205
194 Communication between userspace and the kernel 206 Communication between userspace and the kernel refers to GEM objects
195 using local handles, global names or, more rec 207 using local handles, global names or, more recently, file descriptors.
196 All of those are 32-bit integer values; the us 208 All of those are 32-bit integer values; the usual Linux kernel limits
197 apply to the file descriptors. 209 apply to the file descriptors.
198 210
199 GEM handles are local to a DRM file. Applicati 211 GEM handles are local to a DRM file. Applications get a handle to a GEM
200 object through a driver-specific ioctl, and ca 212 object through a driver-specific ioctl, and can use that handle to refer
201 to the GEM object in other standard or driver- 213 to the GEM object in other standard or driver-specific ioctls. Closing a
202 DRM file handle frees all its GEM handles and 214 DRM file handle frees all its GEM handles and dereferences the
203 associated GEM objects. 215 associated GEM objects.
204 216
205 To create a handle for a GEM object drivers ca !! 217 To create a handle for a GEM object drivers call
206 function takes a pointer to the DRM file and t !! 218 :c:func:`drm_gem_handle_create()`. The function takes a pointer
207 locally unique handle. When the handle is no !! 219 to the DRM file and the GEM object and returns a locally unique handle.
208 with a call to drm_gem_handle_delete(). Finall !! 220 When the handle is no longer needed drivers delete it with a call to
209 handle can be retrieved by a call to drm_gem_o !! 221 :c:func:`drm_gem_handle_delete()`. Finally the GEM object
>> 222 associated with a handle can be retrieved by a call to
>> 223 :c:func:`drm_gem_object_lookup()`.
210 224
211 Handles don't take ownership of GEM objects, t 225 Handles don't take ownership of GEM objects, they only take a reference
212 to the object that will be dropped when the ha 226 to the object that will be dropped when the handle is destroyed. To
213 avoid leaking GEM objects, drivers must make s 227 avoid leaking GEM objects, drivers must make sure they drop the
214 reference(s) they own (such as the initial ref 228 reference(s) they own (such as the initial reference taken at object
215 creation time) as appropriate, without any spe 229 creation time) as appropriate, without any special consideration for the
216 handle. For example, in the particular case of 230 handle. For example, in the particular case of combined GEM object and
217 handle creation in the implementation of the d 231 handle creation in the implementation of the dumb_create operation,
218 drivers must drop the initial reference to the 232 drivers must drop the initial reference to the GEM object before
219 returning the handle. 233 returning the handle.
220 234
221 GEM names are similar in purpose to handles bu 235 GEM names are similar in purpose to handles but are not local to DRM
222 files. They can be passed between processes to 236 files. They can be passed between processes to reference a GEM object
223 globally. Names can't be used directly to refe 237 globally. Names can't be used directly to refer to objects in the DRM
224 API, applications must convert handles to name 238 API, applications must convert handles to names and names to handles
225 using the DRM_IOCTL_GEM_FLINK and DRM_IOCTL_GE 239 using the DRM_IOCTL_GEM_FLINK and DRM_IOCTL_GEM_OPEN ioctls
226 respectively. The conversion is handled by the 240 respectively. The conversion is handled by the DRM core without any
227 driver-specific support. 241 driver-specific support.
228 242
229 GEM also supports buffer sharing with dma-buf 243 GEM also supports buffer sharing with dma-buf file descriptors through
230 PRIME. GEM-based drivers must use the provided 244 PRIME. GEM-based drivers must use the provided helpers functions to
231 implement the exporting and importing correctl 245 implement the exporting and importing correctly. See ?. Since sharing
232 file descriptors is inherently more secure tha 246 file descriptors is inherently more secure than the easily guessable and
233 global GEM names it is the preferred buffer sh 247 global GEM names it is the preferred buffer sharing mechanism. Sharing
234 buffers through GEM names is only supported fo 248 buffers through GEM names is only supported for legacy userspace.
235 Furthermore PRIME also allows cross-device buf 249 Furthermore PRIME also allows cross-device buffer sharing since it is
236 based on dma-bufs. 250 based on dma-bufs.
237 251
238 GEM Objects Mapping 252 GEM Objects Mapping
239 ------------------- 253 -------------------
240 254
241 Because mapping operations are fairly heavywei 255 Because mapping operations are fairly heavyweight GEM favours
242 read/write-like access to buffers, implemented 256 read/write-like access to buffers, implemented through driver-specific
243 ioctls, over mapping buffers to userspace. How 257 ioctls, over mapping buffers to userspace. However, when random access
244 to the buffer is needed (to perform software r 258 to the buffer is needed (to perform software rendering for instance),
245 direct access to the object can be more effici 259 direct access to the object can be more efficient.
246 260
247 The mmap system call can't be used directly to 261 The mmap system call can't be used directly to map GEM objects, as they
248 don't have their own file handle. Two alternat 262 don't have their own file handle. Two alternative methods currently
249 co-exist to map GEM objects to userspace. The 263 co-exist to map GEM objects to userspace. The first method uses a
250 driver-specific ioctl to perform the mapping o 264 driver-specific ioctl to perform the mapping operation, calling
251 do_mmap() under the hood. This is often consid !! 265 :c:func:`do_mmap()` under the hood. This is often considered
252 dubious, seems to be discouraged for new GEM-e 266 dubious, seems to be discouraged for new GEM-enabled drivers, and will
253 thus not be described here. 267 thus not be described here.
254 268
255 The second method uses the mmap system call on 269 The second method uses the mmap system call on the DRM file handle. void
256 \*mmap(void \*addr, size_t length, int prot, i 270 \*mmap(void \*addr, size_t length, int prot, int flags, int fd, off_t
257 offset); DRM identifies the GEM object to be m 271 offset); DRM identifies the GEM object to be mapped by a fake offset
258 passed through the mmap offset argument. Prior 272 passed through the mmap offset argument. Prior to being mapped, a GEM
259 object must thus be associated with a fake off 273 object must thus be associated with a fake offset. To do so, drivers
260 must call drm_gem_create_mmap_offset() on the !! 274 must call :c:func:`drm_gem_create_mmap_offset()` on the object.
261 275
262 Once allocated, the fake offset value must be 276 Once allocated, the fake offset value must be passed to the application
263 in a driver-specific way and can then be used 277 in a driver-specific way and can then be used as the mmap offset
264 argument. 278 argument.
265 279
266 The GEM core provides a helper method drm_gem_ !! 280 The GEM core provides a helper method :c:func:`drm_gem_mmap()` to
267 handle object mapping. The method can be set d 281 handle object mapping. The method can be set directly as the mmap file
268 operation handler. It will look up the GEM obj 282 operation handler. It will look up the GEM object based on the offset
269 value and set the VMA operations to the :c:typ 283 value and set the VMA operations to the :c:type:`struct drm_driver
270 <drm_driver>` gem_vm_ops field. Note that drm_ !! 284 <drm_driver>` gem_vm_ops field. Note that
271 userspace, but relies on the driver-provided f !! 285 :c:func:`drm_gem_mmap()` doesn't map memory to userspace, but
272 individually. !! 286 relies on the driver-provided fault handler to map pages individually.
273 !! 287
274 To use drm_gem_mmap(), drivers must fill the s !! 288 To use :c:func:`drm_gem_mmap()`, drivers must fill the struct
275 <drm_driver>` gem_vm_ops field with a pointer !! 289 :c:type:`struct drm_driver <drm_driver>` gem_vm_ops field
>> 290 with a pointer to VM operations.
276 291
277 The VM operations is a :c:type:`struct vm_oper 292 The VM operations is a :c:type:`struct vm_operations_struct <vm_operations_struct>`
278 made up of several fields, the more interestin 293 made up of several fields, the more interesting ones being:
279 294
280 .. code-block:: c 295 .. code-block:: c
281 296
282 struct vm_operations_struct { 297 struct vm_operations_struct {
283 void (*open)(struct vm_area_st 298 void (*open)(struct vm_area_struct * area);
284 void (*close)(struct vm_area_s 299 void (*close)(struct vm_area_struct * area);
285 vm_fault_t (*fault)(struct vm_ !! 300 int (*fault)(struct vm_fault *vmf);
286 }; 301 };
287 302
288 303
289 The open and close operations must update the 304 The open and close operations must update the GEM object reference
290 count. Drivers can use the drm_gem_vm_open() a !! 305 count. Drivers can use the :c:func:`drm_gem_vm_open()` and
291 functions directly as open and close handlers. !! 306 :c:func:`drm_gem_vm_close()` helper functions directly as open
>> 307 and close handlers.
292 308
293 The fault operation handler is responsible for 309 The fault operation handler is responsible for mapping individual pages
294 to userspace when a page fault occurs. Dependi 310 to userspace when a page fault occurs. Depending on the memory
295 allocation scheme, drivers can allocate pages 311 allocation scheme, drivers can allocate pages at fault time, or can
296 decide to allocate memory for the GEM object a 312 decide to allocate memory for the GEM object at the time the object is
297 created. 313 created.
298 314
299 Drivers that want to map the GEM object upfron 315 Drivers that want to map the GEM object upfront instead of handling page
300 faults can implement their own mmap file opera 316 faults can implement their own mmap file operation handler.
301 317
302 For platforms without MMU the GEM core provide 318 For platforms without MMU the GEM core provides a helper method
303 drm_gem_dma_get_unmapped_area(). The mmap() ro !! 319 :c:func:`drm_gem_cma_get_unmapped_area`. The mmap() routines will call
304 proposed address for the mapping. !! 320 this to get a proposed address for the mapping.
305 321
306 To use drm_gem_dma_get_unmapped_area(), driver !! 322 To use :c:func:`drm_gem_cma_get_unmapped_area`, drivers must fill the
307 :c:type:`struct file_operations <file_operatio !! 323 struct :c:type:`struct file_operations <file_operations>` get_unmapped_area
308 a pointer on drm_gem_dma_get_unmapped_area(). !! 324 field with a pointer on :c:func:`drm_gem_cma_get_unmapped_area`.
309 325
310 More detailed information about get_unmapped_a 326 More detailed information about get_unmapped_area can be found in
311 Documentation/admin-guide/mm/nommu-mmap.rst !! 327 Documentation/nommu-mmap.txt
312 328
313 Memory Coherency 329 Memory Coherency
314 ---------------- 330 ----------------
315 331
316 When mapped to the device or used in a command 332 When mapped to the device or used in a command buffer, backing pages for
317 an object are flushed to memory and marked wri 333 an object are flushed to memory and marked write combined so as to be
318 coherent with the GPU. Likewise, if the CPU ac 334 coherent with the GPU. Likewise, if the CPU accesses an object after the
319 GPU has finished rendering to the object, then 335 GPU has finished rendering to the object, then the object must be made
320 coherent with the CPU's view of memory, usuall 336 coherent with the CPU's view of memory, usually involving GPU cache
321 flushing of various kinds. This core CPU<->GPU 337 flushing of various kinds. This core CPU<->GPU coherency management is
322 provided by a device-specific ioctl, which eva 338 provided by a device-specific ioctl, which evaluates an object's current
323 domain and performs any necessary flushing or 339 domain and performs any necessary flushing or synchronization to put the
324 object into the desired coherency domain (note 340 object into the desired coherency domain (note that the object may be
325 busy, i.e. an active render target; in that ca 341 busy, i.e. an active render target; in that case, setting the domain
326 blocks the client and waits for rendering to c 342 blocks the client and waits for rendering to complete before performing
327 any necessary flushing operations). 343 any necessary flushing operations).
328 344
329 Command Execution 345 Command Execution
330 ----------------- 346 -----------------
331 347
332 Perhaps the most important GEM function for GP 348 Perhaps the most important GEM function for GPU devices is providing a
333 command execution interface to clients. Client 349 command execution interface to clients. Client programs construct
334 command buffers containing references to previ 350 command buffers containing references to previously allocated memory
335 objects, and then submit them to GEM. At that 351 objects, and then submit them to GEM. At that point, GEM takes care to
336 bind all the objects into the GTT, execute the 352 bind all the objects into the GTT, execute the buffer, and provide
337 necessary synchronization between clients acce 353 necessary synchronization between clients accessing the same buffers.
338 This often involves evicting some objects from 354 This often involves evicting some objects from the GTT and re-binding
339 others (a fairly expensive operation), and pro 355 others (a fairly expensive operation), and providing relocation support
340 which hides fixed GTT offsets from clients. Cl 356 which hides fixed GTT offsets from clients. Clients must take care not
341 to submit command buffers that reference more 357 to submit command buffers that reference more objects than can fit in
342 the GTT; otherwise, GEM will reject them and n 358 the GTT; otherwise, GEM will reject them and no rendering will occur.
343 Similarly, if several objects in the buffer re 359 Similarly, if several objects in the buffer require fence registers to
344 be allocated for correct rendering (e.g. 2D bl 360 be allocated for correct rendering (e.g. 2D blits on pre-965 chips),
345 care must be taken not to require more fence r 361 care must be taken not to require more fence registers than are
346 available to the client. Such resource managem 362 available to the client. Such resource management should be abstracted
347 from the client in libdrm. 363 from the client in libdrm.
348 364
349 GEM Function Reference 365 GEM Function Reference
350 ---------------------- 366 ----------------------
351 367
352 .. kernel-doc:: include/drm/drm_gem.h 368 .. kernel-doc:: include/drm/drm_gem.h
353 :internal: 369 :internal:
354 370
355 .. kernel-doc:: drivers/gpu/drm/drm_gem.c 371 .. kernel-doc:: drivers/gpu/drm/drm_gem.c
356 :export: 372 :export:
357 373
358 GEM DMA Helper Functions Reference !! 374 GEM CMA Helper Functions Reference
359 ---------------------------------- 375 ----------------------------------
360 376
361 .. kernel-doc:: drivers/gpu/drm/drm_gem_dma_he !! 377 .. kernel-doc:: drivers/gpu/drm/drm_gem_cma_helper.c
362 :doc: dma helpers !! 378 :doc: cma helpers
363 379
364 .. kernel-doc:: include/drm/drm_gem_dma_helper !! 380 .. kernel-doc:: include/drm/drm_gem_cma_helper.h
365 :internal: 381 :internal:
366 382
367 .. kernel-doc:: drivers/gpu/drm/drm_gem_dma_he !! 383 .. kernel-doc:: drivers/gpu/drm/drm_gem_cma_helper.c
368 :export: <<
369 <<
370 GEM SHMEM Helper Function Reference <<
371 ----------------------------------- <<
372 <<
373 .. kernel-doc:: drivers/gpu/drm/drm_gem_shmem_ <<
374 :doc: overview <<
375 <<
376 .. kernel-doc:: include/drm/drm_gem_shmem_help <<
377 :internal: <<
378 <<
379 .. kernel-doc:: drivers/gpu/drm/drm_gem_shmem_ <<
380 :export: <<
381 <<
382 GEM VRAM Helper Functions Reference <<
383 ----------------------------------- <<
384 <<
385 .. kernel-doc:: drivers/gpu/drm/drm_gem_vram_h <<
386 :doc: overview <<
387 <<
388 .. kernel-doc:: include/drm/drm_gem_vram_helpe <<
389 :internal: <<
390 <<
391 .. kernel-doc:: drivers/gpu/drm/drm_gem_vram_h <<
392 :export: <<
393 <<
394 GEM TTM Helper Functions Reference <<
395 ----------------------------------- <<
396 <<
397 .. kernel-doc:: drivers/gpu/drm/drm_gem_ttm_he <<
398 :doc: overview <<
399 <<
400 .. kernel-doc:: drivers/gpu/drm/drm_gem_ttm_he <<
401 :export: 384 :export:
402 385
403 VMA Offset Manager 386 VMA Offset Manager
404 ================== 387 ==================
405 388
406 .. kernel-doc:: drivers/gpu/drm/drm_vma_manage 389 .. kernel-doc:: drivers/gpu/drm/drm_vma_manager.c
407 :doc: vma offset manager 390 :doc: vma offset manager
408 391
409 .. kernel-doc:: include/drm/drm_vma_manager.h 392 .. kernel-doc:: include/drm/drm_vma_manager.h
410 :internal: 393 :internal:
411 394
412 .. kernel-doc:: drivers/gpu/drm/drm_vma_manage 395 .. kernel-doc:: drivers/gpu/drm/drm_vma_manager.c
413 :export: 396 :export:
414 397
415 .. _prime_buffer_sharing: <<
416 <<
417 PRIME Buffer Sharing 398 PRIME Buffer Sharing
418 ==================== 399 ====================
419 400
420 PRIME is the cross device buffer sharing frame 401 PRIME is the cross device buffer sharing framework in drm, originally
421 created for the OPTIMUS range of multi-gpu pla 402 created for the OPTIMUS range of multi-gpu platforms. To userspace PRIME
422 buffers are dma-buf based file descriptors. 403 buffers are dma-buf based file descriptors.
423 404
424 Overview and Lifetime Rules !! 405 Overview and Driver Interface
425 --------------------------- !! 406 -----------------------------
426 407
427 .. kernel-doc:: drivers/gpu/drm/drm_prime.c !! 408 Similar to GEM global names, PRIME file descriptors are also used to
428 :doc: overview and lifetime rules !! 409 share buffer objects across processes. They offer additional security:
>> 410 as file descriptors must be explicitly sent over UNIX domain sockets to
>> 411 be shared between applications, they can't be guessed like the globally
>> 412 unique GEM names.
>> 413
>> 414 Drivers that support the PRIME API must set the DRIVER_PRIME bit in the
>> 415 struct :c:type:`struct drm_driver <drm_driver>`
>> 416 driver_features field, and implement the prime_handle_to_fd and
>> 417 prime_fd_to_handle operations.
>> 418
>> 419 int (\*prime_handle_to_fd)(struct drm_device \*dev, struct drm_file
>> 420 \*file_priv, uint32_t handle, uint32_t flags, int \*prime_fd); int
>> 421 (\*prime_fd_to_handle)(struct drm_device \*dev, struct drm_file
>> 422 \*file_priv, int prime_fd, uint32_t \*handle); Those two operations
>> 423 convert a handle to a PRIME file descriptor and vice versa. Drivers must
>> 424 use the kernel dma-buf buffer sharing framework to manage the PRIME file
>> 425 descriptors. Similar to the mode setting API PRIME is agnostic to the
>> 426 underlying buffer object manager, as long as handles are 32bit unsigned
>> 427 integers.
>> 428
>> 429 While non-GEM drivers must implement the operations themselves, GEM
>> 430 drivers must use the :c:func:`drm_gem_prime_handle_to_fd()` and
>> 431 :c:func:`drm_gem_prime_fd_to_handle()` helper functions. Those
>> 432 helpers rely on the driver gem_prime_export and gem_prime_import
>> 433 operations to create a dma-buf instance from a GEM object (dma-buf
>> 434 exporter role) and to create a GEM object from a dma-buf instance
>> 435 (dma-buf importer role).
>> 436
>> 437 struct dma_buf \* (\*gem_prime_export)(struct drm_device \*dev,
>> 438 struct drm_gem_object \*obj, int flags); struct drm_gem_object \*
>> 439 (\*gem_prime_import)(struct drm_device \*dev, struct dma_buf
>> 440 \*dma_buf); These two operations are mandatory for GEM drivers that
>> 441 support PRIME.
429 442
430 PRIME Helper Functions 443 PRIME Helper Functions
431 ---------------------- 444 ----------------------
432 445
433 .. kernel-doc:: drivers/gpu/drm/drm_prime.c 446 .. kernel-doc:: drivers/gpu/drm/drm_prime.c
434 :doc: PRIME Helpers 447 :doc: PRIME Helpers
435 448
436 PRIME Function References 449 PRIME Function References
437 ------------------------- 450 -------------------------
438 451
439 .. kernel-doc:: include/drm/drm_prime.h 452 .. kernel-doc:: include/drm/drm_prime.h
440 :internal: 453 :internal:
441 454
442 .. kernel-doc:: drivers/gpu/drm/drm_prime.c 455 .. kernel-doc:: drivers/gpu/drm/drm_prime.c
443 :export: 456 :export:
444 457
445 DRM MM Range Allocator 458 DRM MM Range Allocator
446 ====================== 459 ======================
447 460
448 Overview 461 Overview
449 -------- 462 --------
450 463
451 .. kernel-doc:: drivers/gpu/drm/drm_mm.c 464 .. kernel-doc:: drivers/gpu/drm/drm_mm.c
452 :doc: Overview 465 :doc: Overview
453 466
454 LRU Scan/Eviction Support 467 LRU Scan/Eviction Support
455 ------------------------- 468 -------------------------
456 469
457 .. kernel-doc:: drivers/gpu/drm/drm_mm.c 470 .. kernel-doc:: drivers/gpu/drm/drm_mm.c
458 :doc: lru scan roster 471 :doc: lru scan roster
459 472
460 DRM MM Range Allocator Function References 473 DRM MM Range Allocator Function References
461 ------------------------------------------ 474 ------------------------------------------
462 475
463 .. kernel-doc:: include/drm/drm_mm.h 476 .. kernel-doc:: include/drm/drm_mm.h
464 :internal: 477 :internal:
465 478
466 .. kernel-doc:: drivers/gpu/drm/drm_mm.c 479 .. kernel-doc:: drivers/gpu/drm/drm_mm.c
467 :export: 480 :export:
468 481
469 .. _drm_gpuvm: !! 482 DRM Cache Handling
470 !! 483 ==================
471 DRM GPUVM <<
472 ========= <<
473 <<
474 Overview <<
475 -------- <<
476 <<
477 .. kernel-doc:: drivers/gpu/drm/drm_gpuvm.c <<
478 :doc: Overview <<
479 <<
480 Split and Merge <<
481 --------------- <<
482 <<
483 .. kernel-doc:: drivers/gpu/drm/drm_gpuvm.c <<
484 :doc: Split and Merge <<
485 <<
486 .. _drm_gpuvm_locking: <<
487 <<
488 Locking <<
489 ------- <<
490 <<
491 .. kernel-doc:: drivers/gpu/drm/drm_gpuvm.c <<
492 :doc: Locking <<
493 <<
494 Examples <<
495 -------- <<
496 <<
497 .. kernel-doc:: drivers/gpu/drm/drm_gpuvm.c <<
498 :doc: Examples <<
499 <<
500 DRM GPUVM Function References <<
501 ----------------------------- <<
502 <<
503 .. kernel-doc:: include/drm/drm_gpuvm.h <<
504 :internal: <<
505 <<
506 .. kernel-doc:: drivers/gpu/drm/drm_gpuvm.c <<
507 :export: <<
508 <<
509 DRM Buddy Allocator <<
510 =================== <<
511 <<
512 DRM Buddy Function References <<
513 ----------------------------- <<
514 <<
515 .. kernel-doc:: drivers/gpu/drm/drm_buddy.c <<
516 :export: <<
517 <<
518 DRM Cache Handling and Fast WC memcpy() <<
519 ======================================= <<
520 484
521 .. kernel-doc:: drivers/gpu/drm/drm_cache.c 485 .. kernel-doc:: drivers/gpu/drm/drm_cache.c
522 :export: <<
523 <<
524 .. _drm_sync_objects: <<
525 <<
526 DRM Sync Objects <<
527 ================ <<
528 <<
529 .. kernel-doc:: drivers/gpu/drm/drm_syncobj.c <<
530 :doc: Overview <<
531 <<
532 .. kernel-doc:: include/drm/drm_syncobj.h <<
533 :internal: <<
534 <<
535 .. kernel-doc:: drivers/gpu/drm/drm_syncobj.c <<
536 :export: <<
537 <<
538 DRM Execution context <<
539 ===================== <<
540 <<
541 .. kernel-doc:: drivers/gpu/drm/drm_exec.c <<
542 :doc: Overview <<
543 <<
544 .. kernel-doc:: include/drm/drm_exec.h <<
545 :internal: <<
546 <<
547 .. kernel-doc:: drivers/gpu/drm/drm_exec.c <<
548 :export: <<
549 <<
550 GPU Scheduler <<
551 ============= <<
552 <<
553 Overview <<
554 -------- <<
555 <<
556 .. kernel-doc:: drivers/gpu/drm/scheduler/sche <<
557 :doc: Overview <<
558 <<
559 Flow Control <<
560 ------------ <<
561 <<
562 .. kernel-doc:: drivers/gpu/drm/scheduler/sche <<
563 :doc: Flow Control <<
564 <<
565 Scheduler Function References <<
566 ----------------------------- <<
567 <<
568 .. kernel-doc:: include/drm/gpu_scheduler.h <<
569 :internal: <<
570 <<
571 .. kernel-doc:: drivers/gpu/drm/scheduler/sche <<
572 :export: <<
573 <<
574 .. kernel-doc:: drivers/gpu/drm/scheduler/sche <<
575 :export: 486 :export:
Linux® is a registered trademark of Linus Torvalds in the United States and other countries.
TOMOYO® is a registered trademark of NTT DATA CORPORATION.